Guest commentary by Malte Meinshausen, Zebedee Nicholls, and Piers Forster
Of all the troubling headlines emerging from the release of the Intergovernmental Panel on Climate Change (IPCC) WG1 report, one warning will surely dominate headlines in the next days and weeks: Earth is likely to reach the crucial 1.5℃ warming limit in the early 2030s.
In 2018, the IPCC Special Report on 1.5C warming stated in its summary for policy makers that the world was likely to cross the 1.5℃ threshold between 2030 and 2052, if current warming trends continue.
In this latest AR6, a more comprehensive assessment was undertaken to estimate when a warming level of 1.5℃ might be reached. As a result, some early media reports suggest 1.5ºC warming is now anticipated 10-years earlier than previously assumed (AFR, THE TIMES).
We want to explain here why that is not backed up by a rigorous comparison of the SR1.5 and AR6 reports. In fact, the science in the previous SR1.5 report and the new AR6 report are remarkably consistent.
In a very low emission scenario (i.e. the so-called SSP1-1.9 scenario), our best estimate of when 1.5ºC warming might be reached in the AR6 report is around 2034.5 (the year on which the 20-year period 2025-2044 is centred, as stated in Table 4.5 of that report). In SR1.5 the comparable estimate was 2035 (stated in Table 2.SM.12 of Chapter 2 in the report’s underlying material). These two dates are exactly the same within rounding.
But let’s take a step back before going into the more detailed comparison below.
According to the AR6 findings, Earth’s average temperature in the last decade (2011-2020) was 1.09℃ warmer than pre-industrial times (here defined as 1850-1900). Obviously, this goes most of the way to 1.5℃ of warming and human influence is pretty much the sole contributor to this observed warming. On their own, our greenhouse gas emissions would have caused a much higher warming, were it not for the masking effect of aerosol pollution.
We often use 20-year or 30-year averages to deal with the fact that there is a lot of internal variability in the system (a very likely (5-95%) range of ±0.25℃ in any single year). This is why the discussion about exceedance times for any specified global warming level relates to the long-term global average (such as 20-year periods). We are going to see individual years already exceeding 1.5ºC warming before the long-term average does. In fact at the peak of the 2015-16 El Niño event, global anomalies for individual months already exceeded 1.5℃.
What did the previous SR1.5 report say on the timing of crossing 1.5ºC?
The headline number from SR1.5 – a likely 2030 to 2052 range – was based on the assumption of a linear continuation of then current warming trends on top of the then-assumed current warming. There is also a figure (Figure SR1.5 SPM.1) that includes a red-dashed linear line, hitting the 1.5C mark somewhere between 2039 and 2040.
However, apart from that figure, the central estimate does not get any mention in the SR1.5 report. In fact, the underlying material in Chapter 1 describes that there are multiple lines of evidence supporting the lower 2030 bound of the stated range and supposedly only a single study supporting the upper bound of the likely range, i.e. 2052.
Thus, the SR1.5 provides us basically with a lopsided likely range from 2030 to 2052, with the balance of evidence tilting towards 2030. Remember that this is based on an approach that simply assumes a linear extrapolation of current warming trends.
A simple linear extrapolation is a good first approximation. But certainly not the best method to estimate when 1.5ºC warming might be reached in the real world. For that, we should take into account the evolution of GHG emissions, aerosols and other forcings. Thus, a scenario-based approach is more appropriate.
Chapter 2 in the SR1.5 report did just that. Buried a bit deeper in that report was an assessment of when the low set of scenarios could reach warming of around 1.5ºC based on a large range of 37 scenarios that reach 1.5ºC and don’t exceed it by more than 0.1ºC. On the basis of this, more sophisticated, analysis the time at which 1.5ºC is reached was estimated to be around 2035 (2033-2036, interquartile range across the best-estimates of the 37 scenarios) (SR1.5 Chapter 2, Table 2.SM.12).
What does the AR6 say about when we could reach 1.5ºC?
Chapter 4 provides the assessed temperature projections under the main five SSP scenarios considered in that report. For the very low scenario, SSP1-1.9, the best-estimate finding is that the 2025-2044 period is going to be the first 20-year period, in which an average warming of 1.5ºC is attained. That makes a 2034.5 central point:
As a summary statement, the IPCC AR6 report states that 1.5℃ warming will be reached or exceeded in the early 2030s in all emissions scenarios considered – except the highest emissions scenario, for which the crossing would occur even earlier. That’s also in line with earlier research (Henley and King, 2017) on the high RCP8.5 scenario, which showed crossing times of around 2028.
Under the very low emissions scenario considered in the report, the story is not one about exceedance or “crossing”. Our best estimate is that under the low scenario, we’re not going to see the 1.5ºC level far away in the rear-mirror. That’s why the new IPCC AR6 Summary for Policy Makers talks about “reaching” when it comes to the lowest scenario and 1.5ºC:
The best estimate projection is that warming will reach levels around 1.5ºC (with a potential limited overshoot up to +0.1C) and then drop slightly to 1.4ºC again by the end of the century again under SSP1-1.9. That is pretty much exactly the same finding as the previous SR1.5 scenarios had for 37 scenarios in its scenario class into which SSP1-1.9 would fall, i.e the class “1.5ºC with limited overshoot”. In a nutshell, the new AR6 report does not suggest it is any harder to limit warming to around 1.5ºC than what IPCC previously said. That is surprising to some extent, as an update of historical warming added 0.08ºC to the observed warming to date, but there are many updates in the underlying methodologies, from narrower climate sensitivity ranges, and updated assumptions on aerosols.
However, note that the uncertainty ranges are somewhat large. Partly due to the methodological choices, where future temperatures are merged with historical ones using the 1995-2014 base period, the lower bound is pretty much “now” – that would mean that – under this particular methodology, there is a chance that we are already ‘now’ (2022) sitting in the middle of a 20-year period (2013-2032) that has a 1.5ºC warming as its average (Chapter 4, Table 4.5). On the far end, that uncertainty stretches out quite a bit as well, because we might never actually hit 1.5ºC under the low SSP1-1.9 scenario (Display 3 above). Thus, in simple terms, the uncertainty range is between “now and never” (we acknowledge the trademark rights of our esteemed colleague Dr. Erich Fischer from ETH Zurich, who coined that summary during our IPCC approval deliberations). In some ways this large uncertainty is unrealistic as we can be pretty confident the crossing time isn’t this year, this shows there is still room for improvement on the AR6 crossing time estimates.
In summary, when apples are compared to apples (i.e. best estimates of scenario-based exceedance times), then the SR1.5 and AR6 provide remarkably consistent numbers: 2034.5 versus 2035. Such a remarkably robust scientific finding is boring to report on, so expect a few headlines in the media that compare apples and oranges. Luckily, a rigorous approval process is sensitive to such issues, which is why now a rather powerful footnote accompanies these findings:
The AR6 assessment of when a given global warming level is first exceeded benefits from the consideration of the illustrative scenarios, the multiple lines of evidence entering the assessment of future global surface temperature response to radiative forcing, and the improved estimate of historical warming. The AR6 assessment is thus not directly comparable to the SR1.5 SPM, which reported likely reaching 1.5°C global warming between 2030 and 2052, from a simple linear extrapolation of warming rates of the recent past. When considering scenarios similar to SSP1-1.9 instead of linear extrapolation, the SR1.5 estimate of when 1.5°C global warming is first exceeded is close to the best estimate reported here.”
Footnote 27. AR6 SPM
Does that mean the 1.5ºC goal is lost?
Not quite. Two elements are important to answer that question. For one, would you abolish a speed limit for somebody who drives too fast? Similarly, the 1.5ºC goal in the Paris Agreement is not a betting game of where we will end up with maximum temperatures.
Rather, the 1.5ºC goal is underpinned by an international compromise agreement, where the international community considers the projected impact to outweigh the costs of mitigations getting there. Not explicitly in dollar terms, but implicitly as a value judgement and normative target of which impacts we consider as being too high.
The 1.5ºC target is a goal enshrined in Art. 2 of the Paris Agreement as “pursuing efforts to limit the temperature increase to 1.5ºC”. That Art. 2 is even referred to as a single temperature goal (Rajamani and Werksman, 2018) by other Articles (Art. 4) of the Paris Agreement, even though it includes the “well below 2°C” part. In other words, the “well below 2ºC” part of Art. 2 sets a cap on the peak temperature at all times, while the ambition is to limit warming to below 1.5ºC, if not during peak times, then again thereafter. Whether it was a drafting oversight to refer to Art. 2 as a single temperature target or not, it makes sense to see both 2ºC and 1.5ºC together as one target. After all, they are connected by “and” and not “or” in Art. 2. Thus the 1.5ºC goal is not lost, even with a slight or limited overshoot. Overshooting the 1.5ºC limit of course comes with additional, and sometimes irreversible, impacts, but even the SR1.5 report used the class of scenarios that included the limited overshoot to investigate the 1.5ºC futures in terms of impacts, adaptation and mitigation.
Therefore nothing is new here. In fact, all the main 1.5ºC scenario literature available at the time of the Paris Agreement in 2015, for example the UNEP Emission Gap report (2015) or the UNFCCC Synthesis Reports in 2015 used such a scenario classification. For the 1.5ºC scenarios, the key thing is where temperatures are after the peak by the end of the century. If they are below 1.5ºC, then the scenario is a 1.5ºC scenario. If the peak temperature throughout the 21st century is below 1.5ºC, then it is a “below 1.5C” scenario. If temperatures remained below
1.6ºC at its peak, then the scenario has a so-called ‘limited’ overshoot (the precise SR1.5 definition of limited overshoot was slightly different, choosing a 33% likelihood of staying below 1.5C over all time, but that is another story).
As many nations are basing their current Nationally Determined Contributions and their COP26 negotiating positions on the SR1.5 scenarios and net zero dates, one of the most important conclusions of AR6 is that the SR1.5 science is robust: policy targets do not need to be revised, just acted on.
Despite upward adjustments of historical warming estimates by around 0.08ºC, the new IPCC report provides remarkably robust policy advice. The main temperature projections for the scenarios are similar under a like-with-like comparison, the remaining carbon budgets are very close to previously stated ones and our best estimate of when we experience 1.5ºC warming under a future scenario has stayed pretty much the same: the early or mid 2030s. It seems like the only thing that is changing across 30 years of IPCC reports is that the time is running. And given lackluster mitigation action, time seems to be running out.
B.J. Henley, and A.D. King, “Trajectories toward the 1.5°C Paris target: Modulation by the Interdecadal Pacific Oscillation”, Geophysical Research Letters, vol. 44, pp. 4256-4262, 2017. http://dx.doi.org/10.1002/2017GL073480
L. Rajamani, and J. Werksman, “The legal character and operational relevance of the Paris Agreement’s temperature goal”, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 376, pp. 20160458, 2018. http://dx.doi.org/10.1098/rsta.2016.0458