Will 2023 be the first year above 1.5C?

It is possible, but only in one of the four temperature records that stretch back to the 1800s

Early this year we wrote an analysis of whether next year, 2024, will be the first year whose annual average temperature exceeds 1.5C. At the time 2023 was off to a warm but not record warm start, with a lingering La Nina event suppressing global temperatures over the first three months.

What a difference a few months can make! Ever since June global temperatures have been shattering prior records, with June, July, August, and (soon) September shattering past records for the month by 0.2C to 0.3C. 2023 now looks extremely likely to be the warmest year on record.

Daily global surface temperature anomalies from the JRA-55 reanalysis product. The current month (September) is highlighted in blue.

This has raised the real possibility that 2023 might be the first year whose annual average temperature is 1.5C above preindustrial levels – defined here as the average over the 1850-1900 period (as reliable global temperature records do not go back before then).

In its recent August 2023 temperature report, Berkeley Earth puts the odds of the year as a whole exceeding 1.5C as approximately 55% in their dataset, though they acknowledge that the exceptional warmth of the past few months makes it difficult to find past analogues for the year that are predictive of where temperatures go from here over the next four months.

Berkeley Earth annual global temperatures and 2023 year-to-date (red) and projection (green).

Based on the data through August, we can estimate what the anomaly value of the remaining months would need to be to exceed 1.5C for the year as a whole in the Berkeley Earth dataset. As shown in the figure below, this would require the remaining four months of the year to end up somewhere between July and August temperature anomalies – or around 1.6C above preindustrial levels.

Monthly temperature anomalies relative to a 1850-1900 baseline in Berkeley Earth, along with estimates of required anomalies for 2023 to reach 1.5C in red.

Preindustrial uncertainties

While global temperatures are quite well constrained over the past 50 years, going all the way back to 1850 requires dealing with sparser spatial coverage and combining lots of differing and oft-changing measurement techniques. Because of this, there is wider disagreement between records from different research groups during the 1800s and early 1900s, as shown in the figure below.

Annual global surface temperature changes since 1850. Figure from the WMO State of the Climate 2022 report.

These preindustrial differences result in estimates of warming to-date that vary up to 0.2C across groups, with Berkeley Earth having the most warming since preindustrial and NOAA’s GlobalTemp the least. While these differences have minimal bearing on our understanding of current global warming, they have a large impact on our estimate of the likelihood that 2023 will exceed 1.5C.

Using the NOAA dataset – which has the least warming since preindustrial – the remaining months would need to be an implausible ~2C for the year as a whole to exceed 1.5C

Monthly temperature anomalies relative to a 1850-1900 baseline in NOAA GlobalTemp, along with estimates of required anomalies for 2023 to reach 1.5C in red.

Similarly, the NASA GISTEMP dataset would requires a very unlikely ~1.9C for the remaining months of the year for 2023 to exceed 1.5C.

Monthly temperature anomalies relative to a 1850-1900 baseline in NASA GISEMP, along with estimates of required anomalies for 2023 to reach 1.5C in red.

Finally, in the ERA5 dataset (using HadCRUT5 to fill in the preindustrial period), global temperature anomalies would have to be close to 1.8C for the remaining months of the year (HadCRUT5 itself is now shown as it had not yet reported an August 2023 temperature at the time this post was published).

Monthly temperature anomalies relative to a 1850-1900 baseline in ERA5, along with estimates of required anomalies for 2023 to reach 1.5C in red. Note that HadCRUT5 was used to extend ERA5 back to 1850.

Thus while there is a real chance that Berkeley Earth could have 2023 as the first year above 1.5C, its unlikely that will be true in any other dataset due to differences in estimated preindustrial temperatures.

What does a single year above 1.5C mean?

In the 2016 Paris Agreement the world set an aspirational to “keep global temperatures well below 2C above pre-industrial times while pursuing means to limit the increase to 1.5C.” However, these targets have created a fair bit of confusion, as they refer to long-term average global temperatures rather than short-term year-to-year variability.

A single year above 1.5C (or 2C for that matter) would not mean that the world has passed that temperature target. Given the warming rate of 0.2C per decade since 1970, and the fact that a super El Nino event can add up to 0.2C to a specific year’s global temperatures through natural variability, it is possible for us to see a single year exceeding 1.5C a full decade before the long-term average warming driven by human emissions of greenhouse gases does.

Because of this, we should be a bit cautious about how we interpret the first year passing 1.5C (whether it happens in 2023, 2024, or later)! Most models don’t expect the longer term average to pass 1.5C until the early-to-mid 2030s, but we do expect many individual years to pass 1.5C before then. In fact, earlier this year the UK Met Office gave a 66 percent chance for a single year to pass 1.5C during the period from 2023-2027.