The year 2023 may well come to be seen as a watershed year for the climate: annually averaged global mean temperatures are thought to have reached or exceeded the 1.5 °C warming threshold for the first time since observations began1,2,3. The significance of this moment has reinvigorated appeals by the global climate community for the reduction of greenhouse gas emissions: United Nations Secretary-General António Guterres suggested that this “record global heat should send shivers down the spines of world leaders”4.

Keeping the global context aside for a moment, the impacts in terms of natural disasters associated with the recorded extreme temperatures in 2023 would almost go unnoticed in the parade of record-shattering years that have characterised the 2010s and 2020s. The EM-DAT database, a record of global disasters and associated damages5, suggests 2023 was a fairly average year for natural disasters by the standards of the past 23 years.

EM-DAT records heatwaves under meteorological disasters and droughts and wild/forest fires under climatological disasters. As seen in Fig. 1, neither category particularly stands out with respect to other years during this century. There may have been singular events like the Canadian wildfires, and temperature records may have been broken in multiple countries2, but at least in the EM-DAT database, this did not translate to an unusually large number of natural disasters. While a single year is by no means representative of changes in extreme events, 2023 underscores the importance of understanding how changes in the mean translate to changes in extremes for both global and regional climates.

Fig. 1: Natural disasters in the 21st century.
figure 1

The total number of natural disasters (panel a) does not show a strong trend over the 21st century, and the year 2023 does not stand out. Associated damages are higher in the second half of the record (panel b) but not unusually so in 2023. The red solid and dashed lines represent the mean and ±1 standard deviations, respectively.

Full size image

All in all, the year 2023 is likely to be memorable not as an example of the impacts of an extremely warm climate but as a reminder of how close we are to crossing over entirely into a 1.5 °C world.

Natural variability superimposed on global warming: El Niño

One immediate question about the near-term future is how the natural variability that is superimposed on the long-term trend affects climate warming. This question was explored extensively with respect to decadal Pacific variability earlier this century6. In the current context, the inter-annual variability of the tropical Pacific and its interaction with anthropogenic warming is likely to be of interest.

An anomalously warm tropical Pacific has a dramatic impact on the vertically integrated thermal energy (or enthalpy) of the atmosphere (Fig. 2, panel a). The long-term trend in enthalpy seems fairly flat until the strong El-Niño of 2010, and then quickly ramped up to a higher mean level around the El-Niño of 2015. Nearly all of the warmest years on record have been in the period starting in 2015 (Fig. 2, panel b)2. A consistent drop in enthalpy is observed between 2020 and 2023, reflecting the triple-dip La Niña event that preceded last year’s El Niño. However, the 2-m temperature shows a more consistent upward trend throughout the period 1980–2023 and is punctuated by the El Niño signal, as observed for enthalpy.

Fig. 2: Atmospheric thermal energy and surface temperature.
figure 2

Monthly anomalies of total atmospheric thermal energy (enthalpy, a) and 2-m temperature (b) with respect to the climatology 1980–2023 show that the year 2023 was exceptional in terms of warmth. A new, higher level appears to have been reached after the 2015 El Niño. The dashed red line represents the 5-year running mean of the anomalies in both panels. Data are from the ERA5 dataset9. The colour shades represent the amplitude of the Oceanic Niño Index10, with red colours representing an El Niño-like state and blue colours representing a La Niña-like state.

Full size image

Comparing the seasonal evolution of the zonal mean anomalies (relative to a 1980–2023 climatology) of four of the strongest El Niños of recent times (Fig. 3)—1997, 2010, 2015 and 2023—further argues for the case for a change between 2010 and 2015. The El Niños of 1997 and 2009 appear fairly close to the mean, whereas the later events show higher enthalpy. The fact that such high amplitude El Niños appear close to the mean reflects the dramatic change in enthalpy in the post-2010 period. In 2023, El Niño seemed unremarkable in the boreal spring in terms of its enthalpy signature but quickly established a positive tropical anomaly by the summer and transitioned to unprecedented amplitudes globally in the fall, moving into record-shattering territory7. The jump in September temperatures was very unlikely to have happened due to internal variability alone8. Nevertheless, by winter 2023, enthalpy was quite comparable to the three earlier El Niños as the year ended.

Fig. 3: Four strong El Niños by season.
figure 3

Looking at the zonal mean vertically integrated thermal energy by season for four of the strongest recent El Niños reveals that the year 2023 fell broadly within the envelope of other events in spring (a), then exceeded earlier values in summer (b) and autumn (c) and returned back to the fold in winter (d).

Full size image

The year 2023 represents the first sample of a warmer world that is not a model projection. This year stands out by a substantial margin in terms of its warmth while being unremarkable in terms of impacts. It is imperative that the climate science community participates to disentangle the reasons for this record warmth as well as its impacts (or lack thereof). While extreme event attribution to anthropogenic warming has now become nearly an operational activity (thanks to the World Weather Attribution group), this year is a prime example of the necessity to understand how the combination of global warming and natural variability shapes extremes. Furthermore, we need to understand better if this “plus-1.5 °C/year” represented the median case or was an outlier in terms of impacts.