Record-breaking heatwaves, like the one experienced in western North America in June 2021, cause severe societal and environmental impacts. Current adaptation strategies often focus on events as extreme as those already experienced, neglecting the potential for unprecedented events. This study addresses the need for proactive planning by identifying regions globally that may be particularly vulnerable to future high-impact heatwaves due to a lack of historical extreme events. The research focuses on regions that haven't experienced high-temperature extremes, making them potentially unprepared for such events. This vulnerability is exacerbated by factors like growing populations and limited resources for healthcare and energy. The study's significance lies in its ability to inform policymakers and governments to prepare for events exceeding current records, considering the escalating probability of extremes due to anthropogenic climate change. Understanding the likelihood of extreme heat events is crucial for effective preparation, but the rarity of these events makes assessment challenging. This necessitates the use of statistical methods to estimate return periods, accounting for the limitations of the observational record and potential biases.

Existing research highlights the devastating effects of heatwaves and the importance of preparedness. Studies show that improved preparation can significantly reduce heatwave mortality. Examples include the reduction in deaths after the 2006 European heatwave, compared to the 2003 event, thanks to policy changes implemented after the earlier event. Similarly, improved humanitarian response plans in Bangladesh reduced mortality from Cyclone Amphan in 2020. Several methods exist for measuring heat extremes, including annual maximum daily maximum temperature (TXX), which this study employs due to its global applicability and recommendation by the World Meteorological Organization (WMO). Alternative metrics focus on multiple days above a threshold or percentile, or integrate temperature and humidity. However, the choice of metric influences the results, and the focus on TXX allows for a global-scale comparative analysis. Existing research also utilizes extreme value theory (EVT) for assessing meteorological extremes, accounting for non-stationarity in climate data, though the relationship between regional extremes and global mean surface temperature is not always linear.

The study uses extreme value theory (EVT) to assess the return periods of observed temperature extremes globally. The analysis begins with an example of the technique applied to the western North American heatwave of June 2021, which demonstrated an event beyond the statistical maximum. This event, with temperatures reaching 49.6 °C in Lytton, British Columbia, was shown to be virtually impossible without climate change. For the global assessment, the study uses ERA5 and JRA55 reanalysis datasets, focusing on regions where the two datasets show consistency (1990-2022). The annual maximum daily maximum temperature (TXX) is used as the extreme value metric, and the GEV fit is calculated by excluding only the record year from the dataset to obtain a more robust measure. The statistical maximum is defined as the magnitude of a 1-in-10,000-year event based on the GEV fit excluding the record. This is compared to the 1959-2021 record for each region, calculating the return period where possible. Uncertainties in reanalysis data are addressed by comparing ERA5 and JRA55, using only regions exhibiting consistent extremes in the satellite era (1990 onward). To further validate findings, the study analyzes two large ensembles of global climate models, CanESM5 and MIROC6, each with 50 realizations. GEV fits are applied to each ensemble member individually and by merging all members into a single distribution. The predefined 0.5 Mm² regions from Stone (2019) are used, excluding areas with inconsistent reanalysis data and the Antarctic, resulting in 136 regions for analysis. The study uses bootstrapping for uncertainty estimation in the GEV fit.

The study found that the 2021 western North American heatwave was exceptionally extreme, nearly 2°C beyond any other region. However, the analysis revealed that in 31% (41 of 136) of the regions examined, the current temperature record is statistically implausible based on the GEV fit, meaning these events were highly unlikely based on the pre-existing data. These implausible regions are geographically diverse, suggesting that such extremes could occur anywhere. The analysis of climate model data (CanESM5 and MIROC6) found a similar pattern: 18–26% of regions exhibited implausible extremes in at least one model realization. Regions with shorter return periods for their current record are highlighted in Table 1. These regions, including parts of Russia, Central America, Afghanistan, and others, are identified as being at greater risk of future record-breaking heatwaves. These regions are of particular concern due to factors beyond the statistical likelihood of an event, such as socio-economic factors, like limited resources and rapidly growing populations. The study highlights these regions as especially vulnerable, combining high statistical risk with a rapidly increasing population. While the study shows an increase in the number of exceptional extreme events in more recent decades, it also identifies such events throughout the reanalysis period, suggesting that changing dynamics don’t fully explain the occurrence of these outliers.

The findings highlight the vulnerability of regions that haven't experienced extreme heat events, which may have led to insufficient preparedness. The study's results show that statistically implausible extremes have occurred in a substantial portion of regions, with no clear spatial or temporal pattern. Climate model data support this, indicating that such events could occur anywhere. The June 2021 heatwave stands out as exceptionally extreme, but it is not unique. The vulnerability of a region depends not only on the likelihood of a record-breaking event but also on socioeconomic factors. Developing countries with limited resources and rapidly growing populations are particularly susceptible. While the study focuses on the TXX metric, the use of different metrics could yield variations in results, as different metrics are more sensitive to different impacts. While the study identifies patterns, it doesn’t fully explain the underlying causes of the observed extremes, emphasizing the need for additional research to improve our understanding of heatwave dynamics.

The study identifies regions with a higher statistical likelihood of experiencing record-breaking heatwaves due to the lack of extreme events in the past. Socioeconomic factors further amplify their vulnerability. Both observational and model data support the conclusion that unprecedented heatwaves could occur globally. Future research should focus on refining methods for estimating extreme plausible events, investigating the physical mechanisms driving these events, and improving the consistency of reanalysis data, particularly in data-scarce regions. This necessitates a more comprehensive understanding of heatwave dynamics and the interaction between climate and socioeconomic factors to enhance preparedness and adaptation strategies.

The study acknowledges limitations in the data, including uncertainties in reanalysis products, particularly in data-scarce regions like parts of Africa. The analysis relies primarily on the TXX metric, potentially overlooking the impacts of different heatwave metrics. The study also acknowledges the inherent limitations of EVT in extrapolating beyond the observed data, making the estimation of the most extreme events uncertain.