Scientists find new technique for predicting record-breaking heat waves

In June 2021, temperatures in the Pacific Northwest reached as high as 121 degrees Fahrenheit and were blamed for hundreds of deaths. Many people, including scientists, were surprised by the intensity of the heat, which broke temperature records by a whopping 8 degrees Fahrenheit in some areas. At the time, that level of heat was thought to be beyond the realm of possibility, even by climate models.

Now, research released Tuesday shows that such record-shattering heat waves - dubbed virtually impossible without climate change - can be predicted by existing climate models. The study, which models extreme scenarios for various cities, unlocks a new technique that could help communities better prepare for life-threatening heat waves in the future.

“This is a proof of concept that existing climate models can deliver useful information about the range of possible extreme weather events,” said David Ham, a geoscientific model developer at Imperial College London who was not involved in the study. “Previously, it would not have been believed that they could do so.”

The new research comes as heat waves sweep across the central United States and Europe, including in Chicago and Paris - two cities coincidentally modeled in the study. Current weather forecasts from the National Weather Service showed peak heat indexes in Chicago may reach between 105 and 115 degrees on Wednesday, potentially setting new highs in surrounding locations. Temperatures around Paris already matched or surpassed some all-time records on Monday, according to the French national meteorological service.

Temperatures are record-breaking and life-threatening in areas, and simulations in the study show that Chicago and Paris could be poised to experience even more extreme heat waves - sometimes 2 to 10 degrees Fahrenheit warmer than observed - in the future under slightly different atmospheric conditions. Ham said the study is not necessarily making predictions, but it shows what this new technique could discover.

“The longer we measure, we should expect to see less and less records because we should have seen almost everything that’s possible. But with climate change, that’s quite different,” said Erich Fischer, a climate scientist at ETH Zurich and lead author of the study. “We see a lot more heat records [than] we should be expecting to see by chance, and these records are broken by larger margins.”

The goal of the work, Fischer said, isn’t to fearmonger but to help communities prepare for impending extreme disasters: encourage people to stay out of the heat, fortify energy systems, set up cooling centers, and check on the elderly and other vulnerable populations.

“We don’t want to be alarmist here,” said Fischer. “I hope that people start to use such approaches for planning.”

Fischer explained that a warming world boosts heat waves in a few ways. For one, weather patterns are starting at a warmer baseline, as global temperatures have risen roughly 2 degrees Fahrenheit since the preindustrial era. High-pressure systems responsible for the heat waves, sometimes called heat domes, can also suck in warmer air. Drier land also decreases evaporation and allows for more direct heating of the air and ground.

“The exact same weather system is happening in a substantially warmer background climate,” said Fischer. “Warm air that is sucked in is also warmer. The land certainly is hotter, so the whole thing can actually amplify.”

In the new study, Fischer and his colleagues set out to predict large record-breaking heat waves using current climate models. The team modeled 1,000 years of today’s climate to simulate the Pacific Northwest heat wave and found very intense heat waves occurred, but the heat waves still weren’t as extreme as observed - not terribly surprising.

“When we run climate models, we get snapshots of how the weather could play out in the future, but limited computing resources means we only get a limited set of snapshots,” Ham said. “This means that we might well never see in the models the most extreme weather that we will experience in reality.”

Fischer said running more simulations “may become really expensive,” so they tried a different focused approach to push the climate model to its limits. They picked the worst-case scenario in the previous model runs. Then, they restarted the model two to three weeks before the worst-case scenario occurred to produce thousands of outputs. Not only were they able to reproduce the extreme scenarios, but they produced events that were even hotter.

Fischer said that, by comparing the various scenarios, they are also able to understand the drivers that make one event more extreme than another.

Perhaps shockingly, it didn’t take much to push heat events to record territory. With nothing more than a butterfly effect, just a tiny random perturbation to the initial atmospheric conditions, they were able to match or exceed the Pacific Northwest heat wave, he said.

“It just means that by chance, things have could have turned out even worse,” Fischer said.

Ham said this approach is not a finished technique, but the research “is key in demonstrating that such predictions are possible and showing at least part of how they could be done.” But, he said, by choosing which simulations to run to demonstrate what weather is possible, it’s also harder to say how likely that weather is. Instead, the authors can just label an event as unlikely, or a 1,000-year or even 10,000-year event.

Fischer said this study is just a starting point and they welcome other researchers’ perspectives and modeling techniques to ascertain these new climate-fueled heat waves.