Long after the winds subside and the hard work of rebuilding begins, scientists responsible for predicting hurricanes will continue to replay Hurricane Milton over and over again. They’ll focus on how they failed to predict Milton’s sudden strengthening.

Milton became one of most the powerful hurricanes on record in the Atlantic basin after its winds accelerated by 95 miles per hour (150 kilometers per hour) on Monday, a faster jump than any storm on record other than Hurricane Wilma in 2005. That mark is nearly three times the standard definition of what scientists call “rapid intensification.” It led the normally staid U.S. National Hurricane Center (NHC) to call Milton’s intensification “explosive.” 

Milton is “essentially off the charts,” said Christopher Rozoff, an atmospheric scientist at the National Center for Atmospheric Research. The storm gathered its strength from the Gulf of Mexico, which is abnormally warm for this time of year. It’s a jarring reminder that models don’t necessarily capture what’s becoming a signature of major hurricanes in a world 1.3C (2.3F) hotter than pre-industrial times.

The list of storms that have undergone extreme rapid intensification is deadly and full of multi-billion-dollar disasters. In 2005, Hurricane Katrina underwent two bouts of it, followed by Wilma just weeks later. The latter underwent “ultra-rapid strengthening” from a 70 mph tropical storm to a 185 mph Category 5 behemoth, according to a post-storm report. Hurricanes Maria and Dorian are also on the list as is Helene. 

Eighty percent of major hurricanes—defined as Category 3 and higher—go through this phase. Greenhouse gas pollution is making that more likely in the Atlantic and other cyclone basins by supercharging ocean heat. A snap analysis of Hurricane Helene, for example, found the storm was made 500 times more likely by the record-hot Gulf. That makes climate change an increasingly confounding factor for weather models used to forecast storms. 

“It is extremely difficult for numerical models to get the timing and magnitude of [rapid intensification] correct,” said Gus Alaka, director of the hurricane research division at the U.S. National Oceanic and Atmospheric Administration (NOAA). “But it's way better than it was a decade ago.”

NOAA has operated a program to create better hurricane forecasts since 2007, and the agency last year launched its Hurricane Analysis and Forecast System model, which has demonstrated an improved ability to predict intensification. One-day Atlantic storm intensity forecasts were off by about 11 mph in 1990 and less than 8 mph in 2022, with “a considerable decrease in error” coming since 2010, according to a recent NOAA review.

But storms like Milton show the limits of models, putting the need to improve rapid intensification forecasts ever higher on meteorologists’ to-do list. It’s no small task: By definition, rapid intensification happens over a very short period of time, and it’s driven by localized conditions in the ocean and atmosphere that can turn storms like Milton into monsters.

Understanding what’s going on where the sea meets the sky and how it transforms hurricanes from the inside out is crucial to predicting rapid intensification, said Colin Zarzycki, a hurricane researcher at Pennsylvania State University. “Even the best climate models that we have can't see those processes.”

A lack of observational data has also hampered researchers’ efforts. Reconnaissance data can improve hurricane forecasts by 10% to 20%, Alaka said. Until recently, it’s largely been left to the federal government, which gathers it from sources including satellites, buoys and hurricane-hunting planes.

Over the past few years, the NHC has started to pay a new wave of weather-monitoring startups and tech firms collecting information. “Saildrones”—autonomous robots that can operate in the maelstrom of a hurricane—are collecting wave heights and wind gusts. NOAA has tested or deployed marine drones from Black Swift Technologies, Anduril Industries and Dragoon Technology among other tools to see cyclones from the inside out. 

NOAA is also in talks with WindBorne Systems to gather data from the sky without the need for crewed flights, co-founder John Dean said. The startup deploys specialized weather balloons, including a new version that will be able to measure temperature, pressure, humidity and windspeed from inside hurricanes. By attaching sensors to balloons that can dive inside a storm and are easily controlled and retrieved, “we can gather vertical profiles inside the hurricane,” Dean said, which can provide models with data currently in short supply. 

Artificial intelligence could also help tip off forecasters faster that a storm is likely to intensify. The NHC has incorporated it into its workflow for several years, and researchers at the University of Wisconsin-Madison have had success training AI to spot markers for rapid intensification in storm imagery, said NCAR’s Rozoff.

Climate change is adding urgency. Zarzycki and a Penn State colleague and national lab researchers recently undertook a study to see how past hurricanes would unfold at different levels of warming. They found tropical storms were far more likely to become powerful hurricanes through rapid intensification, including ones that hadn’t done so in the past. “As the world warms, rapid intensification becomes more of the norm,” Zarzycki said, creating even more complications for forecasters.

Hurricane Milton’s abrupt escalation caused heartbreak and concern among atmospheric scientists this week as the storm careened toward Tampa Bay, one of the most vulnerable regions in the U.S. And it raised a new question that will drive research: Is the probability of extreme rapid intensification shifting? 

“I don't know the answer to that question,” Rozoff said, “but I think that's an important question that we should be asking.”

This article was provided by Bloomberg News.