When Justin Smith checked the weather on Tuesday, he wasn’t worried. The report from WINK News on DirecTV said Hurricane Ian was a threat to Tampa, some 100 miles to the north of Fort Myers Beach, where he was honeymooning with his wife, Karinna Smith.
“We were there on our honeymoon,” Smith said. “We were trying to have a good time, not be glued to the TV watching the weather.”
But that meant they missed messages that much of southwest Florida, not just Tampa, was under threat of hurricane conditions as Ian approached. Their first warning that the hurricane was making a turn directly toward Fort Myers Beach was a note posted in the empty hotel lobby Tuesday night, indicating that the hotel was being evacuated.
As conditions became catastrophic on Wednesday, the Smiths survived by taking refuge in a hotel stairwell.
The danger faced by the couple and many others who did not evacuate underscores the challenges of communicating forecasts for storms like Ian. Research shows people often cling to an initial version of forecasts, often missing key updates and changing threats. And meteorologists can struggle to convey the uncertainty in their predictions of a storm’s path and potential, in part because hurricane forecast cones and other tools of communication aren’t as useful as they could be for storms like Ian, whose track toward Florida was difficult to pin down even a day ahead of landfall.
The forecast uncertainty even challenged officials charged with making evacuation decisions in Lee County, where Ian came ashore. They now face questions as to whether they delayed too long.
The National Oceanic and Atmospheric Administration is working to improve communication of uncertain, fast-changing threats, but the task is not straightforward. It involves refining messages and optimizing graphical information to simplify the complex for very diverse audiences and keeping them apprised of important changes. The latter was critical in the case of Ian since small deviations in the predicted track would significantly alter which communities would be affected.
“There were a lot of alternate futures that were possible,” said Kim Klockow-McClain, a research scientist at NOAA’s National Severe Storms Laboratory. “Communication is not as simple when there are a lot of possible outcomes.”
Asked to review how well it communicated storm risks and uncertainties with Ian, National Hurricane Center officials deferred comment to NOAA risk communication experts.
Gina Eosco, a program manager and social scientist at the agency, said it can be a challenge for forecasters to overcome what she called “optimistic bias,” when people focus too much on early forecasts suggesting low risks of storm impacts, and miss updates signaling new and changing hazards.
“It can trick your brain into thinking you can relax and you may not pay as much attention to the forecast,” Eosco said. “And so it’s possible people didn’t hear that message.”
Along with that bias toward optimism, past experiences when dire forecasts did not come to pass can also prevent people from properly interpreting the realities of forecasts and storm conditions, she added.
Laura Myers, a senior research scientist at the University of Alabama who studies communication around disasters, applauded the Hurricane Center’s work tracking the storm and broadcasting its threats. And she echoed Eosco, saying people often “anchor” their expectations to early forecasts and then are caught unaware when storm predictions change.
“[Forecasters] know that people are going to cling to that and then walk away and not come back to the information,” Myers said. People form their own “scale of risk aversion” and go back to their daily lives despite broadcast meteorologists encouraging them to check back in for updates, according to Myers.
“If they are shocked about impact, it’s because they anchored,” she said.
That said, the Hurricane Center’s archive of Ian forecasts shows that, as its predictions of the storm’s path shifted, meteorologists did not begin to emphasize risks to the area around the eventual landfall point until about a day in advance.
It wasn’t until Tuesday morning, while Ian was passing over western Cuba, that the Hurricane Center extended a hurricane warning southward to cover the stretch of southwestern Florida coastline that would soon be devastated. Even then, the centerline for the predicted storm track passed through Tampa and wasn’t over Fort Myers until 11 p.m. that night.
In the days before that, what would become ground zero for Ian’s devastation was at the edge of areas the Hurricane Center warned were in the storm’s path. Areas to its south, including Naples — which endured a record ocean surge — were left out.
That meant some, like the Smiths, were caught unaware by Ian’s intensity. Smith said he received none of the National Weather Service text alerts that are supposed to broadcast imminent hazards to any cellphones in their path. And he said he got no alarm from hotel staff.
“They didn’t knock on our door,” Justin Smith told The Washington Post. “They didn’t call that room. They didn’t do anything. By the time that we found out we didn’t have a rental car or anything like that, so we were kind of stuck.”
While some meteorologists suggested it was a failure of the Hurricane Center, others stressed that that represents a misunderstanding of what the forecast cone actually means. There is a 60 to 70 percent chance a storm’s eye will remain within the cone’s boundaries — meaning in about one out of three cases, the storm will move outside of the cone.
The problem is that the forecast cone is not well-designed for unpredictable storms like Ian, Klockow-McClain said. The width of the cone is based on the Hurricane Center’s past error in storm forecast track predictions, but with Ian, that meant an underestimation of potential error.
“The problem is, with that graphic, we’re communicating about how we’ve done in the past. We’re not saying a whole lot about the uncertainty of the current situation,” Klockow-McClain said.
Researchers say the challenge is engaging with the public so that people understand the broader potential for hurricane impacts even outside the forecast cone. Both Eosco and Myers suggested that more localized warnings could better help people evaluate their personal risk.
“Our research has indicated that most people appreciate the worst-case scenario,” Myers said. “They appreciate knowing there is a chance that they would be included in the impacts and what the impacts might be.”
Myers said more should be done to educate the public about hurricane meteorology and risks outside of active weather events, so that when storms strike, they aren’t overwhelmed with too much information.
“If you don’t do that in advance and don’t do it in as many different ways as you possibly can, you’re going to have issues with understanding,” Myers said. Even then, it is not guaranteed that the warnings will be properly interpreted.
NOAA has invested heavily in efforts to narrow gaps in communication and improve public perception and understanding of forecasts, Eosco said. Before and after storms, NOAA conducts a multi-wave project to increase their understanding of how well people understood risks and what actions they took during a hurricane, she said.
And that goes along with Hurricane Center work to improve graphics and messaging around hurricane risks, including adjusting forecast cone images to include wind field sizes so people understand how far dangerous conditions will extend. And the Center has moved to stress that forecast cones are fallible, and that risks extend throughout them and beyond them, though that message doesn’t always get across to the wider public.
“I’m thrilled that if there has to be a hurricane, that we have the ability to learn something from it so that we should improve our communication for future storms,” Eosco said. “If we can find an opportunity of hope here to learn something from it so we can improve such situations and reduce societal impact, that is the type of opportunity NOAA wants to take.”
Meena Venkataramanan contributed to this report.