Better Prepared? What Meteorologists Learned from Hurricane Sandy
National Weather Service makes number of changes, but challenges remain in improving forecasts and communication with public
October 29, 2012 seemed like just another day for Ron Krauss and his family. The Halloween decorations were up at their Union Beach home, and though Sandy was approaching, Krauss wasn’t terribly worried.
“We said we’re going to stay because nobody was knocking on our door to let us know to leave,” he recalled. “There were warnings, but nobody said ‘get out.’ So after the morning storm came through and we survived it, we said, ‘All right, I guess it was OK,’ not realizing that the worst was yet to come.”
They cooked dinner and were getting ready to go to bed, when Krauss’s son noticed water on the floor of their attached garage. Within seconds, a powerful surge came rushing in from the Raritan Bay across the street, carrying away their car, while the family ran for safety. Then waves crashed through one of their walls and their front door. Krauss spent several hours furiously trying to keep the water out.
“It was like the movie 'Titanic' when the boat goes down,” he said. “We were the Titanic! It was just flight or fight, and we luckily won the fight.”
Not everyone was as lucky. By the end of the storm, all 22 other houses on his block were destroyed, as were 40 more elsewhere in town.
As the largest hurricane on record when it made landfall two years ago, Sandy presented a variety of challenges for forecasters and emergency management officials. Some have since been addressed, but many others remain. Communication problems meant that residents didn’t understand the true nature of the threat or how it would affect them; warnings were misinterpreted; and people failed to take them seriously. Above all, there was a pervasive mistrust of the predictions, with many people falsely believing that the storm couldn’t be as bad as forecasts suggested, since they had never experienced anything of that magnitude. What’s more, for all the work forecasters have done to build credibility, incidents like the false alarm about January’s monster blizzard have been greatly damaging to those efforts.
In the aftermath of Sandy, the National Weather Service has taken steps to rectify some of these problems, starting with an extensive, internal analysis of its performance during the storm. It’s announced a number of policy changes to help ensure that it’s more effective reaching vulnerable populations in advance of storms.
But, as does the challenge of restoring public trust.
All sides seem to agree that priority No. 1 is better communication, but that’s complicated by the fact that predicting the weather is an imprecise science.
Government forecasters use several different, often conflicting, computer models to determine the odds of what will happen in any given storm. But while the data provides useful insights, there’s also an element of human interpretation. During Sandy, a European forecast model ended up correctly predicting the storm’s path while other models failed. Still, when meteorologists put their trust in that same computer system a few weeks ago and predicted a record-breaking blizzard for the region, the storm ended up changing course at the last minute. And they ended up with what they later referred to as a “big forecast miss.”
In the fallout that followed, forecasters were criticized for overhyping the situation, and pundits said they should do a better job conveying the inherent uncertainty of their predictions.
“I think that’s a great idea, and my comment would be, ‘But we’re doing that,’” said, head meteorologist at the National Weather Service’s Mount Holly forecast office. While unable to comment on the actions of private-sector forecasters like local TV stations that may have exaggerated threats from the storm, he noted that the Weather Service already puts the probabilities of its predictions on its website, including best- and worst-case scenarios such as potential snowfall ranges.
He acknowledged that “we need to do a better job of getting that information out there,” but he also pointed to human nature: Many people get excited by forecasts of major storms but tend not to read or listen past the headlines. Fueled by posts on social media, they often fail to fully comprehend the complexity of a forecast, hearing instead an oversimplified version of what is likely to occur.
With Sandy, for example, many residents trusted their personal, often flawed instincts rather than listening to the experts.
Many municipalities had issued mandatory evacuation orders, and officials including Gov. Chris Christie and President Barack Obama were all warning that this was a serious storm, but a great number of people instead placed their trust in the advice of those around them, who told them things would probably be just fine.
“I checked with my neighbors who are lifelong residents. And they’re like, ‘You have nothing to worry about, cause the water never comes up this high!’” said Brigid Conway, who lives on the other side of Union Beach from Ron Krauss.
To Marshall Moss -- a senior meteorologist at Accuweather, a commercial forecasting firm -- it’s no surprise that many people base their reaction to severe weather on their personal experiences of past storms. He said he and his colleagues faced an uphill battle trying to warn the public about Sandy, since many residents were still thinking about Hurricane Irene, which had largely spared the region the year before.
“It’s incumbent on us in the meteorological community to make sure we’re clear in our messaging,” he said, “that we’re accurate and not crying wolf, not issuing forecasts that are too scary when the case isn’t calling for it, but certainly getting the message out when it needs to be.
That’s a difficult balancing act, though, and some critics think the messaging wasn’t as clear as it should have been.
“What we saw in reports that came out after Sandy was the overwhelming majority of people in the impact zone of this storm had no idea what was actually coming at them,” said Kathryn Miles, the author of, an hour-by-hour account of the events leading up to the storm making landfall.
“Some people thought it was a hurricane. Some people thought it was a nor’easter. Some people thought it was a gale,” she said. “I think that it’s pretty easy to tie a direct correlation between that sort of unknowability and the overwhelming occurrence that we saw of people who were disregarding and disobeying evacuation orders.”
One survey taken after the storm found that only a third of coastal residents in New Jersey chose to leave for higher and safer ground. Part of that reluctance may have stemmed from confusion as the storm weakened slightly from a hurricane to a post-tropical storm before making landfall. Government forecasters followed the procedures they’d always followed, calling the storm by its scientific classification, but they realized that could be problematic for some laypeople.
“When we do these label changes … when we switch something from being a hurricane to no longer a hurricane and the National Hurricane Center stops issuing advisories on it, there’s a whole bunch of beliefs that occur that people are operating under that have almost no basis in reality,” explained Szatkowski, the Weather Service forecaster.
For instance, he said, “There’s a belief that the storm has somehow dropped significantly in intensity, but it may have changed very little. You may only be talking about a change in wind speed of five miles per hour, and calling something a hurricane or saying, ‘OK, it’s no longer a hurricane’ is really kind of a technical discussion. But when you tell someone a really bad coastal storm is coming, there’s one reaction. If you tell that same person a hurricane’s coming, there could be a very different reaction, even if when we say it, as scientists, we’re basically trying to tell them about the same thing.”
Over at Accuweather, Marshall Moss and his colleagues made the decision to continue calling Sandy a hurricane even after it was downgraded because “that’s what grabs people’s attention, and that’s what they act to,” he said. “Whether it was truly a warm-core hurricane at the time of landfall or not, you’re splitting hairs. It was a close call. But the public good was better served by using the terminology that would get the action in that circumstance.”
After some soul-searching in the aftermath of the storm, the National Weather Service has instituted a new policy. In the future, to minimize confusion among people who rely on their forecasts, they’ve decided they’ll keep hurricane warnings in effect for dangerous storms, even if they’re technically no longer hurricanes.
The Weather Service has also created storm-surge maps so people in places like Union Beach will have a better sense of the risks they face. And they’ve committed to using more easily understandable language with less scientific terminology. So instead of talking about warm and cold fronts, they’ll focus more on conveying the real-life threats a storm will pose and how it might affect the morning commute, for instance.
To Szatkowski, the change in emphasis addresses a significant problem.
“I think when we said ‘record coastal flooding,’ and ‘eight or 10 feet of storm surge,’ people had trouble visualizing that” he explained. “They had trouble understanding that meant they’re going to get three feet of water in their business or their home and what they needed to do about it … If something goes beyond your life experience, a lot of people are going to struggle with that. A lot of people are going to have a hard time first of all believing it and secondly kind of getting their head around … what does that mean? So do I need to evacuate? How do I protect my property?”
Going forward, he said the biggest lesson for residents is to not get too caught up in the technical details.
“If five feet of storm surge is going to come into your building, don’t worry if it’s being caused by a hurricane, a nor’easter or the Martians,” he said. “Don’t worry about the cause. Worry about the impact. Worry about how bad it’s going to be, when it’s coming, what the timing is, and then take the right action. Make the right decisions based on that. That’s the important thing to decide.”
The challenge, said Miles, is conveying information to the public in the most effective and accessible way possible, but also in a way that’s convincing and that they can fully understand early enough to act.
“What social scientists tell us,” she said, “is that for the average person, it takes about three days to really internalize an evacuation threat such that they’re willing to take action on it and find somewhere for grandma to go or their dog or their cat to go and pick up their families and move and all of the other things that are at stake when they leave their homes. That takes a lot of time.”
While she thinks the recent changes in forecasting protocols sound promising, she’s still worried about the difficulty in connecting with a population that tells pollsters they’re more concerned about distant threats like terrorism and nuclear war than they are about natural disasters, which present a real and present danger.
In addition, Miles is worried about a variety of problems that still have not been fixed, including a legacy of budget cuts to the Weather Service, broken satellite equipment, and outdated computer systems that make forecasters’ jobs difficult.
“We have seen a slight increase in funding for things like hurricane prediction,” she said. “That’s really great. But until we address this larger issue, this larger crisis of staffing, this larger crisis of technology, of network security, we’re really just kind of putting Band-Aids on a much bigger sort of cancerous problem. And I think that’s something as a nation we have to address immediately.”