A hurricane by any other name: How Sandy changed the way we issue storm warnings
The day before Superstorm Sandy struck the Northeast last fall, Sarah Romulo and her family left their home in Rockaway Beach, N.Y. They did not feel comfortable sitting tight, blocks from the boardwalk that faces the Atlantic — not with the warnings they were hearing and with their two kids, a 15-year-old and 2-year-old. With an evacuation order in place, they went to stay with Sarah’s mother-in-law in Queens Village, about 20 kilometers inland.
From there, Sandy felt like “just a little windstorm, it seemed like nothing,” Romulo recalls. But television news showed the unfolding devastation, as the storm made landfall on Oct. 30. “We felt so helpless, not knowing what was happening” with the family’s home and business, a martial arts gym a few blocks from their house. When they got back to their property a few days after the storm, they found that the basement in their house had flooded; meanwhile, their gym had been filled with at least two meters of water.
Despite the warnings, Romulo and her family had considered staying put — and many others did remain at home. Hurricane Irene, which struck in August 2011, had not been so bad. So it’s understandable that some people thought they might be able to ride out Sandy. Unfortunately, Sandy was about to become one of the most devastating storms in decades to strike the U.S., let alone the Northeast.
Somehow, the warnings issued by federal agencies that Sandy would be worse than Irene, worse than anything the Northeast had seen since 1962, seemingly were missed. What went wrong? And how can the responsible federal agencies avoid a repeat?
Early on, scientists at NOAA’s National Hurricane Center (NHC) in Miami, Fla., started tracking the atmospheric activity that would become Sandy. For several decades, weather monitoring, modeling and prediction have become more data-driven, increasing the accuracy of early predictions of a storm’s path and intensity. Instead of tracking a storm just two days before it makes landfall, scientists can start watching it weeks before.
About Oct. 11, Sandy began forming off the coast of Africa at the eastern edge of the Atlantic Basin. Models foreshadowed the westward track the air lens would likely take.
As it migrated, the storm gathered strength, eventually turning toward the Bahamas. By Oct. 25, Sandy passed over Cuba as a Category 3 hurricane. Slightly weakened to a Category 1, Sandy then headed north, up the U.S. East Coast.
While the storm made its way north, just offshore of land, it peaked, waned and intensified again. The storm also took on an odd shape, with winds most intense on its western lobe (usually a hurricane is strongest on the eastern side). As it approached Atlantic City, N.J., Sandy’s wind speeds dropped from more than 150 kilometers per hour (kph) as it passed over a cold patch of the Atlantic that hurried it toward land and cooled it down. The storm was still strong, with estimated 130 kph winds, but it lost its “tropical” characteristics. The scientific definition of a hurricane requires it to have sustained wind speeds of greater than 118 kph for a minute, and for it to meet other criteria such as drawing its heat from the ocean.
Several days before the storm’s last landfall, the NHC team could see that something unusual was about to happen with Sandy: The storm, while maintaining relatively high wind speeds and other hurricane-like characteristics, like the potential to create a storm surge, would morph into a post-tropical cyclone — not necessarily smaller than a hurricane, but a different beast altogether. In addition, there was the potential for the post-tropical storm to become enveloped in a nor’easter — a cyclone-like storm that typically forms along the East Coast, with namesake winds that blow from the northeast.
Sandy’s transition to a post-tropical storm wasn’t just a matter of meteorology, but had practical impacts: How a storm is classified determines which warnings are issued and by whom. The projected change in classification meant that the NHC would follow its own protocol, hammered out over the years with the rest of NOAA’s National Weather Service (NWS) and many external partners, including meteorological news services and local governments. Once the storm could no longer technically be called a hurricane, NHC would step back from the storm warnings system, according to rules laid out by NOAA, the center’s parent organization.
In order to keep the weather service’s warnings consistent throughout the event, NHC could not issue hurricane warnings for Sandy, says James L. Franklin, branch chief of the Hurricane Specialist Unit at the NHC in Miami. The NHC officials were “really, really afraid that at a crucial time, within hours of coming ashore, Sandy would have changed and hurricane warnings would have come down and people would think the threat is over,” Franklin says. So, instead, based on the center’s forecasts, local NWS offices issued warnings about the strength of the storm.
The NHC warnings come in predefined packages of words and relevant statistics composed in a strict format. Private companies and outfits like the Weather Channel and AccuWeather, as well as local TV meteorologists, can download the information and immediately plug it into their own computer programs in order to spit out maps, warnings, and forecasts for television broadcasts. Just as weather forecasts have improved, the entire weather reporting system has been metamorphosing into a sleek, data-driven computer “deliverable.”
Because of this strict structure, Franklin says, if the NHC were to have ignored protocol, the unexpected changes could have thrown a wrench in the “whole community of systems that interact in ways that we don’t control.” If the center had sent out messages under the banner of a “post-tropical storm” while calling for “hurricane warnings” in the same breath, “it might have worked, but we had no way to test it,” he says. “We are obligated to follow the rules that we told [weather reporters, local governments and emergency responders that] we will follow.”
Still, the agency considered for several days how to craft the warnings, Franklin says. For a moment, NHC officials discussed the consequences of continuing to call Post-Tropical Cyclone Sandy a hurricane. Gary Szatkowski, chief meteorologist-in-charge of the NWS office in Mt. Holly, N.J., says he was one of the people within the agency pushing the NHC to be consistent. He says he argued that it was better to avoid labeling Sandy a hurricane when it wasn’t going to be one, technically, once it made landfall. After all, accuracy is one of the main concerns for the agency: a false alarm — even if the storm were as big as a hurricane — would call the agency’s actions and future warnings into question.
In the end, NHC stuck with the NOAA protocol and went with the suite of warnings prescribed for nontropical storms, consistent with the NHC’s forecasts. But that choice meant that local meteorological offices, without the NHC’s hurricane expertise, determined what local warnings to issue; it meant that local government officials had to determine their own evacuation orders from the local meteorologists. There was no unified message, such as would have come from NHC if the team had retained control in a hurricane scenario. Every warning would have come from NHC, labeled as a hurricane from start to finish, had Sandy hit land as a hurricane.
“None of the choices were good,” says Adam Sobel, an applied physicist and mathematician at Columbia University who models climate dynamics and tropical meteorology, including hurricanes. “We have a National Hurricane Center that is supposed to respond to hurricanes, and local offices that have a different mandate and expertise. They don’t necessarily have the same experience in hurricanes. They can say different things, from two offices nearby.”
The final warnings from NHC about Sandy were still quite serious. They made it clear that the storm would be big, and included cautionary notes about potential record coastal flooding due to the storm coinciding with high tide. And Szatkowski, as the local NWS meteorological chief, included his own personal pleas in local reports he gave every morning and afternoon leading up to the storm (see sidebar). He practically begged people to evacuate from the coast if they were asked to do so: “If you are reluctant to evacuate, and you know someone who rode out the ‘62 storm on the barrier islands, ask them if they would do it again,” he said at the time. In March 1962, a powerful nor'easter flooded tens of thousands of homes on the East Coast, with waves higher than 7 meters reported along New Jersey’s shore, according to the U.S. Geological Survey.
But even Szatkowski’s strongly worded message wasn’t a hurricane warning. “We certainly got pushback from our media partners,” he says, for not providing a full “hurricane” warning. One of the most vocal critics was Bryan Norcross of the Weather Channel. The NHC issued “an accurate portrayal of an extreme threat,” says Norcross, the cable network’s hurricane specialist. That threat included the possibility of a surge that was over the threshold of danger for New York City. But at the time he felt the warnings didn’t go far enough, he says.
Norcross, who expressed his ire with the lack of communication during and after the storm, says he acknowledges that some people got the message. Meteorologists in the affected states managed to make clear to some emergency responders and decision-makers that this was going to be a big storm, Norcross says, including Szatkowski and those advising New Jersey Gov. Chris Christie. But somehow, others missed the point. Norcross tells anecdotes about emergency responders trapped atop their own ambulance as the flooding rushed inland. He also says he was less than impressed with New York City’s response to the flood risk, noting that Mayor Michael Bloomberg held a lackadaisical press conference two days before Sandy’s arrival, then issued last-minute evacuation orders as the storm bore down on the city.
“The public messages were very garbled on the level of threat,” Norcross says. He says he thought at the time that NOAA could have let the NHC break protocol: to keep on with the hurricane warnings, even though the storm had become a post-tropical storm wrapped in a nor’easter. But he says he understands where the agency was coming from in its decision to follow protocol. “This whole circumstance of whether [NOAA] should issue a hurricane warning just hasn’t come up in this way in a modern system,” he says.
Receiving the message
Although the message was garbled, the messengers weren’t entirely to blame for what happened. For one thing, Sandy was an unusual storm: Mixed storms like this one usually head out to sea before making landfall, but a high-pressure system acted as a wall that pushed it back to shore, according to NOAA’s final report released earlier this year. The last time this happened, when the physical characteristics of a storm treaded the scientific distinction between hurricane and tropical storm, was the so-called “Perfect Storm” of 1991 (an extra-tropical cyclone wrapped in a nor’easter). But even that event took place mostly at sea. “Storms can lie on the border of a scientific distinction. It probably never has mattered as much as it did in this case,” Sobel says.
Human nature also played a role in augmenting the storm’s devastation. Sobel and others suspect that one reason people ignored some of the warnings about Sandy is because the storm was not labeled a “hurricane.” A hurricane comes with certain expectations. People don’t necessarily know what to expect with a storm by any other name. People also may have had lowered expectations of Sandy because of Hurricane Irene’s mild performance the year before.
Irene was a Category 1 storm when it first struck in North Carolina. As it moved north, the strongest winds stayed offshore, and the winds that struck New Jersey peaked at 55 kph, as Irene blanketed the Northeast with cloud cover nearly 1,000 kilometers wide. But the storm had relatively minor repercussions at the coasts: Irene came ashore at low tide, when its additional meter or two of storm surge would be equivalent to a normal high tide. Instead, the storm brought rain and snow that downed trees and triggered catastrophic flooding inland.
Sandy, however, was even bigger — it stretched 1,600 kilometers across, a swirling behemoth of tropical storm-force winds. The storm also hit at high tide in New York, adding 2 to 3.5 meters to already high water levels, which led to severe inundation at the coasts. Even though the NHC made one of the best storm track and arrival time predictions in its history, Szatkowski and others say, a few hours’ difference in Sandy’s arrival could have meant much less flooding at the coast — equivalent to Hurricane Irene.
Romulo recalls preparing for Irene: “We brought everything of value up to the first floor — all our kids’ clothes, electronics …. We sandbagged the heck out of the gym,” she says. So when nothing happened with Irene, the family’s response to Sandy was less than wholehearted. They decided to leave everything in the basement. They put the gym’s mats about a meter off the ground (the water levels would reach 1.8 meters). “When we left for Sandy, my son just took his Xbox with him, just because he wanted something to do” while weathering the storm at his grandmother’s home, Romulo says.
According to telephone surveys by researchers for the Wharton School of Business at the University of Pennsylvania — conducted as Sandy worked its way north up the coast — more than 90 percent of respondents made some kind of preparations for the storm, whether they heard “hurricane,” “nor’easter,” or “big storm” in the forecasts. But these preparations only amounted to activities such as filling their cars’ gas tanks or buying extra water. Only half of the people who owned storm shutters actually put them up, and only a fifth of the respondents living in areas under evacuation orders said that they planned to go.
The Wharton report, released online (pdf) as a working paper Nov. 20, also found that the majority of people consistently overestimated the wind speeds of the storm, expecting most damage to come from wind. Very few thought that flooding would be the most damaging aspect of Sandy. The researchers suggest from responses to their survey that people were hearing a repeat of “Hurricane Irene”: about a third of those surveyed experienced that storm, and that group expressed the least concern over possible damages. The authors say they were surprised to find that those who had never experienced a hurricane expressed the most worry. Overall, 75 to 85 percent of respondents said they felt safe staying in their homes through the storm, says Ben Orlove, a co-author of the report.
Orlove, a professor at the Columbia University School of International and Public Affairs and co-director of the Center for Research on Environmental Decisions, says multiple issues feed into how people respond to warnings: Too much warning too early could lead to “storm fatigue,” when people tune out the threat. False alarms and memories of past events might lead people to tone down their responses. Negative personal experiences may cause them to go overboard or to not react.
People are not good at calculating risks, he says, especially for events that occur rarely. But such big, damaging storms will only get worse in the future, Orlove and Sobel point out, as more infrastructure is built on the coasts and as sea level rises. Communication has to be clear to make sure people are safe in the future, they say. “What happens in the past informs people’s perceptions of what can happen in the present and the future,” Sobel says. People need to be shown what the hazards could be.
Hindsight Is 20/20
Knowing that there will be a next time, in the wake of Superstorm Sandy, NOAA changed its practices on hurricane warnings. In April, in time for the 2013 Atlantic hurricane season, NOAA agreed to allow the NHC to go off script if necessary.
Where the older protocol would have been to hand off the messaging to local bureaus, Franklin says, now NOAA has broadened the category of storms for which hurricane warnings can be used. Plus, the NHC will maintain its jurisdiction over these storms and continue writing advisories. “That’s something we couldn’t do during Sandy,” Franklin says.
Sobel lauds the move. NOAA and the NHC “correctly perceive that they have tremendous authority. They have to be steady at the wheel” because a lack of consistency would “erode the trust they have built up,” he says. However, while NOAA and the NHC will now have control over the message for the next big storm, Sobel says, the next step is to improve communication of that message. For example, he says, they are rightly recognizing the need to emphasize storm surge and flooding forecasts, not just wind.
Jamie Rhome, head of the storm surge team at NHC, says that the agency is speeding up the release of new maps that will show probabilities of flooding during a storm as clearly as possible. Such maps are set to be released for individual storms starting in 2014, Rhome says, but the NHC might introduce these color-coded topography-linked graphics this year if possible. The surge maps, tested with the help of social scientists, show neighborhoods and homeowners just how high the water might be locally with flooding.
Sobel says he thinks this step, in addition to NOAA’s decision to let the NHC keep control of the message in the future for these kinds of “superstorms,” will help make the risks clearer to people.
“What we’re all learning in this profession is that communication is a tricky thing,” Sobel says. Scientists in general can’t simply gather the data and “dump it over the fence” to be interpreted by the public, he says.
The final report on Sandy, issued by NOAA in May, also recommended a more coordinated online presence: “For future storms like Sandy, NHC should be the principal point of contact responsible for the event, including delivery of a consistent suite of products and a unified communications protocol within NOAA, to key NOAA federal partners and the media.” The final recommendation calls for all NOAA and NWS websites to be up-to-date and well designed to “ensure the most important message is quickly evident.”
The Next Big Storm
Meanwhile, in Rockaway Beach, recovery continues. Sarah and Chris Romulo bought a new house (complete with the required flood insurance, she says), and they found a new gym space, a few blocks from the old one; the old gym space has now been turned into a restaurant. Other new businesses have also opened up as people have returned to the Rockaways. The subway line to the communities there was finally fully functional again in late May, following seven months of limited service and $650 million in repairs.
Yet last year’s damage and memories from Sandy linger. In May, city planners anticipating the coming hurricane season installed huge sand-filled plastic bags to act as berms to protect homes near the beach.
Romulo notes the sand berms with approval, and she voices the cautious hope that the family’s new home — which sits above the projected flooding level for a major storm surge — will be high enough the next time a Sandy-like storm hits. “Maybe I’m just in denial,” she says, “but I can’t believe that anything on that level can happen for a while.”