This is what to tell people to get them to evacuate before a hurricane hits

Hurricane Irma was still several days away from South Florida when officials in Miami-Dade began telling about 650,000 people to leave their homes. “If you do it later, you may be caught in a flood of traffic trying to leave the area,” Mayor Carlos A Gimenez said on Wednesday night. “You may find yourself in a car during a hurricane, which is not the best place to be.” Tens of thousands are expected to remain, though, either because they don’t have the resources or they think they can weather the storm. In storm after storm, people either can’t or don’t follow evacuation orders.

Florida is no stranger to hurricanes. But Irma is the first major one to strike the state directly in more than a decade. We study communication about storms and other disasters. Unfortunately, the best way to tell the public how to stay safe in a storm is not always clear – each disaster is different, forcing different tactics for different circumstances. That means there’s no playbook for emergency managers to follow, and that leaves them with constant decisions to make as a storm approaches.

But there are a few things that we know: appeals to fear can work, and authorities often use them ahead of severe storms, but they can backfire if the risks seem overhyped. Evacuation orders that identify a geographic location, such as flood zones, can help target messages to those facing the greatest peril. Specific risks must be communicated clearly, but ultimately, it’s how individual residents and their households perceive the danger that largely determines whether they evacuate.

One of the tenets of good risk communication – that authorities should speak with one voice – was notably missing before Harvey, when Texas Governor Greg Abbott told Houstonians to evacuate while Mayor Sylvester Turner was simultaneously telling them to shelter in place. Warnings about Irma went more smoothly. Monroe County, Florida, which includes the Florida Keys and Miami-Dade County, began ordering mandatory evacuations on Wednesday based on forecasts of Irma’s track. Making an evacuation “mandatory” can help convince more people to evacuate overall.

But our experimental survey research conducted after Hurricane Sandy has found that mandatory evacuations may also lead more people outside the intended evacuation zone to flee, with the potential to create additional traffic and fill up beds in shelters unnecessarily. On a peninsula such as Florida, where a large population may need to evacuate in the same direction, this is especially dangerous. Even worse is when people must leave through other at-risk areas, as in Florida – where the storm is likely to traverse the entire state.

In the same study, we also found that messages that identify certain geographic areas, such as flood zones or evacuation zones, can help encourage residents to leave those areas while reducing evacuation among those who live outside those areas. This depends, however, on whether people think they live in a flood or evacuation zone – and sometimes, people are wrong. To be effective, messages must use wording that helps people recognise whether the message applies to them. (One example is the mandatory evacuation of zones A and B in Miami-Dade County that was issued on Thursday.) Research shows that many people pay attention to evacuation orders. But first they have to know that they live in areas at risk.

Another approach to motivating evacuation is to use “fear appeal” messages, such as the National Weather Service’s warnings that people who stayed behind during Hurricanes Katrina and Ike would “face certain death.” As Hurricane Harvey approached Texas last month, meteorologists and the news media warned of “catastrophic and life-threatening flooding,” and similar messages have been used about Irma. “I’ll do anything in my power to convince them this is a very serious storm,” Miami Beach Mayor Philip Levine said on Thursday, after learning that some residents were planning to ignore evacuation orders. “This is a nuclear hurricane. They should leave the beach, they must leave the beach.”

Our survey research finds that such warnings may help motivate evacuation, but if they are too strong, some people think the message is “overblown” and thus does not need to be heeded. People who think hurricanes are less risky and whose worldviews favour personal autonomy are more likely to reject fear-based messages, as are people who don’t think they can protect themselves against the threat. So messages that describe the storm’s potential effects without being so overwhelming that they threaten people’s sense of autonomy can be more effective. Messages are often more persuasive when they include concrete, specific information to help people understand what they can do to protect themselves.

How to determine the best communication is not easy. We can look to the language and methods used before prior hurricanes to learn about which messages do and don’t work with various audiences. But each hurricane is unique – with different tracks, different wind and flood hazards, and a different evolution of forecasts – and the messages vary as a result. People access messages from many sources (public officials, meteorologists, friends and family members) over multiple channels (social media, the internet, television news), and these messages evolve with the forecast. This makes it difficult for researchers like us to isolate which features of a message work best. To offset some of these variables, we often conduct experiments where we ask people to imagine hypothetical storms, but those have limitations as well.

Of course, what people hear from authorities isn’t the only factor that affects whether they leave storm zones. When deciding whether, when and how to evacuate, people ask if they have a place to stay with relatives or friends outside of the evacuation zone. Can they afford a hotel? Do they have a car, and money for gas and food? Are they more worried about evacuation traffic or staying to protect their property? Do they have a disability that might make it difficult to move to a different location or stay in a shelter? In the past, having a pet was a major barrier to evacuating for some; after Hurricane Katrina, Congress passed a law requiring states to factor animals into their emergency plans and allow the Federal Emergency Management Agency to shelter them, which has made it much easier to evacuate with pets. People’s decisions are also strongly influenced by what they hear from other people and see others doing – including family members, friends, neighbours and often even friends of friends or new connections they make online.

Another influence: How recent was the last big storm, what were its effects and were they happy with their decision to evacuate or not? We know that people interpret forecast and warning information through the lens of their experience; sometimes, how people feel about their hurricane experiences is more powerful than any message. Good or bad examples of sheltering in place or evacuating can affect how people respond to future storms. The emotions from hurricanes linger. If people stayed at home and were terrified, they’re more likely to leave next time; if they were stuck in traffic or miserable in a shelter, they’re more likely to stay. As one emergency manager told us, “No matter what people did during [Hurricane] Sandy, whether they stayed or they evacuated, they told us, ‘I’m never going to do that again’.”

When it comes time to decide whether to stay or go, the most important consideration is whether people think that evacuating will help protect them and their families from harm. If they think they’re safe in their homes, or just as safe as they would be with their other options, then they don’t have a reason to evacuate. As Harvey clearly showed, evaluating who is safe and who isn’t is complicated by the multiple hazards that hurricanes present – including strong winds, flooding and the risk of tornadoes. Each of these hazards raises its own prediction challenges, and each can have various effects on people in different types of homes in different areas. What is most important ahead of storms such as Harvey and Irma is that forecasters, emergency managers, the media and coastal residents work together to help people understand which risks they face from that specific storm and how they can best protect themselves.

Cara Cuite is an assistant extension specialist at Rutgers, the State University of New Jersey, who studies coastal storm risk communication and public perceptions of novel food technologies. Rebecca Morss is a senior scientist at the National Centre for Atmospheric Research in Boulder, Colorado, who studies meteorological and societal aspects of extreme weather, including hurricanes. This article originally appeared on the Washington Post