UMass Amherst Research Professor Teams with National Weather Service to Build Database on Public Response to Severe Weather Hazards

University of Massachusetts Amherst research professor Brenda Philips has received a grant from the National Oceanic and Atmospheric Administration (NOAA) to work with National Weather Service forecast offices across the country to ascertain the factors that influence people’s responses to severe weather events. The goal of the two-year, $396,855 grant is to build a national multi-year database on human reactions to four types of weather hazards: flash floods, tornados, severe thunderstorms and winter weather events.

“NOAA collects lots of meteorological data, but only collects a very limited amount of data on societal response to weather and warnings,” says Philips, who has worked closely with National Weather Service (NWS) forecasters as part of her research on severe weather warning systems and as co-director of UMass Amherst’s Paros Center for Atmospheric Research (PCAR).

“I’m excited about this surveying effort,” Philips says. “We’re launching surveys 1-14 days after medium-to-high-end events happen so we can collect detailed accounts of how people experienced a specific flood, for example. If we aggregate survey responses over many events, we’ll be able to understand better how different segments of the population experience hazards.”

Finding out how people interpret and react to severe weather is becoming more and more important. A 2020 report from the United Nations Office for Disaster Risk Reduction found climate-related disasters have jumped 83% in the past 20 years, with major floods more than doubling and severe storms up 40%.

“We’re seeing more dramatic flash floods and winter events due to climate change,” she says. “Everybody’s having these 100-year and 200-year floods and seeing really unusual situations. Getting a handle on how different environmental, social and situational factors influence an individual’s risk perception is very important,” she added.

Human response varies depending on the hazard, the type of protective action they are advised to take and also the duration of the warning, Philips says.

“A tornado warning lasts about 45 minutes. It’s very focused and you should go to a protected area, like a basement, immediately; flash flood warnings last at least two or three hours and cover a large geographic area, so drivers just have to be aware that they might encounter a flooded road,” she says.

Under the grant Philips received from the weather program office’s social behavioral and economic sciences program, she will work with nine NWS forecast offices (Boston, Fort Worth, Pittsburgh, Atlanta, Nashville, Little Rock, Philadelphia, Portland and Phoenix) to send surveys soon after weather events to people in the affected areas. Questions address their information sources, preparations prior to the event, whether they sheltered in place in the case of a tornado, or avoided flooded roads for flash flood events.

By the end of the grant period, Philips hopes to have sufficient responses to be able to start to understand the relative importance of key factors such as environmental cues, demographics, situational circumstances and risk communication.

“For flash flooding, I’ve done some pilot studies, and those have shown that it’s the social cues — whether other cars are turning around — are one of the biggest influences for turning around,” she says. “Also, the current situation matters. People drive on flooded roads more often to get to work — they’re more likely to take the risk.”

She also hopes to learn more about any racial, economic and regional factors that need to be addressed as weather offices craft warnings in the future, which could influence what channels they use to do so.

While the weather service uses multiple channels, such as television, radio and social media, “even then, you have people who say I didn’t get the message,” she says.

“To start to formulate policy or prioritize research on warning systems, it’s important to start with the people that NOAA is serving,” Philips added. “This research will give NOAA the tools to do that more effectively.”