For Amy Hessl, professor of geography at West Virginia University, the past can predict the future. Hessl is a dendrochronologist, a scientist who unravels climate histories and trends through the study of tree ring growth patterns.
So when wildfires engulfed the Los Angeles, California area in January 2025, resulting in dozens of deaths and tens of thousands of destroyed structures, Hessl wasn’t taken aback. In fact, she said, ongoing warm air temperatures and variable precipitation – which can be presumed from her studies of climate history trends — will lead to even more extreme fires in the future.
“Long records of past fires — that you can get from old trees that survived past fires, but recorded scars — can tell us a lot about how often fires occurred in the past, prior to European colonization, and what these records often tell us is that fires of pre-colonial periods were, in many cases, less extreme but more frequent than they are today,” Hessl explained.
“This change that we have seen in many places in the world is caused by the interaction between human-caused climate change, the history of land management leading to more abundant and more connected fuels, and people moving to the wildland urban interface — in other words —putting themselves in the way of fire.”
Hessl is just one of many West Virginia University researchers across a broad range of disciplines studying disasters and how to prevent and respond to them.
Her own research on ancient tree rings has implications that stretch beyond Earth.
Danger from above
When the sun blasts Earth with severe geomagnetic storms, it threatens the planet’s technological infrastructure. An extreme storm could destroy power grids and telecommunications for 15 years.
It’s impossible to predict when a geomagnetic storm may occur, but Hessl has found evidence of rare and extreme solar events, called Miyake events, in the rings of ancient trees. Miyake events occurred in 774 CE and 993 CE, and since that time, the research community has pinpointed seven more in the last 14,000 years. Hessl, who received a grant from the National Science Foundation has been digging deeper into these natural records.
“If one of these events were to happen today and you were on a high latitude flight to Norway, you’d probably receive your lifetime dose of radiation on the plane,” Hessl said. “And if you were in space, it potentially could kill you.”
Though Miyake events are exceedingly rare, they’re not out of the question, and a society dependent on satellites would be incapacitated in the wake of one. Most evidence points to the sun as the source of Miyake events, but Hessl hopes tree ring records from locations around the world can help the scientific community determine the causes and how severe past events were. Her research may also help prepare for future devastating space weather, which in turn, could advise the IT community how to prepare and protect vital technologies. She and her colleagues are studying trees in three U.S. locations that date back to the last major Miyake events.
“We’re looking at three species,” Hessl said. “An evergreen conifer from Utah, the bristlecone pine, is the longest-lived tree species in the world. They live for several thousand years.” Her team will be able to compare bristlecone pines with deciduous bald cypress in North Carolina and oak trees preserved in Missouri riverbeds. Any tree that was alive during one of the Miyake events would record it in the chemistry of the rings.
“We're trying to define how extreme those events were and when they occurred,” Hessl said. By studying the way different trees record past events, she hopes to better understand how to prepare for future ones, including those intense enough to threaten technological infrastructure. Though she said a Miyake event is unlikely, the key is advanced preparation.”
Danger from below
While a way to reliably predict or prevent natural disasters is still in the future, by studying the events of the past and collecting vital data, researchers can use what they learn to direct recovery efforts and prepare for the future.
While many parts of the country deal with storms like Hurricane Helene, the western U.S. lives with seismic activity. One of the problems, according to Jaime Toro, professor of structural geology and tectonics in the Department of Geology and Geography, is perception.
“Small earthquakes happen frequently, but big earthquakes are rare,” said Toro, who began his career in geology in the mid 1980s and has focused on geology in Appalachia as well as the tectonic evolution of remote northern mountains. “They may come once a century or two. People are not that good at thinking about this kind of terrible but rare event.”
Several west coast cities sit in the line of fire, including San Francisco, which is built on top of an active fault that has produced large earthquakes in the past, including the 1989 Loma Prieta quake.
“We know it will produce huge earthquakes in the future,” Toro said. “Millions of people are living right there, but nobody really expects it’s going to happen to them.” Another huge seismic zone in the Pacific Northwest has been inactive for more than 500 years but still has the potential for severe quakes and tsunamis. Seattle, Tacoma, Portland and cities along the western seaboard are vulnerable, and it's not a question of if, but when, as evidenced by ancient records of past tsunamis that came between 1,000-2,000 years ago.
“We don't know when the next one's going to come, but it will,” Toro said. “As a society, we really don't think in those terms. Unless something happens every five-to-10, years, we don't worry about it. So how do you prepare for a hazard that comes once in a millennium?”
There’s no good answer, yet.
A disaster for democracy?
Hurricane Helene caused catastrophic flooding in September 2024. Rivers rose in mere hours, overtaking residents who believed they were safe because they had been in the past. The waters swept homes away and swallowed cars with drivers still at the wheel. Mountainsides came roaring down in a torrent of mud, demolishing everything in their path. When the floodwater retreated, western North Carolina was barely recognizable, and more than 100 people had lost their lives.
Natural disasters like Helene disrupt life in large geographic areas, but life beyond those damaged areas goes on, and in 2024 that meant a general election in just a few weeks.
Erik Herron, Eberly Family Distinguished Professor at WVU Eberly College of Arts and Sciences, looks at how these kinds of catastrophes affect elections. He traveled to the hardest-hit areas, hoping to understand how the nightmare scenario would affect voters and their participation in the impending presidential election.
“I saw a tremendous amount of hurricane damage in some places and heard some harrowing stories from voters,” he said. “Roads were washed out, people were trapped. In many cases, people said the place they were supposed to vote was gone. What do you do when the polling place is washed away? How do you make sure people can vote? This was the question we were interested in.”
Counties in North Carolina took various approaches to voting during the crisis. Yancy County had no cell service or internet access, so election officials put up tents where the polling place stood before it was washed away. Alternatively, nearby Avery County consolidated polling places. Past research indicates that travel distances can affect turnout, and the researchers are continuing to study what effects these decisions had on voter experiences.
To understand how the storm affected voter perceptions of the election process, Herron and two colleagues from North Carolina conducted exit polls. They asked voters about their experiences with the hurricane as well as questions about their perceptions of the election’s integrity. They also asked about partisan lean but not about who voters cast a ballot for.
The researchers are still collecting data and will conduct interviews with both election officials and disaster response organizations to study how the flooding affected the process, but some initial findings indicate that confidence in the election was high; however, there was a small difference based on partisanship. Voters who identified as Republicans were slightly more confident in the outcomes than those who identified with other parties. There was no statistical difference between beliefs about election integrity and being in a disaster-affected county.
As for attitudes about the disaster itself, Herron said, “Political ideology and partisanship are predictors of assigning blame for Hurricane Helene’s destruction to climate change.
Experiencing impacts from the storm does not shift political conservative or Republican identifiers’ attributions.”
In the past, Herron studied elections in the Republic of Georgia during the COVID pandemic as well as those held in Ukraine during the ongoing war. Despite contrasting situations, the challenges— the health and safety of the voters and the people running the elections — were the same.
“The most critical issues in an election are making sure voters can participate and not undermining the perception of legitimacy and integrity,” he said. “A wartime environment, a pandemic environment and a post-disaster environment are all really different. But what's common among all of them is that if it's not handled well, it can undermine those perceptions.”
His research during combat in Ukraine in 2014, when Crimea was annexed by Russia, indicated that voter participation near the front lines was limited but increased with distance from combat zones. Election administrators and security were able to stabilize and maintain an environment where people felt safe to participate. Ten years later, Ukraine delayed elections due to increased fighting and missile and drone attacks. Herron found the majority of voters want to wait until after hostilities end to hold elections. They worried about violence and cited concerns that soldiers and refugees would be denied the chance to participate.
“One of the lessons is that in a democratic society, we expect elections to be held. We don't like it when they're postponed or canceled. But when you have a crisis, whether it's war or a pandemic or natural disaster, administrators, voters and candidates all have to adapt, sometimes very quickly, to a different environment to make sure the election has integrity and that people consider the results legitimate.”
The right tools
As weather events like hurricanes become more frequent and more intense, extreme patterns in functional data increase. Predicting them is crucial for managing the risk of all types of natural disasters, but until now, there have been no statistical tools for analyzing such complex data.
Enter Mihyun Kim, assistant professor in the School of Mathematical and Data Sciences. She’s developing tools to predict and model extreme patterns — including weather events and pollution levels — with support from the NSF. Kim’s research focuses on analyzing extreme patterns in functional data, which form a curve and are useful for determining rates of change. Temperatures, for example, may be recorded in daily curves, but functional data provide richer information than simply noting the maximum or average temperature.
“Instead of just noting how much it rained today, functional data would record how the rain changed every minute of the day,” Kim said.
“ My research aims to fill this gap by developing statistical tools to model and predict patterns more precisely. ”
— Mihyun Kim
“Our tool focuses only on the extreme part, not the whole body of data. More specifically, it measures the dependence between two extreme patterns.
“For example, how likely is Location A to experience a similar pattern of temperature as Location B on the same day during a heat wave? The same approach can be applied to other types of data, like daily precipitation patterns or river flow patterns. This tool will be our starting point. Leveraging it, we will create methodologies to predict future extreme patterns later.”
Kim’s research was inspired by the 2016 flood in West Virginia. In mountainous regions, topology can dictate the severity of the damage, which was very concentrated in certain areas of the state. The tool can compare two locations that show a similar data pattern.
“If someone was looking at only one location, then I could say, ‘Hey, there's a high level of dependence between these two locations. So you should also take a look at the second one as well.’ Based on this measure, we can create a time series model, which means we can predict future extreme patterns.”
While the future is uncertain, the ability to adapt to changing situations is important, as is the compassion to look beyond the disaster and focus on those in need. Through their continuing work, WVU researchers are giving communities hope and the tools to prepare, rebuilt and recover.
