91������Ƶ

Water will continue to shape—and reshape—Vermont communities. Sometimes it will come in a deluge, pummeling agricultural fields and swamping downtowns. Sometimes water will be scarce. There will be seasons that feel abnormally normal. This is life in a rapidly changing climate.

Statewide flooding in Vermont in July 2023 caused upwards of $600 million in damages, the bulk of which occurred in communities located along the spine of the Green Mountains, where two-months-worth of rain fell in about 48 hours. In Montpelier, where I call home, more rain fell in one day—5.28 inches—than in any day since record collecting began. Every time I walk my children to school on rainy mornings, they look at the Winooski River and ask, “Is it going to flood again?”

“Yes,” I respond. “But not today.”

As I began to write this story on July 10, 2024—the one-year anniversary of the flooding—a flood and tornado watch was issued for 12 Vermont counties, and rain from the remnants of Hurricane Beryl started falling. On the drive home, fat rain drops bounced off the windshield and I wondered if downtown Montpelier had enough sandbags. Because we will flood again—eventually. I wondered what measures could mitigate the risk to our communities. Meaningful actions require nuanced understanding of the problems we face—and this is where research can help.

Before the July 2023 storm was even over, 91������Ƶ researchers had pulled on muck boots to gather data and help recovery efforts. One year later, in the wake of another disastrous flood, those researchers continue collecting and parsing information to help Vermonters build more resilient systems.

Help From Above

Jarlath O’Neil-Dunne G’04—an ex-Marine who ran up mountains, biked to work in the snow, and played the cowbell at marathons with his band, Sciatica—was “a larger-than-life character who gave everything to his job,” said Paige Brochu ’15, director of UVM’s Spatial Analysis Lab (SAL).

For 12 years, until his sudden death from a heart attack in January, O’Neil-Dunne directed SAL and built up the team from a handful of technicians to about two dozen full-time staff. He stored a go bag in his office and the phone numbers of first responders on his cell phone in case the lab’s drone team was needed during emergencies. When O’Neil-Dunne was at a conference in California in July of 2023, a call came that Vermont was poised to flood, and he began orchestrating an aerial response from his hotel room.

“We deployed as soon as possible, even without direct tasking orders from the emergency operations center,” Brochu said. “If we had wait for official orders, we would not have captured the highwater marks for a lot of the communities impacted.”

SAL’s Unoccupied Aircraft Systems unit used their drones to document the scale of flooding and map landslides and low-lying areas that disappeared under inches of water. This was the type of emergency response by SAL that O’Neil-Dunne envisioned being especially useful.

He first recruited Brochu to the team as a geospatial technician, mapping tree canopies when she was an undergraduate. He persuaded her to come back to UVM after she earned her doctorate in environmental health to work on SAL projects such as helping towns across the nation quantify their urban forest and understand the impact of greening initiatives.

“There is always a why,” Brochu explained. “We don’t just circle trees to circle trees.”

After the July 10-11, 2023, flooding, the drone team made time-stamped imagery publicly available, often within 24 hours of collection, and helped unlock federal relief funds.

“First responders used it to assess damage, not just to property, but also to infrastructure,” Brochu said. “… Drone images are great but if we are not doing anything with them, why are we capturing them?”

When I spoke to Brochu in May, she was about to move into the director’s office and staff were cataloging all their 2023 flooding work to conduct a post-mortem of the team’s response. They wanted to understand public perception of the work.

“Jarlath was a really big proponent of community-engaged work and listening to what communities need,” Brochu said. “One thing the lab did really well, and Jarlath was really good about, was maintaining relationships with state agencies. He had everyone’s cell phone number … so that when there was a request it was immediate. And it was built on personal trust. They knew they could call us and that we would do good work.” 

With summer 2024 approaching, she contemplated what the season would bring.

“I feel like I am waiting for a shoe to drop,” she said, adding that whatever comes, SAL will be ready to deploy. “We will be available because we need to be to support Vermont.”

On July 10, 2024—the first anniversary of the 2023 floods—SAL team members were prepared when the Vermont Emergency Operations Center requested drones to capture a record of flooding. Communities like Barre, Johnson, and Cambridge were once again underwater, and flash flooding decimated parts of Barnet, Lyndonville, and Plainfield. A manufactured home community (MHC) in Berlin, one of several devastated by flooding the year before, was listed as a site of interest for SAL to flyover—and it is one Kelly Hamshaw ’06, G’11, Ph.D. ’24 knows all too well. 

A statewide assessment found that 60 percent of manufactured home communities in Vermont are located within a flood hazard area.

Showing Up

Hamshaw has a long history of knocking on the doors of manufactured homes – also called mobile homes – throughout Vermont. Since her undergraduate years at UVM studying planning and community development, she has worked to better understand the needs and complexities of MHCs across the state. Over the years, Hamshaw has visited over half of Vermont’s nearly 240 MHCs, conducting surveys and tabletop emergency exercises and involving her classes in clean-up efforts after extreme weather events.

So it was fitting that before the 2023 floodwaters fully receded, Hamshaw was distributing notices to residents explaining that if they sold their waterlogged homes for pennies on the dollar, they wouldn’t qualify for Federal Emergency Management Agency (FEMA) assistance. Hamshaw has witnessed this before and knows that many MHC residents can least afford such a loss.

“I’ve been to a lot of communities, I’ve talked to a lot of people,” she said, explaining that the most common person she meets in the parks is “a single woman living on her own over the age of 65.”

In Vermont, many MHCs were built in the ’60s and ’70s and sited along the developing highways as worker housing. The term “manufactured home” replaced “mobile home” as the technical terminology in 1976, Hamshaw explains, so while the language has shifted, our thinking hasn’t quite caught up.

These homes—and populations—are largely not mobile either by choice or by circumstance. Most of the units can’t be moved without significant expense or structural damage. In Vermont, only about one percent are actually moved each year. Residents often elect to live in the parks due to their affordability and proximity to family, schools, doctors, and social networks.

“And [because they do] not have to share walls with people,” Hamshaw said. “That is really important to folks and a lot of people forget that.”

What makes MHCs different from other housing developments when it comes to emergency planning is the land ownership. Residents may own their home, but unless they live in one of the few nonprofit or co-op-owned parks, they have little control over investments in flood mitigation for the property itself because someone else owns the land.

MHCs fill an important niche in the state, Hamshaw explained. “People have realized in the context of our larger housing market, if these communities went away, it would just further compound our already dire circumstances.”

A statewide flood assessment the Vermont Department of Environmental Conservation conducted with UVM SAL imagery found that nearly 60 percent of MHCs are located within a hazard area and nearly 15 percent are in FEMA flood zones. In the months prior to the 2023 flood, Hamshaw conducted virtual and in-person workshops at three Washington County MHC’s vulnerable to flooding, including one in Berlin that flooded a month later. Corinne Cooper, a UVM Extension employee who came to the meetings, was among the first to contact Hamshaw about it. Hamshaw came with fliers and worked with the state’s mobile home program and residents to advocate for debris removal.

“You can do a lot if you show up," Hamshaw said.

Her work continues this fall with state partners and UVM colleagues including Kristen Underwood, Ph.D. ’18, research associate professor of civil and environmental engineering, with a project involving five MHCs at risk of flooding. The team will perform risk assessments for each park and build a toolkit to increase preparedness and resiliency.

“In my view, that really fulfills our land-grant mission here at UVM—taking research, distilling those findings, and sharing those findings with a community that could very much benefit from [them],” Underwood said.

Improving Floodplain Maps

Rivers are dynamic by nature. They swell and move as flows change, occasionally jumping their banks and spilling excess water onto adjacent floodplains. Over centuries, humans have straightened and channeled rivers as we’ve built and farmed alongside them. We’ve dammed rivers. Dredged them. Bermed them. Humans have excelled at moving water efficiently through river corridors and disconnecting them from floodplains that slow water as it sprawls.

Learning where floodplains are, how they operate, and where they may be restored could boost resiliency for communities at risk of flooding. That is one idea behind the state’s Department of Environmental Conservation’s Functioning Floodplain Initiative (FFI), an effort UVM research helps inform, in part through improved mapping. Because Vermont lacks comprehensive floodplain maps. This is especially important in a changing climate that is bringing more frequent and intense floods.

Since 2019, Kristen Underwood and Rebecca Diehl, a research assistant professor in the Department of Geography and Geosciences, have partnered to map Vermont floodplains. In contrast to traditional high-complexity models that require time-intensive computations and field surveys, they use low-complexity models that rely on remote sensing and simplify mathematical assumptions to simulate how water moves through a river corridor.

“We lose information when we do that, but what we gain is the ability to map over very broad regions,” said Underwood.

This allows researchers to answer deeper questions too. Questions like: Where are Vermont’s floodplains and how often are they inundated? And are our floodplains connected?

“In Vermont, one of the big issues with floodplains is that they are fairly disconnected, meaning they don’t get accessed as regularly as they should by floodwaters,” Diehl explained. “[This] limits the role that floodplains can play and all sorts of those co-benefits that they could provide society.”

In 2016, UVM researchers quantified one co-benefit by simulating ten flooding events in Middlebury, including Tropical Storm Irene, and calculating structural damages. They estimated that the Otter Creek floodplains and wetlands upstream reduced damages between 84 and 95 percent for Tropical Storm Irene and help save hundreds of thousands of dollars annually. 

“I call it the poster child for floodplains and flood resilience,” Diehl said.

She and Underwood are exploring if similar effects may be possible in other communities. What floodplain characteristics decrease flood waves? How many floodplains does one need to conserve to make a difference downstream?

“Setting aside all of those social issues of money and people and property rights—do we have the capacity in our river network today to attenuate these floods?” Underwood asked. “And are there certain kinds of reaches or floodplains that are better able to attenuate floods?” 

In 2022, Underwood and Diehl released a mapping tool to show the extent of flooding across the Lake Champlain Basin for storm sizes ranging from a two-year to 500-year event. The floodplain layer was created for planning purposes and research and is not intended to replace FEMA floodplain maps. It was used as a reference layer by the state’s Emergency Operations Center to guide response and short-term recovery during the July 2023 and 2024 floods.

Diehl and Underwood are working to build a dynamic low-complexity model to simulate how a flood wave moves through the river network during a storm. Underwood envisions it could one day be used by first responders and citizens who need to know in real-time who needs to move to higher ground during a flooding event.

In the months before the July 2023 floods, the state mapped all of Vermont with Light Detection and Ranging (Lidar)—a remote sensing technique that produces precise topographical maps using pulses of light. UVM’s Spatial Analysis Lab flew drones to capture Lidar on several stretches of riverways after the 2023 and 2024 floods. Diehl and Underwood and their students are quantifying how the topography changed as a result of erosion and deposition that occurred during these floods, and potential impacts on water quality.

“How did our rivers respond to these extreme events and are we seeing a lot of erosion?” Diehl asked. “I am super interested in thinking about …  How do our floodplains function in light of these extreme events? Can our floodplains keep pace with a July 2023 flood? We know that there is a ton of stuff coming off the landscape.”

On the first anniversary of the Great Vermont Flood of 2023, extreme rainfall struck again, flooding communities such as Plainfield. Image captured by a drone flown by UVM SAL's Unoccupied Aircraft Systems team.

 The Landscape Remembers

On a muggy day in late July, Elizabeth Doran and two undergraduates, Lydia Emry, a senior in the Rubenstein School of Environment and Natural Resources, and junior Andrew Chen, an environmental engineering major, pulled on muck boots near the mouth of the Lewis Creek in North Ferrisburgh. They are there to rebait camera traps and survey high water marks from the recent July 10-11 floods, when up to 10 inches of rain fell across the watershed. Hours earlier, another flood watch was issued for most of Vermont.

Doran, an environmental engineer and assistant research professor, is leading an interdisciplinary team (including Diehl and Underwood) in a three-year mapping study to better understand the role of floodplains as unique ecosystems and habitat for native plants and animals. Her team is testing the presence of various natural communities of flora and fauna to better understand their interaction with flood events, as well as how the system may have changed over time. The team partnered with local non-profit watershed groups, including the Lewis Creek Association, that have been conducting restoration projects in the watershed for decades. The results will help improve floodplain restoration efforts.

“You can't just plant a mature silver maple floodplain forest,” Doran explained. “It doesn't work that way. Saplings need time to grow, and they don't necessarily just survive because you think that's where they should be.” We can and should take action, she explained, “but the landscape kind of remembers what happened to it,” and we have to take that into account as we prioritize where to focus our restoration efforts.

In the spring, the research team placed motion-activated cameras along 49 sites in the Lewis Creek Watershed to document the various wildlife that call it home, particularly otters and amphibians that need both the water and floodplain to thrive—including northern leopard frogs and Jefferson salamanders.

“I have pictures,” Emry said, pulling out her phone. “Leopard frogs look really similar to another species called pickerel frogs.”

“You can hunt for a few minutes in the grass over there and see one,” Chen said.

As we traipsed over grass towards the water’s edge, tiny green missiles shot out of the way of our boots. Emry scooped one up and checked the tiny frog’s white belly to properly identify it.

“Sometimes they calm down and just vibe in your hands,” she said with a smile.

In addition to examining the natural communities along the sites, UVM scientists are studying the geomorphology of the river system. How does a flood wave move through it? Are the floodplains connected to the river? The team has data from before and after the July 2024 flood and can examine if wildlife is affected by flooding.

“Where do amphibians go during the flood?” Doran asked. “Where do otters go? How long before they come back?”

We sloshed through the water, parting waves of tall grass flattened just weeks before. We reached the camera and Andrew swapped out the SD card while Lydia swabbed bark with a musky scent to lure otter.

“Part of the reason why this area didn’t flood much is that we have this huge wetland area,” Doran explained, gesturing upstream. “The water had space to spread out, get really calm and slightly less destructive. … Floodplains have many benefits. They are habitat for wildlife. They can capture nutrients and reduce flood peaks and reduce damage.”

Three kilometers upstream, the team visited a site with significant washouts after the flood. Bridge abutments constricted the water, pushing it towards the bank where it knocked out several mature trees, including one with a camera lashed to its side.

“We did see an otter here one time before the floods,” Emry said, as she applied fresh lure. “I'm super interested to see if we’re still going to see those otters because this landscape has changed so much.”

Lydia Emry, a senior in the Rubenstein School of Environment and Natural Resources, swabs fresh lure on bark to attract otters. 

 Putting Roots Back in the Ground

“All of this is floodplain,” said Andy Jones, gesturing one blustery May afternoon to the fields he has managed since 1993 at the Intervale Community Farm (ICF), one of several farms on a 360-acre site near the Winooski River.

ICF was almost completely flooded during Tropical Storm Irene, and fared worse in July 2023. Jones knew from experience to pick crops early and move heavy equipment to higher ground, but ICF was still $60,000 in the hole when waters receded.

After a flood, fields are soaked with silt and water contaminated with whatever pollutants were upstream. Knowing when they are safe to replant can be tricky without data. After the 2023 floods, UVM’s Agricultural and Environmental Testing Lab performed free soil testing on 166 sites, including ICF, to identify spots potentially laden with heavy metals, hydrocarbons, and bacteria. While initial tests revealed high microbial counts, levels returned to normal within weeks.

“We are organic farmers; We don’t want to make people sick,” Jones said. “… It was helpful to me and to a lot of the other farmers here to know we are not being ridiculous to say, ‘we can go forward, and the food will be safe.’ We may still lose our shirts because it’s going to flood again, but that is a different problem.”

Over the decades Jones has watched nightly temperatures climb and left crops in the fields later into the fall. While this has boosted productivity, the benefits are offset by upticks in extreme events. In recent decades, extreme rain events have increased across New England, pancaking the ground and leaving plants vulnerable to fungal diseases.

“It is hard on the plants,” Jones explained. “… and then we will have long stretches [of nothing]. We are adapting that way. We have invested a lot in irrigation equipment. … We need it more and more even though we are often getting too much rain. It’s a crazy oscillation.”

Extreme weather events and more variable frost dates add uncertainty to farming, an already risky enterprise. But there are ways to increase resilience from extreme weather events.

Interventions such as high tunnels—plastic- or fabric-covered hoop structures—can protect crops from extreme weather events and give farmers more control over growing conditions, said Joshua Faulkner, coordinator of UVM Extension's Farming and Climate Change Program. Building soil health can help too, he said, adding that roots in the ground prevent erosion from runoff, increase organic matter in the soil, and soak up some of the moisture during heavy storms.

“When we have erosion, it is farmers losing topsoil, and it takes in some cases decades, if not hundreds of years, to build topsoil,” said Faulkner. “So we are losing that agricultural resource that is important for food security.”

Back at ICF, Andy Jones considered whether anything good comes from a flood.

“It’s easy when things are going well just to cruise and to keep doing the same thing,” he said. “A flood does force a little bit of introspection.”

Mainly, ICF is looking up.

UVM's SAL team captured this aerial image of flood damage in Addison County. These images can help landowners to receive disaster recovery funds.

“We are trying to go all-in on what we have for the higher grounds—put up a few more tunnels on the sections that didn’t flood,” Jones said. “We are going to invest as much productive capacity on the least flood-prone land.”

He walked out back to show where the new tunnels would stand. We followed a dirt path, and he paused to point downslope to where the river flowed across the farm last summer. We passed one high tunnel filled with tomatoes planted just hours earlier and a second brimming with baby cucumber starts.

“There is a lot of land down here,” Jones said pointing to a copse of trees in the distance. “And the river, it’s always present, the mighty Winooski.” 

We walked back to the parking lot, passing patches of dandelions soaking in the late afternoon sun.

“We can keep doing this as long as it’s working,” he said. “The question that looms is: ‘How often does this have to happen for it to no longer really be a viable business model?’ And I don’t think we know, or anybody really knows.”

On the first anniversary of the Great Vermont Flood of 2023, ICF and the Intervale were inundated once again. Farm staff and volunteers salvaged thousands of pounds of crops from the fields and UVM Extension staff were back out testing flooded fields. ICF staff replanted within weeks and has begun searching for some land out of the floodplain.