About 100 miles from the nation’s capital, near Taylor’s Island and Fishing Creek on Maryland’s Eastern Shore, hundreds of acres of dead trees stand upright, like toothpicks piercing the sky. They are among the mid-Atlantic’s largest “ghost forests,” woodlands rapidly converted to marsh because of sea-level rise.
All along the margins of the mid-Atlantic today, in Maryland, Delaware, and Virginia, as well as many other low-lying parts of the East Coast, rapid sea-level rise is driving a deadly sogginess inland. Frequent floods and higher high tides are pushing marsh into forests and drowning the cedar and pine trees that fringed the shore even a decade ago. Some habitats are changing so fast that the dead trees haven’t had time to fall over. They haunt the landscape, effigies of climate-driven change.
“Ghost forests are the most striking indicator of climate change on the East Coast,” says Matthew Kirwan, a leading authority on the subject. Kirwan works as a coastal geomorphologist at the Virginia Institute of Marine Science (VIMS) in Gloucester Point. From his seaside office, littered with maps and topographic drawings, Kirwan coauthored a 2022 remote sensing study tracking the spread of the mid-Atlantic’s ghost forests since 1984 (1). “The numbers are staggering,” he says. Some 40,000 acres of forest and farmland have converted into ghost forest in roughly the last 30 years.
Yet, most people have never heard of these huge forest die-offs. Ghost forests have remained largely out of the public eye, along marshy backroads in rural areas. Farmers are often the first to notice that the land is slowly drowning, when saltwater affects their crops. It begins with a muddy squelch under their boots, along a corner of a field that used to be dry. The water creeps into lowland areas with flat topography, just slowly enough that most people wouldn’t notice it. Marsh banks are always lined by a few dead trees. Those few trees that died each year suddenly add up to 40,000 acres, Kirwan says. “The satellite perspective allows you to see the history over a couple decades, and in a quantitative way,” he says.
Mapping the die-offs reveals that ghost forests are growing. Their expansion raises questions about the amount of carbon dioxide and methane that decomposing trees could belch into the atmosphere. Field studies are tracking the precise causes of ghost forest formation and how much carbon they stand to lose.
The stakes are high. Storm surge and flooding are already pummeling many seaside towns, and communities are contemplating adaptation measures now. Dying forests will have an impact on economies, tourism, and housing. It may be too late to stop the die-offs. But predicting where they’ll spread next can help towns to proactively adapt, rather than waiting on the tide to wash in.
As recently as 2016, ghost forests were just a whisper in university halls. Lindsey Smart says she remembers hearing about them in college, in the 2010s, but only “in an anecdotal way.” Researchers had noticed progressive tree die-offs in coastal areas, but nobody had mapped the extent of the losses. What was clear from the botany literature is that forests have a limited tolerance for saltwater (2), she says.
During her PhD at North Carolina State between 2014 and 2018, Smart took a keener interest in the ghost forest phenomenon. Braving the hot and soupy North Carolina summers, she’d drive the 3 hours from inland Raleigh to field sites along the coast, leading one of the first efforts to map the spread of ghost forests anywhere on the Eastern Seaboard.
Slamming her truck door and waving off the mosquitoes, Smart waded into coastal bog and brush carrying a measuring tape and a clinometer to measure tree height. She bushwacked her way through more than 100 field sites along the Albemarle-Pamlico Peninsula, a low-lying thumb of eastern North Carolina. The peninsula juts into the Atlantic and is already lapped by rising tides. Healthy forests still grow there, as do marshes. Along the water, in the transition zones between marsh and forest, dead and dying trees spear the sky.
Over several summers, Smart measured various traits of the vegetation in all three habitat types: healthy forest, marsh, and ghost forest. She jotted down the plant species that she saw and wrapped her measuring tape around tree trunks at chest height to find differences in the average width-around (a proxy for biomass) of trees like pines versus marsh species like sweet gum and cypress. Smart found that healthy forests had tall, dense vegetation and were high in biomass. Marshlands had shorter plants, with lower biomass. And ghost forests were tall, with low biomass, and patchy vegetation.
Back in her office, Smart compared field data to satellite images of the peninsula from the prior decade. Three-dimensional topographic scans of Albemarle-Pamlico, captured in LiDAR data from 2001 and 2014, told her the height of the vegetation back then and allowed her to calculate the density and biomass of plants growing on the peninsula during those years. She saw a dramatic die-back of healthy tall and dense forests, and the expansion of patchy ghost forests, when comparing the maps from 2014 and 2001. In results she published in 2020, some 41,300 forested acres, or 15% of the peninsula’s coastal unmanaged public land, had died in just 13 years. That pattern has presumably continued or even sped up since 2014, Smart says, but the extent of Albemarle-Pamlico’s ghost forests has yet to be reevaluated.
Ghost forests now fringe all low-lying coastal areas of the eastern United States, from the humid palmetto thickets of the Gulf Coast to the white-cedar groves of New England. Maps can give a bird’s eye view of the die-offs. But to get into the weeds, and pin down the process by which ghost forests spread on the ground, researchers are also leading observational studies that track various stages of forest death across elevations and field sites. Salt-tolerant marsh species are the first plants to invade flooded forest understories, and they only need a few years to take over.
Community ecologist Keryn Gedan, a professor at George Washington University in DC, has watched ghost forests spread all along the Chesapeake Bay. She first noticed the die-offs about 15 years ago. Gedan moved to the area for a postdoc in tidal ecology in 2010, after finishing a PhD on climate impacts to marshes in Rhode Island.
One day, while chest-deep in the Chesapeake collecting mussels for her research, she looked up and noticed a wall of dead trees lining the bay shore. Marsh shrubs poked up around the dead trunks. Gedan knew from her dissertation that, at least up in New England, some marsh plants were growing like crazy in response to warming. She also knew that, over long timescales, they could push uphill and inland with sea-level rise. Standing there in the Chesapeake, Gedan wondered if she was seeing the same consequence of climate change playing out at high speed hundreds of miles south.
The answer seems to be yes. In the intervening decade, Gedan established a research group tracking the formation and expansion of ghost forests, mostly using observational studies. Every few months, she makes the 90-minute drive to the massive ghost forests outside the capital, in Maryland’s Blackwater National Wildlife Refuge. To make observations at field sites there, her research team takes what’s called a “chrono-sequence” approach, in which they assume that neighboring parts of the landscape represent different stages of the same long-term process.
Blackwater is a very flat, muddy park. There aren’t mountains, or even really hills. But there are subtle elevation differences, of tens of centimeters. Standing on these slight slopes, Gedan has found many more dead trees at lower elevations, where saltwater intrudes, than even 10 or 20 centimeters upslope. Over time, she’s also found faster rates of tree mortality at lower sites, a result she’s presented at conferences, but not yet published. Discovering this pattern across dozens of transects tells her that the same process is likely killing the trees at every site.
Not only has Gedan found more dead trees at low elevations, but the mix of understory plants is also changing there. Marsh grasses and the invasive reed Phragmites are moving into waterlogged soils, newly opened to the sky, where forest canopies used to shade (3). Slightly higher up, where saltwater is just starting to creep in, “we see branches breaking and an increase in shrubs in the understory,” Gedan says. Red flags of forest stress hint that these stands may turn to ghosts in the next decade (4).
To predict where ghost forests may go next, ecologists, hydrologists, and others are pinpointing exactly how forests drown. They know that a combination of gradual sea-level rise over years, paired with punctuated floods during storms and high tides, are primary physical stressors for the trees, explains Holly Michael, a hydrologist at the University of Delaware.
“They talked about handing over their land to the next generation.”
Healthy tree roots ramble through the upper layers of the soil. Sea-level rise is slowly pushing up the water table, carrying the groundwater up and up, closer to the surface. Paired with very high tides and storm surge flooding, ghost forest soils are soggier and saltier than those in healthy forests, Michael says. Many of the dead trees turn out to have their roots wet. But whether it’s the prolonged waterlogging of the soil, the salt, or some feedback between them that ultimately kills the tree remains an area of active research.
Michael is now the lead on a 5-year NSF-funded project, tracking the migration of marshlands into forests and farms. She works at two dozen field sites across habitats from marsh to forest around the Virginia coast, Delaware River, and Maryland’s Chesapeake Bay, as well as in New York and New Jersey. On any of the project’s dozens of subsites, collaborators have installed long, thin PVC pipes, threaded 10 feet down into the soil. They track temperature, salinity, and the elevation of the water table. Other standalone tools monitor soil moisture, water chemistry, and more.
Paired with ecological monitoring of sap flow and other health indices of the trees, Michael hopes to learn about the mechanisms driving the formation of the ghost forests. The work isn’t yet published, but one 5-year goal, she says, is to pinpoint the physical drivers of the die-offs, so that hydrological models—for example, of storm surge or rainfall—can help predict where ecological declines will spread. “It’s much easier to model the hydrology,” Michael says, than to predict where trees will die based on ecology alone.
When change does come, Gedan has found that it isn’t linear. Forests drown in fits and starts. “Wouldn’t it be nice to see salinity in the soil or groundwater just creep up, up, up, like we expect?” Gedan says. But, instead, salinity might peak after a storm or high tide event, and a handful of trees might die, but then, within a few months, the salt levels settle back to normal, and trees begin to recover (5)—that is, until the next major storm bludgeons the ecosystem, usually before the forest has recovered from the first insult.
Tracking the Carbon
Underlying all these investigations is a larger question with big implications: How much carbon will these many decomposing trees release into the atmosphere, perhaps accelerating climate change as part of a positive feedback loop? The dead trunks of drowned trees not only release carbon dioxide as they rot, but can act like hollow straws, moving methane and other greenhouse gases from the soil and belching them into the atmosphere, via their dead, dry wood (6).
Invading marshes, on the other hand, lock up carbon in their deep roots and soils. In fact, they may offset some of the greenhouse gas escape. As marshes develop, their dense root mats suck carbon from the water and create low-oxygen aquatic zones, where any fallen trees will decompose slower than in oxygen-rich environments. Gedan says it takes centuries for a healthy forest to convert into a mature marsh, however. Researchers want to predict emissions in the next decade or two, long before the marsh root mat will grow deep enough to trap much carbon from the rotting trees.
Smart took an early stab at it in her 2020 study (7), investigating how the expanding die-offs between 2001 and 2014 had affected carbon stored aboveground, in tree trunks, leaves, branches, and marsh grasses on North Carolina’s Albemarle-Pamlico Peninsula. First, she used the field estimates of plant biomass as a proxy for the carbon stores at each of her field sites. Then, she ascribed those carbon values to the corresponding pixels, at the location of each field site on the satellite map of the peninsula. Next, a machine-learning algorithm identified all similar pixels on the map (based on the vegetation’s height and density in the LiDAR data) and assigned those pixels similar carbon values. Pixel by pixel, the algorithm estimated carbon storage for the entire peninsula. Comparing the maps from 2001 and 2014, Smart found a 42% decline in carbon storage for areas that had made the transition from healthy forest to ghost forest.
That said, some landscape-scale analyses offer brighter possibilities. In a pair of 2022 studies, Kirwan and postdoctoral ecosystem ecologist Yaping Chen, also at VIMS, tracked the retreat of mid-Atlantic coastal forests using satellite images snapped since 1984 (1, 8). About 64,000 acres of the farmland and forest at low elevations along the coast (less than 2 meters high) had died off and converted to new marsh cover, which is likely to store less carbon, at least in the short run. But uphill and inland from those sites, the same warmer and wetter conditions also led to an increase in the growth of upland forests. Hence, across the entire landscape, there was a net increase in plant biomass, which means, as Kirwan puts it, “more carbon stored in the biomass of plants now than 30 years ago.”
The entire picture of how ghost forests will impact carbon emissions is still developing. Regarding research on ghost forests today, Smart, now a coastal scientist at The Nature Conservancy, says the new emphasis is on “the implications of these ghost forests for carbon storage and their role in the global carbon cycle.”
Regardless, there’s plenty of concern for the people and wildlife who live in areas affected by sea-level rise and associated forest loss, Chen says. Residents will have to figure out how to respond as these forests march inland.
When to Retreat
Studies that hint where ghost forests may expand next could help some towns proactively plan for managed retreat. For instance, Kirwan and Chen’s finding that the die-offs are concentrated less than 2 meters below sea level may help nonprofits like The Nature Conservancy identify specific areas, or even individual properties, where marsh migration is most likely.
Keen to find avenues to adaptation, Smart canvassed coastal neighborhoods in North Carolina to learn about property owners’ perspectives and reported her findings in a July 2023 article (9). She surveyed about 200 landowners, handing out paper questionnaires to anyone living on more than 5 acres. Smart wanted to know how interested these folks might be in different adaptation and mitigation strategies, including federal buyouts to incorporate their land into conservation easements (see “Managed retreat increasingly seen as necessary in response to climate change’s fury”). The marsh would essentially be allowed to reclaim their homes. Smart said she met less resistance than she expected from some groups—for instance, older people on forested land, whose children live in cities out of state. Farmers have been a harder sell and are more interested in salt-tolerant crops, she says.
Certainly not every homeowner will have to move. But even people without flooded foundations or overflowing septic systems may still consider leaving, as the slow march of landscape change makes coastal communities less livable.
“They talked about handing over their land to the next generation,” Smart says of the people she met. As seas rise and hotspots of forest loss spread, researchers and residents will have to grapple with what exactly they’re handing over.