In 2025, scientists deliberately flooded a Maryland forest with hurricane-level water to see how trees turn into a “ghost forest”

What happens when a healthy forest is repeatedly hit by the kind of flooding associated with extreme storms and rising seas? Scientists in Maryland decided to find out by bringing the flood to the trees. At the Smithsonian Environmental Research Centre in Edgewater, researchers created an experimental forest landscape where they could deliberately reproduce extreme freshwater and saltwater flooding. The project, known as TEMPEST, was designed to reveal what happens inside trees and soil before a woodland becomes a so-called “ghost forest,” a landscape where dead, often leafless trees remain standing as wetlands and marshes gradually take over. The experiment began in 2022, with researchers pumping about 80,000 gallons of water into each of two forest plots. By studying the trees' water movement, soil oxygen and salinity, scientists are tracking the earliest signs of stress that could eventually lead to forest collapse. The findings are offering a rare, real-time view of a transformation already occurring across Maryland's coast. As flooding becomes more frequent and persistent, the experiment could help scientists identify when forests reach a point from which recovery is no longer possible.According to the study ‘Short-term experimental flooding impacts soil biogeochemistry but not aboveground vegetation in a coastal forest,’ the experiment was established at the Smithsonian Environmental Research Centre, where researchers divided two 2,000-square-metre forest plots for controlled flooding. One received freshwater, while the other was flooded with saltier water drawn from the nearby Rhode River, allowing scientists to compare the effects of extreme rainfall with storm-surge-like saltwater exposure. A third forest plot was left untouched as a control. The scale of each artificial flood was deliberately extreme. During the first experiment in summer 2022, researchers delivered around 80,000 gallons of water to each plot in a single day, enough to create roughly six inches of water across the experimental area. That was comparable to the rainfall Hurricane Sandy delivered around Annapolis in 2012. Black irrigation pipes stretched across the forest floor, releasing water through hundreds of sprinklers. Around 200 sensors were installed on trees and underground to continuously measure how the forest responded. Researchers also collected soil gases and monitored changes long after the water disappeared.Flooding does not simply drown trees from the outside. Much of the damage begins underground, where saturated soil cuts off oxygen around the roots. Saltwater creates an additional problem because salt makes it harder for trees to extract water from the soil, effectively creating a physiological drought even when their roots are surrounded by water. The researchers found that the effects could persist after the visible flood had disappeared. During the saltwater treatment, soil conductivity, an indicator of salinity, rose to more than 19 times its normal level. Five days later, it was still more than ten times higher than before the flood. Soil oxygen also plunged dramatically in both treatments, falling below 1% of normal by the second day. That lack of oxygen matters because tree roots need oxygen to respire. When prolonged flooding seals the soil away from the atmosphere, roots can essentially begin to suffocate. Some pockets of oxygen-depleted soil also persisted even after overall oxygen levels had recovered.One of the most important discoveries was that the forest did not respond as a single organism. The three dominant tree species in the experimental plots, American beech, red maple and tulip poplar, reacted differently to the same flooding conditions. Scientists measured tree health partly through sap flux, which tracks how quickly water moves through the tree. It provides a useful indication of whether the tree's internal water-transport system is functioning normally. American beeches proved comparatively resilient, with sap flux increasing during both freshwater and saltwater flooding. Red maples and tulip poplars, however, showed declines. Tulip poplars were particularly vulnerable. Trees in the saltwater plot began turning brown and dropping their leaves roughly two months earlier than expected in autumn. The researchers said that detecting species-specific declines after only the second year of flooding was earlier than anticipated, highlighting how little is known about the exact thresholds at which coastal forests begin to fail. These differences could eventually help scientists predict which species disappear first as coastal forests become wetter and saltier.A “ghost forest” does not appear overnight. It is the visible endpoint of a much longer ecological transition. Rising seas, saltwater intrusion and repeated flooding can gradually make coastal soils too wet and saline for trees to survive. As the trees die, their trunks may remain standing for years while marsh vegetation moves into the former woodland. Maryland is already seeing this transformation outside the laboratory. Along the Chesapeake Bay, forests are increasingly being replaced by marshes as rising water and saltwater intrusion move inland. The U.S. Geological Survey reported in 2025 that forests lost to saltwater intrusion along Maryland's coast had expanded dramatically, with thousands of additional acres showing signs of decline. The broader scale of the problem was also highlighted by a 2025 Nature Sustainability study that mapped more than 10 million individual dead trees across the US Atlantic coast. The researchers found that more than six million of those trees were concentrated in low-lying forests, with salinisation emerging as a major driver of mortality. The TEMPEST experiment therefore gives scientists something that satellite images and aerial surveys cannot: a close-up view of the process as it happens. By deliberately recreating extreme floods, researchers can observe changes in soil chemistry, root conditions and tree physiology before a forest visibly becomes a ghost forest. And that may be the experiment's most important lesson. The dead trunks are not the beginning of a ghost forest. They are the final evidence of a transformation that began much earlier, beneath the soil, inside the roots and in the water surrounding them.