Duke’s biggest, oldest laboratory—past, present and future

Duke’s Biggest, Oldest Laboratory—

Past, Present and Future

By Kara Manke

The Robson gravestones are not easy to find. As we wade through the dense underbrush of the Duke Forest, skirting an open area overtopped by power lines, Nicolette Cagle points out the landmarks she uses to find her way.

“At the second transmission pole, I know I am close,” said Cagle, a lecturer at Duke’s Nicholas School of the Environment, whose long blonde hair is tucked inside the hood of a black rain coat. “And then, we just turn and walk deeper into the Forest.”

On this rainy day in February, the three slate gravestones blend in with the browns and tans and grays of the forest floor. But up close, the engraved words still stand out clear as ever.

WILLIAM ROBSON 1783 to 1871 lies in the middle, flanked by his wife ANN ROBSON 1795 to 1872, and his oldest son JOHN ROBSON 1819 to 1842. Nearby lie the likely gravesites of three or four enslaved people they owned, marked only by plain stones half buried by soggy leaves.

The Robson family ran a mill and farm near New Hope Creek in what is now the Korstian division of the Duke Forest, halfway between Duke’s campus and Chapel Hill. A few hundred yards from the gravesite lies the foundation of their old home, a deep depression in the ground, and a few hundred yards beyond that, along the banks of the creek, the stone remains of the Robson Mill that ground grains.  

The Robson gravestones, located in the Korstian division of the Duke Forest, mark the graves of three members of a prominent family who ran a farm and grist mill along the banks of New Hope Creek during the 1800s.

The Robsons were just one of many families that divided up the land of what would become the Duke Forest, though their relative wealth was the exception rather than the rule. Most inhabitants were yeoman farmers, eking out a living on small parcels of land growing corn, wheat and tobacco, and too poor to own slaves.

The degradation of the soil from unsustainable agricultural practices combined with economic collapses of the Civil War and Great Depression drove most of them off the land, providing cheap, fallow land for Duke University to buy up in the 1920s.

Though the landscape appears wilder now, hints of human habitation remain littered all over the Duke Forest. Cagle points them out as we explore the ruins of the Robson plantation: trees grow out of mysteriously wavy ground, the leftovers of old agricultural furrows; rectangular slumps in the dirt mark the locations of collapsed gravesites; non-native plants like daffodils and flowering vinca, popular garden fixtures in the 1800s, grow near the foundations of old homesites.

But perhaps the most glaring echo of the Forest’s agricultural past soars above our heads. The towering loblolly pines are now a fixture of the Duke Forest, and the Southern Piedmont as a whole. These pines are a true sign of a resurgent forest.

“If you see a pine canopy, you probably can bet the land is an old cultivated field or pasture,” said Daniel Richter, a professor of soils and forest ecology in the Nicholas School of the Environment.

Historical aerial photos show the forest cover near New Hope Creek in the Korstian division of the Duke Forest in 1938, 1955 and 1975.
Aerial photo of Duke Forest 1975


Aerial photo of Duke Forest 1955


Aerial photo of Duke Forest 1938


“Some folks may think the Forest looks pristine,” said Judd Edeburn, who directed the Duke Forest for 36 years, from 1978 to 2014. “But it has undergone dramatic changes. It was a very different landscape 80 years ago—very, very different.”

Archaeologists have evidence that humans first inhabited the Piedmont—the hilly plateau region that sits between the Appalachian Mountains and the Carolinas’ coastal plains—more than 10,000 years ago, arriving shortly before the end of the last ice age. As temperatures warmed, cold-tolerant conifer forests slowly transitioned to deciduous forests. First peoples regularly burned the land to drive game, a practice that favored fire-tolerant species such as oaks and hickories.

The arrival of Europeans and Africans in the 1740s drastically changed the landscape as farmers cleared large swaths of the piedmont forest for agricultural fields and pastureland. But soil degradation and worsening economics eventually chased these farmers off the land.

Throughout the 20th century, trees, vegetation and wildlife have slowly reasserted their grip. The pines are tall and thick, and the shade is deep. Predators like bobcats and coyotes have returned, along with more than 30 other species of mammals, 180 bird species and at least 50 species of amphibians and reptiles.

Standing Still, but Always Changing

Like all living beings, forests do not appear fully formed. They must go through stages of growth, starting at birth and bumbling their way through to adulthood. The trees and plants of a young forest do not match those of a mature forest, and neither do the soils in which they grow, nor the wildlife that feeds on them.

But it’s hard for us humans to imagine—much less study—changes that can take more than a lifetime to occur. For forest researchers to understand the full lifespan of these complex ecosystems, they must have the foresight to launch observations and experiments that they may not live to see completed.

Clarence Korstian, the first director of the Duke Forest and founding dean of the School of Forestry, meticulously documented the return of trees to the countryside starting in 1933. In some areas he let the trees grow in naturally, while in others he actively planted and managed the trees.

As his data passed from generation to generation, researchers have added new knowledge about the Forest’s growth, its ecosystems and soils, and the impacts of people and climate change on the land.

Their work transformed the 7,052-acre Duke Forest into the largest research lab on campus. It also created perhaps the longest-running experiment at Duke: an 85-year study into the childhood and adolescence of a forest.

A stand of young loblolly pines were carefully documented and numbered by first Duke Forest director Clarence Korstian.

Many of the trees in the Forest used to bear a rather odd marking: numbers in faint whites or blues, painted at eye level on the rugged trunks.

The enigmatic numbers were in place in the early 1970s when Norman Christensen, now professor emeritus in the division of environmental sciences at the Nicholas School, and Robert Peet, now professor emeritus of biology at the University of North Carolina, began their collaborative studies of the Forest.

“Almost anybody walking in the Forest would have noticed the different places where trees had numbers painted on them,” Christensen said.

But nobody knew what they meant, not even Fred White, who was then the director of the Duke Forest.

That all changed when Benjamin Jayne, dean of the Forestry School at Duke, was cleaning out some filing cabinets and stumbled upon pages and pages of Korstian’s old notes. The documents listed species, tree numbers and tree-stand locations, all in fastidious detail.

“It was immediately obvious that these corresponded to the trees that we were looking at out in the woods,” Christensen said. “It was like finding a Rosetta Stone.”

It was immediately obvious that these corresponded to the trees that we were looking at out in the woods. It was like finding a Rosetta Stone.

When Korstian and soil scientist Theodore Coile began their studies in the 1930s, they had a practical question in mind: How to best manage the growth of this young forest to maximize the quality and quantity of its lumber?

“This was a landscape that was heavily deforested 200 years ago, farmed, and then land was abandoned after the Civil War,” Christensen said. “Most of the forests of the Piedmont were young forests, and this represented enormous economic opportunity. But there was no science regarding how one should go about managing forests of this kind.”

The Forest of 1933 sported a patchwork of land cover. Wild shrubs and pine seedlings dotted the furrowed fields of recently abandoned farmland. Along their borders, stands of young short-leaf and loblolly pine grew from fields left behind at the turn of the century. And in low-lying and hilly areas, hardwoods still dominated, but were widely used for timber and livestock grazing.

To track the growth of this fledgling forest, Korstian and Coile established 87 permanent sample plots, tracts of land varying in size from one-tenth of an acre to one quarter acre. In each plot, they noted the species of each tree, its height and the diameter of its trunk. Sample plots each received a different treatment: in some plots, they planted a lot of trees and let them thin themselves out. In others, they planted only a few trees and let them grow fast. In many, they let the pines establish themselves naturally, but adjusted the density.

  • Text Hover
In the early 1930s, the Duke Forest was a patchwork of young pine stands and fallow fields marked with deep furrows.

Every five years, the team returned to each sample plot and detailed the changes that had taken place, measuring the size of each tree stem.

By 1967, many of the forestry questions had been answered, and the data were filed away. But when the notes resurfaced in the 1970s, forest ecologists including Christensen and Peet had a raft of new questions to ask of the trees. 

In addition to documenting individual trees in the permanent sample plots, Korstian also noted the dominant tree types—pine, hardwood or a mixture—in each former agricultural plot. The dominant tree types were re-characterized every 10 to 15 years, illustrating how the forest cover has changed over time. Yellow plots are dominated by pine, dark green plots are dominated by hardwoods, and light green plots are a mixture of pines and hardwoods. Clear areas are not forested.
  • Text Hover
  • Text Hover
  • Text Hover
  • Text Hover
  • Text Hover
  • Text Hover
  • Text Hover
  • Text Hover

Christensen and Peet were primarily interested in a phenomenon called secondary forest succession, the process by which forests re-establish themselves after land-clearing disturbances like fires, agriculture, pests or storms.

“The Piedmont is an interesting area because it did have this very striking transition from agriculture to forest in the early 20th century,” said Jim Clark, Nicholas Professor of Environmental Science at Duke’s Nicholas School of the Environment, who has been monitoring the growth and reproduction of trees in the Forest since the late 1990s. “So much of the Piedmont, and much of the eastern U.S., is in a state where forests are regenerating following farming or other forms of clearing.”

Understanding forest succession isn’t just important for understanding the ecology of the Duke Forest, but can shed light on the forests in the Piedmont as a whole, Clark said.

Photographs taken in 1935, 1950, 1983 and 2016 document the transition of permanent sample plot 40 from a dense stand of young loblolly pines to mature forest. Scroll down to reveal each photo.

A simple model of forest succession—based heavily on research conducted in the Duke Forest—was described in a 1942 paper by naturalist Henry J. Oosting.

Abandoned farm land is initially colonized by weeds and small shrubs, he wrote. Within a few years, pine seeds drift in on the breeze and take root, soon growing into dense stands of miniature trees packed together like soldiers. While trunks swell and branches stretch outward, smaller and weaker trees are starved for light and nutrients and slowly die out.

As the forests of loblolly and short-leaf pine mature, shade-tolerant hardwood species slowly grow up in the understory. When a pine tree collapses from age, wind, or competition from larger trees, opening a hole in the canopy, these hardwoods thrive under the sudden sunlight. Over the course of decades, the young pine forest slowly transforms into a forest dominated by hardwood, Oosting concluded.

While this model “is certainly a nice cartoon for a textbook,” Peet said, the truth is a lot more complicated.

After unearthing Korstian’s records, Christensen and Peet combed through the data, identifying as many of the old sample plots as they could and updating the data to map the locations of the surviving trees.

With the basic framework of succession in hand, the team wanted to nail down the details. In particular, they wanted to know how trees “thin” themselves—why some survive to make it to the canopy, while others die out.

“We wanted to understand the population biology of trees, to understand why trees die, when they die, why trees get established and what are the things that make it more likely for one tree to get established and another one to die,” said Christensen, who retired from Duke in 2014.

While the textbook models of succession generally held, they found some surprising deviations that showed the continuing impacts of humans on the land. In some ways, the course of succession has changed.

Hundreds of years ago, the landscape was dominated by oaks and hickories, Peet said. But as the Duke Forest has matured from pine to hardwood, these species are not returning in the same numbers. Instead, the regrowth has been dominated by other hardwood species, particularly red maple and beech.

This evolution is likely being driven by changes in land management, Peet said. Indigenous peoples and early settlers would have grazed livestock and periodically burned the forest to maintain open land. “The oaks and hickories are adapted to resprout when fire or animals chip them off and then they spring back from their roots and grow up,” Peet said.

But burning and grazing are no longer in favor, especially in Eastern forests now flanked by residential and urban areas.

“Basically we are slowly transitioning to an entirely different kind of forest,” Peet said.

Basically we are slowly transitioning to an entirely different kind of forest.

Christensen and Peet further enriched the data by creating 230 new sample plots to add to Korstian’s collection in the late 1970s. In these plots, they documented not only woody plants like trees but also the shrubs and herbaceous plants that crowd the forest floor.

They soon noticed a rather unexpected pattern. Peet, who managed plots in the Korstian division of the Forest, was consistently counting more species of plants in his plots than Christensen, who managed plots in the Durham division.

“I got a phone call from Bob and he said, ‘There is a problem here—we’ve each sampled about 30 plots, and I’m finding at least 10 more species in each of my plots on average than you are,’” Christensen said. “He said, ‘You guys just aren’t looking hard enough!’”

Christensen looked up data from soil samples and found that the pH of the soil in the Durham division was generally lower or more acidic than in the Korstian division. More acidity usually indicates less fertile soil.

They consulted geologic maps of the region and found the Duke Forest sits atop two different types of bedrock. To the west lies the Carolina Terrane, a wide shelf of rock that supports the rolling hills of the Piedmont. To the east lies the Triassic Basin composed of sedimentary rock like sandstone and siltstone.

“The differences in rock types can explain an enormous amount of the diversity in the forest, and that was something nobody knew before,” Christensen said.

These different soil types also affect how forests rebound from disruption, and their detailed soil analysis gave the pair an even deeper insight into the process of forest succession.

“Different things grow back at different rates on different soils,” Christensen said. “With those plots we have been able to look at successional change in a much more fine-grained way.” 

Students in Daniel Richter’s soil resources laboratory gather around while he describes the different layers that appear in a soil core sample.

History Hidden Underfoot

Daniel Richter, usually soft-spoken, becomes animated when talking about the history of the soil that lies beneath our feet.

For more than 30 years, Richter has studied the interactions between young Piedmont forests and the soils that sustain them. Though he now conducts most of his research at the Clemson Experimental Forest in South Carolina—a location whose history and ecology parallels that of the Duke Forest—his studies are “built directly on the work of Coile, Korstian, Peet and Christensen,” he said.

The Carolina Terrane and the Triassic Basin have been weathering away for millions of years, transforming the rocks into these thick layers of dirt and clay. But the arrival of agriculture brought rapid changes to the upper layers of the soil, Richter said.

On a chilly January day, Richter crouches over a soil core sampled from the Durham division of the Duke Forest. He points out the different layers that we see, some of which have been accumulating for eons. On the top lies the dark and loamy O-horizon, chock-full of nutrients and decayed organic matter. Below that, the sandy grey and tan layers of the A- and E-horizons. And finally, the B-horizon, a thick burgundy layer of iron-rich clay.

When farming stripped the land of trees and shrubs, the soil lost the protective carpet of plants, leaves and pine needles that hold the fertile O- and A-horizons in place. Heavy rains quickly washed topsoil downhill, leaving behind clay-rich surfaces on the uplands, and filling the lowland floodplains with sediment. Soil researchers estimate that throughout the Piedmont, the land lost an average of 5 to 10 inches of surface soil from agriculture, and that most floodplains are filled with legacy sediments half a meter to two meters in depth.

“These numbers are kind of misleading because they are averages,” Richter said. “There are many places that are more severely degraded, some on the Duke Forest.”

Ninety years later, the Forest is slowly rebuilding much of the soil organic matter—but it is still far from the end of this so-called rebuilding process, Richter said.

“We know from experiments that much of the soil organic matter under a 90-year old pine stand is definitely due to this forest,” Richter says. “But you can also get a sense that it could take many more centuries to really rebuild it.” 

Carbon dating shows that this pine stump, located in a floodplain in Browning’s Branch on the Duke campus, first sprouted in the 1740s and died in the 1840s. Richter claims that erosion from higher-lying fields buried former floodplains, raising the level of the ground by a meter as a result of Piedmont agriculture and and burying the stump in a thick layer of soil. He calls this soil “legacy sediment soil.”

Though agricultural practices are no longer kicking up dirt, humans are still accelerating the flow of topsoil from uplands to floodplains, Richter said. Development like roads and houses prevents water from percolating into the ground, leading to heavier floods and faster erosion.

“Wherever people live, we tend, as a course of how we live, to accelerate erosion on the upland. And where do those sediment particles go? They go downstream,” Richter said. “We are inadvertently changing valley morphology. It’s just a fact of human occupation.” 

We are very inadvertently changing valley morphology. It’s just a fact of human occupation.

This map comes from research by Duke student Emma Maschal, who was studying how human settlement affects waterways and ecology.

Wilder Again, but Never the Same

Even as the Forest approaches adulthood, it will never truly return to its primitive state. Urban development, climate change, invasive species and the overabundance of deer now drive the trajectory of the Forest’s growth, reshaping the landscape into something entirely new.

Duke scientists are now using the Forest as a laboratory to understand how human activities will continue to influence the fate of the wild areas around us.

“There are many examples of how the Forest provides teaching and research opportunities, not just in ecology, but across a variety of disciplines. We have studies that span from the outer reaches of our atmosphere all the way down to the water flowing underneath the Forest floor,” said Sara Childs, a Nicholas School graduate who has been director of the Duke Forest since Judd Edeburn retired in 2014. In 2016, for example, the Forest was home to 48 research projects and welcomed 1,375 students from kindergarten through graduate school.

Graduate student Chase Nuñez is part of a team, led by Jim Clark and lab manager Jordan Luongo, that has been using sites in the Duke Forest to study how climate change will affect masting—or seed production—of trees. Several times a year, he sorts through leaves and animal poop to identify tree seeds that have fallen into his one of his “seed traps”—square, meter-wide PVC baskets that Nuñez buries in the Forest and covers in mesh.

He places these traps in grids, carefully measuring their locations in relation to nearby trees. With the help of some triangulation and mathematical modeling, Nunez can track which seeds came from which tree. The data, which the team has been gathering for more than 25 years, help them monitor how seed production varies from year to year and across locations in the Duke Forest and throughout the Eastern United States.

The team has recently added camera traps to capture different creatures in the act of seed foraging, in hopes of understanding how the variability in seed numbers also changes the behavior of the wildlife that feast upon them.

“If we are trying to preserve entire ecosystems full of critters and trees, it is imperative that we understand exactly how this variability in masting is actually influencing these consumer communities,” Nunez said. 

  • Text Hover
  • Text Hover
  • Text Hover
  • Text Hover
Duke researchers installed camera traps to catch creatures in the act of foraging for seeds. They plan to study how tree seed production influences hungry wildlife.

In a study led by Duke biology professor Emily Bernhardt, researchers are using forest waterways to study how water quality changes when streams flow through urban areas.

Another study, spearheaded by Dean Urban, professor of landscape ecology at Duke’s Nicholas School of the Environment, examines how urban heat islands change the ecology of the Forest.

Clark, a scientist on the Duke FACE experiment that examined the effects of elevated carbon dioxide levels on tree growth, predicts that the next 20 years will bring more changes to the Forest, though the exact nature of these changes may be hard to predict.

“This is the period where we will start to see climate change having an effect,” Clark said. “This is probably going to mean a shift in some of the composition of the Forest, not just trees but probably also wildlife.” 

Every spring, following the first warm rain of the season, hundreds of spotted salamanders gather in the Duke Forest’s gate 9 pond to do as the birds and the bees do—mate. Nicolette Cagle and her class are investigating what all the salamanders are up to when not busy procreating; where they come from, and where they go. Answering that question is key to helping the species survive as roads and urban development enclose the wild areas of the Forest.

The permanent sample plots established by Korstian and Coile, and later Christensen and Peet, continue to generate valuable data about the ever-changing Forest.

After Hurricane Fran decimated parts of the Forest in 1996, Peet and his students resurveyed the sample plots that were still intact—about 90 in all. They surveyed them again in 2009. Besides the continued transition from oak and hickory to a forest dominated by maples, beech, sweetgums and tulip trees, the plots revealed other troubling changes in plant life.

“One thing that emerged was a drop in the number of plant species pretty much uniformly throughout,” Peet said.

The loss in diversity is likely due to the influx of deer, invasive species and human recreation, Peet said. Plants that are tasty to deer, like lilies and orchids, are nearly gone, while invasive species like Japanese stiltgrass and Russian Olive are thriving.

Childs and her team are now exploring whether forest management practices can reverse some of these trends. In one experiment, she said, they are using prescribed burning to try to encourage the resurgence of oaks and hickories, whose mast—large edible acorns and hickory nuts—play a major role in forest ecosystems.

“Oaks and hickories are dominant mast producers and drive a major food chain system in the Forest,” Childs said. “So, if those trees are transitioning out of the Forest, what does that mean for the future?”

“I have seen dramatic changes over the duration of my interaction with the Forest,” Peet said. “And it will certainly keep changing.”

Like the Robsons, we’ve all left some indelible marks on the Duke Forest. Its future, like its past, will remain entwined with the lives of the people who live alongside it.

Story by Kara Manke
Photos by Megan Mendenhall
    and provided by the Office of the Duke Forest
Design by Sam Huntley
Learn more about Duke Forest…

  • Explore Duke Forest
Explore Duke Forest and see what makes Duke’s living laboratory special.
  • Office of the Duke Forest
Learn more about teaching, research, management and events going on now.
  • How to Enjoy Duke Forest
See some tops picks for recreation activities open to the public.