The research team, led by Bryan Shuman of the University of Wyoming, extracted fossil pollen from cores collected at Blanding Lake and Sunfish Pond, two glacially formed lakes sitting near the forest ecotone known as the Tension Zone. This narrow band marks the boundary between northern hardwood and conifer forests dominated by hemlock, beech and birch, and southern oak-hickory forests. By counting pollen grains preserved in layered lake sediment and dating them with radiocarbon measurements, the researchers reconstructed century-by-century changes in forest composition spanning the entire Holocene Epoch.

The hemlock decline at approximately 5,000 years before present is one of the most debated events in North American paleoecology. Explanations have ranged from disease to temperature shifts, but the new evidence points strongly toward climate. At Blanding Lake, temperature-sensitive lipids from bacterial cell membranes, called branched glycerol dialkyl glycerol tetraethers, revealed that mean temperatures above freezing rose by about 1.1 degrees Celsius, with a range of 0.9 to 1.3 degrees depending on the calibration, between roughly 5,150 and 4,880 years before present. The same sediment samples that recorded the warming also captured the hemlock collapse, and lake-level reconstructions showed a pronounced drop in water levels indicating severe drought at the same time.

What followed the hemlock decline was not a simple recovery. Instead, the pollen records document a sequence of abrupt declines in birch and beech, followed by surges in oak and white pine, taxa that had been important several thousand years earlier. At Sunfish Pond, where sampling resolution reached below 130 years between samples, the researchers identified five distinct forest phases: a birch and beech phase from about 5,100 to 4,800 years before present, an oak, hickory and chestnut phase from 4,800 to 4,250, a white pine, hemlock and ironwood phase from 4,250 to 3,800, an oak and white pine phase from 3,800 to 3,200, and finally a chestnut phase from 3,200 to 2,800 years before present.

To test whether these patterns reflected climate forcing or something else, the researchers built simple heuristic simulations of three idealized tree species responding to synthetic climate histories. Each pseudo-species carried a unimodal climate response curve representing its niche, and a lagging response time of 75 years approximated successional timescales in Appalachian forests. The simulations reproduced key features of the pollen records: sequences of abrupt declines as fast as the driving climate forcing, short-lived phases dominated by different taxa, and differences between sites despite consistent underlying low-frequency changes. Because decadal climate variability differed between simulated sites, the timing of abrupt changes shifted, mirroring the asynchrony observed in the actual pollen data.

The researchers then compared their forest records against alternative explanations. Succession following the hemlock decline, they argue, cannot explain the observed patterns. Known examples of post-disturbance succession, such as the chestnut blight at Linsley Pond in Connecticut or stand-replacing fire at Hidden Lake in Colorado, produced unidirectional trajectories of seral stages over 25 to 250 years, far shorter than the 340 to 610-year duration of individual forest phases near the Tension Zone. Moreover, the post-decline sequences showed repeated peaks and declines rather than a single progressive return to a pre-disturbance state.

Compound disturbances and ecological random walks fared no better as explanations. Disturbances like storms, wildfires and drought tend to have small or heterogeneous spatial extents, yet the rise and decline of beech, oak, hickory and pine played out similarly at multiple sites separated by hundreds of kilometers, from Pennsylvania to Massachusetts, New York and Ontario. Seed dispersal distances for these taxa rarely exceed 1,500 to 5,000 meters, limiting connectivity among similar forest changes across such broad regions. Random walks cannot easily explain such correlated patterning, but shifts in regional climate gradients can.

Instead, the forest histories conformed with the interaction of individualistic species climate niches and multiple scales of climate variation. Modern pollen-climate relationships show that hemlock is rare today in the warm, dry climates of southern Michigan, Ohio and western New York, while beech and birch tolerate those conditions. By 4,800 years before present, the regional climate shifted to warmer and drier conditions than today, favoring oak and hickory over beech. As the region ultimately cooled and moistened toward the present, other taxa like white pine and chestnut increased before substantial cooling favored birch and hemlock again after about 2,100 years before present.

These niche-driven changes had a spatial consequence: the Tension Zone itself shifted. Two oak maxima at 4,800 to 4,250 and 3,800 to 3,200 years before present mark northward movements of the northern-temperate forest ecotone during warm-dry phases coinciding with low hemlock abundance. With cooling after about 3,800 years before present, the ecotone shifted southward again as the pine belt and renewed hardwoods replaced the mid-Holocene oak forests. The long-term Holocene moistening trend further modified these patterns, shifting the region with less than 1,000 millimeters of annual precipitation westward out of Pennsylvania into the Ohio River Valley.

The implications for modern forests are sobering. If roughly one degree Celsius of warming triggered the hemlock collapse and a millennium of cascading forest turnover, then current anthropogenic warming will yield complex, rapid forest changes again. The study demonstrates that vegetation at landscape scales is highly sensitive to climate change, and that small differences in species climate relationships can produce individualistic responses even among taxa considered ecologically similar. Both hemlock and beech are mesic species, yet they responded to mid-Holocene climatic variability in fundamentally different ways. Understanding these niche differences, along with direct evidence of forest disturbances and plant physiological responses, will be essential for anticipating how today’s forests will reorganize as climates continue to shift.

Subject of Research: Paleoecology

Article Title: The role of the climate niche in repeated abrupt tree declines and ecotone dynamics in the Appalachian Mountains during the Holocene

Article References: Shuman, B. N., Stefanova, I., Stefanescu, I. C., Ray-Cozzens, T., Johnson, J., & Nelson, D. M. (2026). The role of the climate niche in repeated abrupt tree declines and ecotone dynamics in the Appalachian Mountains during the Holocene. Frontiers of Biogeography, 19, Article e150642. https://doi.org/10.21425/fob.19.150642

Image Credits: AI Generated

DOI: 10.21425/fob.19.150642

Keywords: paleoecology, climate change, Appalachian forests, hemlock decline, fossil pollen, Holocene, biogeography, ecotone, role, climate, niche, repeated