Deer Browse in Southeastern Pennsylvania: Why Piedmont Woodlots Aren't Regenerating

We assessed a three-acre forest fragment in Chadds Ford earlier this year. The canopy was in decline from beech leaf disease, the understory was stiltgrass and browse line, and there was essentially nothing coming up to replace what was dying.

Three acres is not much of a forest. But the failure sequence on that site is the one I keep seeing across Bucks, Delaware and Chester counties, and it's worth laying out, because the order of events matters.

Beech is one of the species that persists under heavy deer pressure. It's browse-tolerant, it root-sprouts, and where chronic browse has run long enough it ends up over-represented in the residual understory. Kain and colleagues documented this at a sixty-year exclosure in Pennsylvania, where browsing had produced a stand dominated by beech and striped maple while the density of trees capable of reaching the canopy dropped by 85 percent.

So on a fragment like the Chadds Ford site, beech is partly there because of the deer. It's the survivor. And beech leaf disease, caused by the nematode Litylenchus crenatae mccannii, is now killing the survivor while browse continues to prevent anything else from establishing underneath it.

That's the whole problem in miniature. The disease is what the client called about. The deer are why there's no answer.

The numbers we borrowed

Almost all of Pennsylvania's foundational deer research came off the Allegheny Plateau. Tilghman (1989) stocked enclosures in the northwest at densities from zero to 80 deer per square mile and found that after five years, seedlings in clearcuts at the lowest densities were nearly twice as tall as those at the highest, with diversity collapsing toward nearly pure black cherry at the top end. She recommended a ceiling near 18 per square mile.

Horsley, Stout, and deCalesta (2003) followed with a ten-year experiment at 4, 8, 15, and 25 deer per square kilometer and found species richness declining linearly with density across every silvicultural treatment.

Those numbers came out of large contiguous forest blocks with active timber programs, real winters, and hunters who could get to the deer. They became our management framework anyway.

Now look at what Abrams and Johnson (2012) measured at Valley Forge in paired plots established in 1992 and resurveyed in 2010: densities running 70 to 93 deer per square kilometer over the preceding decade. Call it 180 to 240 per square mile. Three to four times the highest density Horsley's crew ever tested experimentally.

The understory they described should look familiar. Roughly 500 saplings and 10,000 seedlings per hectare, dominated by red maple and blackgum rather than the oaks standing over them. In the tulip poplar stand, stiltgrass covered 72 percent of the ground.

Valley Forge sits on the same soils, with the same species pool and the same farm-to-forest land use history as the properties I work on. Its neighbors are the same subdivisions.

What a decade of culling actually buys

This is worth sitting with, because it's the response you'll get from any preserve board or township that already runs a program.

Miller and colleagues (2023) assessed regeneration across about forty eastern national parks using twelve years of plot data. Valley Forge and Hopewell Furnace are both in that network. Parks with active deer management for more than a decade, Valley Forge included, still landed in the imminent regeneration failure category, mostly for lack of saplings.

Sharpshooters, federal funding, no access constraints, ten-plus years of it. Seedlings came back. Saplings didn't, or not enough of them. That's the honest ceiling on what reduction alone does once a site is far enough gone.

Why we're a harder case than the plateau

Winter doesn't limit anything here. Forage is subsidized year-round by crop stubble, hay, landscaping, and an enormous quantity of forest edge, which means the woodlot isn't carrying the herd. The landscape is, and the deer come into the woods to eat seedlings on top of a diet they get somewhere else. Density objectives calibrated to forest carrying capacity don't mean much under those conditions.

Then there's access. Kilpatrick, Labonte, and Barclay (2011) tracked suburban deer with hunt-season home ranges from about 17 to 121 hectares, core areas hugging roads, spread across many separate owners. In Delaware County that's a home range crossing a dozen properties, several posted, one of them a township park with a discharge ordinance, plus a state road. Nobody's intentions matter much against that geometry.

And where access does exist, hunting may not close the gap alone. Blossey and colleagues (2019) used standardized red oak seedlings to measure browse pressure across sites under different regimes and found that neither sterilization nor recreational hunting reduced deer or their impacts sufficiently. The protected seedlings did fine everywhere. Growing conditions weren't the problem.

Stiltgrass is our recalcitrant layer

On the plateau, chronic browse produces hay-scented fern. Here it produces stiltgrass, with barberry, wineberry, and lesser celandine behind it.

Averill and colleagues (2016) ran preference trials and found consistent avoidance of garlic mustard, Japanese barberry, and stiltgrass. Knight and colleagues (2009) found more bare ground and more of both garlic mustard and stiltgrass in deer-access plots in a Pennsylvania forest, with natives smaller and fewer of them flowering. Averill's later synthesis across 23 sites (2018) found the same pattern regionally, driven largely by those two species.

The closest thing to a purpose-built study of our conditions is across the river. Morrison (2017) ran factorial deer-exclosure and stiltgrass-addition experiments in suburban New Jersey Piedmont forests, same province, same suburban matrix, same two stressors stacked. Deer were generally the stronger factor.

So the invasive line and the deer line on a management plan aren't two problems. Spray the stiltgrass without touching browse and you've reset the clock.

Fire rewards what's already there

For anyone thinking about prescribed fire in an oak stand around here, where burn windows are narrow and smoke receptors are close, every burn has to earn its slot.

Perles, Niu, Ruth, and Gibbons (2021) crossed prescribed fire with deer exclosure fencing at two oak sites. Where oak seedlings were already well stocked, burning promoted oak over maple. Where chronic overbrowsing had left a depauperate understory, burning favored maple at oak's expense. Same treatment, opposite outcome, decided by what was on the ground beforehand. Fencing on its own more than doubled tree seedling cover.

Nuttle and colleagues (2013) found the mechanism: fire increased the share of top-killed saplings that sprouted, browsing decreased it, and browsing after fire pushed the sprouting community toward browse-tolerant species. Their northern red oak saplings weren't fire tolerant and didn't throw tall sprouts.

On a dry chestnut oak ridge in northern Chester County where oak advance regeneration has hung on, fire is the right call. On a browsed mesic slope where the seedling bank is already red maple and stiltgrass, you're releasing the competitor.

The lag

Nuttle, Ristau, and Royo (2014) found understories still reflecting past deer densities twenty years after those densities ended. Pendergast and colleagues (2016) excluded deer for eleven years and watched density rebound while diversity didn't, because species that are locally gone have nowhere to come back from. Miller's group reads the exclosure literature as suggesting forty to seventy years for full recovery.

The seedling bank depreciates and doesn't re-accrue on its own. On a property that changes hands every twenty years, deferral pushes the payoff onto somebody else and raises their bill.

Where I'd push back on myself

Begley-Miller and colleagues (2024) ran seven years of fencing, herbicide, and lime treatments in central Pennsylvania oak-hickory. All three factors interacted. Where initial seedling abundance was high and lime raised pH above 4.6, unfenced controls performed about as well as fenced-only plots, and competing vegetation wasn't the primary constraint. They argue acid deposition deserves more attention, since soil amendment outlasts a fence.

But Long, Brose, and Horsley (2012) tested lime against fencing across five Pennsylvania oak shelterwoods for six years and got no significant growth response to lime at any rate. The only significant response was fencing: 32 cm inside versus 17 cm outside by the end.

Both are Pennsylvania studies and they disagree. My read is that soil chemistry is real, it binds harder on some sites than others, and it doesn't substitute for dealing with browse. On post-agricultural Piedmont ground with a plow line still in the profile, both hypotheses are live, and the only way to know which is limiting your stand is to test it.

What to do about it

Measure browse instead of guessing at it. Blossey's sentinel method, planting standardized red oak seedlings and scoring them, works even where the understory is too far gone to read by composition. A regeneration inventory broken out by height class will tell you more about a stand's future than any deer count.

Look at advance regeneration before writing a release treatment of any kind. Harvest and fire both reward what's there, and neither one creates it.

If you fence, fence the regeneration opening rather than a demonstration plot. A 20-by-20 exclosure is a good persuasion tool for a supervisors' meeting and a useless silvicultural intervention. The fence stays up until stems clear browse height, which here means most of a decade.

Use the mechanisms that exist. DMAP for qualifying landowners, special deer control permits through the Game Commission for municipalities and institutions, controlled archery where firearms discharge is constrained. None is sufficient alone.

Talk to your neighbors. Given those home range numbers, an eight-acre owner acting alone is managing a fraction of one animal's territory.

And set the timeline honestly with whoever's paying. A client expecting visible change in two seasons will pull the plug right around the time it starts working.

On the Chadds Ford site, the beech will likely come out of the canopy over the next several years regardless of what anyone does about the nematode. What happens in the gaps is still an open question, and it depends almost entirely on decisions about browse that need to be made before the canopy opens rather than after.

References

1. Abrams, M.D., and S.E. Johnson. 2012. Long-term impacts of deer exclosures on mixed-oak forest composition at the Valley Forge National Historical Park, Pennsylvania, USA. Journal of the Torrey Botanical Society 139(2):167–180.

2. Averill, K.M., D.A. Mortensen, E.A.H. Smithwick, and E. Post. 2016. Deer feeding selectivity for invasive plants. Biological Invasions 18:1247–1263.

3. Averill, K.M., et al. 2018. A regional assessment of white-tailed deer effects on plant invasion. AoB PLANTS 10(1):plx047.

4. Begley-Miller, D., et al. 2024. Complex interactions of deer herbivory, soil chemistry, and competing vegetation explain oak–hickory forest tree regeneration in central Pennsylvania, USA. Canadian Journal of Forest Research. doi:10.1139/cjfr-2024-0034

5. Blossey, B., P. Curtis, J. Boulanger, and A. Dávalos. 2019. Red oak seedlings as indicators of deer browse pressure. Ecology and Evolution 9:13085–13103.

6. Carta, L.K., Z.A. Handoo, S. Li, M. Kantor, et al. 2020. Beech leaf disease symptoms caused by newly recognized nematode subspecies Litylenchus crenatae mccannii (Anguinata) described from Fagus grandifolia in North America. Forest Pathology 50(2):e12580.

7. Horsley, S.B., S.L. Stout, and D.S. deCalesta. 2003. White-tailed deer impact on the vegetation dynamics of a northern hardwood forest. Ecological Applications 13(1):98–118.

8. Kain, M., L. Battaglia, A. Royo, and W.P. Carson. Over-browsing in Pennsylvania creates a depauperate forest dominated by an understory tree: results from a 60-year-old deer exclosure. USDA Forest Service Treesearch record 40939.

9. Kilpatrick, H.J., A.M. Labonte, and J.S. Barclay. 2011. Effects of landscape and land-ownership patterns on deer movements in a suburban community. Wildlife Society Bulletin 35(3):227–234.

10. Knight, T.M., J.L. Dunn, L.A. Smith, J. Davis, and S. Kalisz. 2009. Deer facilitate invasive plant success in a Pennsylvania forest understory. Natural Areas Journal 29(2):110–116.

11. Long, R.P., P.H. Brose, and S.B. Horsley. 2012. Responses of northern red oak seedlings to lime and deer exclosure fencing in Pennsylvania. Canadian Journal of Forest Research 42(4):698–709.

12. Miller, K.M., et al. 2023. Overabundant deer and invasive plants drive widespread regeneration debt in eastern United States national parks. Ecological Applications 33(4):e2837.

13. Morrison, J.A. 2017. Effects of white-tailed deer and invasive plants on the herb layer of suburban forests. AoB PLANTS 9(6):plx058.

14. Nuttle, T., A.A. Royo, M.B. Adams, and W.P. Carson. 2013. Historic disturbance regimes promote tree diversity only under low browsing regimes in eastern deciduous forest. Ecological Monographs 83(1):3–17.

15. Nuttle, T., T.E. Ristau, and A.A. Royo. 2014. Long-term biological legacies of herbivore density in a landscape-scale experiment. Journal of Ecology 102:221–228.

16. Pendergast, T.H., S.M. Hanlon, Z.M. Long, A.A. Royo, and W.P. Carson. 2016. The legacy of deer overabundance: long-term delays in herbaceous understory recovery. Canadian Journal of Forest Research 46:362–369.

17. Perles, S.J., X.M. Niu, A.D. Ruth, and L.D. Gibbons. 2021. Initial conditions influence effects of prescribed burns and deer exclosure fences on tree regeneration and understory diversity in Appalachian oak-dominated forests. Forest Ecology and Management 495:119353.

18. Tilghman, N.G. 1989. Impacts of white-tailed deer on forest regeneration in northwestern Pennsylvania. Journal of Wildlife Management 53(3):524–532.