White-tailed deer (Odocoileus virginianus) are among the most impactful herbivores in the eastern United States. Legacy forest effects, those accrued from intense herbivory over time, manifest as low seedling regeneration, high cover of plant species that are infrequently browsed by deer, presence or expansion of nonnative or invasive plant species, few herbaceous species, and diminished capacity for recovery. Interfering vegetation (that is, species that increase in cover and density due to avoidance by deer, such as American beech sprouts, Pennsylvania sedge, and hay-scented fern) increase competition for light and hinder recruitment of trees into the forest canopy.

The lower Hudson Valley in New York has been heavily browsed by white-tailed deer since the early 20th century. The region has some of the lowest tree regeneration rates in New York State as a result of deer browsing and subsequent increases in interfering vegetation. The U.S. Geological Survey and the State University of New York College of Environmental Science and Forestry studied sites where deer hunting is permitted (case sites) and nearby sites where hunting is currently prohibited (control sites) to assess and identify forest structure and composition differences.

Instead of using deer exclosures, which are time-consuming and expensive to install and maintain, we used a case-control study because such studies are well-suited to effects with long latency and rare outcomes. Case-control studies seek to describe the relation between an outcome of interest (in this study, forest understory recovery from chronic herbivory) and forest condition. We inferred recovery by comparing these characteristics on adjacent sites in the lower Hudson Valley with similar forest communities and land uses but different deer population management histories. Case plots were on lands where deer management has taken place annually for several decades. Control plots were on lands where deer populations have not been consistently managed to lowered abundance. We accounted for differences in forest recovery not attributable to deer by first matching case and control plots along several important environmental gradients (slope, aspect, elevation, moisture, canopy openness). By controlling for these gradients, we looked for associations between measured forest conditions and deer herbivory reduction through population management.

We surveyed more than 200 plots in upland forest types across case and control sites where we assessed forest condition by estimating density (number per unit area) and composition and cover (percent) of important vegetation constituents in ground, shrub, subcanopy, and canopy layers of the forest. We recorded 37 tree species, 22 shrub species, 57 herbaceous species, and 19 species of grasses and sedges in our plot surveys, including a number of nonnative and invasive plants. We also estimated the ages of a number of common canopy trees by counting rings from cores extracted from individual stems.

Effects of more than 100 years of chronic deer browsing manifested in low herbaceous ground cover and little to no tree recruitment (saplings) on lands without deer management. In contrast, sustained deer management resulted in forests with conditions that indicated substantial recovery from chronic herbivory in the ground, shrub, and subcanopy layers. Sites with ongoing deer management exhibited greater ground cover of tree seedlings and herbs and less ground cover of interfering vegetation and nonnative species. The well-developed sub-canopy layer of small trees, saplings, and tall shrubs on sites with deer management indicates a high potential for sapling recruitment to the canopy of the future forest.

Of the 25 subcanopy trees sampled on control sites, most were more than 100 years old, indicating little to no regeneration in areas sampled for more than 100 years. The forest canopy, a relic of land uses of bygone days, requires a source of young trees to replace itself as older trees die. Without an abundant layer of young trees in the subcanopy, a forest cannot be sustained over time. Reduction in deer herbivory promotes forest recovery and could benefit Harriman and Bear Mountain State Parks (the control sites for the study), but removal of interfering vegetation may be necessary to mitigate legacy effects where they currently hinder ground layer recovery. To successfully promote a more desirable forest condition that includes elimination of nonnative plant species, promotion of tree recruitment into the forest canopy, and development of diverse and abundant herbaceous cover in ground layer vegetation, future management decisions could include information on herbivory reduction and management of interfering vegetation where necessary.