Decades of chronic overbrowsing has seriously degraded native plant diversity throughout much of the Eastern Deciduous Biome (Rooney 2001; Rooney and Waller 2003; Cote et al. 2004; Kain et al. 2011; Nuttle et al. 2013; Schumacher and Carson 2013; Begley-Miller et al. 2014; Sabo et al. 2019; Burton et al. 2021; VanderMolen and Webster 2021). Here, we evaluated gap-phase regeneration within a forest where deer had been kept close to historic densities for nearly seven decades compared to nearby forests where deer have been overabundant for a similar period. Overall, we only found support for our hypotheses that took into account the abundance of deer. For example, we found that abundance, richness, and diversity of native herbaceous and woody species were 35–65% higher at low deer density. At low deer density, the average height of tree species was more than double that at high deer density, with three native tree species more than 10 times taller. Essentially, at high deer density, tree species were unable to push through the deer filter. Moreover, at low deer density, we found that 11 native species were significantly associated with drier habitats (indicator species analysis), including five canopy tree species, one understory tree, and one understory shrub. Two additional native canopy tree species and one native shrub were associated with wet habitat only at low deer density. No native canopy tree or shrub species was associated with wet or dry sites at high deer densities. Moreover, native tree species abundance was always greater at lower deer density except for browse-resistant black cherry, which was more abundant at high deer density. High deer densities often lead to understories dominated by black cherry (Horsley et al. 2003; Royo et al. 2019; Royo and Carson 2022). Collectively, these findings clearly demonstrate that keeping deer density much closer to their historical levels (3–4 deer/km2; McCabe and McCabe 1997) leads to gap-phase regeneration dominated by native tree and shrub species.
In contrast, at high deer density, exotic species were much more abundant in gaps, whereas native tree species were short in stature and in low abundance. In fact, exotic species were 6-fold more abundant at high deer density (Fig. 2a). Surprisingly, neither gap age nor habitat type (wet vs. dry) had a significant impact on plant diversity, richness, or abundance, except where exotic species were more abundant in older gaps with more deer (Fig. 3). This is worrisome because it indicates that older gaps accrue exotic species during the first 12 years of gap-phase regeneration at high deer density, and high deer density is now the norm across vast regions (Long et al. 2007; Pendergast et al. 2016; Adams and Villarreal 2020).
We expected that exotic richness and diversity would be higher outside of Camp Garfield because exotic propagule pressure increases substantially with human settlement, extensive public roadways, and hiking trails (Stohlgren et al. 2006; Iannone et al. 2015; von der Lippe and Kowarik 2007; Liedtke et al. 2020). Indeed, our gaps outside of Camp Garfield were accessible to the public and were closer to urban and urban-fringe habitats, though they were still in larger tracts of continuous forest. It is important to note that exotic richness and diversity did not differ in areas of low versus high deer density (Fig. 2b, 2c). However, as gaps aged, exotic cover increased, suggesting that competition for light and soil resources between native and exotics may have intensified in older gaps. The finding that the richness and diversity of exotics did not differ in gaps within high versus low deer density was surprising. This may relate to the documented paucity of exotic species in the regional species pool that are both highly shade-tolerant and can also persist in closed canopy mature forests prior to gap formation (Martin et al. 2009; Eschtruth and Battles 2011).
On the relevance of our findings to oak regeneration
Across much of the eastern deciduous forest, oaks have poor regeneration due to a combination of overbrowsing, fire suppression, the small size of canopy gaps, and dense recalcitrant understories layers (Royo et al. 2006; Amatangelo et al. 2011; Nuttle et al. 2013; Thomas-Van Gundy et al. 2014). Nonetheless, our findings demonstrate that where deer were at low density, gaps had a diverse, dense, and robust sapling layer, which included three species of oak. These saplings reached well into the subcanopy (Fig. 4a, 4b). Whether understory fires in these fairly mesic forests would further enhance the regeneration of oaks is unknown but fires were likely far less common historically in beech-maple forests relative to forests elsewhere formerly dominated by oaks (Vankat 1979). Indeed, Iverson et al. (2008) found that canopy gap creation coupled with fire was effective in promoting oak regeneration in dry but not mesic sites in southeastern Ohio, USA in forest outside of the Beech-Maple Association.
On the importance of military installments as refugia from overbrowsing
The forests we sampled occurred on abandoned farmland in 1940. Natural forest regeneration at Camp Garfield and within natural areas outside Camp Garfield occurred at a time when deer densities were beginning to rise dramatically in the eastern US (Leopold et. al. 1947; Iverson and Iverson 1999; Cote et al. 2004). Thus, Camp Garfield represents an 88 km2 area that has undergone secondary succession with atypically low deer densities. Based on current trends outside Camp Garfield, the prognosis for the regeneration of many native plant species is poor, but the prognosis for exotics is more favorable (Eschtruth and Battles 2009; Knight et al. 2009; Kalisz et al. 2014; Averill et al. 2018). Consequently, military bases and posts with extensive forest cover and aggressive deer management may typically provide important deer refugia for numerous browse-sensitive plant species over large areas. Indeed, the top five largest military bases in the continental U.S. total 5,001 km2 (mybaseguide.com). The prospect of large-scale refugia for browse-sensitive plants from deer overbrowsing should be explored for military installments around the U.S. whenever these posts are located in areas of high ungulate density. While not directly relevant to large-scale browsing refugia, the Fort Liberty Military Installation in North Carolina (formerly Fort Bragg) has effectively used prescribed fire to promote and maintain one of the largest remnant long-leaf pine ecosystems in the southeastern U.S. (Allen et al. 2006; Lashley et al. 2014). These efforts have contributed to maintaining open-longleaf pine bird assemblages which include the endangered red-cockaded woodpecker (Picoides borealis) and the threatened Bachman’s sparrow (Aimophila aestivalis; Allen et al. 2006).