EFFECTS OF WHITE-TAILED DEER DENSITY ON PHYSICAL CONDITION AND FOREST VEGETATION IN CUYAHOGA VALLEY NATIONAL PARK

Trowbridge, Jacob

Given that overabundant white-tailed deer (Odocoileus virginianus) can exert lasting negative effects in forested ecosystems, management of deer has been a foremost conservation concern in eastern North America since the last century. Although knowledge of density impacts on body mass, pregnancy rates, and forest regeneration are vital for deer management, relatively few studies have assessed these relationships in the same study area for a >5-year period. I took advantage of a managed culling program to investigate the impacts of deer density on body mass, pregnancy rates, and native forest understory and trillium (Trillium grandiflorium) regeneration at Cuyahoga Valley National Park (CUVA), Ohio, during 2016–2022. Deer densities were estimated using distance sampling in November and deer were culled by sharpshooters the following January–March. Each deer culled was aged, sexed, weighed, and inspected for fetuses (recorded as present or absent). Deer densities ranged from 8.6–18.0 deer/km2 and deer removed each study year varied between 205 and 450. Concurrently, the 1–6 tallest native seedlings per genera per subplot were measured in July–August in 25 fenced and unfenced paired plots and in long-term ecological monitoring (LTEM) plots. Browse was estimated in unfenced paired plots in alternating years. Trillium stem heights in both unfenced controls/fenced exclosures and browse in unfenced control plots were collected in April–June each year. Using a linear mixed model (LME), I analyzed change in body mass of culled deer and found a significant negative relationship with current deer density (P < 0.001), after controlling for age, sex, month, and zone ii of removal. I used a generalized linear model to determine whether pregnancy rates of yearling and adult females were influenced by weight, age, and deer density with a ≤3-year time-lag. The top model for pregnancy rate incorporated a 3-year time-lag in parkwide density (P < 0.001) with a positive relationship; age and weight had a significant effect. I used a generalized linear mixed model (GLMM) to determine whether the mean seedling height per subplot of the 6 most common genera in CUVA were influenced by year and deer density with a ≤3-year time-lag. The top model found mean seedling heights increased with increased time since the initial measurement (P = 0.006) and incorporated an insignificant negative effect of deer density for the previous fall (P = 0.091). I used nonparametric statistical tests to analyze seedling heights, and taxonomic richness and diversity, respectively, in LTEM plots. Heights of all seedlings in LTEM plots (P = 0.010), taxonomic richness (P = 0.029), and diversity (P = 0.036) increased after 4 years of culling. Seedling browse decreased over time in 11/14 genera with ≥10 plots having browse for each genus. I analyzed trillium stem heights with LME and an ANOVA, and I analyzed trillium browse with a t-test, Wilcox test, and a GLMM. Plants were taller in fenced exclosures than unfenced controls (P = 0.004). Stem heights of trillium in unfenced plots were similar during low and high deer density years (P = 0.123). The top model for stem height incorporated a positive 2-year time-lag in spring deer density (P < 0.001) and negative effect of year (P < 0.001). Deer preferentially browsed taller plants (P = 0.029). Browse rates on all plants and large/reproductive plants were nearly greater during high density years (P = 0.062 and P = 0.185, respectively). Deer density as a numeric covariate did not affect browse rates (P = 0.086). Culling ≥30% of the deer population at CUVA improved deer body mass, pregnancy rates, and native forest understory regeneration. Pregnancy rates were driven by deer density 3 years prior, highlighting the importance of long-term monitoring. I suggest there may be a habitat feature iii that has a long-lasting effect on pregnancy rates in this population, making it imperative to concurrently quantify potential habitat influences while managing deer numbers. Trillium stem height appeared to not have a useful relationship with deer density; however, this could be a result of legacy effects, thus continued monitoring of trillium is warranted.