Long term community changes following deer management in a remnant montane Pinus palustris forest

Montane longleaf pine (Pinus palustris) forests are currently dwindling due to fire suppression. Longleaf pine forests require reoccurring natural forest fires for longleaf regeneration and maintenance of the understory community composition. In many areas of the southeastern United States, decades of fire suppression have led to invasion of longleaf-dominated forests by deciduous hardwoods and other pine species. This study shows the impacts that high densities of white-tailed deer have on a fire-suppressed remnant longleaf pine ecosystem within Alabama’s Oak Mountain State Park. The park is divided into two geologic and ecological zones: the lower elevation foothills region and the higher, rocky, southeastern-facing slopes of the ridge of Double Oak Mountain. In order to compare the impacts of deer among these two habitats, we surveyed understory woody plant diversity and rates of deer browsing in 20 plots that had initially been surveyed in 2003. Species richness and diversity of plants on the southeastern slopes of the ridge were 35 and 1.573 respectively. While in the foothills region of the park, Species richness and diversity were 27 and 1.134. Deer browsing pressure was higher in the foothills. The probability of a plant experiencing deer browsing was positively correlated with the proportion of palatable plant species within a plot, indicating that plants may benefit from associational resistance. Comparisons of similarity indices of canopy trees and understory seedlings showed that reproduction of canopy trees may be slightly more successful in the foothills than on the ridge. INTRODUCTION Current land area dominated by longleaf pine (Pinus palustris Mill.) forests in the Southeastern US has dwindled to 3% of its former coverage, due to the impacts of fire suppression, logging, and land use changes (Varner et al., 2005). Longleaf pine recruitment requires frequent forest fire events to eliminate competition for light by non-fire-tolerant trees and shrubs (Outcalt, 2000). Decades of fire suppression in many areas of Alabama have facilitated the incursion of fire susceptible hardwoods like oaks (Quercus spp.) and maples (Acer spp.) and shrubs like blueberry bushes (Vaccinium spp.) into former longleaf pine forests (Varner et al., 2005). In fire-suppressed habitats, longleaf pines are poor light competitors (Brockway & Lewis, 1997) and are often confined to abiotically poor habitats with shallow soil, steep grades, and greater potential for thermal or water stress (Outcalt, 2000; Varner et al., 1998). Fire suppression has impacted not only longleaf pines, but also the historic plant and animal communities that relied on sparse canopy cover and xeric soils. An estimated 20 to 30 species/m of herbaceous plants can be supported within a regularly burned longleaf pine forest (Varner et al., 2005). When periodic fires in longleaf pine forests are suppressed, canopy cover increases and decreases light availability for herbaceous understory species. Oak Mountain State Park, a roughly 10,000 acre reserve in southern Appalachian ridge and valley habitat outside Birmingham, AL, is dominated by historical longleaf pine forests that have been fire-suppressed for decades. Remnant longleaf pine adults, saplings, and seedlings are patchily abundant in the park. Although fire suppression may be playing a strong role in structuring Oak Mountain communities, the influence of white-tailed deer, Odocoileus virginianus (Zimmermann), is also likely to be substantial. Deer browsing can alter forest composition and structure (Barton & Hanley, 2013; Burney & Jacobs, 2013; Kain et al., 2011; Nuttle & Royo, 2013; Pekin et al., 2014) by suppressing palatable plants like greenbriar (Smilax spp.), common persimmon (Diospyros virginiana), and wild grapes (Vitis spp.; Stribling, 1996) and increasing the recruitment of non-palatable plant species (Nuttle & Royo, 2013). Although the population size of white tailed deer within Oak Mountain is unknown, density in the park is thought to be high. The Alabama Department of Conservation and Natural Resources estimated the white tailed deer population within the state to be between 1.5 and 1.75 million individuals in 2000 (ADCNR, 2000). Many studies have looked into the effects of high-density deer populations on forest community composition and diversity (Leland and Fowler 2001). Deer can create disparity between the canopy trees and the understory, and left unchecked will alter forest trajectories (Nuttle et. al. 2014). Fire and deer browsing have the potential to interact to affect plant communities. Nuttle et al. (2013) found that fire events and deer browse decreased plant diversity, but fire events alone increased diversity within a deciduous forest in West Virginia. The interacting effects of deer and fire suppression within a montane longleaf pine ecosystem have not been explored. In Oak Mountain State Park, surveys on understory plants have suggested that deer browsing has altered the abundance of palatable plants and is changing the community structure of Oak Mountain (Kearly, Underwood and Johnson; Meadows and Fincher unpublished manuscripts). After surveys of deer browse in the park were conducted in 1999 and 2002 (Kearly, Underwood and Johnson; Meadows and Fincher unpublished manuscripts), deer population control was initiated using sharpshooters and bow hunting to remove approximately 30 individuals per year. We have performed surveys of understory woody plant diversity and deer browsing rates in order examine long term impacts the deer may be having on plant communities within the fire suppressed montane longleaf pine forests of Oak Mountain and to examine the effectiveness of the current deer population management scheme. We sampled two ecological and geologic regions of the park that contain remnant longleaf stands: the Southeast-facing ridge of Double Oak Mountain, and the tops of the foothills northwest of the mountain in order to test the following hypotheses. Hypotheses: 1. The rate of deer browsing on woody understory plants on the ridge side of the park will be lower and more concentrated on a few species of plants than in the foothills because plant species that can tolerate the stressful conditions on the ridge may be less palatable. Evolution of chemical defenses to deter herbivores may be common in stressful environments where resources to replace important plant parts are scarce (Barton & Hanley, 2013; Stamp, 2003). Stressors include high heat, limited water availability, and shallow soil. 2. As a consequence of lower deer browsing rates and less environmental stress, we predict that woody understory plant diversity will be higher in the foothills area and lower in the ridge area. 3. We predict that comparison between current rates of deer herbivory and corresponding data from the 2000 survey in Oak Mountain will show that since the deer population control measures began, deer herbivory has decreased. 4. We also predict that since 2004 (Duncan Unpublished Data, 2004), woody understory plant diversity has increased due to decreased browsing pressure and a corresponding increase in the density of palatable plant species. 5. Comparisons between 2014 understory diversity and canopy tree diversity in 2004 (Duncan Unpublished Data, 2004) will show that canopy tree recruitment, as measured by the similarity of canopy and understory communities, is greater in ridge plots than in foothill plots because ridge plots are more difficult for deer to access and possess more unpalatable species. METHODS: Twenty plots (50x20m) with the long edges running north to south were established in 2004 using a random stratified design within two areas of Oak Mountain State Park. Ten plots were established on the SE facing slope of Double Oak Mountain and ten on the foothills to the NW of the mountain. The plots are subdivided into 10x10m subplots. Deer browsing rates and woody plant diversity were measured in a 1m wide transect along the eastern and western 50 m edge of the plot, for a total area of 100m per plot (Fig. 1). Any woody plant with a stem between 10cm and 130cm in height within the transect was identified. Height was measured from the ground to the terminal meristem of the plant. Each plant was examined for evidence of deer browsing such as truncated or ripped stems or branches with no apical buds (Fig 2). Deer browsing was recorded as present or absent for each plant. Standing dead woody plants within the transect that were identifiable and showed signs of browsing before death were included in analyses. Canopy tree data and seedling data that was collected in 2004 was used in concordance with my data set for analysis. The seedling data collected in 2004 included seedlings between 0 and 130 cm in height. The seedlings were often surveyed throughout the entire plot, however in some cases where there were many seedlings in a subplot, the subplot was subsampled in a 5x5m quadrat in the upper left corner of the subplot. These subsamples were noted during data collection. The seedling surveys of 2004 did not include deer herbivory measurements. Kearley and Underwood collected unpublished data on deer browsing in Oak Mountain State Park in 2000. A mile long transect was established and circular plots were surveyed at 100 yard intervals. Plants were identified, measured for stem diameter (<5cm), and checked for signs of deer browsing (Kearly and Underwood, 2000 unpublished manuscript). Data from this study was used for comparisons of percents of individuals browsed at the genus level. We used EstimateS (University of Connecticut) to calculate Shannon Weiner diversity indices of foothills and ridge plots and Chao-Jaccard similarity indices among foothill and ridge plots. Any plants that were rare across all sampling plots (n<11) were excluded from the palatability investigations. Palatable species were identified as plant species that were found to have 50% or more individuals browsed. Unpalatable plant species were identified as species with fewer than 50% of individuals that were browsed upon (Table 1). There were a total of 7 palatable species and 11 non palatable species. We used SPSS statistical software (IBM Corp. version 19) to compare deer browsing rates and plant diversity among years and among the foothills and ridge sites using two sample t-tests. Regression analysis allowed us to examine the relationship between the density of palatable individuals in a plot and the rate of browsing on all woody plants or on particular species of interest.