Skip to main content
Dryad logo

Data for: The long-term impacts of deer herbivory in determining temperate forest stand and canopy structural complexity

Citation

Reed, Samuel et al. (2021), Data for: The long-term impacts of deer herbivory in determining temperate forest stand and canopy structural complexity, Dryad, Dataset, https://doi.org/10.5061/dryad.b8gtht7dn

Abstract

1. Ungulates place immense consumptive pressure on forest vegetation globally, leaving legacies of reduced biodiversity and simplified vegetative structure. However, what remains unresolved is whether browse-induced changes occurring early in succession ultimately manifest themselves in the developed forest canopy. Understanding the development and persistence of these legacies is critical as canopy structure is an important determinant of forest ecosystem functions like carbon sequestration and wildlife habitat.

2. We measured how white-tailed deer (Odocoileus virginianus) browse during stand initiation affected canopy structure, tree species richness, diversity, stem density, and basal area on Pennsylvania’s Allegheny Plateau using a portable canopy LiDAR system. We capitalized on an historic deer enclosure experiment where forests were subjected to four deer densities (4, 8, 15, and 25 deer/km2) for ten years following stand initiation.

3. Deer browsing impacts on the forest canopy are apparent nearly four decades since stand initiation. The highest deer density treatment experienced a significant reduction in tree species diversity, density, and basal area with stands becoming dominated by black cherry (Prunus serotina). Reductions in overstory diversity and tree density resulted in a more open canopy with low leaf area and high horizontal leaf variability. Canopies were tallest at the lowest and highest deer densities.

4. Synthesis and Applications: Using a portable canopy LiDAR system and a former deer enclosure experiment, we show that high deer browsing pressure during stand initiation can have a decades-long impact on stand and canopy structure. High deer densities led to stands with lower species diversity and tree density, which resulted in canopies that were taller and less dense. Managers should consider the lasting legacy of ungulate herbivory on canopy structure, as canopy structure influences several important management goals, such as forest carbon sequestration, maintenance of diverse understory communities, and creation of wildlife habitat.

Methods

All data was collected in 2016 within the formerly clearcut locations of a replicated deer enclosure experiment. Deer populations were maintained from 1979 to 1990 at four distant sites, each of which was partitioned by deer densities of 4, 8, 15, and 25 deer/km2, in Elk County (710 m elevation; 41°34’22” N, 78°28’30” W), Warren County (550 m elevation; 41°38’48” N, 79°08’11” W), Forest County (550 m elevation; 41°34’40” N, 79°06’19” W), and McKean County (670 m elevation; 41°38’21” N, 78°19’33” W). More information regarding experimental design can be found in the corresponding paper (Reed et al. 2022).

We placed three parallel 30 x 5 m belt transects in each clearcut treatment, all separated by a minimum of 30 m from one another. All trees >5 cm diameter at breast height (DBH) were identified to species and measured for DBH. Canopy structural complexity was measured along each transect using a portable canopy LiDAR system. Canopy structural data were processed using the "FORESTR" R package (Atkins et al. 2018b). 

We have uploaded all raw species and DBH values collected, along with basal area and predicted live biomass calculated from Chojnacky et al. (2014). Further, we have uploaded the full suite of canopy structural data as generated by FORESTR. Although our paper only uses a few well-established canopy structural metrics calculated from the package, other metrics we have included in this dataset may prove ecologically valuable as more canopy research is published.

Usage Notes

A) One treatment (15 deer/km2) at State Game Land 30 was harvested prior to our study, reducing our sample size to 15 treatment areas.

B) All statistical methods are included in the manuscript (Reed et al. 2022)

C) All naming conventions and metadata included in README file