Patuxent Home

Accession Number 5004527

Title Impacts of white-tailed deer on vegetation within parks in Montgomery County,

Maryland: A statistical analysis of deer exclosure data collected during 1995-98

Project Description The Maryland National Capital Park and Planning Commission, Montgomery County, Maryland

(M-NCPPC) conducted a field study of the effects of white-tailed deer (Odocoileus virginianus) on

the vegetation in several park locations during 1995-1998. The design of this study is described in

the Request for Proposal (RFP) from M-NCPPC, dated 19 Dec 2000, and is summarized below.

Eight 20 m x 20 m exclosures were set up with paired control plots at 8 park locations in 1995.

Four additional plots without exclosures were set up in 3 parks. Within each of the larger plots,

tree species, diameter at breast height (DBH), and vigor class were recorded for all individuals

defined as trees by Storm and Ross (1992). Other measurements specified by Storm and Ross

include seedling species, height class, and number of stems on four 2 m x 2 m subplots nested

within each of the larger plots, shrub species and cm coverage (line intercept), herb/vine class,

percent coverage, and maximum height class, summarized by 2 m x 2 m subplot, canopy

measurements (four densiometer readings, one on each quarter of the 20 m x 20 plots), and

browse measurements (species, browsed stems, recorded on transects of varying lengths, 1.7 m

radius). One complication of this data set is caused by unequal sampling effort over years. In

1995, data were collected on all 20 m x 20 m plots, four 2 m x 2 m nested subplots, all the

quarter plots of those nested subplots, and also some of the center 2 m x 2 m subplots. In 1996,

however, only half of the 1 m x 1 m quarter subplots were measured. In 1997, only half of the 20 m

x 20 m plots were measured, while in 1998, all 20 m x 20 m plots were measured, as well as four

subplots, and all of the quarter plots of those subplots. Some statistical analyses can handle

missing data, but they may not be as powerful and they may be more difficult to interpret. Another

complication in this data set is the four additional 20 m x 20 m plots that were set up without

exclosures. These data may cause problems for some analyses because the paired deer

exclosures are missing from these plots. All attempts will be made to include these data in as

many analyses as possible, but they may be eliminated from some analyses if complications are

great. These plots will be included in some analyses, however, such as sample size calculations.

The primary goal of this study is to test for differences in each variable between exclosed and

control sites. The design warrants a repeated measures analysis of variance (ANOVA). Proc

Mixed in SAS software (Littell et al. 1996) will be used to analyze this paired, repeated measures

design, with the null hypothesis for each variable being no difference in vegetative growth and no

change over time between exclosed study plots and the paired, unexclosed control sites. In

particular, the density, height, and species diversity of seedlings will be analyzed, along with cover

and species diversity of shrub layer, cover of the understory at various heights, and cover and

diversity of ground vegetation. Other variables will also be analyzed. Diagnostics will include

comparison of at least two variance covariance structures for the ANOVAs, testing of residuals for

normality, and investigations of transformations and heteroscedasticity. Summary statistics of

means and standard deviations (SDs), will be provided in tables for each variable analyzed and the

relevant factors. Tukey pairwise comparisons of least square means will be used to sort out

yearly differences in any variable for which year is significant (P < 0.05). If categorical variables are

analyzed, Proc Genmod (Stokes et al. 1995) will be used to provide the analogous analysis to

the paired, repeated measures ANOVA. To help with decisions about future data collection efforts,

sample size (number of plots and subplots), size of plots, and the variables to be measured will

be addressed. Sample size software (nQuery Advisor 4.0, Statistical Solutions) will be used to es

Keywords exclosure, montgomery county maryland, odocoileus virginianus, white-tailed deer,

Principal Jeff S Hatfield, USGS Patuxent Wildlife Research Center:;


Navigate PWRC SIS projects by Accession Number: First Previous Next Last

Return to SIS Projects Listing

Nav Bar link SIS List Nav Bar Link Patuxent Nav Bar Link Biological Resources Nav Bar Link USGS Nav Bar Link Interior Navigation Bar