Fig 1.
Land-use history (a) and abundance of invasive species in Powdermill Nature Reserve in 2006.
Point sizes correspond to the number of invasive species present ranging from 1 to 3 invasive species present (b) and abundance of R. multiflora (c) and B. thunbergii (d). Blue lines represent rivers, while brown lines represent major roads in the region.
Fig 2.
Plot-level odds of (a) R. multiflora and (b) B. thunbergii occurrence and (c) plot-level log odds of finding multiple invasive species.
A total of 4388 10-m diameter plots with different land-use histories in a deciduous forest in eastern North America were included in this analysis except in the invasive species count analysis in which only plots with at least one invasive species were included (n = 792). The odds of occurrence for each invasive species in a given land-use category relative to always forested plots is given by each point with 95% confidence intervals. Sample sizes are given in parentheses. Among land-use histories, always forested plots (i.e. forested continuously since at least 1939) was used as the reference group and asterisks indicate a significant difference from always forested plots (*p < 0.05, **p < 0.01, ***p < 0.001).
Fig 3.
Probability of occurrence of R. multiflora versus distance to a road.
Individual points represent plot-level observations of R. multiflora’s occurrence (1 = present, 0 = not present,n = 4388 plots). The best fit line was fitted using the generalized linear model of occurrence versus log distance to road and provides the probability of R. multiflora occurrence at varying distances from roadways.
Fig 4.
Odds of high abundance for (a) R. multiflora, (b) B. thunbergii with different land-use histories.
Each point indicates the odds that each invasive species will be highly abundant (>10% cover) in a given land-use history ± 95% confidence intervals (n = 1074 plots for R. multiflora and n = 558 plots for B. thunbergii). Asterisks indicate a significant difference from always forested (i.e. forested continuously since at least 1939) plots (*p < 0.05, **p < 0.01, ***p < 0.001) and sample sizes are given in parentheses. Plots with R. multiflora or B. thunbergii cover class of 1–2 were categorized as low abundance plots and plots with cover classes of 3–6 were categorized as high abundance plots. Analysis of the always developed and developed to forest land-use categories were not complete for B. thunbergii due to small sample size.