Site Selection
Two geographically distinct groups of redcedar woodlands were sampled to answer differing sets of questions: one set of sample sites aimed to capture the spatial structure of redcedar-dominated stands across a range of size classes and the second evaluated the interactions between redcedar and codominant deciduous species. The locations of the first set of 15 surveys, which we term the ‘Midwest’ dataset, center in the Ohio River valley and include samples from Ohio, Pennsylvania, and Kentucky (Figure 1). Survey sites in the Midwestern dataset were selected by examining aerial imagery of wooded lands and grasslands to determine if redcedar were present. Redcedar is an evergreen conifer that is easily identifiable in aerial images (e.g. Wang et al., 2017).
The second set of sample plots, termed the ‘Oklahoma’ dataset were located in the Cross Timbers region of central Oklahoma, near Stillwater (Figure 1). The Cross Timbers represent the transition zone between the eastern deciduous forest and the Great Plains (Francaviglia 2000; Stambaugh et al. 2009). To evaluate the interaction between encroaching redcedar and other trees in the Cross Timbers, point-pattern data were recorded at 13 neighboring woodlands with varying proportions of codominant redcedar and Quercus species, and other deciduous trees.

Sampling Design

From a random starting point, meter tapes were laid out in a 10 m x 20 m (x, y) grid aligned along the dominant environmental gradient (i.e., slope) (Wiegand et al. 2013). All trees within the sample plot had the following attributes recorded: coordinates (to nearest 0.1 m), species name, alive or dead, and diameter at breast height (DBH) (trees > 1.3 m). Plots were extended in 10 m increments along the x-axis until a minimum of ~50 trees were recorded (average sample > 88 trees). All samples were collected from March 2020-December 2021.