Supplement 1. R scripts and an example data set for conducting the power analysis simulations described in the main text.

File List Miller_and_Mitchell_Power_Analysis_Code.r (md5: e0161858aaaeac3e81a2c755640b9feb) Tree_BA.csv (md5: e7e52c80094e578260cc7d4bade00208) Description Miller_and_Mitchell_Power_Analysis_Code.R - This script runs a bootstrap power analysis based on a mixed effects model of sample data (plot is the random effect, time and site (park) are fixed effects). The simulation determines power to detect a uniform percentage per sampling cycle change in the value of a metric as a linear trend in a mixed-effects model. The sample sizes tested by the script do not have to be the same as the number of samples in the data file; any desired number of samples will be bootstrapped from the actual data. The script will report power for a uniform trend across all parks (model with no interaction), as well as power for a trend that occurs only at one park (model with an interaction effect, where simulated effect occurs and power is tested for each park in turn). This script requires a comma delimited (.csv) file with the following headings: ID (unique alphanumeric value for each row of data; does not need to be called "ID") Plot (text, not numeric only, e.g.: "ACAD1" not "1") Park (text) Year (year of sample, numeric) Metric (metric to be evaluated, numeric) The data in the file must have two measurements for each plot, with the initial measurement of all plots collected prior to any second measurements (separate data collection cycles). Tree_BA.csv – This is an example of the data sets used for the power analysis in this article. Data sets must be formatted as demonstrated in this data set for the simulation to work properly.