|
|
Locating roadkill hotspots is key to mitigate road impacts on wildlife, but study design can affect hotspot definition. The DeadCanMove R package contains functions to estimate how sampling scheme may affect the spatial accuracy of hotspot identification.
From a baseline data frame of roadkilled individuals recorded daily at different road stretches (like the roadkills sample dataset provided with the package), you can simulate varying sampling schemes (i.e. surveys at increasing intervals), determine roadkill hotspots based on each sampling scheme, and compare hotspot patterns between sampling schemes and the baseline data.
To install DeadCanMove directly from R-Forge, paste the following command in the R console when connected to the internet:
install.packages("DeadCanMove", repos="http://R-Forge.R-project.org")
This normally requires the latest R version. Otherwise, you can download the compressed source files (either .tar.gz or .zip) and then install the package from your disk (something like "Packages - Install packages from local zip files", or "Tools - Install packages - Install from: Package Archive File", or "Packages & Data - Package installer, Packages repository - Local binary package", ... depending on your R menu interface).
Then, each time you start a new R session, use the following commands to load the package and maybe check out its help files:
library(DeadCanMove)
help("DeadCanMove")
?hotspots.comparison
The following commands will load the package and the roadkills sample dataset, get a look at a part of these data, and see how roadkill patterns obtained from increasingly sparse sampling relate to those obtained from the whole (baseline) data, as is done in Santos et al. (2015):
library(DeadCanMove)
data(roadkills)
roadkills[780:790, 1:10]
hotspots.comparison(dataset = roadkills, sampl.columns = 4:ncol(roadkills), sampl.intervals = 1:30, region.column = "segment", group.column = "taxon", include.all.together = TRUE, confidence = 0.95, min.total.events = 80, min.hotspot.threshold = 2, comp.method = "Phi", plot = TRUE, sep.plots = FALSE, omit.baseline.interval = TRUE, pch = 20, ylim = c(0, 1))
This package is a companion to the following article, which should be cited when you use it:
Santos S.M., Marques J.T., Lourenço A., Medinas D., Barbosa A.M., Beja P., Mira A. (2015) Sampling effects on the identification of roadkill hotspots: implications for survey design. Journal of Environmental Management, 162: 87-95. DOI: 10.1016/j.jenvman.2015.07.037
To see how to cite the package itself, load it in R and type citation(package="DeadCanMove")
The R-Forge project summary page is found here.