changeRangeR (version 1.1.0)

mcpSDM: SDM-based Minimum Convex Hull Polygon

Description

Implements the technique to estimate IUCN's extent of occurrence (EOO) geographic range estimate of species threat level by delineating a minimum convex polygon (i.e., convex hull) around a thresholded SDM prediction, first described by Syfert et al. (2014) <doi:10.1016/j.biocon.2014.06.012>. For each increment of 0.01 between a user-specified threshold and the maximum SDM prediction value, the prediction is thresholded to this value to make a binary raster. This raster is then converted to points, which are used to delineate a trial MCP. Each trial MCP is spatially intersected with the original MCP (based on the occurrence coordinates) and the original occurrence points. The Jaccard similarity index is calculated to determine geographic similarity between the trial and observed MCP. The trial MCP is also spatially intersected with the original occurrence points to determine how many were omitted. The "best" MCP is the one that has the highest JSI and also omits the least original occurrence points.

Usage

mcpSDM(p, xy, ch.orig, thr)

Value

a list of 5 objects.

Arguments

p

Raster* object of a continuous species distribution model prediction to base hull calculation on

xy

Matrix or Data frame of occurrence coordinates

ch.orig

SpatialPolygons object of original minimum convex hull based on occurrence locality coordinates

thr

Numeric threshold used to convert the continuous SDM prediction to a binary range map; this is then used to delineate the hull

References

Syfert, M. M., Joppa, L., Smith, M. J., Coomes, D. A., Bachman, S. P., & Brummitt, N. A. (2014). Using species distribution models to inform IUCN Red List assessments. Biological Conservation, 177, 174–184. https://doi.org/10.1016/j.biocon.2014.06.012

Examples

# \donttest{
# create continuous raster
p <- raster::raster(nrows=108, ncols=108, xmn=-50, xmx=50)
raster::values(p)<- runif(n = (108*108))
raster::crs(p) <- "+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs"
# create occurrences
xy <- dismo::randomPoints(p, 4)
# create original convex hull
ch.orig <- mcp(xy, crs = "+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs")
# set threshold
thr <- 0.5
# mcpSDM
mcpSDM(p, xy, ch.orig, thr)
# }