Range bagging SDM model building method

The model building component of an implementation of the range bagging species distribution modelling (SDM) method (Drake, 2015).

Usage

rangebag(
  x,
  p,
  n_models = 100,
  n_dim = 2,
  sample_prop = 0.5,
  limit_occur = TRUE,
  ...
)

# S3 method for class 'Raster'
rangebag(
  x,
  p,
  n_models = 100,
  n_dim = 2,
  sample_prop = 0.5,
  limit_occur = TRUE,
  ...
)

# S3 method for class 'SpatRaster'
rangebag(
  x,
  p,
  n_models = 100,
  n_dim = 2,
  sample_prop = 0.5,
  limit_occur = TRUE,
  ...
)

# S3 method for class 'data.frame'
rangebag(
  x,
  p,
  n_models = 100,
  n_dim = 2,
  sample_prop = 0.5,
  limit_occur = TRUE,
  ...
)

Arguments

x: Climate (or environmental) data as a terra::SpatRaster or raster::Raster* (with any CRS), or a data.frame with WGS84 lon and lat columns.
p: Species occurrence data as a data.frame (or matrix) with WGS84 lon and lat columns. Any points outside the extent of x will be ignored.
n_models: Number of convex hull models to build in sampled environment space (default = 100).
n_dim: Number of dimensions (variables) of sampled convex hull models (default = 2).
sample_prop: Proportion of environment data rows sampled for fitting each convex hull model (default = 0.5).
limit_occur: Logical to indicate whether to limit occurrence data to one per environment data cell (default = TRUE).
...: Additional parameters.

Value

A "Rangebag" model S4 object containing slots:

method: SDM method: "rangebag".
variables: List of climate (or environmental) variable names.
presence: The selected climate data corresponding to occurrences points.
coordinates: The coordinates for the selected climate data.
ch_models: A list of convex hull models (vertices).

References

Drake, J. M. (2015). Range bagging: a new method for ecological niche modelling from presence-only data. Journal of the Royal Society Interface, 12(107), 20150086. doi:10.1098/rsif.2015.0086