boundary_null_distrib {BoundaryStats}

Null distribution for overlap statistics

Description

Creates custom probability distributions for two boundary statistics (number of subgraphs and length of the longest subgraph). Given a SpatRaster object, simulates n iterations of random raster surfaces from a neutral model.

Usage

boundary_null_distrib(
  x,
  convert = FALSE,
  cat = FALSE,
  threshold = 0.2,
  n_iterations = 10,
  model = "random",
  p = 0.5,
  progress = TRUE
)

Arguments

x

A SpatRaster object.

convert

TRUE if x contains numeric trait data that needs to be converted to boundary intensities. default = FALSE.

cat

TRUE if the input SpatRaster contains a categorical variable. default = FALSE.

threshold

A value between 0 and 1. The proportion of cells to keep as boundary elements. default = 0.2.

n_iterations

An integer indicating the number of iterations for the function. A value of 100 or 1000 is recommended to produce sufficient resolution for downstream statistical tests. default = 10.

model

Neutral model to use. Options: 'random' (stochastic), 'gaussian' (Gaussian random field), and 'random_cluster' (modified random clusters method)

p

If using modified random clusters, proportion of cells to be marked in percolated raster.Higher values of p produce larger clusters. Default: p = 0.5

progress

If progress = TRUE (default) a progress bar will be displayed.

Value

A list of two probability distribution functions for boundary statistics.

Author(s)

Amy Luo

References

Saura, S. & Martínez-Millán, J. (2000). Landscape patterns simulation with a modified random clusters method. Landscape Ecology, 15:661-678.

Examples

data(T.cristatus)
T.cristatus <- terra::rast(T.cristatus_matrix, crs = T.cristatus_crs)
terra::ext(T.cristatus) <- T.cristatus_ext

T.crist_bound_null <- boundary_null_distrib(T.cristatus, cat = TRUE, n_iterations = 100,
model = 'random_cluster')