Skip to contents

Compute overlap for utilisation distributions

Source: R/hr_tt_ud_overlap.R
hr_tt_ud_overlap.Rd

This function computes the overlap between pairs of utilisation distributions (UDs) using various methods. If x is a SpatRaster, it computes the overlap between x and y. If x is an hr_tt_ud_tbl, it computes the overlap for all unique pairs of UDs in the table. If cond_level is set, it computes the conditional overlap for the specified level, which is the overlap between the UDs estimated within a given isopleth level (e.g. 50%) of the UD, rather than the full UD. This can be useful for comparing the core areas of the UDs.

Usage

hr_tt_ud_overlap(x, ..., method = c("ba", "vi", "udoi", "earth_mover"))

# S3 method for class 'SpatRaster'
hr_tt_ud_overlap(
  x,
  y,
  ...,
  method = c("ba", "vi", "udoi", "earth_mover"),
  cond_level = NULL
)

# S3 method for class 'hr_tt_ud_tbl'
hr_tt_ud_overlap(
  x,
  ...,
  method = c("ba", "vi", "udoi", "earth_mover"),
  cond_level = NULL
)

hr_ud_overlap(...)

Arguments

x

A SpatRaster of the utilisation distribution (with a layer ud), or a tibble of UDs of class hr_tt_ud_tbl (e.g. as created with hr_tt_kde()).

...

Additional arguments (not currently used)

method

A character string specifying the method to use for overlap calculation. Options are "ba" (Bhattacharyya's Affinity), "vi" (Volume of Intersection), "udoi" (Utilisation Distribution Overlap Index), and "earth_mover" (Earth Mover's Distance). "earth_mover" returns a distance, where zero indicates identical UDs, and requires the suggested package emdist. Default is "ba".

y

A SpatRaster of the utilisation distribution, if x is a single UD. Else, if x is tibble of UDs, y is not used.

cond_level

Optional, the level for which the the conditional overlap is computed.

Value

A numeric value representing the overlap between the two UDs according to the specified method, or a matrix of such values if x is a tibble of multiple UDs.

Details

When x is an hr_tt_ud_tbl, each UD is validated, converted to cell values, and conditionally masked once before all pairwise comparisons are calculated. This avoids repeated raster reads and cumulative-distribution calculations for UDs that occur in multiple pairs.

Examples

example_kde <- hr_tt_kde(example_tt)
hr_tt_ud_overlap(example_kde)
#>           a         b         c
#> a 1.0000000 0.2287909 0.1684975
#> b 0.2287909 1.0000000 0.5419589
#> c 0.1684975 0.5419589 1.0000000