Skip to contents

Descriptors of centrality, border proximity, and isolation in attribute space

Source: R/lets_attrcells.R
lets.attrcells.Rd

Computes per-cell descriptors for an attribute-space presence–absence matrix (as returned by lets.attrpam), using the two numeric attributes as a 2D coordinate system. The descriptors include:

  • Richness: number of taxa in each attribute cell (from the richness raster).

  • Distances to attribute midpoints: distance to the weighted midpoint (weights = richness) and to the unweighted midpoint, both computed in standardized attribute space; returned as negated values so that larger numbers indicate greater centrality.

  • Border proximity: three proxies — minimum distance to any zero-richness cell; robust mean of the closest perc fraction of zero-richness distances; and minimum distance to the convex-hull (MCP) border of occupied attribute cells.

  • Attribute isolation: richness-weighted mean distance from each cell to all other cells in standardized attribute space.

Usage

lets.attrcells(x, perc = 0.2)

Arguments

x

A list produced by lets.attrpam containing:

  • $PAM_attribute: data.frame with columns Cell_attr (first), the two attribute coordinates (second and third), followed by taxa columns (0/1).

  • $Attr_Richness_Raster: SpatRaster with richness per attribute cell.

perc

Numeric in (0, 1]; fraction used in the robust border metric (mean of the n smallest zero-richness distances). Default 0.2.

Value

A data.frame with one row per attribute cell and the columns:

  • Cell_attr: attribute cell identifier (1..n).

  • Richness: taxa count for the cell (from $Attr_Richness_Raster).

  • Weighted Mean Distance to midpoint, Mean Distance to midpoint: negated distances (larger = more central) in standardized attribute space.

  • Minimum Zero Distance, Minimum 10% Zero Distance, Distance to MCP border: border proximity proxies.

  • Frequency Weighted Distance: richness-weighted mean distance to all other cells.

Details

Attribute-cell descriptors for an attribute-space PAM (AttrPAM)

Attributes (columns 2–3 of $PAM_attribute) are standardized (z-scores) prior to distance calculations. Cells with NA richness are treated as zero when identifying zero-richness neighbors, and later masked in plotting.

Examples

if (FALSE) { # \dontrun{
# Example with simulated traits
n <- 2000
Species <- paste0("sp", 1:n)
trait_a <- rnorm(n)
trait_b <- trait_a * 0.2 + rnorm(n)  # correlated trait
df <- data.frame(Species, trait_a, trait_b)

# Build AttrPAM
x <- lets.attrpam(df, n_bins = 30)

# Compute descriptors
y <- lets.attrcells(x)

# Plot descriptors
lets.plot.attrcells(x, y)
} # }