Skip to contents

Further FRAGSTATS landscape metrics: shape, core, proximity and information theory

Source: R/LandscapeMetricsExtra.R
PatchStructure.Rd

PatchStructure: per patch frac \(2\ln(0.25p)/\ln a\) (1 for single cells), gyrate (mean distance of cell centres from the centroid, m), circle \(1 - a/a_c\) (smallest circle enclosing the cell corners), contig (3x3 template: self 1, rook 2, diagonal 1; \(((\sum t)/n - 1)/12\)), core (ha; cells more than edge_depth rook steps from outside the patch, landscape boundary counted unless consider_boundary) and cai. ClassStructure: class means and area-weighted means (frac, shape, para, gyrate), circle_mn, contig_mn, core_mn, cai_mn, tca, cpland and pafrac \(2/b\) from the regression of \(\ln a\) on \(\ln p\) (NA below 10 patches). ProximityMetrics: PROX \(\sum a_j/h_{ij}^2\) over same-class patches within search_radius (edge-to-edge cell-centre distance \(h\)), SIMI with a class similarity matrix, and CONNECT (percentage of patch pairs within threshold). LandscapeInformation: marginal entropy, joint entropy, conditional entropy, mutual information and relative mutual information of the rook co-occurrence matrix (McGarigal and Marks 1995; Gustafson and Parker 1994; Nowosad and Stepinski 2019; Hesselbarth et al. 2019). Identical to the Python arm morie.fn.lsmextra.

Usage

PatchStructure(
  landscape,
  res = 1,
  directions = 8,
  edge_depth = 1,
  consider_boundary = FALSE,
  circle_method = "exact"
)

ClassStructure(
  landscape,
  res = 1,
  directions = 8,
  edge_depth = 1,
  consider_boundary = FALSE,
  circle_method = "exact"
)

ProximityMetrics(
  landscape,
  search_radius,
  res = 1,
  directions = 8,
  similarity = NULL,
  threshold = NULL
)

LandscapeInformation(landscape, base = "log2")

Arguments

landscape

Integer matrix of classes (NA allowed).

res

Cell size (m).

directions

4 or 8 (patch connectivity).

edge_depth

Edge depth in cells for core area.

consider_boundary

Logical; if TRUE the landscape boundary is not an edge.

circle_method

"exact" (smallest circle enclosing every cell corner) or "landscapemetrics" (corners of the cells in the extreme rows and columns only, as landscapemetrics; that circle can miss other corners).

search_radius

PROX/SIMI search radius (m).

similarity

Class-by-class similarity matrix (ascending classes); default identity.

threshold

CONNECT distance threshold (m); default search_radius.

base

Logarithm base: "log2", "log" or "log10".

Value

List of metrics.

References

McGarigal, K. and Marks, B. J. (1995). FRAGSTATS. USDA Forest Service General Technical Report PNW-351. Gustafson, E. J. and Parker, G. R. (1994). Using an index of habitat patch proximity for landscape design. Landscape and Urban Planning, 29, 117-130. Nowosad, J. and Stepinski, T. F. (2019). Information theory as a consistent framework for quantification and classification of landscape patterns. Landscape Ecology, 34(9), 2091-2101.

Examples

m <- matrix(c(1,1,1,1,1, 1,1,1,1,2, 1,1,1,2,2, 2,1,1,1,1), 4, byrow = TRUE)
round(PatchStructure(m)$contig, 6)
#> [1] 0.604167 0.000000 0.277778
round(LandscapeInformation(m)$condent, 6)
#> [1] 0.689275