Calculate forest (or similar) fragmentation connectivity, density, and fragmentation class as per Riitters et al. 2020 Conservation Ecology 4:3. (also see errata). Forest density, connectivity, and fragmentation indices can be calculated for binary (forest/non-forest) and continuous rasters.

You can install the package using:

remotes::install_github('adamlilith/fragmentation', dependencies = TRUE)

You may need to install the remotes package first.

Riitters et al. define 5 classes of fragmentation, plus an "edge" case, based on the density and connectivity of cells with/without forest. These are:

1. Patch (density < 0.4)
2. Transitional (0.4 <= density < 0.6)
3. Perforated (density >= 0.6 & density - connectivity > 0)
4. Edge (density >= 0.6 & density - connectivity < 0)
5. Undetermined (density >= 0.6 & density == connectivity)
6. Interior (density == 1)

The fifth class, "undetermined", is the edge case and rarely occurs. Users of the functions in fragmentation can decide to use the "undetermined" class or to assign these rare cases to one of the other classes.

The package contains two functions and one raster used for examples. The functions are:

• fragBinary: Calculate fragmentation for rasters with binary values (forest/not forest)
• fragCont: Calculate fragmentation for rasters with continuous values (e.g., percent forest)

Here is an example of input (top left) and output (all others) from the fragBinary function for an area located in the eastern portion of the Republic of the Congo:

Adam