Similarity Statistic for Quantifying Niche Overlap using Warren's-I
The overlap function computes the I similarity statistic (Warren et al. 2008) of two overlapping niche estimates. Similarity is based on the Hellenger distance. It is assumed that the input data share the same extent and cellsize and all values are positive.
The I similarity statistic sums the pair-wise differences between two predictions to create a single value representing the similarity of the two distributions. The I similarity statistic ranges from a value of 0, where two distributions have no overlap, to 1 where two distributions are identical (Warren et al., 2008). The function is based on code from Jeremy VanDerWal
overlap(x, y)
| x | A matrix, rasterLayer or sp raster class object |
|---|---|
| y | A matrix, rasterLayer or sp raster class object with the same dimensions of x |
A value representing the I similarity statistic
Warren, D. L., R. E. Glor, M. Turelli, and D. Funk. (2008). Environmental Niche Equivalency versus Conservatism: Quantitative Approaches to Niche Evolution. Evolution 62:2868-2883.
Jeffrey Evans <jeffrey_evans@tnc.org> and Jeremy VanDerWal
# add degree of separation in two matrices p1 <- abs(matrix(1:50,nr=50,nc=50) + runif(n = 2500, min = -1, max = 1)) p2 <- abs(matrix(1:50,nr=50,nc=50) + rnorm(n = 2500, mean = 1, sd = 1)) # High overlap/similarity ( I <- overlap(p1,p2) )#> $original.I #> [1] 0.9815073 #> #> $corrected.I #> [1] 0.999316 #>