try this overlap metric

From above:

Median pairwise overlap in a community can range from zero, with no overlap among any species pairs, to one, with complete overlap of all species pairs. Overlap values approach-ing zero indicate a high degree of niche partitioning. To cal-culate the overlap statistic, we first generated a kernel density estimate for the trait distribution for each population. We used the ‘density()’ function with the ‘bw.nrd0’ method in the base R distribution to select the appropriate bandwidth for the kernel. The overlap statistic was not sensitive to the method used to select the optimal bandwidth (for our data-set, the minimum pairwise correlation among all bandwidth selection methods was 0.96). The density functions for each species are normalized so that the integral of the function is one. We calculated the pairwise overlap for each species pair in the community by taking the integral of the minimum of the two functions and weighted each pairwise overlap by the harmonic mean of the abundance of both species in the pair. The community-level niche overlap statistic (hereafter, simply ‘overlap’) is the abundance-weighted median pairwise overlap of all species pairs in the community. We calculated the pairwise overlap between species in the same way as Mouillot et al. (2005), but our community-level metric dif-fers in that we used the median instead of the mean to reduce the influence of outliers. In addition, we weighted each value by the harmonic mean abundance of the species pair to reduce the influence of rare species on the value of the metric

and weight it somehow according to abundance. They used harmonic mean, which might make sense.

diazrenata / isds

try this overlap metric #9