cschloss
commented
5 years ago Hi Carrie,
You are absolutely right that edge effects are an issue with the endemism metrics, since they do not account for the global range size of non-CA-restricted taxa. There actually are some important areas right near the border that have concentrations of rare CA-restricted plants, but taxa that are rare in CA and abundant elsewhere also tend to occur at higher rates near the border.
It's an interesting idea to combine the full flora species richness and CA-restricted species endemism data in that way. I think your interpretation of what that E/D metric would represent is correct. The caveat is that because you're normalizing by total diversity, this index would give the same value to a location with 10 total species including 5 CA-restricted endemics and a location with 1000 total species including 500 CA-restricted endemics, if I'm understanding correctly. Depending on how you plan to use the data that might or might not be desirable, but it would likely include a lot of areas that are low diversity in absolute terms by both measures. I tend to think of metrics that are normalized by total diversity as more interesting from an ecological and biogeographic standpoint than a conservation standpoint, but that could be off-base.
Maybe another option could be to take the top 20% of sites by species richness, and also the top 20% by CA-restricted species endemism? This would of course yield less than 40% due to overlap between the two criteria, but if you were going for 20% total you cold find the percentage x such that the top x% of both species richness and CA species endemism together yield 20% of grid cells.
Just brainstorming here of course. Happy to find a time this week to chat via phone if that would be helpful.
Checking with Matt Kling on proposal:
We are using the Species Richness – Diversity data to identify areas with high plant richness. Because we have a categorical approach, we are thresholding ‘high’ at the top 20% of plant diversity in the state. I would like to supplement this data with areas that may be missed with this measure, but are important because they have a high richness of range-restricted plants (your measure of mean endemism). The problem is that with the full suite of species the mean endemism metric seems to find high richness of plants whose ranges may actually be extensive, but cross state boundaries. I’m guessing this because I am noticing an edge effect in the results. This makes me want to pull in the Em of the CA-Only species subset – but because it is normalized by diversity, I would like to use the diversity measure for all species to normalize because those are the areas that will be highlighted in our analysis, rather than the diversity of the subset of CA species which we are not currently using.
So the calculation would be E/D where E is the Endemism of CA-Restricted species and D is the Diversity of all species. I believe that this would capture areas that have MORE range-restricted CA-restricted species than you would expect to see, just given high plant species richness. This would help us make sure we were not missing range restricted endemics as we might if we prioritized just on plant richness.