How to create a species distribution model when I have only 5 occurence points?

102 views

Skip to first unread message

Lucie Fialova

unread,

Mar 20, 2017, 7:27:19 AM3/20/17

to Maxent

Hi everyone,

I focused on "Biodiversity distribution pattern of epiphytic mosses from Orthotrichaceae family". I am beginner and I need your advice.

Jus now I am solving a specific problem. I have to make species distribution model of one species of moss but I have only 5 occurence points from diffrent parts of the World. I want to use Maxent algorithm, but I am not sure if I can. I read that if I have very small database (less than 20 occurence points) Maxent can only provide a poor model.

I don't know how to solve this problem. Should I use diffrent species distribution modeling methods?

Thank you so much.

Best regards, Lucie Fialová.

Jamie M. Kass

unread,

Mar 21, 2017, 10:54:35 PM3/21/17

to Maxent

If you have only 5 points for your species, you probably won't be able to make a good species distribution model regardless of what algorithm you use (or any good statistical model for that matter -- a very low sample size usually means you can't complete a decent analysis). You might be able to make a model that may be somewhat helpful in answering where new records may be found, etc. If you want to experiment with Maxent, try tuning models (trying different parameter combinations) using the jackknife partitioning method (see Shcheglovitova and Anderson 2014). But really try to get at least 10 points or so...