Analysis of sympatry networks: theoretical foundations, software implementation and applications

DOS SANTOS D.A.

Miami

Encuentro; 6th IBS Miami Meeting; 2013

International Biogeography Society

Locality points are the empirical evidence about the distribution of species. Those records become informative when sympatry relationships can be inferred from them across various taxa. Traditionally, areas of distribution are created by enclosing the points within polygonal figures. Then, the strength of sympatric association is estimated through the amount of overlap between those elementary geometric objects. Although commonly accepted, this strategy is intrinsically unrealistic and potentially misleading. The innovative contribution of this work is: 1) to provide quantitative guidelines for measuring the strength of the sympatric association between species, but based solely on dot maps; 2) to model the totality of pairwise scores into a weighted sympatry network. Here, I present a network approach to identify cohesive groups of species connected by sympatry links, called units of co-occurrence. These units fulfill the following duality: 1) within-group connectivity sustained by co-extensive sympatry and 2) between-group allopatry (or weak sympatry). I have implemented the new approach on the R package SyNet freely available on the web. I discuss here an unpublished analysis of 6,500 records of aquatic insects from southern South America. The main findings are: 1) Patagonia and Northwestern Argentina are complex entities composed of many units of co-occurrence which are nested and latitudinally organized. 2) The biotic vicinity of Uruguay is located at the Northeastern Argentina instead of Buenos Aires province. So, the inclusion of Uruguay into a Pampas region should be revised. 3) The trace of the pleistocenic arc (firstly proposed for trees) is recovered here.