NDM-VNDM: Programs for identification of Areas of Endemism

GOLOBOFF, PABLO

Bruselas, Belgica

Congreso; X Meeting of the Global Biodiversity Information Facility; 2005

Global Biodiversity Information Facility (GBIF)

The notion of areas of endemism can be traced back to the early eighteen hundreds.  A species (or taxon) is said to be endemic of an area when it is found in that area, and nowhere else.  If the limits in distribution of different groups of plants and animals are determined by the same factors, it is expected that the distributions of different groups will show similar patterns, and that some areas or regions –areas of endemism-- will consistently house endemic taxa.  Although the notion of endemism is an old one, attempts to identify them by formal, quantitative means, have begun only recently.  The most widely used method simply records presence/absence data in a grid, and subjects the resulting data matrix to analysis by phylogenetic algorithms.  Phylogenetic algorithms, however, are specifically designed for reconstructing phylogeny, and have many problems when used to determine areas of endemism.  Recent work by Szumik et al. (2002) and Szumik & Goloboff (2004), has proposed criteria designed specifically for recognition of areas of endemism. These criteria are implemented in two interacting computer programs, NDM and VNDM.  The programs take as input locality data, and convert them into a grid of presence/absence (optionally, probable presence can be used as a third category). The criterion used is very simple: the cells in the grid can be grouped in different ways;  for a given set of cells (or "area"), a given species can be considered as "endemic" of that area if it is more or less evenly distributed across the area, and not found in distant cells.  Thus, choosing sets of cells for which many taxa are endemic provides a natural way to identify areas of endemism.  The problem here is computational, because the program has to examine, in principle, all possible cell combinations; for a grid of c columns and r rows, there are  SUM[i=2,i=c.r-1] (c.r)! / i! (c.r-i)! possible combinations of cells.  Even for a modest c=10, r=20, there are more than 22 x 10^9 possible sets of cells.  For large data sets, it is impossible to actually examine every possible cell combination, and a trial-and-error technique (similar to techniques used in phylogenetic analysis, where the author has extensive experience) is used by the program NDM. The procedure uses promising sets of cells as starting points, modifies them by deleting/adding cells, reevaluates the degree of endemicity, and keeps the sets of cells with highest endemicity scores. Two additional problems posed by endemicity analysis are discussed: partially overlapping areas, combination of several areas into one (akin to consensus techniques in phylogenetic analysis).  An analysis of a real data set is presented to illustrate the approach.  A future project will try to apply a similar approach to the identification (and sequencing) of distributional barriers, which is one of the main problems in historical biogeography.