Influence of different species range types on the perception of macroecological patterns

Abstract

In the face of increasing availability and use of distribution data, large-scale approaches of mapping species distribution patterns have become a central component of development of large-scale conservation policies. Particularly in tropical regions and for non-vertebrate taxa, knowledge on distribution patterns at large spatial extents remains woefully limited. Datasets are often geographically and taxonomically incomplete, have presence-only character and lack abundance information. One intermediate step for the application of such data common to most approaches is the construction of species geographic ranges. In this study, we investigated the effects of different methods for constructing species ranges on range attributes and species richness. We selected the Neotropical palm genus Bactris as a typical example for a diverse tropical plant taxon that is comparatively well researched, but under collected. For 48 Bactris species, we compared point-to-grid ranges, expert ranges, convex polygons and modelled ranges. Range attributes and resulting species richness patterns differed tremendously and were constrained by incompleteness of the respective data type and by the errors associated to the method itself. The consequences of applying different methods to construct species ranges highlighted here can aid in selecting appropriate methods for analysing distribution data at large geographic scales.