Quantitative panbiogeography: was the congruence problem solved?

Abstract

Panbiogeography represents the spatial congruence among species distributions by means of generalized tracks. Some critics have suggested the method fails to objectively evaluate congruence, being neither repeatable nor falsifiable. The MartiTracks software was proposed to address spatial congruence using geometric properties as a counterpoint to the manual procedures so far employed in generalized track obtainment. To evaluate whether MartiTracks is a reliable alternative to the congruence problem in the quantitative panbiogeographic approach, we tested the software parameters with three analysis schemes under two real datasets. Then, we proceeded to a comparison of the results to those produced from Parsimony Analysis of Endemicity (PAE) and Clique Analysis, two quantitative methods which are based in predefined biogeographic areas or in the employment of grid cells. For PAE we used both analytical units, while Clique Analysis was restricted to grid cells. Through this, we aimed to comparatively evaluate the criteria of spatial congruence in different approaches. For each dataset and method, significantly different tracks resulted, highlighting the disparate congruence criteria among panbiogeographic approaches. Despite PAE ending up as the most reliable of the tools tested, it is still far from solving panbiogeographic congruence. The main focus of this paper, MartiTracks, is indeed a tool that makes minimum spanning tree construction a repeatable and easy-to-visualize process, but stumbles upon its obscure procedures of generalized track obtainment, congruence criteria, subjective parameter definition, the unclear implications of employing said parameters, and dubious results. Our results suggest that the subjectivity of the parameter setup process substantially influences the results, biasing them to the user-desired level of congruence. That the software produces fast and easy-to-visualize results does not make it a definitive solution to the problem of quantitative panbiogeographic approaches.