Towards a Dynamic Analysis of Weighted Networks in Biogeography

    loading  Checking for direct PDF access through Ovid

Abstract

An improvement to the Network Analysis Method (NAM) in Biogeography based on weighted inference and dynamic exploration of sympatry networks is proposed. Intricate distributions of species result in a reticulated structure of spatial associations. Species are geographically connected through sympatry links forming an overall natural network in biogeography. Spatial records are the signals that provide evidence to infer these sympatry links in the network. Punctual data are independent of a priori area determination. NAM is oriented to detect groups of species embedded into the global network that are internally sustained by sympatric cohesiveness but weakly connected (or disconnected) to outgroup entities. These groups, called units of co-occurrence (UCs), are segregated through the iterative removal of intermediary species according to their betweenness scores. Instances of analysis of the original NAM are improved through the following changes and extensions: (i) inference of weighted sympatry networks using new measures sensitive to the strength of overlap and topological resemblance between set of points; (ii) construction of a basal network discriminating major from minor sympatry associations; (iii) evaluation of the entire process of iterative removal of intermediary species for the selection of UCs found on different subnetworks; (iv) network partitioning based on the intrinsic cohesiveness of the UCs; (v) production of a graphical tool (cleavogram) depicting the structural changes of the network along the removal process. Improvements are tested using real and hypothetical data sets. Resolution of patterns is notably increased due to a more accurate recognition of allopatric patterns and the possibility of segregating spatially overlapped UCs. As in original NAM, spatial expressions of UCs are building blocks for biogeography supported by strictly endemic and connected species through sympatry paths.

Related Topics

    loading  Loading Related Articles