A modularity-based approach for identifying biodiversity management units

Abstract

Background: Taxon- and/or ecosystem-based definitions of management units typically focus on conspicuous species and physical habitat limits; these definitions implicitly assume that these classification systems are related to the mechanisms that determine biodiversity persistence. However, ecological theory shows that this assumption may not be supported. Herein, we introduce the use of modularity analysis for objectively identifying management units and topological roles that land cover type plays on species movement through the landscape. Methods: As a case study, we used a coastal system in Uruguay, with 28 land cover types and five taxa (from plants to mammals). A modularity-based approach was used to identify subsets of habitats with biotic affinity, termed modules, across the different taxonomic groups. Modularity detects the tendency of some land cover types to have a higher probability of the mutual interchange of individuals than other land cover types. Based on this approach, pairs of habitats that co-occur in the same module across taxa were considered in the same biodiversity management units (BMU). In addition, the topological role of each habitat was determined based on the occurrence of species through the landscape. Results: Our approach determined three management units that combine land cover types usually considered independent, but instead are interrelated by an occurrence-based ecological network as proxies of the potential flow of individual and land use. For each selected taxon, the specific topological role of each habitat was determined. Conclusions: This approach provides an objective way of delineating spatial units for conservation assessment. We showed that land cover types within these spatial units could be identified as refuges for specific types of biodiversity, sources of propagules for neighboring or overall landscapes, or stepping-stones connecting sub-regions. The preservation of these topological roles might help maintain the mechanisms that drive biodiversity in the system. Interestingly, the role of land cover type was strongly contingent on the taxa being considered. The method is comprehensible, applicable to policy and decision-makers, and well-connected with ecological theory. Moreover, this approach complements existing methods, introduces novel quantitative uses of available information, determines criteria for land cover classification and identifies management units that are not evident through other approaches.