Analysis of endemism of world arthropod distribution data supports biogeographic regions and many established subdivisions
Jonathan Liria,Claudia A Szumik,Pablo A Goloboff
DOI: https://doi.org/10.1111/cla.12448
Cladistics
Abstract:We analyzed 769 242 occurrence records for 115 424 species of terrestrial arthropods, from three biodiversity repositories (Global Biodiversity Information Facility (GBIF), Natural History Museum, London, and "Sistema de Informação Distribuído para Coleções Biológicas" (SpeciesLink)), to test the use of global-scale data points for quantitative assessments of areas of endemism. The data include Insecta (105,941 species), Arachnida (7984 species), Myriapoda (1229) and terrestrial crustaceans (270 Branchiopoda). The species were assigned to 14 543 higher taxonomic groups because such groups often characterize larger areas of endemism. Putative areas of endemism were visualized as sets of cells displaying unique groups of species without the assumption of hierarchical relationships. Yet, the use of 10° grid cells recovered many large areas broadly corresponding to biogeographic Regions (Nearctic, Neotropical, Panamanian, Palaearctic, Afrotropical, Australian, Oceanian and Oriental) albeit with the limits poorly defined. An analysis of 5° grids resulted in 306 sets included in the different biogeographic Realms: Afrotropical, Australian, Madagascan, Nearctic, Neotropical, Oceanian, Oriental, Palaearctic, Saharo-Arabian and Sino-Japanese. The Panamanian Realm comprises 89 partly overlapping sets, crossing the Nearctic and Neotropical boundaries. A total of 7338 species of Insecta were endemic to some areas (Sino-Japanese, Afrotropical, Panamanian, Palaearctic, among others), followed by Arachnida (412 spp) and 105 species in other clades ranked as "classes". Six sets were supported only by genera, except for Panamanian sets that were supported by genera and families. Many of the species in the dataset are included in IUCN red lists, but probably most of those have distributions more restricted than global areas of endemism; only 102 appear as endemic to some area (Neartic, Madagascan, Panamanian, Afrotropical, among others). The results show that data from global databases can be used to identify areas of endemism on a worldwide basis but-owing to their incompleteness-only at a relatively coarse level. At the level of resolution currently allowed by such databases, such global studies are only complementary to studies where areas are determined subjectively by systematists (instead of actual point records), or studies using point records in datasets for specific taxonomic groups curated and compiled by specialists.