Climatic refugia and reduced extinction correlate with underdispersion in mammals and birds in Africa
Data files
Apr 18, 2022 version files 1.82 GB
-
fixed_birds_pam.csv
-
mammal_pam_coords.csv
-
README.txt
Abstract
Macroevolutionary patterns, often inferred from metrics of community relatedness, are often used to ascertain major evolutionary processes shaping communities. These patterns have been shown to be informative of biogeographic barriers, of habitat suitability and invasibility (especially with regards to environmental filtering), and of regions that function as evolutionary cradles (i.e., sources of diversification) or museums (i.e., regions of reduced extinction). Here, we analysed continental datasets of mammal and bird distributions to identify primary drivers of community evolution on the African continent for mostly-endothermic vertebrates. We find that underdispersion (i.e., relatively low phylogenetic diversity compared to species richness) closely correlates with specific ecoregions that have been identified as climatic refugia in the literature, regardless of whether these specific regions have been touted as cradles or museums. Using theoretical models of identical communities that differ only with respect to extinction rates, we find that even small suppressions of extinction rates can result in underdispersed communities, supporting the hypothesis that climatic stability can lead to underdispersion. We posit that large-scale patterns of under- and overdispersion between regions of similar species richness are more reflective of a particular region’s extinction potential, and that the very nature of refugia can lead to underdispersion via the steady accumulation of species richness through diversification within the same ecoregion during climatic cycles. Thus, patterns of environmental filtering can be obfuscated by environments that coincide with biogeographic refugia, and considerations of regional biogeographic history are paramount for inferring macroevolutionary processes. --
Methods
Full data processing methodology is available in the associated manuscript.
Usage notes
Codes are detailed in the attached pdf. This pipeline was formatted to run on a Linux operating system; filepaths have been redacted and adjustments will need to be made for other operating systems.