Aim: Island biogeography theory states that species richness increases with habitat diversity and decreases with isolation from source pools. However, ecological theory must incorporate effects of human activity to explain contemporary patterns of biodiversity. We contemporized island biogeography theory by conceptualizing island trajectories of how species richness changes over time with accelerating land development and economic trade, which increase extinction and immigration rates, respectively. With this contemporized theory, we then articulate and empirically assess expected relationships of native, introduced, and total species richness with natural and anthropogenic metrics of habitat diversity and isolation from source pools.

Location: Greater Caribbean region.

Time period: Database finalized in 2020.

Methods: We built a database of 1075 native and introduced reptiles and amphibians (herps) for 840 Caribbean islands. For each island, we calculated natural and anthropogenic metrics of island habitat diversity and isolation from source pools and used linear model averaging to assess the expected relationships under the contemporized theory for 15 major herp clades.

Results: Natural habitat diversity metrics exhibited positive relationships with native and introduced species richness, strengthening total species richness–area relationships across herp clades. Geographic isolation exhibited negative relationships with native and positive relationships with introduced species richness, weakening total species richness–isolation relationships. Economic area, based on developed land, and economic isolation, based on maritime trade, exhibited negative relationships with native species richness, but positive and negative relationships, respectively, with introduced species richness. Total species richness relationships with these two anthropogenic metrics were strongest in clades with many introduced species.

Main conclusion: A contemporized island biogeographic theory that includes the effects of land development and economic trade on species extinction and immigration explained current Caribbean herp species richness patterns. As human activity continues to accelerate, the contemporized theory we articulate here will increasingly predict island biogeography of the Anthropocene.

We built a database of 840 islands based on the Global Administrative Areas shapefile (version 3.6, GADM, 2012) grouped into 72 banks. Banks were based on historical land connections and underwater topography (GEBCO Compilation Group, 2020) and included Bermuda.

Species Richness Data

We performed an extensive literature search of island-level herp records (searches completed June 2020; see Methods Supp. Info. Section 1 of the manuscript). Native status was given to extant species indicated in the literature as being native to an island. Introduced status was given to non-native species established on an island indicated from the literature or if a non-native species had multiple records in Global Biodiversity Information Facility, which included data from multiple sources (see Methods Supp. Info. Section 1 of the manuscript), that spanned multiple geographic locations and years on an island. We used several sources to standardize species names and taxonomy including caribherp.org (Hedges, 2020), Reptile Database (Uetz et al., 2020), Amphibian Species of the World 6.1 (Frost, 2020), the Integrated Taxonomic Information System (ITIS, accessed June 2020), and GBIF. Our search yielded 1075 extant species of herps on 72 banks. Species were grouped into clades defined by taxonomic class, order, suborder, family, and genus. We identified 154 reptilian and 42 amphibian clades. Native, introduced, and total species richness values per bank per clade were calculated. Not all clades contained enough species on enough banks to fit robust regressions. Based on a power analysis of the fit of the all-herps model, only 15 clades were retained for analysis. 

Island Bank Data

We calculated four common natural habitat diversity metrics: geographic area, island spread, topographical complexity, and natural area; and one anthropogenic metric: economic area. Geographic area was the sum of contemporary land area of each island in each bank from the Global Administrative Areas shapefile (version 3.6, GADM, 2012). Island spread was one minus the total bank land area divided by bank extent estimated as the area of the minimum convex polygon around the islands. Topographic complexity was the standard deviation of terrestrial elevation of each bank from STRM Digital Elevation Data (90m resolution; version 4, Jarvis et al., 2008). Natural area and economic area were the total proportion of bank area covered by natural or economic land cover types in the Annual International Geosphere-Biosphere Programme classification layer of the MODIS Land Cover Type Yearly Global 500m data (Friedl & Sulla-Menashe, 2015). Natural area included all forest, grassland, wetland, savanna, and shrubland land cover types. Economic area included all cropland and urban land cover types. The “cropland/natural vegetation mosaics” type was split evenly between the two metrics. Bare ground and permanent water cover types were excluded.

We calculated a natural and an anthropogenic metric of source pool isolation: geographic isolation and economic isolation. Mainland South and Central America plus the continental islands Cuba and Hispaniola have acted as natural source pools for all Caribbean herps. We used the minimum distance from the geographic centroid of a bank to the shoreline of the nearest mainland or continental island source as geographic isolation (see Methods Supp. Info. Section 3 of the manuscript for comparison to other isolation metrics). We estimated economic isolation from individual ship dockings at ports in the greater Caribbean region in March, June, September, and December of 1979, 1991, 2003, and 2015 for cargo, cruise, and passenger ships (Loyd’s, 2020). We took the inverse of one plus the summed total number of ship visits from outside a bank to docks within each bank across all months, years, and ship types as our economic isolation metric.Bank population density was estimated as the number of people living on a bank averaged over every five years from 2000 to 2015 (Population Count version 4, CIESIN, 2018) divided by bank area.

References:

CIESIN. (2018). Gridded Population of the World, Version 4 (GPWv4): Population Count, Revision 11 [Data set]. Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC). https://doi.org/10.7927/H4JW8BX5

Friedl, M., & Sulla-Menashe, D. (2015). MCD12Q1 MODIS/Terra+Aqua Land Cover Type Yearly L3 Global 500m SIN Grid V006 [Data set]. NASA EOSDIS Land Processes DAAC. https://doi.org/10.5067/MODIS/MCD12Q1.006

Frost, D. R. (2020). Amphibian Species of the World. Amphibian Species of the World: An Online Reference. Version 6.1. https://amphibiansoftheworld.amnh.org/

GADM. (2012). Global Administrative Areas (3.6) [Digital Geospatial Data]. University of California, Berkley. www.gadm.org

GBIF. (2020). Global Biodiversity Information Facility.

GEBCO Compilation Group. (2020). GEBCO 2020 Grid [Map]. doi:10.5285/a29c5465-b138-234d-e053-6c86abc040b9

Hedges, S. B. (2020). Caribherp. Caribherp: Amphibians and Reptiles of Caribbean Islands. http://www.caribherp.org/

Jarvis, A., Reuter, H. I., Nelson, A., & Guervara, E. (2008). Hole-filled seamless SRTM data V4 [Map]. International Centre for Tropical Agriculture (CIAT). http://srtm.csi.cgiar.org

Loyd’s. (2020). Lloyd’s of London. https://www.lloyds.com/

Uetz, P., Freed, P., & Hošek, J. (2020). The Reptile Database. http://reptile-database.org/

All data are provided in csv format and can be opened with Microsoft Excel or any text editor.