Rarity, richness, and randomness:
assessing the sampling effect of biodiversity at the local scale
The presence of rare and endemic species is often used to identify areas of conservation priority. However, it is unclear if rare species diversity varies independently of overall species richness or if it is instead an artifact of a sampling effect of biodiversity. We created a null model to test whether, through a random sampling effect, the presence of locally rare species is primarily a function of overall species richness. We conducted a thorough literature review for biodiversity datasets from various taxa occupying environmentally-bound sites (e.g., ponds and islands) and defined rare species as those occurring at 10% or fewer sites in a dataset. To develop the null model, we used a 2×2 swapping algorithm that simulated random combinations of species at different sites, with fixed species occupancy and local richness. In other words, each site retained the same number of species and each species occupied the same number of sites in the simulations. Our simulation results showed a strong, positive curvilinear relationship between total species richness and the number of expected rare species. Across all sites, expected rare species richness from the simulation analysis was a remarkably strong predictor of the observed number of rare species per site (RÂ²=0.92; p<0.001), providing strong support for the sampling effect of biodiversity in determining the number of rare species per site. Within individual systems, however, our results were more variable, including cases of strong (gastropods: RÂ² = 0.99) and weak relationships (aquatic invertebrates: RÂ² = 0.094) between expected and observed rarity values for certain taxa. Ultimately, our study suggests that at a local level, management efforts directed towards habitats with high species richness are also likely to capture large proportions of rare species of high conservation concern and provides additional evidence for the sampling effect of biodiversity.