Including phylogenetic and functional originality metrics in rarity indices provides a new prospective on which bee species are most rare

Background/Question/Methods

: Many indices can suggest which species are rare, or within the 1st to 25th percentile of highest rarity rankings, and in need of conservation intervention. Most species rarity indices include metrics to address commonness (numeric rarity) and range size (geographic rarity), but characteristics such as specialization are also associated with vulnerable species. Expertise on these specialized relationships (e.g., host, habitat) are invaluable for species assessments, but their application to all species is infeasible for bees and other incompletely documented taxa. Would an alternative calculation of species rarity rankings without habitat generalizations that is insensitive to incomplete life histories change our understanding of which species were most rare? We developed SORI (the Species Originality and Rarity Index), a species rarity index based on species traits and comprised of originality (phylogenetic, functional) and rarity (numeric, geographic, phenological) metrics from the bee taxon. We then compared relative positionings of 225 bee species in a three-year dataset from the National Forests of the Great Lakes Basin in the USA among SORI, and other index approaches that included numeric and geographic rarity (classic approach), as well as the classic approach with the addition of phenological rarity (R-metric approach) or the habitat sensitivity index (habitat approach).

Results/Conclusions

: When rankings were binned by the finest resolution (deciles), SORI rankings differed from classic and R-metric rankings, and further differed from the classic index at the coarsest resolution (quartiles). There were no differences between the R-metric and classic indices, suggesting that phenology did not provide much unique information on species rarity. We deduced that originality metrics provided enough new information to alter species positioning and consequently our interpretation of species relative rarity. Including a habitat sensitivity index altered the overall positioning of species from the classic index as well, but only across deciles. No general differences across bins existed between the habitat index and either the R-metric index or SORI. However, the distribution of position shifts for species across indices showed that the habitat sensitivity index only changed species ranks for the middle ~50% of species as ranked in the classic index. Alternatively, SORI altered species positioning in every decile, including those that represent rare species and was consequently the only index that provided an alternative composition of rare species from the classic approach. Thus, the inclusion of phylogenetic and functional originality metrics in SORI provided a new and more informed perspective on which species were the most rare.