Human activities, such as habitat loss and land use, threaten thousands of species with extinction. However, there is a large amount of variability in predictions of extinction risk due to lack of information on species responses to anthropogenic changes and discrepancies between historical and current extinction patterns. Functional traits have emerged as a valuable framework for predicting responses to change. Trait differences may also reflect differences in species likelihood of extinction and have been used to examine contemporary patterns of species loss. Phylogenetic signatures in extinction events can also provide insights into mechanisms of species loss. Finally, incorporating historical data on species extinctions could be a valuable resource to detect patterns in species extinctions. Examination of a large sample of specimens can potentially reveal the traits that influence the distribution of taxa across habitats and inform their vulnerability to local extinction. We use historical botanical data from Kalamazoo County, MI, to examine patterns of plant extinction events. We use a model testing the role of rarity (the number of MI counties a species is found in), habitat types, and species characteristics, as well as phylogenetic regression, to investigate the role of functional groups and phylogenetic patterns in local biodiversity loss.
Results/Conclusions
Across Kalamazoo County, extinction is influenced by rarity (the number of MI counties a species is found in) and habitat type. Prairies experience higher rates of extinction than any other habitat (Χ2=19.45, p<0.0001). 12.8% of the county’s prairie species disappeared from the 1890s-1990s. The prairie species were classified into a number of functional groups. We found that forbs and vines tend to disappear more than other growth forms (Χ2=15.93, p<0.01) and habitat specialists are more susceptible to local extinction (Χ2=17.57, p<0.0001). Rare species are also at higher risk of extinction (Χ2=5.32, p<0.05). Accounting for phylogenetic relationships does not change these results, and no plant families have a higher risk of extinction than others (Tukey test, p>0.05). This nonrandom distribution of extinction reflects the idea that certain species attributes are beneficial or detrimental in the face of anthropogenic change. Lack of phylogenetic signal suggests that factors unrelated to traits, such as habitat loss or other anthropogenic changes, may shape extinction risk. Understanding the traits and phylogenetic patterns associated with extinction risk will help inform predictions of future extinction risk for threatened species.