ABUNDANCE AND SELECTIVE EXTINCTION: MODELING THE EFFECTS ON EVENNESS

Extinction studies typically assess declines in richness, thus overlooking variations in evenness, the other key component of biodiversity. Because evenness reflects the relative abundance distribution (RAD) of a biota, variation in evenness across extinction boundaries should highlight changes in community structure. Quantitative null models of changes from pre- to post-extinction conditions in several evenness metrics (Shannon Index H, Buzas and Gibson E, Pielou's J, and Hurlbert's PIE) as a function of percent extinction are presented. Evenness patterns are examined for communities that initially exhibit either lognormal or geometric RADs. Several conditions of extinction with respect to taxon abundance are tested, including non-selective extinction, extinction targeting abundant taxa, and extinction targeting rare taxa.

For both initial RADs, expected changes in evenness metrics from pre- to post-extinction are similar in direction and magnitude. In all the models, taxonomic richness, H, and PIE decline across the extinction boundary, regardless of the selectivity of extinction. E increases across the extinction boundary regardless of the selectivity condition; however, the increase will be most substantial for the non-selective condition. J has unique trends across the extinction boundary for each extinction selectivity condition, with a large increase in J when abundant taxa are targeted, a decrease in J when rare taxa are targeted, and a small increase in J when taxa are not selected with respect to abundance, respectively. These trends in evenness metrics suggest it may be possible to determine whether extinction was selective, and if so, whether rare or abundant taxa were more likely to become extinct, by comparing these metrics from pre- and post-extinction communities.

The model results were compared to data collected across a Late Ordovician regional extinction boundary from the Appalachian Basin of the eastern United States. Field data collected across the regional extinction boundary yield a decrease in J, which corresponds to extinction that targets rare taxa.