DOES ECOLOGICAL CHANGE SCALE WITH PERCENT EXTINCTION? QUANTIFYING THE DIFFERENCE BETWEEN TAXONOMIC LOSS AND FUNCTIONAL ECOLOGY
Using data sets from the Eastern U.S., we quantify the relationship between taxonomic and ecological change across three extinction events of increasing genus extinction intensity: the early Late Ordovician M4/M5 extinction, the Mid-Late Devonian mass extinctions (Givetian/Frasnian and Frasnian/Famennian boundaries), and the end-Ordovician mass extinction. Taxonomic and ecological changes during each extinction were contrasted by classifying organisms into genera and guilds, followed by analysis with additive diversity partitioning, ordination, and relative abundance distributions.
In terms of additive diversity partitioning, taxonomic extinction is similar for both the Ordovician/Silurian and Mid-Late Devonian extinction events, but the Devonian shows much greater effects in terms of guilds. When controlling for the effects of extinction alone (i.e., excluding origination/immigration), guild diversity remains high after the Ordovician/Silurian event, but decreases drastically after the Devonian extinctions. The M4/M5 shows minor extinction for both genera and guilds. In terms of ordination, the greatest taxonomic separation was observed across the Ordovician/Silurian boundary while the greatest ecological separation was observed across the Devonian extinctions. In terms of relative abundance distribution analyses, the total change in relative abundance was similar for the Ordovician/Silurian and Devonian extinctions in terms of genera and guilds. Taken together, this suggests that while many taxa go extinct across the Ordovician/Silurian boundary, post-extinction taxa fill the same ecological roles as pre-extinction taxa. In contrast, post-extinction taxa in the Late Devonian only inhabited a fraction of the ecological roles as pre-extinction taxa. These results underscore that the magnitude of taxonomic change during an extinction can be a poor predictor of the amount of ecological change.