Paper No. 172-12
Presentation Time: 11:00 AM
USING PALEOENM TO PREDICT PATTERNS OF SURVIVORSHIP IN THE HELL CREEK FORMATION ECOSYSTEMS ACROSS THE K/PG MASS EXTINCTION
The Cretaceous/Paleogene (K/Pg) extinction is the latest of the big five mass extinctions. Patterns of survivorship often correlate with geographic range size, which is a product of abiotic habitat, biotic interactions, and dispersal. Here we investigated changes in abiotic environmental habitat size in geographic and environmental space for a diverse suite of taxa from the Hell Creek biota. Paleoenvironmental conditions in the Maastrichtian and Danian were characterized using sedimentological proxies; these data were combined with taxon occurrences and analyzed using a paleoecological niche modeling (PaleoENM) approach, which correlates occurrences with the spatiotemporally specific environmental factors in which they occur, to provide a prediction of suitable abiotic habitat (SH). The Maxent modeling algorithm was used to analyze niche changes in geographic space and the Ecospat package in R was used to analyze changes in environmental space. PaleoENMs were trained in the Maastrichtian, then projected into the Danian to determine the availability and size of a taxon’s SH after the extinction interval. 13 out of 38 total genera demonstrate good model performance (AUC ≥ 0.85) and are interpreted here. Results show no consistent pattern between Danian SH and survivorship. All extinct taxa experience increases in available SH, whereas surviving taxa show a mix of increases and decreases in SH. Ecospat analysis reveals that niche occupation decreased in nearly all modeled species, which reflects an overall decrease in size of environmental space in the Danian compared to the Maastrichtian. This suggests that geographic range size is potentially decoupled from environmental habitat size across this interval; i.e., while abiotic habitat diversity decreased across the boundary, the geographic area of remaining SH increased. These results suggest that extinction mechanisms in these taxa were not related to changes in SH, and instead reflected (1) the proximate causes of the K/Pg extinction and (2) alternative recovery conditions such as SH accessibility and/or the Danian biotic environment.