Changes in Late Cretaceous–early Tertiary benthic marine assemblages: analyses from the North American coastal plain shallow shelf

Abstract

The Mesozoic–Cenozoic transition is generally seen as a pivotal time in the evolution of benthic marine assemblages but the details of the timing and drivers of these changes are poorly known. The Atlantic and Gulf Coastal Plains of the United States contain assemblages preserved as original aragonitic and calcitic material in unconsolidated sediments. This makes coastal plain assemblages ideally suited to paleoecological analyses. Data derived from bulk samples of the Coffee Formation (lower/middle Campanian: Mississippi) as well as published faunal lists from comparable samples of the Severn (Maastrichtian: Maryland), Providence (Maastrichtian: Georgia and Alabama), Stone City (Eocene: Texas), and Gosport (Eocene: Alabama) Formations are used to assess changes in taxonomic diversity and ecomorphological group (life habit and trophic group) composition through this time interval.These analyses find a significant decrease in rarefied-sample species richness from the Campanian through the Eocene, but no change in evenness. With the notable exception of the Stone City Formation, increases in carnivore (neogastropod) richness and abundance occur before the Campanian. Epifaunal suspension-feeding species are a smaller proportion of the sample richness in Eocene samples than in Cretaceous samples. Decreased relative epifaunal suspension-feeder biomass but unchanged relative numbers of epifaunal suspension-feeder individuals suggests a relative decrease in epifaunal suspension-feeder size. Infaunal suspension feeders increase in richness and abundance through the interval. The proportion of drilled bivalves and gastropods does not change through the interval. Changes found in the structure of local shallow-shelf benthic assemblages from the Campanian through the Eocene are generally small relative to the variability between samples. Formation-level variation between assemblages is high relative to the magnitude of the temporal signal, emphasizing the need for investigators to include multiple formations per interval in tests of temporal trends.