Multivariate hierarchical analyses of Miocene mollusk assemblages of Europe: Paleogeographic, paleoecological, and biostratigraphic implications

Abstract

Mollusk-dominated fossil assemblages from the marine Miocene bioprovinces of Europe (the Boreal province and the Paratethys) were analyzed by using bulk samples collected hierarchically across several spatiotemporal scales. The study explores quantitative patterns recorded by benthic mollusk assemblages (including implications for paleogeographic, biostratigraphic, and paleoenvironmental patterns), and the relationship between spatiotemporal scale of analysis and levels of variability in paleontological data. A total of 8455 specimens from 23 samples were collected from siliciclastic sediments (clay and sand) at 12 localities from France, Belgium, the Netherlands, Germany, Poland, Austria, and Hungary. With the exception of one sample of Serravallian age from the southeastern North Atlantic, all samples came from Burdigalian and Langhian deposits of the Paratethys and the Boreal province. All specimens were identified to species level (297 species total), classified in terms of ecology, and analyzed by using multivariate and computer-intensive methods. The analyses focused on (1) bivalve/gastropod ratios, (2) diversity levels, (3) multivariate taxonomic and ecological patterns, and (4) multivariate Euclidean distances between samples compared at five different scales (within sites, within localities, within facies, among facies, among provinces). The bivalve/gastropod ratio is a highly volatile parameter; thus, its informative value is highly limited. The estimates of alpha diversity (sample species richness) standardized by rarefaction indicate that fossil assemblages of the Boreal province display higher diversity levels than those of the Paratethys. This result is consistent with the general tendency of isolated epicontinental basins to display impoverished fauna and may be due to geographic barriers limiting larval dispersal and to environmental instabilities inherent to marginal environments of epicontinental seaways. Rarefactions of pooled data indicate that, for both provinces, the taxonomic turnover among sites (beta diversity) accounts for 30%–40% of species diversity. Multivariate analyses indicate that biogeographic variation between the provinces is the most important factor controlling differences among samples, whereas facies type and stratigraphic position of samples play a secondary role. The strong biogeographic signature is consistent with recent paleogeographic reconstructions that suggest a lack of a direct marine connection between the two provinces at that time. The relatively poorer discrimination of mollusk assemblages across facies and biostratigraphic stages reflects the overriding role of the middle Miocene biogeography in controlling the composition of mollusk assemblages. The hierarchical multivariate analysis shows that with the increase in the scale of observation (within sites, within localities, within facies, among facies, among provinces), the average multivariate distance between samples increases continuously—new heterogeneities that add notable amounts of new variation exist at each scale of observation. The highest increase is observed for the “among provinces” comparison level, again indicating the major role of geographic sources of variation in the data. The hierarchical approach provides a rigorous way to quantify the effects of scale-dependent sources of variability on paleontological patterns.