Disentangling the drivers of temporal and spatial biotic patterns

Abstract:

Environmental changes in time and across space are multivariate, thus understanding the drivers of biotic responses to paleoclimate events requires the incorporation of multiple proxies and selection of the variables most associated with the biotic patterns. Here, two case studies, one examining paleoecological change leading into the Early Miocene warming and one examining global diversity patterns in modern bivalves, illustrate the utility of multivariate data sets for understanding biotic patterns.

We create a multivariate time series of benthic foraminiferal faunal composition and environmental variables (δ¹³C, Δδ¹³C, δ¹⁸O, δ¹⁵N, sediment grain size) from the Early Miocene Astoria Formation spanning ~18-20 mya. We then use multivariate statistics and maximum likelihood model selection to disentangle the potential drivers of the faunal changes. We find that d¹⁵N values and age are the most parsimonious correlates with major changes in foraminiferal composition, suggesting oxygenation is primarily affecting the foraminiferal community. Failure to include δ¹⁵N in the analysis still yields significant and supported relationships with Δδ¹³C, which would lead to the incorrect interpretation that the benthic foraminifera are responding primarily to organic carbon flux rather than oxygenation.

Similarly, we examine the environmental factors associated with global diversity patterns. Using occurrence data for modern bivalves and a multivariate oceanographic data set, we identify the modern environmental factors most associated with diversity. However, inclusion of spatial variables in addition to environmental variables in the analysis reveals a well-supported relationship between proximity to diversity hotspots and diversity, suggesting historical processes also play a key role in diversity patterns.

Because environmental variables can be coupled in time and in space, it is important to consider multiple environmental, temporal, and spatial variables, and their interactions, to disentangle the drivers of biotic patterns. Studies that incorporate multiple variables can be powerful tools for identifying the drivers of biotic patterns and projecting biotic responses to future climate changes.