An integrated Bayesian stratigraphic correlation approach for the Cambrian Explosion

The Cambrian Explosion is characterised by a large diversification of life. The precise nature of this major evolutionary event is heavily debated, featuring anomalously fast versus more gradual evolutionary scenarios. Our ability to distinguish between such scenarios hinges on the quality of global correlations and corresponding timescales. With Cambrian temporal uncertainties often in the order of millions of years, establishing such correlations and timelines is a challenging task. Here, we present a novel approach to this problem based on a probabilistic Bayesian conceptual framework. Major advantages of the Bayesian approach include the consideration of multiple information sources in a single analysis and explicit uncertainty formulations.

In the absence of good index fossils, early Cambrian correlations rely heavily on carbon isotope chemostratigraphy and ‘expert-based’ correlations. Inspired by approaches in the radiocarbon community, we have been exploring representations of stable carbon isotope variations using random walk and spline fitting models. Implementation is undertaken using Markov-chain Monte-Carlo (MCMC) approaches. Temporal calibration is mainly dependent on published state-of-the-art U-Pb zircon dating. Our model also allows for the use of different sedimentary facies. Simultaneous analysis of several sections and multiple stratigraphic variables will allow each section to inform the correlation of every other, leading to a single, objectively derived and quantitative reference curve. Ultimately, the aim is to have a coupled Bayesian model setup of both stratigraphy and morphological evolution of the fossil record. These models will better inform us on the origins of diverse animal-dominated ecosystems and their impact on Earth processes.