MODELLING THE LIFE-ENVIRONMENT INTERFACE IN ANCIENT SHELF SEAS

The co-evolution of life and environment is a dynamic system of feedbacks. Yet studying events through a series of stratigraphic horizons reduces dynamic feedback loops to proxy correlations and invites speculation as to the cause-and-effect relationships. Models can suggest hypotheses to test ecosystem dynamics and the effects of changes to life or the environment on the other. Much of the evolution of life took place in localized shelf sea environments. Evolving biota and redox conditions created feedbacks which are hypothesized to have increased the ecospace for life to radiate - and sometimes perhaps brought about its own demise. As ecological complexity increased, these feedbacks created stable ecological networks which were robust to perturbations, but could also collapse in times of environmental stress when perturbation thresholds were exceeded. A particular modelling challenge is to connect these localized dynamics to global Earth system dynamics over long timescales. A hierarchy of models is needed to separate spatial and temporal scales and allow for the construction of models specific enough to be supported by limited geological data.

We introduce a 1D column model of an ocean shelf sea in the PALEO framework to represent the ecological dynamics of key functional groups representative of a Mesozoic ecological network. The model takes a Bayesian approach to parametrizing the ecological network, drawing from distributions that result in stable systems. This model demonstrates that ecological dynamics and nutrient cycling can be modelled together at the finest scales, while being computationally viable over geological timescales. Ongoing work integrating this model with data from critical time intervals of environmental stress and extinction in the Mesozoic ocean will provide specific hypotheses for the feedback relationship between environmental change and ecological collapse, in particular with regard to anoxia and temperature. The model will be both extended and constrained when applied to those intervals by networks of ecological functions. It is anticipated that model outcomes will yield insight into ecosystem functioning and stability, or conditions necessary to stability, during such critical intervals.