Abiotic controls on warm-water carbonates through geological times

Description

Methods

The approach relies on (1) paleo-environmental ocean variables from the HadCM3 (BRIDGE) coupled atmosphere-ocean-vegetation Hadley Centre climate models, (2) the observed modern photozoan-T factory defined after Michel et al., (3) the updated continental margins and paleocoastlines dataset available on Zenodo based on the paleogeography maps from the PALEOMAP project, and (4) sediment flux derived from global scale landscape evolution model for the Phanerozoic HydroShare. In addition, sensitivity tests are making use of the fossil records from the Paleo Reefs PARED database.

Data Acquisition

Paleo-environmental oceanic variables from the HadCM3 climate models  for both smoothed and Foster et al. CO2 curves are available from the Bristol Research Initiative for the Dynamic Global Environment (BRIDGE) website: (BRIDGE) https://www.paleo.bristol.ac.uk}{https://www.paleo.bristol.ac.uk.

The observed modern photozoan-T factory is available from Laugié et al. 2019. The updated continental margins and paleo-coastlines dataset from Kocsis et al., 2021 are compiled in a Zenodo repository available from https://doi.org/10.5281/zenodo.3903163.

Finally, sediment and water flux derived from the global scale landscape evolution model Salles et al. 2023 are available for the Phanerozoic from HydroShare.

Scientific software

The scientific software used in this study, Biomod2, is written in R language and is available from CRAN from the following url: https://cran.r-project.org/web/packages/biomod2/index.html and the software documentation can be found in https://biomodhub.github.io/biomod2. Examples of R and Python scripts, as well as Jupyter Notebooks used in our pre- and post-processing workflows, are available from the following GitHub link: https://github.com/Geodels/paleoReef.

Table of contents

All provided data are available as geotiff files.

The dataset contains the environmental variables for each time slice (approximately every 5 Myr) and for 2 CO2 scenarios (Foster & Smooth curves - Foster_ensemble.zip and Smooth_ensemble.zip):

1. currents
2. ocean stream
3. SSS max
4. SSS min
5. SST max
6. SST min
7. sediment influx

Then the results from the species distribution model (BIOMOD2) ensemble models are provided for each CO2 scenarios and contains four variables (FosterCO2.zip and Smooth_CO2.zip): 

1. median of probabilities (median),
2. mean of probabilities (mean),
3. weighted mean of probabilities according to single evaluation scores (wmean) and
4. committee averaging score (ca):

Additional information are provided in the companion GitHub repository.