147-2 Determining Carbonate Chemistry Changes across the Toarcian Oceanic Anoxic Event using Ooid Sizes and Machine Learning

Session: Climate, Ocean and Environmental Changes Through Earth History: From Marine and Terrestrial Proxies to Model Assessments (Posters)

Presenting Author:

Authors:

Abstract:

The Toarcian Oceanic Anoxic Event (TOAE) is a global climate perturbation in the Early Jurassic Period. The environment changed rapidly due to the Karoo-Ferrar large igneous province eruptions, which released abundant greenhouse gases, such as CO₂, into the atmosphere, causing global temperature increases. It is also speculated that ocean acidification occurred during the TOAE; however, current data are insufficient for a definitive conclusion. Reconstructing ocean acidification in deep time is difficult due to the corrosive nature of acidification.

The scientific community has primarily tried to identify paleo-acidification using dissolution horizons (i.e., carbonate gaps) or boron and strontium stable isotopes. Here, we present another proxy for ancient carbonate chemistry: using ooid size as a proxy for acidification across the TOAE. Ooids are sand-sized (≤2 mm)  spherical carbonate grains made via carbonate precipitation around a central nucleus, which can be a skeletal fragment, grain, or other material. The carbonate saturation state (Ω) is proportional to ooid size; therefore, we can use ooid grain size to model Ω in deep time.

Oolites from the High Atlas Mountains in Morocco were chosen because several field sites preserve high-resolution records of ooid grainstones from before, during, and after the TOAE interval. 

Thin sections of oolites from Morocco were photographed using a Zeiss Imager M2m petrographic light microscope. These microphotographs were analyzed with ‘segmenteverygrain’, a Python package, which outlined what it identified as ooids (i.e., circular grains). Initially, ‘segmenteverygrain’ was inaccurate because it was not trained on oolites. After manually selecting ooids from an initial test, an overlay file with ooids more accurately outlined was saved and sent to a machine learning notebook. This process created a new model for ‘segmenteverygrain’ to use as a basis for ooid selection. After ooids were selected, the program measured major and minor axes of each grain. 

The use of ‘segmenteverygrain’ to measure ooid size can allow faster data production and future use in paleo-ocean acidification modeling. This work provides data for modeling ocean acidification across the TOAE and sets a precedent for further research into ocean acidification across other intervals in deep time.

Geological Society of America Abstracts with Program. Vol. 57, No. 6, 2025

doi: 10.1130/abs/2025AM-9418

© Copyright 2025 The Geological Society of America (GSA), all rights reserved.