Global predictions of organic carbon burial using machine learning

durch Naveen Kumar Parameswaran (FB2-MG)

Mittwoch 20.08.2025, 13:15 → 14:30 Europe/Berlin

8A-002 - Hörsaal Ostufer / Lecture Hall East (GEOMAR - Standort Ostufer / GEOMAR - East Shore)

Abstract:

Predicting Total Organic Carbon concentrations(TOC %) and Mass Accumulation Rates(MAR) in marine sediments is essential for quantifying total organic carbon (TOC) burial, rain rates, and burial efficiency—key parameters for understanding the ocean’s role in the global carbon cycle. In this study, we utilize various oceanographic datasets encompassing geological, biological, and physical features. We apply a range of machine learning approaches, including tree-based models (Random Forest, XGBoost, CatBoost) and deep learning methods (neural networks, multi-task learning, TabPFN), to predict TOC % and MAR. By incorporating degradation rates from Jørgensen et al. (2021), we generate global maps of TOC burial, rain rates, and burial efficiency. Our results provide robust estimates of TOC burial and related fluxes, offering valuable insights into carbon sequestration processes in marine sediments.