98 D1. Recent Advances in Economic Geology (Posters)

Monday, 23 September 2024
8:00 AM - 5:30 PM
Hall D

PREDICTIVE MODELING OF CRITICAL METAL ABUNDANCE IN SEAFLOOR MASSIVE SULFIDE DEPOSITS: A MACHINE LEARNING APPROACH

Paper number: 98-4
Program: Technical Sessions
Time: 8:00 AM – 5:30 PM

Authors

Abstract

Seafloor massive sulfide (SMS) deposits exhibit complex geochemical variability, requiring a deeper understanding of the factors influencing their metal content. This study utilizes a global geochemical dataset of SMS deposits to develop robust classification and prediction models using a machine learning (ML) approach.

We present a novel multi-stage ensemble learning framework to predict critical metal concentrations within SMS systems. We employ K-Means clustering to partition the dataset, incorporating both trace metal concentrations and geological features (spreading rate, depth), and create a new classification scheme. A Random Forest classification model then leverages this classification scheme to assign the appropriate class label for new vent field data lacking trace metal content. Finally, an XGBoost regression model predicts critical metal compositions for SMS associated hydrothermal vent systems.

Our results offer an informative framework for estimating critical metal abundance in SMS deposits while revealing important insights into the influence of spreading rate, depth, hydrothermal fluid temperature, and sedimentation. Notably, Co content reveals a strong correlation with depth, hydrothermal fluid temperature, and potential association with pyrite formation. These findings demonstrate the potential of ML for unraveling the complex geochemical processes governing the formation and enrichment of critical metals within SMS deposits. This framework provides valuable insights and paves the way for a more comprehensive understanding of these critical marine resources.

Geological Society of America Abstracts with Programs. Vol. 56, No. 5

doi: 10.1130/abs/2024AM-403189

© Copyright 2024 The Geological Society of America (GSA), all rights reserved. Permission is hereby granted to the author(s) of this abstract to reproduce and distribute it freely, for noncommercial purposes. Permission is hereby granted to any individual scientist to download a single copy of this electronic file and reproduce up to 20 paper copies for noncommercial purposes advancing science and education, including classroom use, providing all reproductions include the complete content shown here, including the author information. All other forms of reproduction and/or transmittal are prohibited without written permission from GSA Copyright Permissions.