Predicting the geomechanical response of marine sediments to hydrate dissociation using machine learning and fluid flow modeling ...

• Main Authors: Carty, Olin Rico, 0000-0002-4017-4721
• Format: Text
• Language: English
• Published: The University of Texas at Austin 2021
• Subjects: Methane hydrates; Geospatial machine learning; Methane hydrate
• Online Access: https://dx.doi.org/10.26153/tsw/48613; https://repositories.lib.utexas.edu/handle/2152/121787

With increasing global and ocean temperatures, the dissociation of methane hydrates at the feather edge of the of the gas hydrate stability zone (GHSZ) has become a greater concern to the scientific community. Possible responses to hydrate dissociation including seafloor methane venting and slope failure have been associated with gas generation and hydrate dissociation and have been used to create predictive maps of hydrate and gas locations around the world. Recently there has been a more concerted effort to model seafloor characteristics using machine learning methods to better estimate the global location of hydrate and gas formation. Seafloor total organic carbon (TOC) can be used to predict where methane hydrate and gas are likely to occur beneath the seafloor. I used a k-nearest neighbor machine learning model to predict global TOC at the seafloor. Within the region around the U.S. Atlantic margin (29°N–45°N and 82°W–66°W), I focused specifically on an area with high TOC predictions along the ...