Optimizing an Inner-Continental Shelf Geologic Framework Investigation through Data Repurposing and Machine Learning

Abstract

The U.S. Geological Survey (USGS) and the National Oceanic Atmospheric Administration (NOAA) have collected approximately 5400 km2 of geophysical and hydrographic data on the Atlantic continental shelf between Delaware and Virginia over the past decade and a half. Although originally acquired for different objectives, the comprehensive coverage and variety of data (bathymetry, backscatter, imagery and physical samples) presents an opportunity to merge collections and create high-resolution, broad-scale geologic maps of the seafloor. This compilation of data repurposes hydrographic data, expands the area of geologic investigation, highlights the versatility of mapping data, and creates new geologic products that would not have been independently possible. The data are classified using a variety of machine learning algorithms, including unsupervised and supervised methods. Four unique classes were targeted for classification, and source data include bathymetry, backscatter, slope, curvature, and shaded-relief. A random forest classifier used on all five source data layers was found to be the most accurate method for these data. Geomorphologic and sediment texture maps are derived from the classified acoustic data using over 200 ground truth samples. The geologic data products can be used to identify sediment sources, inform resource management, link seafloor environments to sediment texture, improve our understanding of the seafloor structure and sediment pathways, and demonstrate how ocean mapping resources can be useful beyond their original intent to maximize the footprint and scientific impact of a study.

Quantifying the Physical Impact of Bottom Trawling Based on High-Resolution Bathymetric Data

“…However, the training of a model to recognize trawl marks would be difficult because of the different shapes and scales involved (from holes with few decimeters in diameter to several-kilometer-long lines) [27]. In recent years, new methods based on morphological parameters to classify seafloor features on larger spatial scales were developed [28,29]. Since trawl marks are significant local depressions compared to the surrounding seafloor, parameters such as slope and bathymetric positioning index can be used for their visualization.…”

“…However, the morphological appearances of trawl marks comprise both linear depressions and isolated holes. Therefore, it is difficult to automatically derive a local trawling intensity from several interacting parameters, as has been suggested for ridges and valley bottoms [28][29][30]. A manual interpretation, on the other hand, is infeasible, especially when considering regular monitoring approaches to map the physical integrity of the seafloor over time.…”

Predictive model of overburden deformation: based on machine learning and distributed optical fiber sensing technology

“…At present, machine learning has been used to solve geological and geotechnical engineering problems popularly and performed well in a variety of practical cases. For example, the support vector machine is used to build a three-dimensional geological model (Guo et al, 2019), predict penetration rates of the tunnel boring machine (Mahdevari et al, 2014) and form a landslide evaluation model to divide the regional landslide sensitivity (Colkesen et al, 2016;Lin et al, 2017;Hu et al, 2020;Dou et al, 2020); also random forest and logistic regression are applied in the geological mapping (Cracknell and Reading, 2014;Pendleton et al, 2019), landslide sensitivity analysis (Bui et al, 2016;Chen et al, 2018;Ðuri c et al, 2019) and mineral prospect modeling (Xiong and Zuo, 2018). In addition, the Bayesian network is used to evaluate gas explosion in a coal mine (Tong et al, 2018) and the clustering algorithm is used to process geological data (Kapageridis, 2015).…”