Data driven applications in coastal geomorphology
Date
2023
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Our ability to digitally capture coastal processes and landforms has progressed immensely in the last several decades. Satellite-based, drone-based, surface vessel-based, and underwater vehicle-based platforms carrying sensors like multispectral cameras, LiDAR, and sonar allow us to image the texture and topography of the subaerial and subaqueous coastal landscape at high resolution and accuracy. Consequently, our improvements in data collection have exceeded our ability to analyze and discern patterns from said datasets. In this dissertation, I will present several applications of using data-driven methods (e.g., convolutional neural networks) to analyze coastal processes and landforms. This includes the detection and characterization of widespread sandy depressions on the Atlantic Coastal Plain (Carolina Bays), the detection and characterization of seabed fluid-escape depressions on the continental shelf (pockmarks), and satellite-based analysis/prediction of shoreline change.
Description
Keywords
Machine learning, Shoreline, Coastal processes, Neural networks, Topography