Global Seafloor Mapping with Satellites and Ships Assisted by Machine Learning
Abstract
"Marine gravity and bathymetry are foundational data, providing basic infrastructure for military,scientific, economic, educational, and political work. Auxiliary products such as the gravity-to-topography correlation coefficient, are used by the by the Naval Research Laboratory as part of their global seafloor characterization program. Moreover, our predicted bathymetry model is used as the foundation layer for the GEBCO/Seabed 2030 project and we are collaborating with the four GEBCO data centers to expand and improve the coastal and deep ocean bathymetry sounding data -base and to highlight uncharted areas for future surveys. Satellite altimetry provides the best means for deriving these models at global scale and over the last 25 years, we have developed the necessary infrastructure to process all available altimetry data for optimal gravity field recovery and bathymetry prediction. Over the next 3 years, three currently operating satellite altimeters CryoSat-2, SARAL/Altika and Sentinel-3 will provide a wealth of new marine gravity data. In addition, the Surface Water and Ocean Topography (SWOT) mission, due to launch in September 2021, will begin providing next Generation, swath altimetry data that will revolutionize space based gravity field recovery. We will also continue to assemble what is likely the largest set of edited global depth soundings (shallow and deep ocean) from a wide variety of sources. These soundings are currently being used as a training set to develop machine learning algorithms for assisting in the editing of newly acquired data and in the recovery of previously misclassified data that were erroneously discarded. We have recently engaged in collaboration with the Naval Research Laboratory Stennis Space Center (NRLSSC), who are working on highly complementary machine learning applications for improving the accuracy of our seafloor prediction workflow and we plan to solidify and expand this collaboration over our next research cycle. We propose to : Improve the accuracy and spatial resolution of the global marine gravity using new satellite altimeter data collected by SARAL/Altika, CryoSat-2, Sentinel-3 and eventually SWOT. Refine machine-learning algorithms to recover sounding data that were mis-classified as erroneous. Use these improved gravity maps along with our global compilation of soundings to refine a 15-arc second bathymetry model (SRTM15+V3). Better connect the mid- and deep-ocean bathymetry with the higher resolution coastal bathymetry being compiled by other groups (e.g., GEBCO and NOAA). Work with NRL scientists to develop a machine learning algorithm to produce synthetic seafloor with appropriate roughness in areas having no soundings. Prepare the next generation of scientists for ocean research. We will coordinate our research efforts with the Naval Oceanographic Office, the Naval Research Lab, The National Geospatial-Intelligence Agency and make thesegrids available to Navy labs, defense contractors, and the general public."
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 29, 2020
- Source ID
- N000142012350
Entities
People
- David T. Sandwell
Organizations
- Office of Naval Research
- United States Navy
- University of California, San Diego