Using Acoustic Tomography to Improve Ocean Predictions in Undersampled Regions

Abstract

"The Philippine Sea is a region characterized by strong mean flows, baroclinic instability,high eddy kinetic energy from both sub- and mesoscale eddies, and significant energy fluxfrom internal tides. Furthermore, it borders an important geopolitical region with Taiwan,Japan, the Philippines and China all disputing territorial claims.This project will utilize an advanced machine-learning technique (4D-Var) to estimate thestate of the ocean using acoustic travel-times and other physical observations to quantifythe impact of each observation on weekly predictions of the Philippine Sea/Luzon Straitocean estimates. It will combine the existing Philippine Sea Acoustic Experiment data(from 2010) with a validated and published numerical model and implement a techniqueto directly assimilate acoustic travel-times with given ray paths. It has been shown by thePI that the observations can be directly quantified to understand the informational contentof the ocean that was sampled. This can be extended to determine how each observationimpacts the prediction from the analysis.There are several research goals of this project including: implementation of method todirectly assimilate acoustic travel-times in 4D-Var; quantifying the role of each observationtype (sea surface height, sea surface temperature, moorings, gliders, acoustic travel-times,Argo, etc.) in improving our estimate of the western Philippine Sea; quantifying the roleof each observation in improving the predictability of the western Philippine Sea; and, toquantify how observations sample various dynamical processes present in the region.This proposal brings together existing models and data with the intent to perform basicresearch on the effect of constraining a model with acoustic travel-times. A new graduatestudent will be recruited and trained in ocean acoustics, numerical modeling, machine learning,and ocean state estimation. The graduate student will be able to conduct the bulk ofthe research thanks to the existing resources and methodologies."

Document Details

Document Type
DoD Grant Award
Publication Date
Mar 15, 2021
Source ID
N000142112211

Entities

People

  • Brian A Powell

Organizations

  • Office of Naval Research
  • United States Navy
  • University of HawaiĘ»i System

Tags

Fields of Study

  • Environmental science

Readers

  • Asian Economic Studies
  • Data Mining and Knowledge Discovery.
  • Ocean-Atmosphere Mesoscale Modeling, Data Assimilation, and Flux Boundary Layers

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • Autonomy