Synergy between Observation Systems Oceanic in Turbulent Regions

TL;DR

This paper explores data-driven models, including neural networks, to better understand ocean dynamics in turbulent regions, addressing limitations of current observation systems in capturing three-dimensional oceanic data.

Contribution

It introduces novel deep regression approaches for modeling complex ocean behaviors in Gulf Stream and Kuroshio regions, enhancing understanding of internal ocean structures.

Findings

Deep regression neural networks effectively model ocean salinity and temperature.

Results demonstrate improved spatial and temporal understanding of turbulent ocean regions.

Open-source code facilitates further research and validation.

Abstract

Ocean dynamics constitute a source of incertitude in determining the ocean's role in complex climatic phenomena. Current observation systems have limitations in achieving sufficiently statistical precision for three-dimensional oceanic data. It is crucial knowledge to describe the behavior of internal ocean structures. We present the data-driven approaches which explore latent class regressions and deep regression neural networks in modeling ocean dynamics in the extensions of Gulf Stream and Kuroshio currents. The obtained results show a promising data-driven direction for understanding the ocean's characteristics, including salinity and temperature, in both spatial and temporal dimensions in the turbulent regions. Our source codes are publicly available at https://github.com/v18nguye/gulfstream-lrm and at https://github.com/sagudelor/Kuroshio.