【报告摘要】

Subsurface ocean observations remain sparse despite their importance for understanding ocean dynamics. This seminar presents an AI-based framework that reconstructs three-dimensional (3D) ocean structures from satellite

and surface observations, enabling new insights into physical processes across the East Sea (Sea of Japan). First, reconstructed 3D temperature and salinity fields were used to investigate the life cycle of the Ulleung Warm

Eddy. A new Wobbling Ratio (WR) was developed to quantitatively identify eddy evolution from growth to decay, providing a practical alternative to conventional energy-budget analyses. Second, the reconstructed subsurface

temperature fields were applied to examine the three-dimensional morphology and long-term variability of the East Sea Subpolar Front. The results reveal persistent frontal tilting and significant long-term changes linked to

variability in the East Korea Warm Current. Third, reconstructed temperature and salinity fields were used to develop a novel Z-score-based upwelling index (ZUI) that incorporates subsurface ocean conditions. Compared with

conventional sea surface temperature-based indices, the proposed index substantially improves the detection of coastal upwelling events. Together, these studies demonstrate that AI can move beyond prediction to reconstruct

hidden ocean structures and derive physically meaningful oceanographic indices. The integration of satellite observations, in-situ measurements, and machine learning provides a unified framework for investigating

mesoscale eddies, basin-scale fronts, and coastal upwelling, offering new opportunities for next-generation ocean monitoring and process-based oceanographic research.

【专家简介】

Young-Heon Jo is a Professor in the Department of Oceanography at Pusan National University, Republic of Korea. He received his M.S. in Meteorology from Florida State University and his Ph.D. in Oceanography from the

University of Delaware, where he specialized in satellite remote sensing and physical oceanography.His research focuses on satellite oceanography, artificial intelligence (AI), and ocean observation technologies to

better understand marine physical processes. His work spans satellite remote sensing, three-dimensional reconstruction of subsurface ocean structure, mesoscale eddies, ocean fronts, coastal upwelling, marine heatwaves,

and ocean forecasting. In recent years, he has developed AI-based methods to reconstruct subsurface temperature and salinity fields from satellite observations, enabling new insights into ocean dynamics that are difficult to

observe directly.

Professor Jo is also developing next-generation ocean observing systems that integrate satellite observations with autonomous platforms, including drones, Helikite-based aerial imaging systems, and intelligent marine sensor

networks for monitoring coastal and offshore environments. His research aims to combine AI, remote sensing, and innovative observation technologies to improve marine environmental monitoring, ecosystem assessment, and ocean

prediction. He has published extensively in the fields of physical oceanography, satellite remote sensing, and marine AI, and actively collaborates with researchers in Asia, Europe, and the United States. His current research seeks to

advance integrated ocean observing systems that bridge surface observations and subsurface ocean processes for next-generation marine science.