Summer Research Project Grant Report
Cross-disciplinary Domains: Physical Oceanography, Data Science
Research Project Objective
This study aims to develop a Python-based machine learning framework that leverages existing publicly available global ocean datasets to enhance hydrographic observations in the central Pacific Ocean. Specifically, the primary objective is to reconstruct dissolved oxygen alongside newly collected profiles from the Nova Canton Trough. The second objective, forming part of ongoing and future work, is to extend the depth range of Essential Ocean Variables in the region beyond the historical sampling limit of 6000 m. The results shown here focus on the first objective.
Background
Essential Ocean Variables (EOVs) represent the minimum set of physical, biogeochemical, and biological parameters required to understand ocean variability and inform policy, as defined by the Global Ocean Observing System (GOOS). These include, but are not limited to temperature, salinity, pressure, oxygen, and nutrients at all ocean depths (ref). A key contributor to the collection of these variables is the Global Ocean Ship-Based Hydrographic Investigations Programme (GO-SHIP), which has coordinated repeat hydrographic surveys globally since the late 1980s. Despite these global efforts, observational coverage remains uneven, with a strong bias toward regions such as the North Atlantic. In contrast, the central Pacific remains comparatively under-sampled and less well understood.
The Nova Canton Trough (NCT), located near the equator north of Samoa, is one such region where depths exceed 6000 m and the environment is particularly remote. Reaching depths exceeding 8000 m and situated along the pathway of Antarctic Bottom Water, the NCT is a location with a potentially important role in the global overturning circulation (Kawabe et al., 2010). However, the NCT has not previously been studied or characterised from a holistic oceanographic perspective; existing work is limited, largely geological in focus, and predominantly over two decades old (Rosendahl et al., 1975; Larson, 1997). In addition, it is unclear whether the proportion of AABW is consistent throughout the trough or varies along the circulation pathway. This knowledge gap provided one of the key motivations for the 2024 research expedition to the Nova Canton Trough.
