Advancing understanding of Southern Ocean phytoplankton phenology with 4D data product

The Southern Ocean plays a major role in the global carbon cycle, with air-sea carbon dioxide (CO₂) exchanges influenced by phytoplankton phenology. However, during summer, global ocean biogeochemistry models struggle to capture the interplay between biological and physical processes and their combined effects on air-sea CO₂ fluxes. Improved constraints on phytoplankton seasonal dynamics and the mechanistic drivers of their spatial and temporal variability are crucial for refining these models. In this study, we used a 24-year (1998–2022) database of weekly vertical profiles of chlorophyll-a and particulate organic carbon concentrations. This observation-based 4D dataset was generated by matching up and merging satellite and hydrological data using a machine learning methodology trained on BioGeoChemical-Argo observations. A Functional Principal Component Analysis, coupled with a Gaussian Mixture Model, was applied to this dataset, enabling a revised regionalisation of the Southern Ocean based on the 3D seasonal variability of phytoplankton biomass and particulate organic carbon concentrations. Distinct latitudinal differences mostly align with physical oceanic fronts, while zonal differences emerge within the Antarctic Circumpolar Current region. These spatial patterns reflect regional disparities in key phytoplankton growth drivers, including light and nutrient availability, vertical mixing, and stratification. We further explore the interannual variability of this bioregionalisation, shedding light on the environmental drivers of phytoplankton seasonal dynamics and their potential biogeochemical impacts in the context of a changing climate.