09:45–10:00 | ABSTRACT 1124
RESOLVING COASTAL CO2 DYNAMICS IN THE NORTH SEA AT UNPRECEDENTED RESOLUTION COMBINING OCEAN COLOUR SATELLITE AND MACHINE LEARNING APPROACHES
The coastal ocean is a key player in the global carbon cycle, accounting for ~20% of the ocean’s CO₂ uptake. However, coastal CO₂ budgets remain poorly constrained due to the complex interplay of biological and physicochemical processes that drive strong spatiotemporal variability in seawater partial pressure of CO₂ (pCO₂), the primary control on air–sea CO₂ fluxes (FCO₂). Here, we investigate FCO₂ dynamics in the North Sea, a temperate shelf sea, at unprecedented spatial (1 km) and temporal (daily) resolution. By combining the dense coverage of in situ pCO₂ observations in the North Sea from the Surface Ocean CO₂ Atlas (SOCAT) with satellite-derived ocean colour reflectance from the ESA Ocean Colour Climate Change Initiative (OC-CCI), we resolve fine-scale pCO₂ dynamics over 2014-2023. Using a machine learning framework, we reconstruct pCO₂ across time and space from 96,931 observations using 11 satellite-derived predictors, including our regionally-optimized satellite retrievals of chlorophyll-a, suspended particulate matter, and particulate organic carbon, with a validation R² of 0.9. From ocean colour reflectance, we identify four distinct provinces, characterized by differing influences of primary productivity, temperature, riverine inputs, and sediment resuspension on pCO₂ dynamics. We show that episodic events such as storms increase particulate organic carbon concentrations by 30%, leading to an 11% rise in pCO₂ through coupled changes in biological activity, vertical mixing, and carbonate chemistry. This study provides new insights into the processes governing coastal carbon dynamics and presents a scalable framework for integrating satellite and in situ observations to constrain CO₂ fluxes in coastal systems worldwide.
Andrea van Langen Rosón*, Ghent University, [email protected], https://orcid.org/0009-0006-2636-1391
Clémence Goyens, Ghent University, [email protected]
Alizée Roobaert, Flanders Marine Institute, [email protected]
Peter Landschützer, Flanders Marine Institute, peter.landschü[email protected]
Griet Neukermans, Ghent University, [email protected]
