The Yangtze Estuary provides a critical link between the land and ocean, transporting vast amounts of terrestrial particulate organic carbon (POC) to the marginal East China Sea, thereby exerting significant influence on the coastal carbon cycles. Here, using paired in situ remote-sensing reflectance (Rrs) and POC data collected during 2021-2023, we developed a robust POC retrieval algorithm to accurately estimate POC from ocean color remote sensing using the convolutional neural network (CNN) method (with mean absolute percent differences, MAPD, less than 15%). Application of the algorithm to multi-year satellite Sentinel-3 OLCI imagery captured the impact of Yangtze inputs, wind regimes, and climate extremes (both typhoons and extreme floods) on POC fluxes at the estuarine scale. Results have important implications for improved predictions of coastal biogeochemical fluxes in river delta-marginal sea systems.

satellite ocean color, particulate organic carbon, remote sensing retrieval, extreme events