POSTER SESSION 2
10:30–12:30, Tuesday, September 15
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
ABSTRACT 1320 | POSTER 202
UNLOCKING BIO-OPTICAL RELATIONSHIPS USING ONEARGO: A MACHINE LEARNING FRAMEWORK
Bio-optical relationships allow the transformation of high spatiotemporal resolution bio-optical measurements—such as chlorophyll-a fluorescence (fluo) and particulate backscattering (bbp)—into key biogeochemical variables like chlorophyll-a concentration (Chla) and particulate organic carbon (POC). However, these relationships exhibit strong variability across vertical, regional, and seasonal scales, driven by changing biogeochemical regimes. The OneArgo program, with its global fleet of BGC-Argo profiling floats measuring bio-optical properties, provides an unprecedented opportunity to refine these relationships at a global scale.
Here, we present a machine learning-based framework to re-evaluate and refine bio-optical conversions. By leveraging recently developed three-dimensional BGC-Argo-based observation products as emulators of float bio-optical measurements, our approach enables systematic comparisons with ship-based reference datasets, despite their spatial and temporal limitations. This methodology allows us to address persistent uncertainties, improve algorithms for deriving Chla from fluo and POC from bbp, and quantify associated biases.
Our results demonstrate how this framework unlocks new avenues for understanding bio-optical variability, enhancing the reliability of biogeochemical variables derived from floats. Ultimately, this method paves the way for more accurate and robust global-scale estimates of ocean productivity and carbon cycling from OneArgo, while opening opportunities to explore previously unresolved bio-optical dynamics.
Raphaëlle Sauzede, CNRS – Sorbonne Université – Institut de la Mer de Villefranche, [email protected], https://orcid.org/0000-0001-9992-5334
Julia Uitz, CNRS – Sorbonne Université – Laboratoire d’Océanographie de Villefranche, [email protected], https://orcid.org/0000-0001-5461-0216
Catherine Schmechtig, CNRS – Sorbonne Université – OSU Ecce Terra, [email protected], https://orcid.org/0000-0002-1230-164X
Hervé Claustre, CNRS – Sorbonne Université – Laboratoire d’Océanographie de Villefranche, [email protected], https://orcid.org/0000-0001-6243-0258
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
