POSTER SESSION 2
10:30–12:30, Tuesday, September 15
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
ABSTRACT 1182 | POSTER 138
A CLASSIFICATION-BASED ALGORITHM FOR GLOBAL RETRIEVAL AND VARIABILITY ANALYSIS OF PARTICULATE ORGANIC CARBON OVER OPEN AND COASTAL WATERS FOR THE LAST 28 YEARS
This study performed in the frame of the OCROC project (Copernicus2 1st Service Evolution Call for Tenders) investigates the global distribution and temporal variability of particulate organic carbon (POC) derived from satellite ocean color observations in relation to biological and physical drivers. A new classification-based approach combines different algorithms according to the optical water classes assigned to each pixel. Match-up exercises performed on 1160 data points collected in both open and coastal waters yield a median absolute percentage difference (MAPD) of 31% and a median bias (MB) of 0.71 mg m–3. The class-based approach also enables the attribution of pixel-level uncertainty. Consistent retrieval performance is observed over a 20-year match-up time series at seven coastal sites, demonstrating the algorithm’s ability to capture POC temporal variability in optically complex environments (MAPD = 36%; MB = 4.4 mg m–3). Global POC time series analyses over the past 28 years reveal a dominant seasonal contribution, with strong regional contrasts in amplitude, reflecting differences in ecosystem dynamics and environmental forcing. Residual variability is more modest but becomes significant in tropical regions and some coastal zones, where transient processes (e.g., mesoscale activity, episodic blooms) play an important role. Long-term variability is also significant and appears linked to large-scale climate modes. Trend analysis between 60°S and 60°N highlights spatially heterogeneous patterns, with 20.6% of regions showing increasing trends (median: 0.49 ± 0.33% yr–1) and 28.5% decreasing trends of similar magnitude. Comparisons with chlorophyll spatio-temporal patterns provide insights into potential shifts in marine ecosystems.
Hubert Loisel, LOG (France), [email protected]
Daniel S. F. Jorge, LOG, [email protected]
Marine Bretagnon, ACRI-ST, [email protected]
Lucile Duforêt-Gaurier, LOG, [email protected]
Aurélien Prat, ACRI-ST, [email protected]
Vincent Vantrepotte, LOG, [email protected]
Manh Tran Duy, LOG, [email protected]
Antoine Mangin, ACRI-ST, [email protected]
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
