POSTER SESSION 1
15:30–17:30, Monday, September 14
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
ABSTRACT 1234 | POSTER 233
MULTI-SENSOR OCEAN COLOUR PRODUCTS TO SUPPORT THE COPERNICUS MARINE HIGH-RESOLUTION COASTAL SERVICE
Satellite-derived ocean colour products are essential for coastal ecosystem monitoring and management. The current Earth Observation (EO) constellation combines medium-resolution ocean colour sensors (Sentinel-3/OLCI) with fine-scale land imagers (Sentinel-2/MSI), each offering specific advantages in spatial coverage, revisit frequency, and spectral quality. Sentinel-3/OLCI provides daily coverage but lacks the spatial detail required for fine-scale coastal processes, while Sentinel-2/MSI resolves these features at 10–60 m, though with a 3–5-day revisit that limits timeliness. Cloud cover further reduces data availability from both sensors, resulting in substantial gaps.
We present an approach to generate daily, cloud-free, multi-sensor ocean colour maps by combining Sentinel-2 and Sentinel-3 observations using DINEOF (Data Interpolating Empirical Orthogonal Functions). A key step is harmonising chlorophyll-a (CHL), suspended particulate matter (SPM), and turbidity (TUR) products between Sentinel-2/MSI and Sentinel-3/OLCI. Because differences in spectral bands and viewing geometry hinder direct algorithm transfer, we employ LightGBM machine learning to adapt advanced OLCI retrieval algorithms (e.g. band ratios, switching schemes, neural networks) to MSI for CHL, SPM, and TUR. DINEOF then exploits daily OLCI fields to gap-fill the higher-resolution, 5-daily MSI products, yielding daily 100 m multi-sensor products that preserve both MSI’s spatial detail and OLCI’s temporal continuity.
The method will be demonstrated and validated in multiple regions (North Sea, Black Sea, Mediterranean Sea, Baltic Sea) in view of integration into the operational high-resolution Copernicus Marine Service, with first releases of the multi-sensor products planned for early 2027.
Dimitry Van der Zande, Royal Belgian Institute of Natural Sciences / OD Nature, [email protected], https://orcid.org/0000-0002-2742-0721
Kerstin Stelzer, Brockmann Consult, [email protected]
Carole Lebreton, Brockmann Consult, [email protected]
Martin Böttcher, Brockmann Consult, [email protected]
Aida Alvera-Azcarate, University of Liege, [email protected]
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
