Oral Session 2

Evaluation of Sentinel-3A and 3B OLCI reflectance products over dynamic coastal waters of Salish Sea, Canada

This research encompasses 61 and 39 days of coincident occurrences of Sentinel-3A and 3B – OLCI cloud-free images and in situ measured reflectance data, covering various seasonal dynamics, in 2016, 2018, 2019, and 2020 in the Salish Sea, west coast of Canada. We used Sentinel-3 Level-2 remote sensing reflectance [Rrs(λ)] processed with POLYMER (v4.9), 3×3-pixel grid box with a time difference of ±3h between satellite and in situ observations. In situ hyperspectral reflectance data [Rrs(λ)in situ] were acquired using autonomous shipborne radiometric sensors (SAS-Solar Tracker) installed on British Columbia Ferries. Along transact Rrs(λ)in situ, with a spatial resolution of 13m, were spatially binned to a 900m resolution, equivalent to the size of a 3×3-pixel grid box of S3A-OLCI, ship superstructure, wind speed-dependent sky radiance, bi-directional reflection distribution function (BRDF), and spectral quality assurance and spatial variability corresponding to inhomogeneous transitional water masses, resulting in 80% of high-quality data (N = 1791). The matchup analysis yielded correlation statistics for S3A and S3B as r = 0.90, 0.81; 0.94, 0.90, and 0.89, 0.85 at the blue, green, and red spectral bands, respectively. Prior to the spatial quality assurance, the matchup results showed a decline in statistics for S3A and S3B OLCI as r = 0.75, 0.41, 0.78, 0.70, and 0.70, 0.64, respectively. Our results suggest that high-quality Rrs(λ)in situ data is essential for evaluating satellite Rrs(λ) in coastal dynamic waters and that POLYMER-derived S3A and S3B Rrs(λ) have low BIAS for all visible wavelengths.

*Maycira Costa, University of Victoria, Canada

Alba Martín Lorenzo, Universidad de Las Palmas de Gran Canaria, Spain

José Fco. López Feliciano, Universidad de Las Palmas de Gran Canaria, Spain

Code of Conduct