An Empirical Algorithm for Estimating the Absorption of Colored Dissolved Organic Matter from Sentinel-2 (MSI) and Landsat-8 (OLI) Observations of Coastal Waters
Vu Son Nguyen,Hubert Loisel,Vincent Vantrepotte,Xavier Mériaux,Dinh Lan Tran
DOI: https://doi.org/10.3390/rs16214061
IF: 5
2024-11-01
Remote Sensing
Abstract:Sentinel-2/MSI and Landsat-8/OLI sensors enable the mapping of ocean color-related bio-optical parameters of surface coastal and inland waters. While many algorithms have been developed to estimate the Chlorophyll-a concentration, Chl-a, and the suspended particulate matter, SPM, from OLI and MSI data, the absorption by colored dissolved organic matter, acdom, a key parameter to monitor the concentration of dissolved organic matter, has received less attention. Herein we present an inverse model (hereafter referred to as AquaCDOM) for estimating acdom at the wavelength 412 nm (acdom (412)), within the surface layer of coastal waters, from measurements of ocean remote sensing reflectance, Rrs (λ), for these two high spatial resolution (around 20 m) sensors. Combined with a water class-based approach, several empirical algorithms were tested on a mixed dataset of synthetic and in situ data collected from global coastal waters. The selection of the final algorithms was performed with an independent validation dataset, using in situ, synthetic, and satellite Rrs (λ) measurements, but also by testing their respective sensitivity to typical noise introduced by atmospheric correction algorithms. It was found that the proposed algorithms could estimate acdom (412) with a median absolute percentage difference of ~30% and a median bias of 0.002 m−1 from the in situ and synthetic datasets. While similar performances have been shown with two other algorithms based on different methodological developments, we have shown that AquaCDOM is much less sensitive to atmospheric correction uncertainties, mainly due to the use of band ratios in its formulation. After the application of the top-of-atmosphere gains and of the same atmospheric correction algorithm, excellent agreement has been found between the OLI- and MSI-derived acdom (412) values for various coastal areas, enabling the application of these algorithms for time series analysis. An example application of our algorithms for the time series analysis of acdom (412) is provided for a coastal transect in the south of Vietnam.
environmental sciences,imaging science & photographic technology,remote sensing,geosciences, multidisciplinary