A novel GSM and fluorescence coupled full-spectral chlorophyll a algorithm for waters with high CDM content

• Published in: Remote sensing of environment Vol. 321; p. 114667
• Main Authors: Li, Juan, Matsuoka, Atsushi, Devred, Emmanuel, Hooker, Stanford B., Pang, Xiaoping, Babin, Marcel
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.05.2025
• Subjects: algorithms; Arctic Ocean; Arctic region; chlorophyll; Chlorophyll [formula omitted] concentration; color; Colored detrital material; data collection; domain; environment; eutrophication; Fluorescence; particulates; phytoplankton; remote sensing; river deltas; Arctic Ocean; Arctic region; Arctic Ocean; Fluorescence; Chlorophyll a concentration; Colored detrital material
• ISSN: 0034-4257
• DOI: 10.1016/j.rse.2025.114667

Standard ocean colour algorithms exploiting only shorter visible wavelengths (less than 560 nm) perform poorly in the Arctic Ocean (AO) due to the interference from colored detrital material (CDM). The incorporation of longer wavelengths, which are less susceptible to interference from CDM, could prove beneficial in retrieving water properties, particularly in Arctic waters with high CDM content. Similarly, algorithms that exploit only the red region of the spectrum, such as fluorescence-based approaches, are also unsuitable for these waters. This is due to the difficulty in accurately describing the background elastic scattering signal. In this study, we propose an algorithm that accounts for elastic scattering and fluorescence of phytoplankton in the full visible spectral domain by coupling a tuned version of the Garver-Siegel-Maritorena (GSM) algorithm (GSMA) for the AO with an optimized fluorescence emission model. Our novel algorithm, FGSM, demonstrate comparable overall performance to an empirical algorithm derived for chlorophyll a concentration (Chl) estimates in the AO (AO.emp), with a mean absolute difference (MAD) of 1.83. In addition, FGSM outperforms both the GSMA and the fluorescence line height (FLH) algorithms, with an improvement in the MAD of Chl estimates up to 41 %. Assessments conducted using both in situ datasets and satellite data at the Lena River Delta, a region characterized by high productivity and the presence of coastal CDM, revealed that for eutrophic waters where Chl is generally high, FGSM significantly mitigate the underestimation of Chl by AO.emp and GSMA, and exhibit enhanced robustness to produce more retrievals than the other semi-analytical algorithms. FGSM also demonstrates superior performance compared to the other algorithms assessed in this study for waters with high suspended particulate matter (SPM). Further validations for Arctic waters, particularly turbid coastal waters, are still expected in the future. •A novel full-spectral chlorophyll a algorithm, FGSM, was proposed.•Phytoplankton fluorescence property was considered in the FGSM.•FGSM outperforms other algorithms evaluated for waters with high CDM or SPM.•FGSM was validated very robust for turbid waters.