Estimation of chlorophyll-a concentration in waters over the continental shelf of the Bay of Biscay: a comparison of remote sensing algorithms

• Published in: International Journal of Remote Sensing
• Main Author: Novona, Stéfani
• Other Authors: Marine Research Division
• Format: Article in Journal/Newspaper
• Language: English
• Published: HAL CCSD 2011
• Subjects: Physical Sciences; [SPI]Engineering Sciences [physics]; North East Atlantic
• Online Access: https://hal.archives-ouvertes.fr/hal-00744816; https://hal.archives-ouvertes.fr/hal-00744816/document; https://hal.archives-ouvertes.fr/hal-00744816/file/PEER_author_10.1080%252F01431161.2010.540588.pdf; https://doi.org/10.1080/01431161.2010.540588

International audience Ocean colour imagery is used increasingly as a tool to assess water quality via chlorophyll-a concentration estimations in European waters. The Bay of Biscay is affected by major river discharges, which alter the constituents of the marine waters. Chlorophyll-a algorithms, designed for use at global scales, are less accurate due to the variability of optically-active in- water constituents. Hence, regionally-parameterized empirical algorithms are necessary. The main objective of the present study was to develop a regional algorithm to retrieve chlorophyll-a concentration (Chla) in surface water using in-situ Remote Sensing Reflectance (Rrs), for a subsequent application to MERIS satellite images. To address this objective, a platform was developed initially and a measurement procedure adapted for the field HR4000CG Spectrometer. Subsequently, the procedure was tested during a survey over the southeastern Bay of Biscay (North-East Atlantic Ocean), to establish a MERIS Chla algorithm for the area, by comparing different global remote sensing Chla algorithms, with band ratios. Results validated with the jackknife resampling procedure show a satisfactory relationship between the 510/560 nm Rrs ratio (r2jac = 0.681). This ratio is better correlated to Chla than those obtained with established Chla remote sensing algorithms. High content in coloured dissolved organic matter (CDOM>0.4 m-1) and suspended particulate matter (SPM>2.8 mg.l-1) influenced this relationship, with yellow substances having a stronger effect.