La télédétection passive de la chlorophylle dans l'Estuaire maritime du Saint-Laurent optimisation spectrale des algorithmes et application au SeaWiFs
This work is one more step toward passive remote sensing as a way to estimate the water chlorophyll content in the St-Lawrence Estuary. In July 1993, during the LIDAR-93 oceanographic cruise, optical and biochemical data were collected to identify the best remote sensing algorithms for chlorophyll in this area. Many algorithms were tested on radiometric data acquired with two portable spectroradiometers aboard ship and a CASI imaging spectoradiometer during low altitude flights. A good consistency was obtained between both data types. An exhaustive search on the visible spectrum has allow to identify some simple and multiple band ratios as best algorithms. Among these simple band ratios, the best results were obtained with 540 and 580 nm bands. Some blue-red ratios (420 to 520 nm/660 to 700 nm) also gave good results. Estimates precision were about 40 to 50%. The best multiple band ratios allow estimations of water chlorophyll content with an accuracy of «30%. The estimates are better in the central and western region of the Estuary. However, all algorithms overestimate chlorophyll content in the eastern part of the Estuary. It seems that these overestimates are due to the vertical distribution of Chl in the water column. With simulated SeaWiFs bands, the estimates accuracy is about 50% for simple band ratios and 40% for multiple band ratios. Poorer accuracy obtained with SeaWiFs bands seems due most to the lack of one more band in the red part of the spectrum than to the spectral resolution. Simulations with CAM5S software, on the impact of errors on atmospheric aerosol optical thickness estimations, show that 490/670 nm would be the best band ratio to use with SeaWiFs sensor for the St-Lawrence Lower Estuary. An accuracy of «0,02 on aerosol optical depth would be enough to have a global accuracy of about 60% on water chlorophyll content estimates in the Lower St-Lawrence Estuary.