Derivation and analysis of the fluorescence algorithms to estimate phytoplankton pigment concentrations in optically complex coastal waters

This paper presents an analysis of several in-water algorithms developed using a large number of bio-optical data collected in Korean and neighbouring ocean waters. The upward and downward light fields measured in these waters were normalized to provide remote sensing reflectance (R-rs) from which the variations in sun-induced chlorophyll-a (Chl-a) fluorescence from phytoplankton biomass were analysed and the fluorescence line height (Delta Flu) was measured using a baseline method. Delta Flu measurements were related to in situ Chl-a concentrations (used as an index for quantifying the algal material) to obtain the Delta Flu(681), Delta Flu(688) and Delta Flu((area)) algorithms, which were compared with those from standard spectral ratios of the remote sensing reflectance, R-rs(444)/R-rs(554) and R-rs(490)/R-rs(554). Comparison revealed the correlation coefficients (r(2)) 0.88-0.90 for the fluorescence algorithms and 0.79-0.85 for the spectral ratios algorithms. A validation analysis using a new data set illustrated the merits of these algorithms in Case II waters. Fluorescence-based algorithms tended to reproduce in situ Chl-a concentrations (from 0.1 to 82 mg m(-3)), whereas standard spectral ratios algorithms gave inconsistent estimates, probably caused by the interference of inorganic sediment and dissolved organic matter in these waters. For measurements of the sun-induced Chl-a fluorescence signal in coastal waters, this study discusses the requirements and provides the optimal channels to be adopted in the design of the new generation ocean colour sensors planned in the near future.