Statistical approach for the retrieval of phytoplankton community structures from in situ fluorescence measurements.

摘要: Knowledge of phytoplankton community structures is important to the understanding various marine biogeochemical processes and ecosystem. Fluorescence excitation spectra (F(λ)) provide great potential for studying communities because their spectral variability depends on changes in pigment compositions related distinct groups. Commercial spectrofluorometers have been developed analyze by measuring field F(λ), but estimations using default methods are not always accurate strong dependence norm spectra, which obtained culturing pure algae a given group assumed be constant. In this study, we proposed novel approach estimating chlorophyll (Chl a) fractions brown algae, cyanobacteria, green cryptophytes based data set collected East China Sea (ECS) Tsushima Strait (TS), with concurrent measurements vivo F(λ) derived from pigments analysis. The new blends statistical features computing band ratios continuum-removed without requiring priori knowledge spectra. model evaluations indicate that our yields good Chl fractions, root-mean-square errors 0.117, 0.078, 0.072 0.060 cryptophytes, respectively. analysis shows models generally robust uncertainty F(λ). We recommend site-specific more estimations. To develop ECS TS, approximately 26 samples sufficient approach, conclusion needs validated additional regions. Overall, provides useful technical basis