Phytoplankton Pigment Communities Can be Modeled Using Unique Relationships With Spectral Absorption Signatures in a Dynamic Coastal Environment

Abstract

AbstractUnderstanding the roles of phytoplankton community composition in the functioning of marine ecosystems and ocean biogeochemical cycles is important for many ocean science problems of societal relevance. Remote sensing currently offers the only feasible method for continuously assessing phytoplankton community structure on regional to global scales. However, methods are presently hindered by the limited spectral resolution of most satellite sensors and by uncertainties associated with deriving quantitative indices of phytoplankton community structure from phytoplankton pigment concentrations. Here we analyze a data set of concurrent phytoplankton pigment concentrations and phytoplankton absorption coefficient spectra from the Santa Barbara Channel, California, to develop novel optical oceanographic models for retrieving metrics of phytoplankton community composition. Cluster and Empirical Orthogonal Function analyses of phytoplankton pigment concentrations are used to define up to five phytoplankton pigment communities as a representation of phytoplankton functional types. Unique statistical relationships are found between phytoplankton pigment communities and absorption features isolated using spectral derivative analysis and are the basis of predictive models. Model performance is substantially better for phytoplankton pigment community indices compared with determinations of the contributions of individual pigments or taxa to chlorophyll a. These results highlight the application of data‐driven chemotaxonomic approaches for developing and validating bio‐optical algorithms and illustrate the potential and limitations for retrieving phytoplankton community composition from hyperspectral satellite ocean color observations.