Dynamics of phytoplankton communities in the Baltic Sea: insights from a multidimensional analysis of pigment and spectral data: part II, spectral dataset

The use of hyperspectral satellite missions opens new opportunities for integrated approaches to the study of phytoplankton communities. The Baltic Sea, with its distinct mixture of marine and freshwater characteristics, is a natural laboratory for understanding marine ecosystems. In this study, we analyzed a dataset from the Baltic Sea containing simultaneous phytoplankton pigment concentrations and absorption spectra. We applied spectral derivative analysis and unsupervised machine learning techniques to identify the unique statistical relationships among phytoplankton pigments and inherent optical properties. The statistical analysis of the absorption spectra provides the basis for a predictive model to assess pigment concentrations from optical measurements. Additionally, we compare our results to know assessment methods, such as Gaussian spectral decomposition, that link the spectral analysis with phytoplankton pigment content. This study investigates the potential of statistical, data-driven analytical approaches in the development and validation of models for retrieving phytoplankton community composition. The integration of these findings with existing research contributes to the advancement of remote sensing capabilities for monitoring marine ecosystems in the Baltic Sea.