PACE satellite detection of marine phytoplankton taxonomic groups based on light absorption

作者：Deyong Sun, Zan Li, Shengqiang Wang, Zixu Ye, Hailong Zhang · 发表于：Optics Express · 年份：2026 · DOI：10.1364/oe.608779 · 研究领域：Marine and coastal ecosystems、Water Quality Monitoring and Analysis、Marine Biology and Ecology Research

As primary producers in aquatic ecosystems, phytoplankton play a fundamental role in regulating marine ecosystem structure and biogeochemical cycling. Understanding phytoplankton community composition is therefore essential for oceanographic research. However, the group-specific optical absorption features are often subtle and embedded within the broad pigment absorption envelope, making them difficult to resolve without high spectral resolution or derivative-based analysis. In this study, phytoplankton group composition was first derived using the CHEMTAX algorithm based on diagnostic pigment measurements, yielding six phytoplankton taxonomic groups (PTGs). A matrix decomposition framework was then applied to hyperspectral phytoplankton absorption data to estimate group-specific absorption spectra from total absorption ( a ph ). The retrieved specific absorption spectra were evaluated through reconstruction of the a ph against in situ observations, showing good agreement with coefficients of determination ( R 2 ) exceeding 0.80 across most spectral bands. On this basis, an absorption-based Random Forest model was developed by integrating a ph and group-specific absorption spectra to estimate PTG biomass. Our model was subsequently applied to the current PACE hyperspectral a ph data as a case study to evaluate its applicability under existing satellite-derived products and to demonstrate the potential of absorption-based approaches for future large-scale monitoring of phytopl...