On the possibility of recognising different phytoplankton functional groups and size classes in optically complex waters

Abstract

Knowledge on the phytoplankton functional types (PFTs) and size classes (PSCs) is critical for many purposes from early warning of potentially harmful blooms to carbon cycle studies. Feasibility of retrieving this information has been demonstrated in clear natural waters where phytoplankton is the main optically active substance. We studied the possibility of retrieving PFTs and PSCs in optically complex waters, namely in the Baltic Sea, using multiple methods: (1) recognising the group-specific pigments, (2) linking particle backscattering to particle size classes, (3) using look-up-tables and (4) applying empirical algorithms linking biomass to percentage presence of PFTs and PSCs. Most of the phytoplankton pigments characteristic to a certain functional type absorb light at the same wavelengths in the blue part (400-500 nm) of spectrum. These pigments are optically indistinguishable from one another, even when using pure pigments and hyperspectral sensors - let alone when trying to detect them within natural phytoplankton assemblages using sensors such as OLCI on Sentinel-3 or MSI on Sentinel-2. In the Baltic Sea PSCs do not correlate with water backscattering as most of the particulate material were mineral particles and not all organic particulate matter were phytoplankton. The laboratory based empirical algorithms linking percentage of each PFT or PSC to phytoplankton biomass (chlorophyll-a concentration) performed poorly. Mapping of PFTs and PSCs in optically complex waters where water reflectance is defined primarily by the amount of CDOM and SPIM rather than phytoplankton is essentially unfeasible.