Cyanobacterial Biomass Estimation in Subtropical and Temperate Reservoirs by Sentinel-2

The study of cyanobacterial biovolume, the relative abundance of phytoplankton in terms of biomass, using remote sensing is not common. Its estimation can be done in two ways, indirectly through the determination of the phycocyanin concentration, since it is their main pigment, or directly by determining their abundance or biovolume. This work aims to improve the expertise for the cyanobacterial biovolume directly estimation with Sentinel-2 imagery. To develop the algorithm empirically 43 georeferenced samples were collected, 20 from the sub-tropical zone and 23 from the temperate zone, which cyanobacterial biovolumes were determined. Images were resampled, atmospherically corrected with the Case2 eXtreme neuronal net and the remote sensing reflectance were extracted. The best results were retrieved with the NDB5B4, obtaining: R2 of 0.92, RMSE of 1.71 mm3 L−1 and NRMSE of 7.6%. To demonstrate its consistency and functionality, the algorithm was applied to images thorough two years, to reproduce both temporal and spatial variations of the cyanobacterial biomass. The highest values were recorded in the subtropical zone. A tail-dam asymmetry was observed in both climatic zones. In the temperate zone, the annual variation registered a minimum and a maximum annual value, while more variations were registered in the subtropical zone.