Integration of Sentinel-2 and LANDSAT for Cyanobacteria Scum Distribution Mapping in Hypereutrophic Curonian Lagoon

Cyanobacteria blooms have recently been one of the major impact factors compromising lakes and lagoons ecosystems equilibrium. Extremely developed bloom conditions appear as a thick surface cyanobacteria layer (scum) completely covering the water surface. In this study a series of Sentinel-2A (MSI) and LANDSAT-8 (OLI) satellites images acquired in summer periods of most recent years were used to investigate the distribution of cyanobacteria scum in the hypereutrophic Curonian Lagoon (Lithuanian and Russian part). Surface accumulations of algae can be highly complex and variable phenomena: some algae produce thick foams, while others show marked changes in their primary and secondary pigments, and others form aggregations of colloidal inorganic material with complex derived chemical constituents. Presence/absence of cyanobacteria scum was estimated by semi-empirical band ratio algorithm applied to images after atmospheric correction with the 6SV. Simply thresholds algorithms based on band ratio use the bands across the red and near-infrared spectral regions. Radiometric measurements acquired with WISP-3, resampled according to spectral characteristics of MSI bands, compared with atmospheric corrected Rrs products acquired synchronously, show a good agreement. Considering the ratio between two bands (705 nm and 665 nm) the in situ average value was 1.55 (±0.16) respect to 1.61 (±0.08) from MSI. In addition, in order to compare OLI and MSI products, we compared band ratio calculated between band 4 and 3 on one MSI and one OLI synchronous images. Again good agreement was shown by the analysis (R2=0.93). We discovered MSI data to be a valuable tool that can be integrated together with EO data of other missions for operational monitoring of cyanobacteria blooms. The analysis indicated pronounced spatial variability of cyanobacteria scum mainly occurring in early July-late October with the max area of 40,000 ha. Different accumulation features is mainly influenced by ambient weather conditions and hydrodynamics. Respect to the ecological role of scum event, its formation, temporal and spatial evolution is of great interest and therefore the retrospective analysis covering more than 30 years, i.e. nearly entire series of Landsat missions and integrating Sentinel-2 data, will be performed in the near future.