Monitoring of primary producers composition, cover and abundance in Mantua Lakes system from Sentinel-2 and landsat-8 data

Earth Observation (EO) data are a useful tool for monitoring water quality and primary producers in freshwater ecosystems. We used Sentinel-2A (S-2) images to assess phytoplankton and macrophytes composition, abundance and distribution in a shallow eutrophic fluvial lake system (Mantua Lakes, Italy). Field measurements acquired from 2014 to 2016 (14 dates, 15 sites) were used for calibration and validation of EO products. 21 S-2 images (5 synchronous to in situ data) were corrected for atmospheric effects with 6SV code. For phytoplankton biomass (chlorophyll-a) we applied a combined semi-empirical algorithms and spectral inversion techniques (bio-optical modelling) to S-2 corrected images (MAE=4.21; R2=0.92). Phytoplankton functional types dominance were detected from two S-2 images (in situ counts were available) with bio-optical modelling inversion by using specific phytoplankton absorption and back-scattering coefficients. Four macrophyte community types (helophy te, emergent, floating and submerged-floating association) were mapped using a rule-based hierarchical classifier fed with multi-temporal seasonal WAVI index (OA Kappa 87.9%; 0.85). Moreover, fractional cover and leaf area index were estimated using semi-empirical algorithms (WAVI and MCARI respectively) with a validation of MAPE<20% for both parameters. Chl-a maps showed an increase of concentrations from upstream to downstream in summer season, in particular in stagnant water due to cyanobacteria growth. Emergent and floating macrophytes were dominant in the lakes. Lower values of biophysical parameters were found in 2016 compared to 2015. All the products obtained were shared with local water management authority to improve the plan for monitoring and safety. This research is part of the EU FP7 INFORM (Grant No. 606865).