Assessment of primary producers composition, abundance and distribution in Mantua Lakes system from airborne hyperspectral data

Spectral reflectance data acquired from remote can be effectively used as ecosystem monitoring tools. In this work, we use airborne hyperspectral data to assess composition, abundance and distribution of primary producer communities in Mantua Lakes (Northern Italy), a shallow eutrophic fluvial lake system characterized by the coexistence of phytoplankton and macrophytes. In this context, our main objective is to explore hyperspectral data capabilities for: 1) distinguishing different macrophyte groups (emergent, submersed, floating-leaved and free floating), 2) identifying phytoplankton functional types (PFTs), and 3) analyzing the complex relationships between primary producers and physico-chemical factors (e.g., water temperature, water velocity, solar irradiance, nutrient availability) in the study site. Two campaigns (2011/09/21, 2014/09/27) with the Airborne Prism Experiment (APEX) sensor were performed. For calibration and validation of spectral products, in situ historical dataset (acquired since 2006) and new spectral data were gathered. APEX and in situ data were combined to develop semi-empirical algorithms and spectral inversion techniques (bio-optical modelling) for mapping water pigments - chlorophyll a (chl-a), phycocianin (PC), and phycoerythrin (PE) - and macrophyte canopy biophysical parameters. The APEX-derived chl-a concentration maps were based on semi-empiric methods which make use of a 'two-bands adaptive' model (r=0.95) and bio-optical modelling inversion for low chl-a concentrations (r=0.96). Further validation was performed on PC pigments obtained from APEX imagery of 2011 (r=0.93) and of a proxy of PFTs product obtained from 2014 data. In 2011, two cyanobacteria species dominated by PC pigments were relevant in the algal composition. The APEX-derived map of PFTs showed a gradient, consistent with field observations, where diatoms progressively decreased from west to east and contemporarily Chrysophyta and cyanobacteria species dominated by PE increased. Semi-empirical models based on spectral proxies of spectral vegetation indices were used to derive macrophyte canopy parameters maps: i.e. fractional cover (fC) (r=0.94), Leaf Area Index (LAI) (r=0.96), and biomass (r=0.92). These products were analyzed to assess cover, abundance, and spatial variability of macrophyte groups and PFTs, and to investigate the relationships with physico- chemical parameters in the Mantua lakes system. This research activity is part of the EU FP7 INFORM (Grant No. 606865, http://www.copernicus-inform.eu/). APEX data gathering was by EUFAR (www.eufar.com).