Assessment of Macrophytes Seasonal Dynamics Using Dense Time Series of Satellite Data

Macrophytes communities are a key component of shallow inland water ecosystems, and physical features and bio-geochemical processes regulating their coexistence with other primary producers strongly influence the carbon cycle and budget in such environments. The study of macrophytes requires a multidisciplinary approach from ecology to botany and environmental sciences. In this context, the new generation of Earth Observation (EO) platforms, and in particular Sentinel-2, are crucial step forward towards operational monitoring. Through the use of Spot-5 Take-5 data acquired from April to September 2015 over a wetland ecosystem, we evaluated the potential of high spatial (10m) and temporal (5 days revisit) resolution for assessing macrophytes coverage and inter-annual dynamics in sight of Sentinel-2 data available from late 2015. The study area is Mantua lakes system, located in Northern Italy, a shallow hypertrophic ecosystems which hosts different macrophytes groups (emergent, floating-leaved, free-floating species) and is a test site for the EU FP7 INFORM project. We set up a multitemporal experiment based on Spot-5 Take-5 data for studying the macrophytes status and dynamics during the growing season. We carried out in situ sampling contemporary to satellite acquisition over five dates in (12 May, 11 June, 16 July, 31 July, and 09 September) along the 2015 aquatic plant growing season, covering three emergent-floating-leaved species (Nelumbo nucifera, Nuphar lutea, and Trapa natans). We acquired data of plant biomass, plant/leaf density, water physico-chemical parameters, leaf pigments (chlorophylls and carotenoids), as well as canopy and leaf spectroradiometric response, adding up to a total of 15 sampling stations for N. nucifera and T. natans, and 14 sampling stations for N. lutea (3 replicates per each station). Spot-5 Take-5 surface corrected reflectance (L2A) data processed by THEIA land data center were used for deriving dense time series of specific spectral vegetation indices (e.g. SAVI, GNDVI, RGRI, TVI), which were integrated with in situ data through semi-empirical regression models for producing: maps of macrophytes groups, their spatial and temporal patterns, phenology and bio-physical parameters (biomass and leaf pigments). Monitored macrophytes exhibited marked spatial and temporal patterns of variability evident from spectral vegetation indices time series analysis: i) N. nucifera and other invasive species (e.g. Ludwigia hexapetala) biomass and fractional cover grow very fast in 5-6 weeks from the beginning of May to half of June, when they reach a plateau and peak of season conditions coinciding with the flowering and reproductive stage; ii) N. lutea shows a gradual development in density and biomass along three months from middle of April to middle of July, reaching biomass values far lower than N. nucifera; iii) T. natans is the species that shows the stronger heterogeneity in the area, with some stands emerging (middle of May) and reaching maturity (August) earlier in the season, and other characterized by late blooming and biomass increment (starting from the second half of June). Further studies are on-going on the capabilities of Spot-5 Take-5 data in providing effective information not only of bio-physical parameters and their evolution thought the growing season, but also on monitoring the macrophyte physiological status and its temporal dynamics.