Characterization of submerged macrophytes in a shallow lake using a combined approach of hyperspectral and LiDAR data

Aquatic vegetation is an important component of wetland ecosystems because of the key role in ecological functions of these environments. In the case of shallow lakes, such as the Trasimeno (central Italy), suspended sediment particles, frequently water level variation and algal bloom can lead to the reduction of underwater light penetration with feedback effect on submerged plants that sometimes drastically retreat. Macrophytes are an essential part of the productive zone in the Trasimeno lake and, by an equilibrium with algal growth, they contribute to the uptake of nutrients and pollutants, favour bottom quality, composition and stability and provide structural, food, shelter and breeding habitats for aquatic communities. The presence, abundance and species composition often represent a good limnological indicator of quality status of wetland areas. This study is a contribution on the knowledge of macrophytes of shallow lakes using a combined approach of airborne sensors such as MIVIS (Multi-spectral IR and Visible Imaging Spectrometer) and LiDAR (Light Detection and Ranging). The primary requirements for this ecosystem analysis are spatial and spectral resolution, accurate calibration and a suitable temporal baseline. Spectral Mixture Analysis (SMA) classifies individual mixed pixels according to the distribution of spectrally pure endmember fractions and provides a tool for discrimination and classification of wetlands areas. Physical and biological compositions are the primary determinants of the inherent reflectance properties at a landscape scale. We use SMA to discriminate among different types of reflectance properties on the basis of endmember fraction distributions. An in depth field validation has been also performed during the airborne survey that was on the 12 may 2009 and a spectral library of macrophytes has been collected. Field observations are necessary to calibrate and validate the spectral mixture model and subsequent thematic classification. The purposes of the field observations are twofold. First, field observations allow us to identify the biophysical habitat and its properties associated with the spectral endmembers. Secondly, field reconnaissance and laboratory analysis allows us to identify significant parameters involved in optical interpretation of the water body. As a first results the spatial distribution and variability in macrophytes presence, the healthy status and the pattern of species clusters obtained with the application of mixing model on hyperspectral MIVIS data is achieved. The effectiveness of classification with hyperspectral data that has undergone SMA is then enhanced by the use of LiDAR bathymetry. In this way the morphological characterization has been integrated into a detailed macrophytes map and has provided a great instrument for the monitoring of wetlands ecosystem with high-resolution biophysical parameters. The results compared with historical data coverage, show a decrease of the area occupied by macrophytes. These achievements should be carefully taken into account by local authorities as they demonstrate a significant loss of valuable submersed vegetation which is the probable result of an earlier nutrient enrichment.