The Need of a Phenological Spectral Library of Submersed Macrophytes for Lake Monitoring
Wolf, Patrick; Roessler, Sebastian; Schneider, Thomas; Melzer, Arnulf
TU Munich, GERMANY

When monitor submerse macrophytes in German freshwater lakes by remote sensing to estimate the lakes' ecological states, it is essential to know how distinct the reflectance spectra of the species are. In other words: Do the spectral signatures of submersed macrophytes change during growing season? Or do they vary between years or lakes?
Therefore, systematic spectral measurements upon four submersed macrophytes were carried out during the vegetation periods (May to October) of 2010 and 2011, respectively. At Lake Starnberg (48.0°N, 11.3°E) and Lake Tegernsee (47.7°N, 11.7°E) in the south of the city of Munich (Germany), the spectral signals of the invasive species Elodea nuttallii and Najas marina and the indigenous species Chara aspera and Potamogeton perfoliatus were collected by underwater RAMSES-spectroradiometers. To calculate remote sensing reflectance, downwelling irradiance and upwelling radiance were measured simultaneously twenty times just upon the surfaces of pure stands of the mentioned plants. This process was repeated up to seven times a day, five days a year, depending on cloud coverage and waves on the water surface. Depending on the distance of the populations to the shore, the sensors were launched from a terrestrial platform or a fixed boat. Additionally, photographs and biometrical data were collected for documentation. The latter consisted of plant density, canopy height and biomass and phenological characteristics. The setup of the study enabled the comparison of the reflectance spectra of a particulate plant within one season as well as from one year to the other and as well as between two different lakes.
The spectra of the particulate species offered identical or very similar shapes from one year to the other. This argues for a certain phenology of the specific plants and for the used method to be qualified to spectrally detect these changes and to collect representative spectral signatures. Hence, for each plant, specific results could be formulated. Chara aspera produced a nearly identical spectral shape along the sampling period. For Potamogeton perfoliatus, the decay of the plants was clearly visible. By measurements of Elodea nuttallii at two lakes, the influence of sediment covering the plant canopy could be defined. Similarities and differences of the spectral shapes of the invasive Najas marina could be linked to the specific phenology. Male and female plants of different heights occurred simultaneously next to each other. As Najas marina is a thermophile macrophyte, it might be used as an indicator for increasing water temperatures. Additionally, the reflectance spectra of different degrees of coverage were collected for all macrophytes. The dependency of the spectral response on the plant phenology emphasizes the necessity of multi-temporal in situ data. Furthermore, the integration of the in situ measurements into a coupled reflection-/growth-model allows validation and interpretation of remote sensing data at every recording date. As the in situ spectra were measured with hyperspectral sensors in a range from 320nm to 950nm with a 3.3nm step, the data can be rescaled to the spectral resolution of numerous remote sensing sensors (e.g. RapidEye, Sentinel-2, Landsat 7 and OLI). To sum up, the use of a spectral library, which contains "phenological spectral fingerprints" of submersed macrophytes improves lake monitoring.