Multispectral remote sensing of invasive aquatic plants using RapidEye

Some invasive aquatic plants are thought to benefit from recent climate warming. Their frequent observation is only feasible using satellite data. Earlier studies used airborne hyperspectral sensors like HyMap to discriminate different plant species—this is a very cost intensive method with an environmental impact through numerous flights. A fully automated processing chain to assess the distribution of invasive species, as well as their abundances and biomass is restricted to spaceborne sensors. The research presented here investigates the ability of multi-seasonal multispectral remote sensing data (RapidEye) to differentiate between vegetated and non vegetated areas, as well as species composition at a subpixel level. This is done by combining in situ derived information about the apparent optical properties of the water body and reflectance measurements of Elodea nuttallii and Najas marina. The information about the current state of the water body is used to produce depth-invariant combinations of spectral bands using a simple physically based semiempirical method. In situ reflectance measurements were processed with the same method and subsequently used in a Matched Filtering spectral unmixing approach. The results show a good separation between vegetated and bare litoral areas. A reliable differentiation of plant species still requires further method development.