Spectral discrimination of marsh vegetation types in the New Jersey Meadowlands, USA

Abstract

This study investigates the fall season similarity of reflectance spectra among salt marsh species to identify and map marsh vegetation types at species level using hyperspectral remote sensing. The medians of the reflectance spectra collected from canopies of dominant marsh vegetation (Phragmites australis, Spartina patens, Spartina alterniflora, and Distichlis spicata) in the New Jersey Meadowlands were compared using a set of statistical metrics. Results show that these marsh species are distinct and separable spectrally in the near infrared (NIR) region of the spectrum. The two Spartina species are similar in spectra and are the most difficult pairs to separate. However, the distribution of red-edge parameter (maximum inflection) suggests that red-edge may be useful for discriminating these two species. The results of this study can be applied to classify marsh vegetation at species level using remote sensing and to map ecotypes along salinity or oxygen gradients as a way to assess coastal wetlands health and condition.