Spectral discrimination of giant reed (Arundo donax L.): A seasonal study in riparian areas

The giant reed (Arundo donax L) is amongst the one hundred worst invasive alien species of the world, and it is responsible for biodiversity loss and failure of ecosystem functions in riparian habitats. In this work, field spectroradiometry was used to assess the spectral separability of the giant reed from the adjacent vegetation and from the common reed, a native similar species. The study was conducted at different phenological periods and also for the giant reed stands regenerated after mechanical cutting (giant reed_RAC). A hierarchical procedure using Kruskal-Wallis test followed by Classification and Regression Trees (CART) was used to select the minimum number of optimal bands that discriminate the giant reed from the adjacent vegetation. A new approach was used to identify sets of wavelengths - wavezones - that maximize the spectral separability beyond the minimum number of optimal bands. Jeffries Matusita and Bhattacharya distance were used to evaluate the spectral separability using the minimum optimal bands and in three simulated satellite images, namely Landsat, IKONOS and SPOT. Giant reed was spectrally separable from the adjacent vegetation, both at the vegetative and the senescent period, exception made to the common reed at the vegetative period. The red edge region was repeatedly selected, although the visible region was also important to separate the giant reed from the herbaceous vegetation and the mid infrared region to the discrimination from the woody vegetation. The highest separability was obtained for the giant reed_RAC stands, due to its highly homogeneous, dense and dark-green stands. Results are discussed by relating the phenological, morphological and structural features of the giant reed stands and the adjacent vegetation with their optical traits. Weaknesses and strengths of the giant reed spectral discrimination are highlighted and implications of imagery selection for mapping purposes are argued based on present results. (C) 2013 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS) Published by Elsevier B.V. All rights reserved.