Spectral discrimination of macrophyte species during different seasons in a tropical wetland using in-situ hyperspectral remote sensing

Wetlands, one of the most productive ecosystems on Earth, perform myriad ecological functions and provide a host of ecological services. Despite their ecological and economic values, wetlands have experienced significant degradation during the last century and the trend continues. Hyperspectral sensors provide opportunities to map and monitor macrophyte species within wetlands for their management and conservation. In this study, an attempt has been made to evaluate the potential of narrowband spectroradiometer data in discriminating wetland macrophytes during different seasons. main objectives of the research were (1) to determine whether macrophyte species could be discriminated based on in-situ hyperspectral reflectance collected over different seasons and at each measured waveband (400-950nm), (2) to compare the effectiveness of spectral reflectance and spectral indices in discriminating macrophyte species, and (3) to identify spectral wavelengths that are most sensitive in discriminating macrophyte species. Spectral characteristics of dominant wetland macrophyte species were collected seasonally using SVC GER 1500 portable spectroradiometer over the 400 to 1050nm spectral range at 1.5nm interval, at the Bhinda was wetland in the state of Haryana, India. Hyperspectral observations were pre-processed and subjected to statistical analysis, which involved a two-step approach including feature selection (ANOVA and KW test) and feature extraction (LDA and PCA). Statistical analysis revealed that the most influential wavelengths for discrimination were distributed along the spectral profile from visible to the near-infrared regions. The results suggest that hyperspectral data can be used discriminate wetland macrophyte species working as an effective tool for advanced mapping and monitoring of wetlands.