Assessment of plant species diversity based on hyperspectral indices at a fine scale.

Fast and nondestructive approaches of measuring plant species diversity have been a subject of excessive scientific curiosity and disquiet to environmentalists and field ecologists worldwide. In this study, we measured the hyperspectral reflectances and plant species diversity indices at a fine scale (0.8 meter) in central Hunshandak Sandland of Inner Mongolia, China. The first-order derivative value (FD) at each waveband and 37 hyperspectral indices were used to assess plant species diversity. Results demonstrated that the stepwise linear regression of FD can accurately estimate the Simpson (R2 = 0.83), Pielou (R2 = 0.87) and Shannon-Wiener index (R2 = 0.88). Stepwise linear regression of FD (R2 = 0.81, R2 = 0.82) and spectral vegetation indices (R2 = 0.51, R2 = 0.58) significantly predicted the Margalef and Gleason index. It was proposed that the Simpson, Pielou and Shannon-Wiener indices, which are widely used as plant species diversity indicators, can be precisely estimated through hyperspectral indices at a fine scale. This research promotes the development of methods for assessment of plant diversity using hyperspectral data.