Assessing nutritional status of Festuca arundinacea by monitoring photosynthetic pigments from hyperspectral data

Abstract

The dynamics of photosynthetic pigment concentrations and composition provide important information about plant nutritional status. For this reason, hyperspectral remote sensing techniques for quantifying plant pigments on a large scale have received much attention. The spectral absorbance properties of photosynthetic pigments are manifested in the reflectance spectra of the canopy level. This offers the opportunity of using measurements of reflected radiation as a non-destructive method for analyzing nutritional status in grass. The objectives of this study were to analyze the relationships between photosynthetic pigments (chlorophyll, carotenoids, and chlorophyll/carotenoid ratio), canopy spectral characteristics, and vegetation indices (VIs) derived from broadband and narrowband analysis in tall fescue ( Festuca arundinacea). The overall goal was to determine a sensitive indicator for assessing grass nutritional status using hyperspectral remote sensing techniques. Canopy spectral measurements from each treatment (Control, Low fertility and High fertility) were taken in the field using a FieldSpec ® FR spectroradiometer. A large number (i.e. 22,500) of two-band combinations in the Normalized Difference Vegetation Index (NDVI) and the Ratio Vegetation Index (RVI) were then used for a linear regression analysis against photosynthetic pigments. Obvious differences in spectral reflectance existed between treatments within certain wavelength regions (400–1000 nm). In addition, the techniques of derivative analysis increased the separation of grasses with different fertility levels, providing the possibility of monitoring grass nutritional status. Unlike chlorophyll and carotenoid values individually, the ratio of chlorophyll/carotenoid was strongly correlated with canopy spectral reflectance (500–730 nm) ( P < 0.01). Further investigation of the relationships between photosynthetic pigments and traditional broadband vegetation indices suggested that wavelengths of 610–690 and 752–1000 nm, the regions of the red and NIR channels of several multi-spectral sensors in place on the current generation of earth-orbiting satellites, were not the optimum wavelengths for NDVI and RVI analysis. However, since the NDVI was closely related to chlorophyll/carotenoid ratio, using the combinations of λ 1 at 540–560 nm and λ 2 at 750–950 nm produced a coefficient of determination R 2 > 0.75. Thus, it was possible to map variation in grass photosynthetic pigments using hyperspectral remote sensing. This provided an insight for building new models and vegetation indices to monitor grass nutritional status by following the dynamics of the chlorophyll/carotenoid ratio.