Exploring a best vegetation index to explain the seasonal variation of a forest photosynthesis using a hyper-spectral camera equipped with liquid crystal tunable filter

Abstract

Liquid crystal tunable filter ( LCTF ) can change the transmissible wavelength by changing the applied voltage to the filter, which enables the drastic increase in the observable wavelength resolution in a small size system and is considered to be a powerful tool for the spectral earth observation from flying units or microsatellites. However, there is limited knowledge about its season-long application for the vegetation monitoring and the prediction of the ecosystem photosynthetic capacity. We compared the seasonal variation of spectral reflectance obtained by a LCTF camera with that obtained by a popular spectral radiometer in a cool-temperate young larch plantation in northern Hokkaido, Japan. Then we tried to find the best normalized difference spectral index ( NDSI ) to explain the seasonal variation of the ecosystem photosynthetic capacity using all pairs of two reflectances observed in the range of wavelength between 500 and 770 nm with 10-nm intervals ( 28 wavelength bands ) by the LCTF. The best NDSI among all combinations ( 28 × 27 ) of two reflectances was NDSI[770, 720] for the maximum gross primary production at light saturation and NDSI[530, 600] for the initial slope of the light-response curve, which reflect the red edge shift owing to the change in the chlorophyll content and relative strength of the light absorbance in the visible red wavelength region compared with that in the green wavelength region, respectively. Predicted daily gross primary production of the plantation using these NDSI agreed well with the observed values. NDSI[530, 600] was better to distinguish each vegetation type of the studied plantation.