A new hyperspectral index for chlorophyll estimation of a forest canopy: Area under curve normalised to maximal band depth between 650-725 nm

摘要

Total chlorophyll (Cab) content of a forest canopy is used as indicator for the current state of a forest stand, and also as an input for various physiological vegetation models (i.e. models of photosynthesis, evapo-transpiration, etc.). Recent hyperspectral remote sensing allows retrieving the Cab concentration of vegetation using the appropriate optical indices, and/or by means of biochemical information, scaled up from leaf to canopy level within radiative transfer (RT) models. Plenty of chlorophyll optical indices can be found in the literature for the leaf level, nevertheless, only some of them were proposed for a complex vegetation canopy like a forest stand. A robust chlorophyll optical index at the canopy level should be driven by the Cab concentration without negative influence of other factors represented by soil background or understory, canopy closure, canopy structure (e.g. leaf area index (LAI), clumping of leaves), etc.\ud\udA new optical index named Area under curve Normalised to Maximal Band depth between 650-725 nm (ANMB₆₅₀-₇₂₅) is proposed to estimate the chlorophyll content of a Norway spruce (Picea abies, /L./ Karst.) crown. This index was designed to exploit modifications of a vegetation reflectance signature invoked within the red-edge wavelengths mainly by the changes in leaf chlorophyll content. ANMB₆₅₀-₇₂₅ is based on the reflectance continuum removal of the chlorophyll absorption feature between wavelengths of 650-725 nm. Suitability of the index and sensitivity on disturbing factors was tested using a 3D Discrete Anisotropic Radiative Transfer (DART) model coupled with a leaf radiative transfer model PROSPECT adjusted for spruce needles. The results of the ANMB₆₅₀-₇₂₅ abilities within a coniferous forest canopy were compared with the performance of the chlorophyll indices ratio TCARI/OSAVI.\ud\udTest results, carried out with the DART model simulating hyperspectral data with 0.9 m pixel size, showed a strong linear regression of the ANMB₆₅₀-₇₂₅ on spruce crown Cab concentration (R²=0.9798) and its quite strong resistance against varying canopy structural features such as LAI and canopy closure. The root mean square error (RMSE) between real and the ANMB₆₅₀-₇₂₅ estimated Cab concentrations was only 9.53 mg/cm² while the RMSE generated from prediction of the TCARI/OSAVI was two times higher (18.83 mg/cm²). Chlorophyll retrieval using the ANMB₆₅₀-₇₂₅ index remained stable also after introduction of two reflectance signal disturbing features: a) 20% of the spectral information of epiphytic lichen (Pseudevernia sp.) regularly distributed within the spruce canopy, and b) simulation of the sensor noise (computed for a signal to noise ratio equal to 5). RMSE of predicted Cab concentration after the introduction of lichens appeared to be 10.51 mg/cm² and the combined influence of lichen presence and sensor noise in the image caused an increase of the RMSE to 12.13 mg/cm².