Physically-based retrievals of Norway spruce canopy variables from very high spatial resolution hyperspectral data

Abstract

This study was conducted to answer two research questions: (1) what is the spatial variability of the leaf optical properties between 400–1600 nm (hemispherical-directional reflectance, transmittance, absorption) within young Norway spruce crowns, and (2) how to design a suitable physically-based approach retrieving the total chlorophyll content of a complex coniferous canopy from very high spatial resolution (0.4 m) hyperspectral data? It was proved that sun exposed needles of current age-class statistically differ (α-level = 0.01) from rest of the needles in reflectance between 510-760 nm. Last four age- classes of sun exposed needles were also found to be significantly different from almost all age-classes of sun shaded needles in transmittance from 760-1350 nm. An operational estimation of chlorophyll a+b content (Cab) from an airborne AISA Eagle hyperspectral image was proposed by means of a PROSPECT- DART inversion employing an artificial neural network (ANN). A spatial pattern of estimated Cab was successfully validated against the Cab map produced by a vegetation index ANCB650-720. Coefficients of determination (R²) between ground measured and retrieved Cab were 0.81 and 0.83, respectively, with root mean square errors (RMSE) of 2.72 μg cm⁻² for ANN and 3.27 μg cm⁻² for ANCB650-720.