The aim of current study was to investigate the spectral separability trends of different tree species due to varying foliage spectral reflectance properties during the growing season. Five tree species prevailing in Lithuania were chosen for the study: aspen (Populus Tremula L.), black alder (Alnus Glutinosa (L.) Gaertn.), Norway spruce (Picea Abies L.), Scots pine (Pinus Sylvestris L.) and silver birch (Betula Pendula Roth). The hyperspectral reflectance data was collected under laboratory conditions scanning the foliage samples from two healthy middle aged stands. Hyperspectral scanning was implemented using Themis Vision Systems LLC VNIR 400H hyperspectral imaging camera in 400-1000 nm range. Principal component analysis and the Jeffries-Matusita distance measure were applied for the analysis of hyperspectral data. The wavelengths providing the best separability between tree species were determined. They were discovered to vary during the growing season: in late spring – early summer the most informative wavelengths were concentrated in blue and near infrared spectral zones, in summer they shifted towards green and red zones, and in autumn they moved further to longer waves - the red and near infrared - spectral zones. The investigated tree species were determined to be spectrally separable during the whole growing season, but the particular periods were revealed to contribute for improved spectral separability between certain tree species. The separability between coniferous species was best in September, while deciduous species were best separable when the samples were collected in August.

The aim of the study was to investigate the properties of hyperspectral reflectance data of Scots pine (Pinus Sylvestris L.) and Norway spruce (Picea Abies L.). The hyperspectral reflectance data was obtained under laboratory conditions from the last season’s needles of healthy 20 year-old trees from the same site. Hyperspectral data was acquired using Themis Vision Systems LLC VNIR 400H portable scanning hyperspectral imaging camera in 400-1000 nm range. Methods of analysis of variance, discriminant analysis and principal component analysis were applied for the hyperspectral data analysis. Differences between Scots pine and Norway spruce reflection data were examined. The most informative spectral range for Norway spruce – Scots pine spectral separation was determined at 666.5 nm – 668.4 nm, most informative waveband - 667.1 nm. Reflectance variations among individual trees of the same species as well as differences in spectral response between needles from northern – southern crown exposition were tested. A significant variation in spectral response of needles of Norway spruce was detected across the whole measured spectral range (955 wavebands) for each sample tree. However, significant variation of spectral response of needles of Scots pine was detected only in 356 out of 955 wavebands for each sample tree. Depending on the crown exposition to the North or South, the reflectance of Scots pine needles differed significantly in 900 spectral bands. No significant differences were detected in 833 wavebands for Norway spruce.