Spectral reflectance properties of healthy and stressed coniferous trees

Abstract

This study investigates the properties of hyperspectral reflectance of healthy and stressed coniferous trees. Two coniferous tree species which naturally grow in Lithuania, Scots pine (Pinus sylvestris L.) and Norway spruce (Picea abies (L.) Karst.), as well as an introduced species, Siberian pine (Pinus sibirica Du Tour), were selected for the study. Hyperspectral reflectance data were collected under laboratory conditions by scanning the needles of healthy (no foliar loss) and stressed Norway spruce (foliar loss 66-70%), Scots pine (foliar loss 71-75%) and Siberian pine (foliar loss 86-90%) trees using a Themis Vision Systems VNIR 400H hyperspectral imaging camera. The spectrometer of the camera covers the spectral range of 400-1000 nm with the sampling interval of 0.6 nm. Simultaneously, the chlorophyll a and b content in the needles was determined by spectrophotometrically measuring the needles’ absorbance of ethanol extracts. The statistical analyses included principal component analysis, analysis of variance and partial least squares regression techniques. Relatively large spectral differences between healthy and stressed trees were detected for Norway spruce needles: 884 out of 955 wavebands indicated a statistically different reflectance (p<0.05). The reflectance associated with the stress level was statistically different (p<0.05) in 767 and 698 out of 955 wavebands for Scots pine and Siberian pine, respectively. The most informative wavelengths for spectral separation between the needles taken from healthy and stressed trees were found in the following spectral ranges: 701.0-715.7 nm for Norway spruce, 706.1-718.2 nm for Scots pine, and 862.3-893.1 nm for Siberian pine. The relationship between the spectral reflectance properties of the needles and their chlorophyll content was also determined for each species. Waveband ranges (as well as single bands) most sensitive to changes in chlorophyll content were: 709.9-722.1 nm (715.6 nm) for Norway spruce; 709.3-721.4 nm (715.0 nm) for Scots pine; 710.6-722.7 nm (720.1 nm) for Siberian pine. In general, the study revealed that narrow-band based hyperspectral imaging has the potential for accurately detecting stress in coniferous trees.