Application of Infrared Spectroscopy in the Identification of Fallen Trees from the Amazon Rainforest (Myristicaceae)

Abstract

Abstract The infrared spectrum carries the so-called fingerprint of a plant, a characteristic that can help the identification of wood from an integrative perspective. Fourier-transform near-infrared spectroscopy (FT-NIRS) was evaluated for its potential use in the taxonomic identification of natural fallen trees from the Amazon rainforest (Myristicaceae). Trees found on the forest floor (Amazonas, Brazil) were surveyed, and eight individuals were selected for the study. The material was removed from the trunk for anatomical identification based on xylem structure and obtaining NIR spectra (10,000 to 4,000 cm-1). A total of 1,080 spectra were collected and used in the multivariate analysis. Absorption bands in the NIR region and chemometrics (calibration models) were used to interpret the spectra, which indicated similarity trends at the genus and species levels. Linear discriminant analysis (LDA) was able to differentiate Iryanthera ulei and Virola (five species), which were misclassified in the field as a single species (Virola sp.). The accuracy of the results obtained by chemometric models (LDA 7 and 14) indicates the robustness of NIR spectroscopy, which can be used as a new tool in the integrative taxonomy of plant material from tropical forests that are difficult to identify and can analyzing many tests quickly and securely. Study Implications: The taxonomic identification of natural fallen trees found on the floor of tropical forests is quite complex given the lack of flowers, fruits, and leaves in field collection. Fourier-transform near-infrared spectroscopy demonstrated its capability to distinguish between wood samples from fallen trees belonging to the Myristicaceae family. The chemometric model successfully discriminated the samples, which had been erroneously lumped together in the field under the single species designation of Virola sp. This technique has been previously overlooked for providing differentiation. Our study highlights its efficacy to address misclassifications and allow taxonomic identification.