Identification of wood species from Côte d’ivoire forest using near-infrared spectroscopy

Abstract

The illegal timber trade results from the confusion whether intentional or not between a species that is banned from harvesting, trade, or destruction and a non-protected species, it was important to find an alternative means of rapidly and reliably identifying species in the Ivorian forest. Near infrared spectroscopy (NIR), being non-destructive, was used to identify five (05) timbers from the Besso classified forest (Côte d’Ivoire). Chemometric tools were used to decode the spectra and reveal trends in the forest species studied. The PLS-DA models effectively differentiated between wood and wood flour. The PLS-DA classification model shows that it is possible to accurately differentiate Côte d’Ivoire woods, with accuracy rates ranging from 0.92 to 0.99 for different types of wood, such as Khaya ivorensis, Mansonia altissima and Nauclea diderrichii. However, the differentiation of certain species, such as Milicia excelsa and Milicia regia, remains more complex due to their similarity, although wood flour from these woods showed better classification results. The reduction in the particle size of wood flour could explain this improvement. The performance of the PLS-R and PLS-DA models is encouraging, with high cross-validation coefficients (R² between 0.79 and 0.92), underlining the effectiveness of NIR spectroscopy combined with chemometric methods to identify and classify wood accurately and quickly. These preliminary results could help control agencies to better monitor timber marketing by developing a more elaborate database of timber species in Côte d’Ivoire, due to the simplicity and reliability of the technology.