Application of near-infrared spectroscopy (NIRS) in the chemical characterization of Corymbia hybrid wood

Interest in Corymbia wood stems from its diverse applications. This study evaluated the use of near-infrared spectroscopy (NIRS) as a nondestructive method for characterizing Corymbia wood. Were developed predictive models for wood chemical properties using 100 samples from 1059 trees across 34 clonal materials from three experimental plantations, each three years old. Spectra were obtained from nondestructive sawdust samples covering the 1110-2500 nm range. The 100 calibration trees were analyzed for total lignin, extractives, holocellulose, and ash content using conventional methods. Partial least squares regression (PLS-R) was used to develop and validate NIRS models. Results showed satisfactory predictions for lignin (R2cv = 0.63; RMSEcv = 0.58%; REp = 2.85%), extractives (R2cv = 0.72; RMSEcv = 0.60%; REp = 14.12%), holocellulose (R2cv = 0.61; RMSEcv = 1.05%; REp = 1.09%), and ash (R2cv = 0.42; RMSEcv = 0.22%; REp = 18.56%). These models predicted chemical properties of all 1059 trees accurately. Clonal materials 12 and 17 had high lignin, extractives, and ash, while materials 28, 33, and 34 had lower values. NIR spectroscopy proves to be an effective and efficient tool for assessing wood quality, particularly useful in large-scale genetic improvement programs.