Rapid non-destructive estimation of Pinus taeda L. physical, morphological and chemical properties using near infrared spectroscopy
Jones, Paul David
MetadataShow full item record
Sixty percent of the nations timber products are produced from trees grown in the Southern United States, with plantation grown Pinus taeda L. (loblolly pine) being the major fiber source. Density, microfibril angle, stiffness, tracheid coarseness, specific surface, wall thickness, perimeter, and radial and tangential diameter from 120 radial strips of Pinus taeda L. (loblolly pine) trees grown on 14 sites in 3 physiographic regions of Georgia (USA) were measured by SilviScan. NIR spectra were also collected in 10 mm increments from the radial longitudinal surface of each strip and split into calibration (9 sites, 729 spectra) and prediction sets (6 sites, 225 spectra). NIR spectra were correlated using untreated and mathematically treated (first and second derivative and multiplicative scatter correction) with tracheid properties to develop calibrations for the estimation of these properties. Strong correlations were obtained for 2 properties related to density, the strongest R being 0.83 (density), 0.90 (MFA), 0.93 (stiffness), 0.80 (coarseness), 0.78 (specific surface) and 0.84 (wall thickness). The addition of a small number of cores from the prediction set (one core per new site) to the calibration set improved the accuracy of predictions and, importantly, minimized the differences obtained with the various math treatments. Also, seventeen P. taeda radial strips, representing seven different sites were selected and NIR spectra were obtained from the radial longitudinal face of each strip. The spectra were obtained in 12.5 mm sections from predetermined positions that represented juvenile wood (close to pith), transition wood (zone between juvenile and mature wood) and mature wood (close to bark). For these sections, cellulose, hemicellulose, lignin (acid soluble and insoluble), arabinan, galactan, glucan, mannan, and xylan contents were determined by standard analytical chemistry methods. Calibrations were developed for each chemical constituent using the NIR spectra, wood chemistry data and partial least squares (PLS) regression. Relationships were variable with the best results being obtained for cellulose, glucan, xylan, mannan and lignin. Prediction errors were high and may be a consequence of the relatively small number of samples available for NIR analysis and the diverse origins of the samples in the test set.