Spectral Pre-Processing and Multivariate Calibration Methods for the Prediction of Wood Density in Chinese White Poplar by Visible and Near Infrared Spectroscopy

Wood density is a key indicator for tree functionality and end utilization. Appropriate chemometric methods play an important role in the successful prediction of wood density by visible and near infrared (Vis-NIR) spectroscopy. The objective of this study was to select appropriate pre-processing, variable selection and multivariate calibration techniques to improve the prediction accuracy of density in Chinese white poplar (Populus tomentosa carriere) wood. The Vis-NIR spectra were de-noised using four methods (lifting wavelet transform, LWT; wavelet transform, WT; multiplicative scatter correction, MSC; and standard normal variate, SNV), and four variable selection techniques, including successive projections algorithm (SPA), uninformative variables elimination (UVE), competitive adaptive reweighted sampling (CARS) and iteratively retains informative variables (IRIV), were compared to simplify the dimension of the high-dimensional spectral matrix. The non-linear models of generalized regression neural network (GRNN) and support vector machine (SVM) were performed using these selected variables. The results showed that the best prediction was obtained by GRNN models combined with the LWT and CARS method for Chinese white poplar wood density (Rp2 = 0.870; RMSEP = 13 Kg/m3; RPDp = 2.774).

Effect of Initial Planting Density on Growth Traits and Wood Properties of Triploid Chinese White Poplar (Populus tometosa) Plantation

Planting density primarily affects the yield and wood quality of plantations. There are multiple reports on the effects of planting density on growth traits and wood properties in young triploid Chinese white poplar (Populus tomentosa) plantations. Nevertheless, assessment of the effects of initial planting density is lacking for plantations older than ten years. Here, an 11-year-old plant density trial (2490, 1665, 1110, 832, 624, 499, and 416 trees/hm2) established with four hybrid clones (S86, B301, B331 and 1316) in northern China was used to determine the effect of initial planting density on growth traits (diameter at breast height (DBH), tree height (H), stem volume (SV) and stand wood volume (SWV)), basic wood density (BWD), and fiber properties (fiber length (FL), fiber width (FW), and the ratio of fiber length to width (FL/FW)). A total of 84 trees from four clones were sampled. In this study, the initial planting density had a highly significant effect on growth traits (p < 0.001) and had a moderate effect on FL. Overall, the reduction in initial planting density led to the increase in DBH, H, SV, and FL/FW. Triploid hybrid clones planted at 416 trees/hm2 had the largest DBH, H, SV, FL/FW and the smallest SWV and FW. Clonal effects were also significant (p < 0.05) for all studied traits except for FL. Clone S86 had a higher growth rate and the largest BWD and FW. Clones–initial planting densities interaction was insignificant for all growth traits and wood properties. A weak and positive estimated correlation between BWD and growth traits (H, SV, SWV) within each planting density was seen. Our results demonstrate that an appropriate reduction in initial density in triploid Chinese white poplar plantations with long rotation is a suitable strategy to promote tree growth and retain excellent wood processing characteristics.

Visible-Near Infrared Spectroscopy and Chemometric Methods for Wood Density Prediction and Origin/Species Identification

This study aimed to rapidly and accurately identify geographical origin, tree species, and model wood density using visible and near infrared (Vis-NIR) spectroscopy coupled with chemometric methods. A total of 280 samples with two origins (Jilin and Heilongjiang province, China), and three species, Dahurian larch (Larix gmelinii (Rupr.) Rupr.), Japanese elm (Ulmus davidiana Planch. var. japonica Nakai), and Chinese white poplar (Populus tomentosa carriere), were collected for classification and prediction analysis. The spectral data were de-noised using lifting wavelet transform (LWT) and linear and nonlinear models were built from the de-noised spectra using partial least squares (PLS) and particle swarm optimization (PSO)-support vector machine (SVM) methods, respectively. The response surface methodology (RSM) was applied to analyze the best combined parameters of PSO-SVM. The PSO-SVM model was employed for discrimination of origin and species. The identification accuracy for tree species using wavelet coefficients were better than models developed using raw spectra, and the accuracy of geographical origin and species was greater than 98% for the prediction dataset. The prediction accuracy of density using wavelet coefficients was better than that of constructed spectra. The PSO-SVM models optimized by RSM obtained the best results with coefficients of determination of the calibration set of 0.953, 0.974, 0.959, and 0.837 for Dahurian larch, Japanese elm, Chinese white poplar (Jilin), and Chinese white poplar (Heilongjiang), respectively. The results showed the feasibility of Vis-NIR spectroscopy coupled with chemometric methods for determining wood property and geographical origin with simple, rapid, and non-destructive advantages.

Variation in growth, leaf, and wood property traits of Chinese white poplar (Populus tomentosa), a major industrial tree species in Northern China

The natural phenotypic variation in Chinese white poplar (Populus tomentosa Carr.), which is distributed across a wide geographical area of northern China (30°N–40°N, 105°E–125°E), is a potential source of beneficial variation for poplar breeding. Thirteen traits related to growth, leaf, and wood properties were quantified in 460 P. tomentosa individuals grown in a common garden plot. There was considerable range-wide phenotypic variation in all traits across individuals according to the patterns of ANOVA among hierarchical groups (populations and regions, respectively). A clear sexual dimorphism for seven traits was examined. In total, 32 trait–trait phenotypic correlations (P ≤ 0.05), 10 trait–geographical factor correlations (P ≤ 0.05), and a highly interrelated structure network were identified, which was further supported by principal component analysis (PCA). These associations can be used in multiple-trait selective breeding programs for advantageous phenotypic traits. A hierarchical cluster analysis was used to classify four groups (southeastern, central, northeastern, and southwestern populations) among the natural populations using these 13 phenotypic traits. This study provides important perspectives into the use of direct breeding to potentially improve economic traits and provides a starting point for genome-wide association studies in P. tomentosa in the near future.