A method for rapid non-destructive determination, correction and prediction of standing tree density and a method for analyzing tree phenotypes and genetic correlations

Abstract

The application discloses a method for rapid nondestructive determination, correction and prediction of standing tree density and a method for analyzing forest tree phenotype and genetic correlation, and belongs to the technical field of forest tree breeding. The method comprises the following steps: 1) determining wood resistance values of standing trees by using a needle instrument; and determining the real density MWD of wood cores by using an X-ray density instrument or a volume method; 2) performing trend elimination processing on the wood resistance values in step 1), and correcting errors caused by tree bark to obtain a corrected wood resistance data set IML TB ; 3) taking IML TB as an independent variable, taking MWD as a dependent variable, constructing a prediction model, and fitting a linear regression model; and 4) predicting the standing tree density by using the linear regression model. The method for rapid determination of wood density provided by the application can be generalized, and is of great significance for wood quality evaluation of spruce and Chinese fir and other coniferous wood tree species germplasm resources and selection breeding of forest wood properties.

Description

Technical Field

[0001] This application relates to the fields of standing tree density measurement and genetic breeding technology, specifically to a method for rapid and non-destructive measurement, correction and prediction of standing tree density, and a method for analyzing the correlation between forest tree phenotype and genetics. Background Technology

[0002] Timber density, defined as the mass per unit volume of timber under specific moisture content conditions, is one of the most reliable indicators for predicting timber's mechanical properties, processing performance, and ultimate economic value (Zobel and Buitjeenen, 1989). Against the backdrop of increasing societal demand for timber products and escalating timber resource scarcity, improving timber utilization efficiency has become a key strategy. Accurate measurement of timber density is not only fundamental to optimizing material selection and achieving "suitable materials for suitable applications," but also crucial for improving the accuracy of forest carbon storage estimation (Mo et al. 2024). However, in actual forestry production, how to quickly and accurately assess timber properties, especially timber density, remains a significant technical challenge.

[0003] Traditional methods for measuring the density of standing timber primarily rely on the volumetric method, which requires drilling core samples and determining their absolute density (Gao et al., 2012). While the volumetric method yields accurate results, it suffers from drawbacks such as being destructive, time-consuming, and labor-intensive. Another high-precision technique is X-ray density measurement (Li & Wu, 2005; Dierickx et al., 2024), which is based on the difference in X-ray absorption by woods of different densities to indirectly estimate density (Eberhardt & Samuelson, 2015). However, this method also requires laboratory sample analysis, resulting in limitations such as high cost, low throughput, and high labor input, making it difficult to meet the practical needs for rapid, economical, and non-destructive testing of standing timber density.

[0004] For coniferous species like Norway spruce, which have a growth cycle of several decades, early selection has irreplaceable value in their genetic improvement process (Chen et al., 2015; Lenz et al., 2013; Chen et al., 2014). Currently, the academic community has conducted extensive systematic research on the response mechanisms of early selection to growth traits in many coniferous species (Wu et al., 2007a). In recent years, the research scope of early selection has gradually expanded to the field of wood quality traits (Hong et al., 2015), which also highlights the necessity of using non-destructive testing techniques such as micro-drill resistance testing during the young age of trees. From the perspective of breeding efficiency, the breeding cycle of Norway spruce from sapling to mature tree is relatively long. If the determination of wood density and growth traits is postponed until the trees mature, the selection cycle for superior individual trees will be significantly prolonged, thus severely restricting the breeding process. Conversely, using non-destructive testing techniques to screen individual trees with high potential for wood density during their juvenile stage can advance the trait measurement point by 10-20 years, significantly accelerating the generational succession in breeding. Furthermore, from the perspective of genetic correlations between traits, relevant studies on coniferous species generally show a high genetic correlation between wood density in juvenile and mature trees (Wu et al., 2007b; Lenz et al., 2011). For example, Chen et al. (2014) found that the age-age correlation of wood density is typically extremely high, with values ​​very close to 1. Therefore, in the Norway spruce wood quality improvement project, conducting early selection based on non-destructive testing techniques is not only a practical need to shorten the breeding cycle and reduce cultivation costs, but also a key technical approach to achieve synergistic improvement of wood density and growth traits. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the first technical problem this application aims to solve is to provide a method for rapid and non-destructive measurement, correction, and prediction of standing tree density, meeting the application requirements for rapid and non-destructive measurement and prediction of standing tree density. Another technical problem this application aims to solve is to provide a method for analyzing forest tree phenotypic and genetic correlations based on the aforementioned measurement data, meeting the needs of forest tree breeding.

[0006] A method for rapid, non-destructive determination, correction, and prediction of standing tree density, comprising the following steps:

[0007] 1) The original resistance data of living timber in forests were determined using a needle penetration tester; the average true density (MWD) of the wood core was determined using an X-ray densitometer or a volumetric method.

[0008] 2) Trend elimination processing was performed on all the raw wood resistance data obtained in step 1), and the error caused by the bark was corrected to obtain the corrected wood resistance dataset IML.TB ;

[0009] 3) Using the IML dataset TB Using as the independent variable and the average true density of timber (MWD) as the dependent variable, a prediction model is constructed. The fitted univariate linear regression model is as follows:

[0010]

[0011] 4) Predict the density of living trees using a univariate linear regression model.

[0012] In step 1), the acupuncture device is an IML-RESI PD300.

[0013] In step 1), the X-ray densitometer is Itrax.

[0014] In step 2), the trend elimination process is as follows: linear detrending processing is applied to the borehole resistance curve, the difference between the readings at the beginning and end of the curve is calculated, the upward slope is obtained by combining the total length of the curve, and then each measuring point on the curve is corrected according to the slope to eliminate the upward trend caused by drill bit friction.

[0015] In step 2), the error caused by the bark is corrected by truncating the effective section of the curve. Specifically, the bark sections with significantly lower resistance at both ends of the curve and the first incomplete growth ring section adjacent to the bark are removed, and only the stable section corresponding to the pure xylem in the middle is retained.

[0016] A method for analyzing the phenotypic and genetic correlation of forest trees based on the aforementioned method includes the following steps:

[0017] 1) Use a needle-punch instrument to measure the original resistance data of living timber; use an X-ray densitometer to measure the average true density (MWD) of the wood core; measure growth traits and calculate trunk volume to obtain phenotypic values;

[0018] 2) The raw wood resistance data obtained in step 1) are all subjected to trend elimination processing, and the error caused by bark is corrected to obtain the corrected wood density dataset IML. TB ;

[0019] 3) Extract forest tree DNA, perform genotyping, and fill in missing genotypes; construct a genome phylogenetic matrix G using LDAK software based on whole-genome single nucleotide polymorphism markers;

[0020] 4) The phenotypic values ​​are corrected using a spatial model, which is as follows: In the formula: The overall mean of the data. For spatial autocorrelation residuals, denoted as independent random residuals; b is the fixed effects vector containing the population mean and block effects, and X is its corresponding design matrix; u is the random additive genetic effects vector, and Z is its corresponding design matrix.

[0021] 5) Conduct phenotypic and genetic correlation analysis of trees.

[0022] In step 4), the first two principal components PC1 and PC2, derived from the genomic kinship matrix G, are incorporated into the model as fixed effects, and the fitted linear mixture model is as follows:

[0023]

[0024] In the formula, y is the vector of phenotypic observations; μ is the population mean; β1 and β2 are the fixed effects coefficients of principal component 1 (PC1) and principal component 2 (PC2), respectively; X is the correlation matrix corresponding to G; and e is the vector of random residual effects.

[0025] In step 5), the formula for calculating narrow-sense heritability is:

[0026]

[0027] In the formula, For additive inheritance variance, This represents the residual variance.

[0028] In step 5), the formula for calculating the genetic correlation coefficient is as follows:

[0029]

[0030] In the formula, Represents additive genetic covariance, characterizing the shared effect between genetic factors controlling variation in trait 1 and trait 2; and These are the additive inheritance variances of trait 1 and trait 2, respectively;

[0031] The formula for calculating the phenotypic correlation coefficient based on variance components is as follows:

[0032]

[0033] In the formula, For residual covariance, and These are the residual variances corresponding to the two traits.

[0034] Beneficial effects: Compared with the prior art, the technical advantages of this application are as follows:

[0035] 1) This application verifies the genetic reliability of IML needle acupuncture correction data on Norway spruce and Chinese fir, respectively: IML after correction in Norway spruceTB Heritability reached 0.64, highly consistent with the X-ray baseline density heritability of 0.65; after correction, the IML of cedar was... TB Heritability increased from 0.25 to 0.29, and genetic signaling stabilized.

[0036] 2) IML TB The results showed significant correlations with the baseline density phenotype and genetic correlations, and a negative genetic correlation with growth traits, consistent with the general genetic laws of coniferous species, proving that the test results can be directly used for genetic assessment.

[0037] 3) This application combines the resistance data of Norway spruce and Chinese fir to form a general prediction model that can be generalized to coniferous timber species such as spruce and Chinese fir, breaking through the limitation of a single tree species. Attached Figure Description

[0038] Figure 1 The following is a graph of the drilling resistance of the same tree: (A) Measured by IML-RESI PD300, the black curve represents the original resistance data, and the red curve represents the data after trend elimination and bark correction;

[0039] Figure 2 This is a graph showing the correlation between the surface-weighted average density (MWD) of Norway spruce Itrax and the wood density estimated by the IML-RESI needle punch.

[0040] Figure 3 This is a graph showing the relationship between the average drilling resistance and wood density of Norway spruce and fir. Detailed Implementation

[0041] The present application will now be described in detail with reference to specific embodiments.

[0042] Example 1

[0043] Norway spruce (Picea abies L.) is a key coniferous species in Northern and Central Europe, possessing significant economic and ecological value (Zhang 1997; Blair et al. 1976). Its wood is characterized by uniform grain, excellent processing properties, and outstanding resonance characteristics, making it widely used in high-value-added sectors such as construction, furniture manufacturing, and pulp production (Hannrup et al., 2004). Wood density is a core indicator of wood quality, closely related to key properties such as calorific value and mechanical strength (Chave et al., 2009; Preston et al., 2006). However, a negative genetic correlation between growth traits and timber density is commonly observed in coniferous species, including Norway spruce (Picea abies) (Chen et al., 2014), Scots pine (Pinus sylvestris) (Hong et al., 2014), white spruce (Picea glauca (Moench) Voss) (Lenz et al., 2011), twisted pine (Pinus contorta) (Hayatgheibi et al., 2017), and radiata pine (Pinus radiata) (Baltunis et al., 2007; Wu et al., 2008). Therefore, one of the core objectives of forest tree breeding strategies is to achieve synergistic improvement of growth traits and timber quality traits.

[0044] The experimental material used in this embodiment was a Norway spruce progeny testing forest established in Häggenås, northern Sweden, in 1987. This testing forest comprised 723 freely pollinated families, 2 controlled-pollination families, and several control plots, with each family having an average of 14 offshoots. A completely randomized block design (CRD) was employed, with individual trees as plot units and a spacing of 2.2 m × 1.5 m. To address environmental variability observed within the experimental site, the research team, following the standard operating procedures of the Swedish forestry breeding project, added a secondary partition to the initial completely randomized block design to improve the accuracy of the experiment.

[0045] 1. Measurement of wood density and resistance data

[0046] In 2022, one core sample was extracted from each of 723 families at the Häggenås experimental site, approximately 1.3 meters above the ground, using a 5 mm growth cone. These cores were then scanned using an Itrax X-ray densitometer. This instrument utilizes a micro-focused X-ray beam to reconstruct high-resolution wood density profiles by precisely measuring the intensity attenuation of X-rays after penetrating wood tissues of different densities. Wood density can be obtained from these profiles. In this embodiment, the average wood density measured by Itrax will serve as a benchmark for evaluating subsequent indirect measurement methods.

[0047] Indirect non-destructive assessment of wood density was performed using an IML-RESI PD300 needle punch, which drilled from the bark to the bark on the opposite side, with the drill holes approximately 3 cm from the growth cone sampling point. This ensured that the measurement location was close to the reference measurement point while minimizing potential interference with the wood structure from core drilling. Knots, bark cracks, and other visible trunk damage were avoided during drilling. The drill bit was held perpendicular to the trunk surface and advanced at a constant feed rate. After each measurement, the drilling resistance profile was checked in real-time on the device interface. Measurements showing abnormal fluctuations were considered invalid, and remeasurements were immediately performed at adjacent locations to obtain a representative profile. All valid data were then exported for further processing. The raw measurement data are publicly available on the Zenodo platform at https: / / doi.org / 10.5281 / zenodo.19453457.

[0048] 2. Data processing method for acupuncture instruments

[0049] The IML-RESI needle penetration instrument performs testing by inserting a drill bit into the wood. Its drill speed and feed rate can be manually adjusted by the operator and remain constant throughout the drilling process. The sampling resolution is 10 data points per millimeter. As the drilling depth increases, the cumulative effect of systematic friction causes the resistance profile curve to show an overall upward trend. Figure 1 ).

[0050] To eliminate this bias and obtain a more accurate estimate of wood density, assuming that the resistance change follows a linear trend, trend elimination correction is performed on the resistance profile curve. The slope of the curve is calculated by comparing the difference between the readings at the beginning and end with the total length, and point-by-point correction is applied to eliminate the upward trend caused by friction. For the resistance curve after trend elimination, low-resistance bark sections on both sides of the curve are removed using threshold judgment and cumulative counting methods. Then, invalid data other than peak values ​​are removed using a difference algorithm, retaining the valid data section of pure xylem in the middle. Three types of datasets are generated for the data from the two acupuncture instruments (Table 1):

[0051] (1) Original resistance dataset (IML) R );

[0052] (2) Trend-eliminating corrected dataset (IML) T );

[0053] (3) Trend elimination correction and bark data removal dataset (IML) TB ).

[0054] 3. Growth trait determination (the original determination data has been made public on the Zenodo platform, the public URL is https: / / doi.org / 10.5281 / zenodo.19453457)

[0055] Growth traits included tree height (HGT), diameter at breast height (DBH), and trunk volume (VOL). Tree height and DBH were measured in 2023, when the trees in the experimental forest were 34 years old. Trunk volume was calculated using a single-tree volume function, with DBH (in cm) and tree height (in m) as input variables, and was calculated using a decimal logarithmic modified binary power function model.

[0056] V = 10 b0 ×DBH b1 × (DBH+20) b2 ×HGT b3 ×(HGT-1.3) b4

[0057] In the formula, V is the volume of the tree trunk, in dm. 3 b0 ~ b4 are the fitting parameters specific to Norway spruce. In this embodiment, the parameters for the southern Swedish region are selected: b0 = -1.02039, b1 = 2.00128, b2 = -0.47473, b3 = 2.87138, b4 = -1.61803.

[0058] 4. DNA extraction

[0059] From 723 families of plants in the Häggenås forest, individual plants with collected core samples were selected, and fresh needles were collected for DNA extraction.

[0060] 5. Genotyping

[0061] All resequencing plants were used as experimental subjects, and single nucleotide polymorphism (SNP) detection was performed using bcftools software. The specific procedure is as follows:

[0062] First, the sequencing sequences were aligned with the reference genome using BWA software. Then, the generated BAM files were formatted and sorted using samtoolsview, and repetitive sequences were removed. Finally, preliminary SNP detection was performed using bcftools software. After detection using bcftools software and multi-step quality control, only biallelic SNPs were selected to construct a variant dataset. The SNPs were then further detected and genotyped using STITCH software.

[0063] 6. Spatial Analysis

[0064] This embodiment uses a spatial model to correct phenotypic values. This model introduces an autoregressive spatial component and independent error terms into a linear mixture model, replacing the single structure of the traditional model that only contains independent error terms.

[0065]

[0066] For spatial autocorrelation residuals, denoted as independent random residuals; b is the fixed effects vector containing the population mean and block effects, and X is its corresponding design matrix; u is the random additive genetic effects vector, and Z is its corresponding design matrix.

[0067] 7. Statistical Analysis

[0068] Variance components were estimated using the Genomic Best Linear Unbiased Prediction (GBLUP) method. Based on genome-wide single nucleotide polymorphism (SNP) markers, a binary format genome kinship matrix file was pre-calculated using the LDAK algorithm. This file was read and parsed in R, including diagonal, off-diagonal, and individual ID information, to reconstruct the complete G matrix. The G matrix was then optimized by matching individual IDs with phenotypic data, symmetry correction, and positive definiteness. The sparse inverse matrix was calculated, and the first five principal components were extracted using principal component analysis to correct the population structure. The optimized G matrix inverse was then substituted into ASReml software to construct a GBLUP mixed linear model to estimate variance components, trait heritability, and genetic correlation coefficients between traits. To eliminate the interference of population structure on the analysis results, the first two principal components (PC1 and PC2) derived from the genome kinship matrix (G matrix) were included as fixed effects in the model. The fitted linear mixed model is as follows:

[0069]

[0070] In the formula, y is the vector of phenotypic observations; μ is the population mean (intercept); β1 and β2 are the fixed effects coefficients of PC1 and PC2, respectively; X is the correlation matrix corresponding to G; and e is the vector of random residual effects. This embodiment uses ASReml-R software to fit the above model and estimates the variance components using the maximum likelihood method (REML). The formula for calculating narrow heritability is as follows:

[0071]

[0072] In the formula, For additive inheritance variance, This represents the residual variance. This embodiment employs a bivariate mixed linear model to systematically analyze the genetic relationships among traits. Bivariate analysis based on the genomic kinship matrix can accurately estimate the additive genetic covariance and residual covariance among traits, and then calculate the genetic correlation coefficient and phenotypic correlation coefficient through variance components. The formula for calculating the genetic correlation coefficient is as follows:

[0073]

[0074] In the formula, Represents additive genetic covariance, characterizing the shared effect between genetic factors controlling variation in trait 1 and trait 2; and These represent the additive genetic variances of trait 1 and trait 2, respectively. The formula for calculating the phenotypic correlation coefficient based on the variance components is as follows:

[0075]

[0076] In the formula, For residual covariance, and These are the residual variances corresponding to the two traits.

[0077] 9. Results

[0078] 1) Variation characteristics of the tested traits

[0079] The statistical results of wood density-related indicators and growth trait indicators measured by the Itrax densitometer and IML-RESI needle penetration analyzer are summarized in Table 1. The area-weighted total mean wood density (MWD) measured by the Itrax densitometer ranged from 267.2 to 701.4 kg m⁻³, with an average of 500.1 kg m⁻³. The uncorrected wood density values ​​from the IML-RESI needle penetration analyzer (IML...) RThe density ranged from 1092.0 to 2616.0, with an average of 1725.6. After trend elimination correction, the average density decreased to 1688.9; however, after further removing bark data, the average density increased to 2027.1. The correlation results between the surface-weighted average density (MWD) of Norway spruce Itrax and the wood density estimated by the IML-RESI needle punch are as follows... Figure 2 As shown.

[0080] The coefficient of variation (CV) for wood density (MWD) was 9.4%, lower than that of the index measured by the IML needle penetration tester (12.0%~14.8%). Among the growth traits, trunk volume (VOL) showed the highest degree of variation (coefficient of variation 66.2%), followed by diameter at breast height (DBH, coefficient of variation 27.8%).

[0081] Table 1 lists the variables and their descriptive statistical characteristics, including minimum, maximum, mean, standard deviation, phenotypic coefficient of variation, and narrow-sense heritability of individual trees.

[0082]

[0083] Note: 1 Itrax overall average density (measured by X-ray densitometer); 2 : Overall timber density predicted by the model; 3 Unadjusted average density (raw data): IML-RESI / RINNTECH raw borehole resistance curve average; 4 Average density after detrending: The average value of the borehole resistance curve after linear detrending processing; 5 Detrended and bark-removed average density: The average value of the borehole resistance curve after linear detrending and bark removal.

[0084] 2) Heredity

[0085] Based on the genome phylogenetic matrix constructed from whole-genome resequencing data, this embodiment estimates the narrow-sense heritability (h0.05) of wood density and growth traits. 2 The results are shown in Table 1. Original measurement data from the IML acupuncture instrument (IML) R The narrow-sense heritability was 0.48; after trend elimination correction (IML) T ) and subsequent bark data removal processing (IML) TB After this, the narrow-sense heritability increased to 0.51 and 0.64, respectively, and was closest to the baseline heritability (0.65) measured by the Itrax densitometer. These results indicate that trend elimination and bark data removal correction can significantly improve heritability estimates, and the IML-RESI needle acupuncture instrument provides the most reliable wood density estimation data.

[0086] 3) Phenotypic and genetic correlation between wood density and X-ray reference density based on micro-drilling resistance method

[0087] The phenotypic and genetic correlation analysis results between wood resistance determined by the borehole resistance method and the baseline mean wood density (MWD) measured by the Itrax X-ray densitometer are shown in the figure. Figure 3 Table 2. Uncorrected density data from the IML acupuncture device showed a moderate phenotypic correlation with the baseline MWD. = 0.54); after trend elimination correction and trend elimination combined with bark data removal correction, the phenotypic correlation coefficients of the two increased to 0.57 and 0.67, respectively.

[0088] Table 2. Genetic and phenotypic correlation coefficients (SE: standard errors) between Itrax face-weighted average density (MWD) and wood resistance measured by the IML-RESI instrument.

[0089]

[0090] At the level of genetic correlation, the genetic correlation coefficients between each IML dataset and the benchmark MWD are as follows: Uncorrected data (IML) R 0.95, Trend Elimination Correction Data (IML) T 0.98, Trend Elimination Combined with Bark Removal Correction Data (IML) TB The correlation coefficient was 0.79. Trend elimination correction slightly increased the genetic correlation coefficient. Although subsequent bark data removal led to a decrease in the correlation coefficient, it improved the estimation accuracy by reducing the standard error.

[0091] In summary, the wood resistance values ​​measured by the IML-RESI needle punch instrument show a strong phenotypic and genetic correlation with the baseline MWD, indicating that this model of needle punch instrument has a certain degree of reliability in wood density estimation.

[0092] 4) Genetic correlation between indirect wood density measurements and wood density components and growth traits

[0093] Table 3 shows the results of the correlation analysis between the total mean wood density (MWD and IML measurements) and wood density, phenotypic characteristics, and genetic traits. Genetic correlation analysis indicated that the baseline density (MWD) and all IML needle-prick measurements showed significant negative genetic correlations with growth traits such as tree height, diameter at breast height (DBH), and trunk volume, with correlation coefficients ranging from -0.37 to -0.99. This is entirely consistent with the general negative genetic correlation between growth traits and wood density in coniferous species. These results confirm that the wood resistance data measured by the IML needle-prick instrument can accurately reflect the true genetic variation in wood density and is highly reliable for forest genetic assessment and breeding selection.

[0094] Table 3. Genetic correlation coefficients between estimated mean density (MWD, IML) of timber and growth traits (standard errors are in parentheses).

[0095]

[0096] Example 2

[0097] This embodiment uses Chinese fir as the subject and employs the method of Example 1 to rapidly and non-destructively determine, correct, and predict the standing density of Chinese fir. Based on the predicted data, an analysis of tree phenotypes and genetic correlations was conducted. The main original measurement data are published on the Zenodo platform at https: / / doi.org / 10.5281 / zenodo.19453457. The main process and results are as follows:

[0098] 1. Test materials

[0099] The experimental materials used in this embodiment are: superior clones from the third-generation Cunninghamia lanceolata germplasm resource bank and second-generation Cunninghamia lanceolata progeny experimental forest.

[0100] Twenty-one clones were selected from the third-generation seed resource bank in the back mountain of Yangkou State-owned Forest Farm, Fujian Province. One to three wood cores were selected from each clone, for a total of 56 wood cores. The resistance data of the needle penetration instrument was measured using an IML-RESI needle penetration instrument, and the density of the wood cores was determined using the volumetric method.

[0101] This embodiment selected a second-generation progeny test forest of Chinese fir from Laizhou Forest Farm, containing 80 freely pollinated families and one control group. The experiment adopted a balanced lattice design with 10 replicates, each replicate containing 9 incomplete blocks, each block containing 9 plots, and each plot containing 4 trees planted in a single row. Based on pedigree records, 887 individuals from 60 families were selected for acupuncture resistance data measurement.

[0102] 2. Measurement of wood density data

[0103] The measurement and processing of needle penetration resistance data is the same as that of Norway spruce, but the reference density uses the volumetric method. Using a growth cone to sample the spruce, cylindrical core samples are obtained. After drying, weighing, and volume calculation of the samples, the wood density value is calculated.

[0104] 3. Variation characteristics and heritability of the tested traits

[0105] Uncorrected wood density values ​​measured by IML-RESI needle penetration test (IML) R The distribution range is 281.7~1894.2, with an average of 868.9; after trend elimination correction, the average (IML) is... TThe mean value decreased slightly to 847.2, indicating that the friction effect has a minor impact on the overall mean. It is worth noting that after further removing bark sections, the mean value (IML) decreased slightly. TB The value significantly increased to 949.2. This phenomenon is consistent with the physical characteristics of wood density measurement, namely, after removing the low-density bark edges, the remaining xylem core section exhibits a higher average resistance characteristic. Regarding variability, the corrected IML... TB The coefficient of variation (CV) for the data was 24.3%, and the coefficient of variation for diameter at breast height (DBH) was 22.8%.

[0106] A genome phylogenetic matrix was constructed based on 50K gene chip typing data. In this embodiment, the narrow-sense heritability (h0.05) of Chinese fir wood density and growth traits was estimated. 2 The results are shown in Table 4. Original measurement data from the IML acupuncture instrument (IML) R The narrow-sense heritability of ) was 0.25; after trend elimination correction (IML) T ) and subsequent bark data removal processing (IML) TB After that, its narrow-sense heritability increased to 0.29. This indicates that the tested population has considerable genetic variation in the wood density trait, providing a material basis for carrying out genetic improvement.

[0107] Table 4 lists the variables and descriptive statistical characteristics of Chinese fir, including minimum, maximum, mean, standard deviation, phenotypic coefficient of variation, and narrow-sense heritability of individual trees.

[0108]

[0109] 4. Genetic correlation between indirect wood density measurements and growth traits

[0110] Table 5 shows the total mean wood density index (IML). R IML T IML TB A negative genetic correlation was found between IML and diameter at breast height (DBH), with a correlation coefficient ranging from -0.17 to -0.30. Specific results are shown in Table 5. Among these, the IML values ​​after trend elimination and bark correction... TB The genetic correlation coefficient was -0.30 (0.17), with a lower standard error and better estimation accuracy. The above results are consistent with the genetic correlation between the density and growth traits of Norway spruce wood in Example 1, further confirming that the negative genetic correlation between wood density and growth traits is a common genetic characteristic of coniferous species, providing a reliable genetic basis for breeding strategies to synergistically improve forest growth and wood quality.

[0111] Table 5 Genetic correlation coefficients between IML needle penetration test values ​​and wood density and growth traits (standard error is in parentheses).

[0112]

[0113] Example 3: Modeling and Generalization Validation of Combined Norwegian Spruce and Chinese Fir Datasets

[0114] 1. Data merging

[0115] The corrected resistance datasets IML from Example 1 (Norwegian spruce) and Example 2 (cedar) were used. TB The variables were merged into a unified independent variable; the corresponding baseline density (MWD / Den) was unified as the dependent variable, and after removing outliers, a mixed tree species generalization dataset was constructed and published on the Zenodo platform at https: / / doi.org / 10.5281 / zenodo.19453457.

[0116] 2. Construction of a general prediction model across tree species

[0117] After the merger, IML TB Using the baseline density as the dependent variable, fit a univariate linear regression model:

[0118] MWD = 0.1389 × IML TB +218

[0119] The coefficient of determination R² = 0.6048 was used to evaluate the model fit. Figure 3 The results show that there is a good linear correlation between the predicted density value and the actual density value. The model has high interpretability and stable and reliable prediction, and can be used for rapid and non-destructive estimation of the density of standing timber.

[0120] 3. Conclusion

[0121] This method first verifies the reliability of heritability and correlation in a single tree species, and then uses multi-species data merging to model and achieve generalized application across coniferous species such as Norway spruce and Chinese fir. It can quickly, non-destructively, and accurately determine the density of living trees and use it for genetic evaluation, which can significantly shorten the breeding cycle and improve selection efficiency.

Claims

1. A method for rapid, non-destructive determination, correction, and prediction of standing tree density, characterized in that, Includes the following steps: 1) Use a needle punch to measure the original resistance data of living trees; use an X-ray densitometer or volumetric method to measure the average true wood density (MWD) of the core. 2) The raw wood resistance data obtained in step 1) are all subjected to trend elimination processing, and the error caused by bark is corrected to obtain the corrected wood density dataset IML. TB ; 3) Using the IML dataset TB Using as the independent variable and the average true density of timber (MWD) as the dependent variable, a prediction model is constructed. The fitted univariate linear regression model is as follows: ； 4) The density of living trees was predicted using a univariate linear regression model to obtain the prediction results data.

2. The method for rapid, non-destructive determination, correction, and prediction of standing tree density according to claim 1, characterized in that, In step 1), the acupuncture device is an IML-RESI PD300.

3. The method for rapid, non-destructive determination, correction, and prediction of standing tree density according to claim 1, characterized in that, In step 1), the average wood density of the core is determined using an Itrax X-ray densitometer or a volumetric method.

4. The method for rapid and non-destructive determination, correction, and prediction of standing tree density according to claim 1, characterized in that, In step 2), the trend elimination process is as follows: Perform linear detrending on the original resistance curve, according to Calculate the linear ascending slope of a single curve, where... This is the reading at the end of the original resistance curve; This is the reading at the beginning of the original resistance curve; This represents the total length of the curve. Press again By correcting each measuring point on the curve point by point, the resistance value after detrending is obtained, where... Let i be the resistance value after trend reversal at point i; This represents the initial resistance value at point i. Let be the length from point i to the starting point of the curve.

5. The method for rapid and non-destructive determination, correction, and prediction of standing tree density according to claim 1, characterized in that, In step 2), the bark error correction process is a data algorithm process, specifically: taking the resistance curve after trend elimination as the object, the low resistance bark sections at both ends of the curve are identified by combining the double-sided threshold judgment with the cumulative counting method. First, the measurement points with resistance values ​​lower than the overall average value of the curve are marked as invalid points. The invalid bark sections on the left are removed by cumulative counting. Then, the complete and valid curve is sorted in reverse and the threshold judgment and cumulative counting are repeated to remove the invalid bark sections on the right. Subsequently, the resistance peak is located by the curve difference algorithm and invalid data outside the peak are removed. Finally, the valid data section corresponding to the pure wood in the middle is retained to complete the bark error correction.

6. A method for analyzing the phenotypic and genetic correlation of forest trees based on the method described in claim 1, characterized in that, Includes the following steps: 1) Use a needle-punch instrument to measure the original resistance data of living trees; use an X-ray densitometer or volumetric method to measure the average true wood density (MWD) of the core; measure growth traits and calculate the trunk volume to obtain phenotypic values; 2) The raw wood resistance data obtained in step 1) are all subjected to trend elimination processing, and the error caused by bark is corrected to obtain the corrected wood density dataset IML. TB ; 3) Extract forest tree DNA, perform genotyping, and fill in missing genotypes; construct a genome phylogenetic matrix G using LDAK software based on whole-genome single nucleotide polymorphism markers; 4) A spatial model was used to correct the phenotypic values. The spatial model used was: In the formula: For spatial autocorrelation residuals, denoted as independent random residuals; b is the fixed effects vector containing the population mean and block effects, and X is its corresponding design matrix; u is the random additive genetic effects vector, and Z is its corresponding design matrix; μ is the population mean of the trait. 7.5) Conduct phenotypic and genetic correlation analysis of trees.

8. The analytical method according to claim 6, characterized in that, In step 4), the first two principal components PC1 and PC2, derived from the genomic kinship matrix G, are incorporated into the model as fixed effects, and the fitted linear mixture model is as follows: ； In the formula, y is the vector of phenotypic observations; μ is the population mean; β1 and β2 are the fixed effects coefficients of PC1 and PC2, respectively; X is the correlation matrix corresponding to G; and e is the vector of random residual effects.

9. The analytical method according to claim 6, characterized in that, In step 5), the formula for calculating narrow-sense heritability is: ； In the formula, For additive inheritance variance, This represents the residual variance.

10. The analytical method according to claim 6, characterized in that, In step 5), the formula for calculating the genetic correlation coefficient is as follows: ； In the formula, Represents additive genetic covariance, characterizing the shared effect between genetic factors controlling variation in trait 1 and trait 2; and These are the additive inheritance variances of trait 1 and trait 2, respectively; The formula for calculating the phenotypic correlation coefficient based on variance components is as follows: ； In the formula, For residual covariance, and These are the residual variances corresponding to the two traits.

Images