A method for predicting the growth of diameter at breast height of Chinese fir based on near-natural transformation of Chinese fir and Phoebe fujianensis stands

Abstract

The present application provides a method for predicting the growth of breast diameter of Chinese fir based on a near-naturally transformed forest stand of Chinese fir. The present application introduces a model through the process of acquiring breast diameter data of Chinese fir under near-natural transformation of Chinese fir, and evaluates and calculates various variables that affect the growth of breast diameter. The nonlinear mixed effect model is used to construct a breast diameter growth model for Chinese fir, and the optimal breast diameter mixed effect growth model is constructed by considering the random effects of different parameters. This greatly improves the accuracy of the estimation of the breast diameter growth model, improves the accuracy and efficiency of the breast diameter prediction, and is of great significance to improving the quality and yield of Chinese fir plantations.

Description

Technical Field

[0001] The present application relates to the field of forestry technology, specifically to the field of forest stand growth and harvest statistical model technology, and more specifically to a method, system and device for predicting the breast height growth of Chinese fir based on a nearly naturally transformed forest stand of Chinese fir. Background Art

[0002] Cunninghamia lanceolata (Lamb.) Hook. is an important fast-growing artificial timber species unique to my country. Due to the long-term single-management of Cunninghamia lanceolata plantations, the ecological problems of Cunninghamia lanceolata plantations have become increasingly prominent, such as soil fertility degradation, decreased timber production, single structural function, frequent pests and diseases, depletion of soil carbon reserves, and changes in soil microbial communities. In order to overcome the above problems, the near-natural transformation of Cunninghamia lanceolata plantations has been rapidly developed in my country. Phoebe bournei is a tall evergreen tree of the Lauraceae family and a precious broad-leaved timber species unique to my country. It has excellent wood properties and relatively high economic value. Since Phoebe bournei is relatively shade-tolerant in its early growth, it is usually introduced into Cunninghamia lanceolata forests as a regeneration species in the near-natural transformation of Cunninghamia lanceolata plantations, so that the coniferous forests of Cunninghamia lanceolata are transformed into mixed coniferous and broad-leaved forests. At present, the near-natural transformation model of Phoebe bournei is the main technical model for the near-natural transformation of Cunninghamia lanceolata plantations in my country.

[0003] In forest management and afforestation decision-making, DBH (diameter at breast height) is a commonly used key single tree parameter. In forest surveys, DBH is easy to measure with a small error. DBH is closely correlated with tree height, timber volume, site, etc. By constructing a relationship model between DBH and forest age, forest stand, site quality and near-natural transformation mode, the growth change of DBH of Chinese fir is predicted, providing theoretical support for the long-term prediction of the growth of Chinese fir species under the near-natural transformation mode of Chinese fir and Phoebe fujianensis, which is currently commonly used in my country, and guiding the designation of reasonable near-natural transformation management strategies.

[0004] At present, there are two main methods to obtain DBH growth data: one is to use the forest stand survey data obtained from fixed forest plots to extract the DBH data of individual trees corresponding to each forest age stage; the other is to use the tree ring width data to convert it into DBH growth data, and then calculate the DBH growth of individual trees. However, due to the long time interval of forest stand survey data, inaccurate growth variables and low precision of the calculation process, the current DBH prediction has low accuracy and low prediction efficiency.

[0005] The above statements are only used to provide background technical information related to the present application. Unless otherwise stated herein, the contents described in this section are not prior art for the contents of other parts of the present application. Summary of the invention

[0006] The present invention proposes a method for predicting the growth of the breast diameter of Chinese fir based on a nearly naturally transformed forest stand of Chinese fir and Phoebe korea. The present application involves the process of obtaining the breast diameter data of Chinese fir under the nearly naturally transformed Chinese fir and Phoebe korea, introducing various variables that affect the breast diameter growth into a model after evaluation and calculation, and adopts a nonlinear mixed effect model to construct a breast diameter growth model for Chinese fir, and considers the random effects of different parameters to construct an optimal breast diameter mixed effect growth model, thereby greatly improving the accuracy of the breast diameter growth model estimation, and improving the accuracy and efficiency of the breast diameter prediction, which is of great significance to improving the quality and yield of Chinese fir plantations.

[0007] In addition, this application uses a single survey data combined with annual ring width data to establish a model, which can complete the construction of the breast diameter growth model of Chinese fir under the near-natural transformation mode of Chinese fir and Phoebe chinensis in the absence of consecutive years of re-measurement data, thereby providing a reference for the subsequent prediction of the yield of Chinese fir species under the near-natural transformation mode of Chinese fir and the screening of suitable near-natural transformation modes of Chinese fir and Phoebe chinensis.

[0008] According to a first aspect of an embodiment of the present application, a method for predicting the growth of diameter at breast height of Chinese fir is provided, comprising:

[0009] Obtain the breast diameter data of Chinese fir under near-natural transformation of Chinese fir and Phoebe fujianensis as the model data set;

[0010] Construct multiple candidate growth models for DBH of Chinese fir, and conduct model evaluation to determine the basic growth model for DBH;

[0011] Screen the multi-category independent variables of the model and determine the effective independent variables; add the effective independent variables to the basic model of DBH growth to build a DBH growth model based on multiple variables and dummy variables;

[0012] According to the DBH growth model, the optimal DBH mixed-effect growth model was constructed considering the random effects of different parameters. The DBH growth of Chinese fir was predicted using the DBH mixed-effect growth model.

[0013] In some embodiments of the present application, obtaining the breast diameter data of Cunninghamia lanceolata under near-natural transformation includes the following steps:

[0014] Based on the survey data of each tree, standard trees of various natural transformation sites of Cunninghamia lanceolata and Phoebe fujianensis were determined, and several trees with diameters closest to the standard trees were selected as tree samples;

[0015] Using Swedish growth cones Haglof Sweden, cores were drilled from the tree samples at breast height in the east-west and north-south directions, with a total of two core samples from each tree.

[0016] The tree core samples were dried and polished for dating analysis, and the tree ring width was accurately measured using the LinTAB 6 tree ring analyzer and TASP-Win software;

[0017] TSAP software was used to cross-date all samples.

[0018] In some embodiments of the present application, a plurality of candidate growth models for the diameter at breast height of Chinese fir are constructed, and model evaluation is performed to determine a basic growth model for the diameter at breast height, including the following steps:

[0019] Select multiple growth equations as candidate equations for the growth model of the diameter at breast height of Chinese fir, and construct multiple candidate growth models of the diameter at breast height of Chinese fir; fit the candidate growth models of the diameter at breast height of Chinese fir respectively and compare the model fitting effects;

[0020] Using the coefficient of determination R 2 The first evaluation formula consisting of three statistical indicators, root mean square error (RMSE) and mean absolute error (MAE), was used to evaluate the model and determine the basic model of breast diameter growth.

[0021] Among them, the specific calculation formula of the first evaluation formula is as follows:

[0022] ;

[0023] ;

[0024] .

[0025] Among them, n is the number of samples, is the measured value of the DBH of the jth tree in the i-th plot, is the estimated value of DBH of the jth tree in the i-th plot, is the average breast diameter.

[0026] In some embodiments of the present application, screening multiple categories of independent variables of the model and determining effective independent variables include the following steps:

[0027] The independent variables of the model are determined from four aspects: forest stand factor, competition factor, pattern factor and site factor;

[0028] During the modeling process, independent variables with weak correlation with DBH and insignificant parameters were removed, and independent variables with collinearity with other variables were screened to determine the effective independent variables.

[0029] The specific steps of screening are as follows: (1) Use the linear regression method to screen out insignificant factors, calculate the variance inflation factor (VIF) of each variable, and remove factors with a value greater than 5; (2) Use the correlation matrix to screen out factors with an absolute correlation value less than 0.1; (3) Finally, calculate the VIF value again, and perform a second screening to obtain the final selectable factors.

[0030] In some implementations of the present application, effective independent variables are added to the basic model of DBH growth, and a DBH growth model based on multiple variables and dummy variables is constructed, including:

[0031] The nonlinear mixed effect model is used to construct the DBH growth model of Chinese fir, and its general form is:

[0032] ;

[0033] Among them, D ij is the number of trees in the i-th plot t The diameter at breast height at the age;

[0034] Among them, f (β, u i , t) describes D ij The relationship between μ and t, that is, the function between DBH and forest age; i is the standard random effect parameter estimate; D is the random effect variance-covariance matrix; R ij is the variance-covariance matrix within the sample plot; μ i ~N(0, t) means that the random parameter has an expectation of 0 and a variance of D Normal distribution; E ij is the error term, with an expected value of 0 and a variance of R ij The normal distribution of E ij With μ i Independent of each other.

[0035] In some embodiments of the present application, according to the DBH growth model, an optimal DBH mixed effect growth model is constructed by considering the random effects of different parameters, including:

[0036] According to the DBH growth model of Chinese fir, the nlme function was used and the plots were used as random variables to construct multiple random effect models configured with multiple parameters. The optimal random effect model was determined based on the model evaluation indicators.

[0037] According to the optimal random effects model, the first-order autoregression method was used to eliminate the autocorrelation of the model, and the exponential function, power function and constant plus power function were used to eliminate the heteroscedasticity between data respectively; and the second evaluation formula was used to evaluate the goodness of fit of the model, and the model with the highest convergence and fitting accuracy was selected as the optimal model based on a certain variable; at the same time, the likelihood ratio test was performed on multiple random effects models, and the models with significant differences were selected; finally, the mixed effect growth model of the breast diameter of Chinese fir was determined.

[0038] Among them, the second evaluation formula includes evaluating the Akaike information criterion AIC value and the Bayesian information criterion BIC value. The smaller the value, the better. At the same time, the larger the log-likelihood value Loglik, the better.

[0039] According to a second aspect of an embodiment of the present application, a system for predicting the growth of diameter at breast height of Chinese fir is provided, comprising:

[0040] The data set module is used to obtain the breast diameter data of Cunninghamia lanceolata under near-natural transformation as a model data set;

[0041] The basic model module is used to construct multiple candidate growth models for the diameter at breast height of Chinese fir and to evaluate the models to determine the basic growth model for the diameter at breast height;

[0042] The mixed effect model module is used to screen the multi-category independent variables of the model and determine the effective independent variables; and add the effective independent variables to the basic model of DBH growth to build a DBH growth model based on multiple variables and dummy variables;

[0043] The random effect adaptation module is used to construct the optimal DBH mixed effect growth model based on the DBH growth model and consider the random effects of different parameters;

[0044] The prediction module is used to predict the growth of the diameter at breast height of Chinese fir by using the mixed-effect growth model of the diameter at breast height of Chinese fir.

[0045] According to a third aspect of an embodiment of the present application, a device for predicting the growth of a Chinese fir diameter at breast height is provided, comprising: a storage unit for storing executable instructions; and a processing unit for connecting to the storage to execute the executable instructions to complete a method for predicting the growth of a Chinese fir diameter at breast height.

[0046] According to a fourth aspect of the embodiments of the present application, a computer-readable storage medium is provided, on which a computer program is stored; the computer program is executed by a processor to implement a method for predicting the growth of diameter at breast height of Chinese fir.

[0047] The method, system and device for predicting the growth of diameter at breast height of Chinese fir of the present application include constructing multiple candidate growth models of diameter at breast height of Chinese fir, and evaluating the models to determine the basic model of diameter at breast height growth; screening the multi-category independent variables of the model to determine the effective independent variables; and adding the effective independent variables to the basic model of diameter at breast height growth to construct a growth model of diameter at breast height based on multiple variables and dummy variables; and constructing the optimal mixed-effect growth model of diameter at breast height based on the growth model of diameter at breast height by considering the random effects of different parameters. The present application introduces the model after evaluating and calculating the data of diameter at breast height of Chinese fir under the near-natural transformation of Chinese fir and each variable affecting the growth of diameter at breast height, and constructs the growth model of diameter at breast height of Chinese fir by using a nonlinear mixed-effect model, and constructs the optimal mixed-effect growth model of diameter at breast height by considering the random effects of different parameters, thereby greatly improving the accuracy of the estimation of the growth model of diameter at breast height, and improving the accuracy and efficiency of the prediction of diameter at breast height, which is of great significance to improving the quality and yield of Chinese fir plantations. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0049] Figure 1 FIG. 2 shows a schematic diagram of the steps of a method for predicting the growth of diameter at breast height of Chinese fir according to an embodiment of the present application;

[0050] Figure 2 FIG. 2 shows a schematic diagram of steps for obtaining the breast diameter data of Chinese fir according to an embodiment of the present application;

[0051] Figure 3 A schematic diagram of the steps of determining a basic model for breast diameter growth according to an embodiment of the present application is shown in FIG.

[0052] Figure 4 A schematic diagram of a technical route for determining a basic model for breast diameter growth according to an embodiment of the present application is shown in FIG.

[0053] Figure 5 A comparison diagram of the growth fitting curves of single tree diameter at breast height based on 8 basic models according to an embodiment of the present application is shown in FIG.

[0054] Figure 6 : shows a schematic diagram of the structure of the Chinese fir diameter at breast height growth prediction system according to an embodiment of the present application;

[0055] Figure 7 Schematic diagram of the structure of the Chinese fir breast diameter growth prediction device 400 according to an embodiment of the present application is shown in FIG. DETAILED DESCRIPTION

[0056] Regarding this application, it is considered that there are currently two main methods for obtaining breast diameter growth data: one is to use the forest stand survey data obtained from fixed forest plots to extract the breast diameter data of individual trees corresponding to each forest age stage; the other is to use the tree annual ring width data, convert it into breast diameter growth data, and then calculate the breast diameter growth of individual trees. However, due to the longer time interval of the forest stand survey data, the annual ring data is in units of one year, which describes the annual changes of the breast diameter of individual trees with the age of the forest, and can better detect the impact of various forest stand indicators on tree growth. In addition, this application also considers introducing the independent variables that affect the breast diameter growth into the model after calculation and analysis, so as to further improve the accuracy of the breast diameter growth model estimation.

[0057] To achieve the above object, the method for estimating the breast diameter growth model of Chinese fir based on a near-natural transformation forest stand of Chinese fir and improving the estimation accuracy of the breast diameter growth model provided by the present invention comprises the following steps: obtaining breast diameter data of Chinese fir under near-natural transformation of Chinese fir and Phoebe fujianensis as a model data set;

[0058] Construct multiple candidate growth models for DBH of Chinese fir, and conduct model evaluation to determine the basic growth model for DBH;

[0059] Screen the multi-category independent variables of the model and determine the effective independent variables; add the effective independent variables to the basic model of DBH growth to build a DBH growth model based on multiple variables and dummy variables;

[0060] According to the DBH growth model, the optimal DBH mixed-effect growth model was constructed considering the random effects of different parameters. The DBH growth of Chinese fir was predicted using the DBH mixed-effect growth model.

[0061] In addition, the present application also provides a system and device for predicting the growth of diameter at breast height of Chinese fir.

[0062] This application goes through the process of acquiring the breast diameter data of Chinese fir under the near-natural transformation of Chinese fir, introducing the model after evaluating and calculating various variables that affect the breast diameter growth, and constructing the breast diameter growth model of Chinese fir using a nonlinear mixed effects model, and constructing the optimal breast diameter mixed effects growth model considering the random effects of different parameters, which greatly improves the accuracy of the breast diameter growth model estimation, improves the accuracy and efficiency of the breast diameter prediction, and is of great significance to improving the quality and yield of Chinese fir plantations.

[0063] In order to make the technical solutions and advantages in the embodiments of the present application more clearly understood, the exemplary embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than an exhaustive list of all the embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0064] Example 1

[0065] Figure 1 Schematic diagram of the steps of the method for predicting the growth of diameter at breast height of Chinese fir according to an embodiment of the present application is shown in FIG.

[0066] like Figure 1 As shown, the method for predicting the growth of diameter at breast height of Chinese fir includes the following steps:

[0067] S1: Obtain the breast diameter data of Chinese fir under near-natural transformation of Chinese fir and Phoebe fujianensis as the model data set;

[0068] S2: Construct multiple candidate DBH growth models for Chinese fir, and conduct model evaluation to determine the basic DBH growth model;

[0069] S3: Screen the multi-category independent variables of the model and determine the effective independent variables; add the effective independent variables to the basic model of DBH growth to build a DBH growth model based on multi-variables and dummy variables;

[0070] S4: Based on the DBH growth model, the optimal DBH mixed-effect growth model is constructed by considering the random effects of different parameters;

[0071] S5: Finally, the DBH growth of Chinese fir was predicted by the mixed-effect growth model of DBH of Chinese fir.

[0072] Figure 2 A schematic diagram of the steps for obtaining the breast diameter data of Chinese fir according to an embodiment of the present application is shown in FIG.

[0073] like Figure 2 As shown, in S1, obtaining the breast diameter data of Chinese fir under near-natural transformation of Phoebe fujianensis includes the following steps:

[0074] S11: Determine the standard trees of each site of near-natural transformation of Cunninghamia lanceolata and Phoebe fujianensis according to the survey data of each tree, and select multiple trees with a diameter at breast height closest to the standard trees as tree samples;

[0075] S12: Using Swedish growth cones Haglof Sweden, cores were drilled from the tree samples at breast height in the east-west and north-south directions, with a total of two core samples from each tree;

[0076] S13: The tree core samples were dried and polished for dating analysis, and the tree ring width was accurately measured using the LinTAB 6 tree ring analyzer and TASP-Win software;

[0077] S14: Use TSAP software to cross-date all samples.

[0078] Specifically, when obtaining the breast diameter data of Chinese fir under the near-natural transformation of Chinese fir, in order to better understand the breast diameter growth dynamics of Chinese fir under the near-natural transformation of Chinese fir, this application calculates the standard trees of each sample plot of Chinese fir under the near-natural transformation of Chinese fir based on the survey data of each tree, and in each sample plot, select five trees whose breast diameter is closest to the standard tree.

[0079] Using Haglof Sweden growth cones, a core was drilled from each tree at breast height (1.3m) in the east-west and north-south directions, with a total of two cores from each tree. When sampling, the corer inside the growth cone should be removed first. After that, the cone is fixed on the cone handle. Make sure that the growth cone is perpendicular to the trunk and slowly tighten the cone handle in a clockwise manner until the cone is at least halfway into the trunk. At this point, use a coring spoon to remove the core, seal it in a plastic tube, and number it for subsequent transportation to the laboratory for analysis. The coring site is then sealed to prevent the sampled tree from dying.

[0080] After the tree core sample is taken, it is stored in a plastic tube and needs to be placed in a wooden groove of customized width in time and glued in the wooden groove with wood glue. When pasting the specimen, make sure that the dark side is facing up so that the tree's tracheal cells or tracheids are perpendicular to the surface of the wooden groove for subsequent microscopic observation.

[0081] The samples were then dried and polished. During the sample pretreatment stage, low-grit sandpaper was used for preliminary polishing, followed by finer high-grit sandpaper for detailed polishing until the core side was smooth and the annual ring boundaries were clearly visible.

[0082] The samples were then dated under a microscope, and the tree ring width was accurately measured using the LinTAB 6 tree ring analyzer (Rinntech, Germany, with an accuracy of 0.01 mm) and TASP-Win software. During the measurement process, the crosshairs of the microscope were always aligned precisely with the edge of the tree ring.

[0083] After completing the measurement of all sample cores, TSAP software was used to perform cross-dating on all samples.

[0084] In order to ensure the accuracy of the measurement and cross-dating results, the COFECHA program was further used for verification.

[0085] In addition, CDendro9.8.1 software was used to adjust and screen tree cores that differed greatly from the main sequence.

[0086] Figure 3 A schematic diagram of the steps for determining the basic model of breast diameter growth according to an embodiment of the present application is shown in FIG.

[0087] like Figure 3As shown, in S2, multiple candidate growth models of the diameter at breast height of Chinese fir are constructed, and model evaluation is performed to determine the basic growth model of the diameter at breast height, including the following steps:

[0088] S21: selecting multiple growth equations as candidate equations for the breast height growth model of Chinese fir, and constructing multiple candidate breast height growth models for Chinese fir; fitting the candidate breast height growth models for Chinese fir respectively and comparing the model fitting effects;

[0089] S22: Use the first evaluation formula consisting of three statistical indicators, namely, the coefficient of determination R2, the root mean square error RMSE and the mean absolute error MAE, to evaluate the model and determine the basic model of breast diameter growth.

[0090] In the specific implementation, first, combining the growth law of trees and the characteristics of the model, eight commonly used growth equations as shown in Table 1 are selected as alternative equations for the breast diameter growth model of Chinese fir, thereby forming eight alternative models as shown in Table 1: Bertalanffy, Chapman-Richards, Compertz, Korf, Logistic, Mitscherlich, Power and Weibull.

[0091] Table 1 Eight alternative DBH growth equations

[0092]

[0093] In Table 1, D For age t DBH (cm) at breast height; t is age, in years; d m is the limit of DBH (asymptotic DBH), where 100 cm is set as the approximate maximum DBH value of Chinese fir; a and b are the model parameters to be estimated.

[0094] Then, the alternative breast diameter growth models of Chinese fir were fitted respectively and the model fitting effects were compared, and the model with the best fitting effect was selected as the basic breast diameter growth model of Chinese fir.

[0095] The eight alternative models in Table 1 were fitted separately and the model fitting effects were compared. The determination coefficient R 2 The first evaluation formula consisting of three statistical indicators: root mean square error RMSE and mean absolute error MAE (Mean Absolute Error) is used for evaluation.

[0096] Where R 2The closer it is to 1, the smaller the RMSE and MAE values ​​are, indicating that the prediction effect of the selected model is more accurate. In this way, the most suitable basic model for the growth of the diameter at breast height of Chinese fir is obtained.

[0097] The specific calculation formula of the first evaluation formula is as follows:

[0098] ;

[0099] ;

[0100] .

[0101] Among them, n is the number of samples, is the measured value of the DBH of the jth tree in the i-th plot, is the estimated value of DBH of the jth tree in the i-th plot, is the average breast diameter.

[0102] As further described, step S3 screens the multi-category independent variables of the model and determines the effective independent variables, including:

[0103] Firstly, the independent variables of the model are determined to come from four aspects: forest stand factor, competition factor, pattern factor and site factor;

[0104] Then, the independent variables with weak correlation with DBH and insignificant parameters were filtered out during the modeling process, and the effective independent variables were determined after screening the independent variables that were collinear with other variables.

[0105] In order to reflect the impact of environmental factors and near-natural transformation models on the growth of the DBH of Chinese fir, this application mainly selects the independent variables of the model from the following four aspects when constructing the single tree DBH growth model:

[0106] (1) Stand factors: The average breast diameter of the stand Dq and the average height of the dominant trees Hd are commonly used stand information factors. In addition, there are the fir composition coefficient fir%, which describes the proportion of fir trees per hectare of cross-sectional area in the near-natural transformation of fir trees, and the proportion of fir trees per hectare in the near-natural transformation of fir trees, such as pFir.

[0107] (2) Competition factors: This study only included stand competition factors that were not related to distance, including basal area per hectare (BA), number of trees per hectare (N), and stand density index (SDI). In addition, the number of trees per hectare (N) of Chinese fir was also considered. fir And the number of Phoebe bournei plants per hectare is Nn.

[0108] (3) Pattern factor: To reflect the near-natural transformation pattern, the near-natural transformation pattern is used as a dummy variable factor, and Z1, Z2, and Z3 are set to 1, 0, 0 / 0, 1, 0 / 0, 0, 1 / 0, 0, 0 respectively.

[0109] (4) Site factor: The site index SI is introduced here. The site index is obtained by taking the average height of the dominant trees in the sample plot at the base age (20 years for Chinese fir) and using the site index table. Since site conditions have a great influence on the growth of trees, it is considered to include the site index in the fixed effect or introduce it into the model as a dummy variable.

[0110] During the modeling process, it is necessary to remove independent variables that have a weak correlation with DBH, insignificant parameters, and collinearity with other variables (VIF < 5).

[0111] The specific steps of screening are as follows: (1) Use the linear regression method to screen insignificant factors, calculate the variance inflation factor (VIF) of each variable, and remove factors with a value greater than 5; (2) Use the correlation matrix to screen out factors with an absolute correlation value less than 0.1; (3) Finally, calculate the VIF value again and screen the final selectable factors again. (4) At the same time, in order to facilitate subsequent actual production applications, select 1 to 3 factors, and try to be different categories of factors mentioned above.

[0112] The cor() function of the nlme package was then used to calculate the correlation between each variable and DBH and forest age, and the summary() function was used to screen variables with P < 0.05 and strong correlation for subsequent modeling.

[0113] The effective variables obtained by screening were added to the optimal DBH growth basic model obtained by S2 to construct a mixed-effect growth model of DBH of Chinese fir containing multiple variables.

[0114] When constructing a nonlinear mixed effects model, it is important to note that the regression function is not only related to the fixed effects, but also includes the nonlinear relationship with the random effects.

[0115] In the following implementation, a nonlinear mixed effect model is used in S3 to construct a breast diameter growth model for Chinese fir, and its general form is:

[0116] ;

[0117] Among them, D ij is the number of trees in the i-th plot t The diameter at breast height at the age;

[0118] Among them, f (β, u i , t) describes D ij The relationship between μ and t, that is, the function between DBH and forest age; i

[0119] is the standard random effect parameter estimate; D is the random effect variance-covariance matrix; R ij is the variance-covariance matrix within the plot; μi ~N(0, t) means that the random parameter has an expectation of 0 and a variance of D Normal distribution; E ij is the error term, with an expected value of 0 and a variance of R ij The normal distribution of E ij With μ i Independent of each other.

[0120] Then, S4 was implemented to construct the optimal DBH mixed effect growth model based on the DBH growth model and considering the random effects of different parameters. It includes:

[0121] 1) Based on the DBH growth model of Chinese fir, the nlme function was used and the plots were used as random variables to construct multiple random effect models configured with multiple parameters; and the optimal random effect model was determined based on the model evaluation indicators;

[0122] 2) According to the optimal random effects model, the first-order autoregression method was used to eliminate the autocorrelation of the model, and the exponential function, power function and constant plus power function were used to eliminate the heteroscedasticity between the data respectively; the second evaluation formula was used to evaluate the goodness of fit of the model, and the model with the highest convergence and fitting accuracy was selected as the optimal model based on a certain variable; at the same time, the likelihood ratio test was performed on multiple random effects models, and the models with significant differences were selected; finally, the optimal DBH mixed effect growth model was determined.

[0123] Among them, the second evaluation formula includes evaluating the Akaike information criterion AIC value and the Bayesian information criterion BIC value. The smaller the value, the better. At the same time, the larger the log-likelihood value Loglik, the better.

[0124] Among them, the formulas for AIC and BIC are as follows:

[0125] (10)

[0126] (11)

[0127] Where p is the number of model parameters, n is the number of sample plants, and logn is the logarithm of n.

[0128] Where loglik is the log-likelihood value.

[0129] In specific implementation, the following three steps are considered when constructing a mixed effects model:

[0130] (1) Random effect parameters: In the model setting, the sample plots are regarded as random effects, and random effect models with various parameter configurations are tried. The most appropriate model is determined by comparing the evaluation criteria of the models.

[0131] (2) Error variance-covariance structure (R): The following formula is used to eliminate the autocorrelation and heteroscedasticity between data. Preliminary analysis shows that the model residuals have heteroscedasticity, which is systematically related to tree age. Therefore, three different functions (exponential function VarExp (E ij ), power function VarPower (E ij ) and the constant plus power function VarConstPower (E ij ))Eliminate the heteroskedasticity among data.

[0132] The formula is as follows:

[0133] ;

[0134] ;

[0135] ;

[0136] ;

[0137] Where: represents the model error variance; is a diagonal matrix used to capture the heteroskedasticity of the random effects; is a variance matrix describing the autocorrelations of the random effects. is the predicted value of DBH of the j-th plant in the ith plot based on the fixed effect parameters; λ, λ1, and λ2 are the parameters to be estimated.

[0138] In addition, to account for the autocorrelation in repeated measurement data, a nonlinear mixed effects model (nlme) was used, which included the first-order autoregressive correlation results to describe the correlation structure within the group. The fitting method of the mixed effects model used restricted maximum likelihood estimation.

[0139] (3) Random effects variance-covariance structure (D): The variance and covariance of the random effects are modeled as generalized positive definite matrices.

[0140] Akaike information criterion (AIC) and Bayesian information criterion (BIC) were used as measurement criteria to select the best performing mixed effects model by comparing the fit and convergence effects of various model combinations and their corresponding parameter settings.

[0141] According to the number of random effects included, the selected variance-covariance structure can be divided into four types: 1×1, 2×2, 3×3 and 4×4. The variance-covariance matrix corresponding to two random parameters can be expressed as:

[0142] ;

[0143] Where: μ1 and μ3 are random parameters; 、 are the variances of μ1 and μ3 respectively; is the covariance of μ1 and μ3.

[0144] In summary, in order to evaluate the growth changes of the breast diameter of Chinese fir under different near-natural transformation modes, this application constructed four types of models: (1) the basic model BDG including only DBH~forest age (step S2); (2) the breast diameter growth model GDG of Chinese fir containing forest stands and competition factors (step S3); (3) the breast diameter growth model DGDG of Chinese fir with the near-natural transformation mode Mode as a dummy variable (step S3); (4) the nonlinear mixed effect growth model type RDGDG of the breast diameter of Chinese fir based on the random effect of sample plots (step S4).

[0145] The nonlinear mixed effect model uses AIC, BIC and Loglik to evaluate the goodness of fit of the model, and the model with the highest convergence and fitting accuracy is selected as the optimal model based on a certain variable.

[0146] The obtained nonlinear mixed effect growth model of Chinese fir was subjected to likelihood ratio test (LRT), and the model with significant difference was selected ( P < 0.05) to avoid the problem of over-parameterization of the model.

[0147] Specifically, it is judged by comparing the AIC (Akaike Information Criterion) and BIC (Bayesian Information Criterion) values ​​of the model. The smaller the AIC and BIC values ​​of these two indicators, the better. At the same time, Loglik (log likelihood value) should also be considered. The larger the Loglik value, the better.

[0148] Among them, the formulas for AIC and BIC are as follows:

[0149] ;

[0150] ;

[0151] In the formula, p is the number of model parameters, n is the number of sample plants.

[0152] Compared with the prior art, the beneficial effect of the present application is that, on the basis of high time series accuracy, the present invention proposes a modeling method based on the near-natural transformation mode of Cunninghamia lanceolata commonly used in my country, which can not only improve the estimation accuracy of the growth model of the diameter at breast height of Cunninghamia lanceolata, but also comprehensively consider the forest stand, site, competition index and near-natural transformation mode. This method is of great significance for accurately predicting the changes in the growth of the diameter at breast height of Cunninghamia lanceolata and the growth of wood species under the forest stand of the near-natural transformation mode of Cunninghamia lanceolata, selecting the optimal near-natural transformation mode of Cunninghamia lanceolata, and accurately improving the quality and yield of Cunninghamia lanceolata plantations.

[0153] In summary, the method for predicting the growth of diameter at breast height of Chinese fir in the present application includes constructing multiple alternative growth models of diameter at breast height of Chinese fir, and performing model evaluation to determine the basic model of diameter at breast height growth; screening the multi-category independent variables of the model to determine the effective independent variables; and adding the effective independent variables to the basic model of diameter at breast height growth to construct a diameter at breast height growth model based on multiple variables and dummy variables; according to the diameter at breast height growth model, considering the random effects of different parameters to construct the optimal mixed-effect growth model of diameter at breast height.

[0154] This application goes through the process of acquiring the breast diameter data of Chinese fir under the near-natural transformation of Chinese fir, introducing the model after evaluating and calculating various variables that affect the breast diameter growth, and constructing the breast diameter growth model of Chinese fir using a nonlinear mixed effects model, and constructing the optimal breast diameter mixed effects growth model considering the random effects of different parameters, which greatly improves the accuracy of the breast diameter growth model estimation, improves the accuracy and efficiency of the breast diameter prediction, and is of great significance to improving the quality and yield of Chinese fir plantations.

[0155] In addition, this application uses a single survey data combined with annual ring width data to establish a model, which can complete the construction of the breast diameter growth model of Chinese fir under the near-natural transformation mode of Chinese fir and Phoebe chinensis in the absence of consecutive years of re-measurement data, thereby providing a reference for the subsequent prediction of the yield of Chinese fir species under the near-natural transformation mode of Chinese fir and the screening of suitable near-natural transformation modes of Chinese fir and Phoebe chinensis.

[0156] Example 2

[0157] This embodiment further illustrates Embodiment 1 in conjunction with a specific implementation scenario.

[0158] The research object of this embodiment is a Chinese fir plantation after near-natural transformation of Chinese fir Phoebe fujianensis in a certain forest farm.

[0159] The experimental site was planted with Chinese fir seedlings in 2004, and thinning was carried out once in 2015. In the same year, Phoebe bournei was replanted under the forest. By the end of 2022, a total of 30 plots of 20×30 m2 (0.06 hm2) were set up, including 27 near-natural transformation plots and 3 pure forest control plots. The density of Chinese fir retention and the density of Phoebe bournei planted under each model are shown in Table 2.

[0160] Table 2 The near-natural transformation model of Chinese fir and Phoebe chinensis in a certain forest farm

[0161]

[0162] The embodiment of the present application provides a method for estimating a breast height growth model of Chinese fir based on a mixed effects model, which consists of three parts: breast height growth model construction, nonlinear mixed effects model construction and random effects adaptation.

[0163] Figure 4 A schematic diagram of the technical route for determining the basic model of breast diameter growth according to an embodiment of the present application is shown in FIG.

[0164] like Figure 4 The specific technical roadmap is as follows:

[0165] 1. Obtain survey data on near-natural transformation of Chinese fir and Phoebe chinensis stands.

[0166] First, obtain the wood core data of Chinese fir. In 2022, measure each tree in 30 near-natural transformation plots in the forest farm, and measure the diameter at breast height (1.3m, Diameter at BrDBH), tree height, crown width (crown radius in the four directions of east, west, south and north), height below live branches (length from the base of the trunk to the first normal live branch), etc. of each Chinese fir tree.

[0167] According to the average diameter of the fir trees in each plot, the cores of the five average fir trees in each plot were sampled in two directions, with a total of ten core strips in each plot. The cores were taken out with a core spoon and sealed in a plastic tube, and numbered for subsequent transportation to the laboratory for analysis. The core sampling site was then sealed to prevent the sampled trees from dying. The sampling work was completed in February 2023, and a total of 318 core samples from 159 trees were collected.

[0168] The historical DBH growth data were calculated based on 318 sets of tree ring data. The specific methods are as follows: (1) The east-west, south-north, and north-south tree ring data were averaged according to the growth year to obtain the annual average growth value D corresponding to each year. 平 ; (2) Add the average annual growth rate D year by year in the form of D~A 平 , and the annual growth rate D 连 ; (3) Based on the measured DBH value D obtained from the survey of each tree 实 , check the total growth D of each tree calculated in the last year (2022) 总 ; (4) Measurement value D caused by uncertainties such as bark removal and measurement error 总 <D 实 Therefore, the study used D 实 / D 总 Find the difference ratio p比 (p 比 ≥1), and D 连 ×P 比 The final annual DBH growth D f .

[0169] The specific calculation formula is as follows:

[0170] ;

[0171] ;

[0172] ;

[0173] .

[0174] Where: D 练ijn represents the growth of the jth tree in the i-th plot when the age is n; D 平ijn is the jth tree in the ith plot at age n The average annual growth rate measured at fijn is the jth tree in the ith plot at age n The annual growth of DBH obtained when the tree was 2022; n is 1, 2, 3, …, A; A is the final age of the tree core after dating and measurement (2022 is the last year of growth); D 总ij is the total growth of the jth tree in the ith plot calculated from the tree ring measurements; p 比ij is the difference ratio obtained for the jth tree in the ith plot. 比ij Except for , all other units are in millimeters (mm).

[0175] The annual growth of DBH (DBH~A, in cm) was calculated according to the above method. All data were used to select the basic model, with a total of 3869 pairs of data.

[0176] In addition, in order to explore the impact of the near-natural transformation mode of Cunninghamia lanceolata on the DBH growth of Cunninghamia lanceolata, the DBH growth data of Cunninghamia lanceolata in all years after the transformation, that is, 7 years of DBH vs t data from 2015 to 2022, a total of 1153 pairs, were used to construct a nonlinear mixed-effect DBH growth model for Cunninghamia lanceolata. 2 / 3 DBH & t data pairs were randomly selected to form the modeling data set (769 pairs), and the remaining data pairs constituted the test data set (384 pairs).

[0177] 2. Calculate forest stand indicators.

[0178] According to the survey data of each tree, the total density of each plot (N, plant / hm2) was calculated. 2 ; including the number of Chinese fir and Phoebe bournei), basal area per hectare (BA, m2 / hm 2 ), average breast diameter of the stand (Dq, cm), average height of dominant trees (Hd, the average of the top 6 trees in the sample plot), stand density index (SDI) and other stand factors.

[0179] The average breast diameter of a stand is the diameter corresponding to the average cross-sectional area of ​​the stand, and is a basic indicator of tree thickness. The average height of dominant trees is often used to identify site quality and compare stand growth under different site qualities.

[0180] In order to reflect the impact of different site qualities on the growth of breast diameter of Chinese fir, this study introduced the site index (SI) as an influencing variable. According to the benchmark age of Chinese fir t0=20, the average height of the six tallest Chinese fir trees was calculated as the average height of the dominant trees, namely the site index.

[0181] The stand density index (SDI) is a stand density index proposed by Reineke based on the perfect standing degree. Its advantage is that it can not only reflect the number of trees in the forest, but also the size of the trees, and is not affected by the forest age and site conditions. Its mathematical expression is:

[0182] ;

[0183] Where: N is the number of plants per hectare, Dq is the average breast diameter of the stand, D0 is the standard diameter of the stand, and the value of Chinese fir is 20 cm; b is the Reineke self-thinning coefficient.

[0184] In order to reflect the near-natural transformation model of Cunninghamia lanceolata and the tree species diversity in the sample plots, the number of Cunninghamia lanceolata plants per hectare (N fir ) and the number of Phoebe bournei per hectare (N n ); the tree species composition coefficients include the fir composition coefficient (fir%) and the Phoebe bournei composition coefficient (nm%). The tree species composition coefficient reflects the proportion of a specific tree species in the total volume or total cross-sectional area of ​​the forest. Since the cross-sectional area at breast height is easier to measure than the volume, the cross-sectional area is often used to determine the tree species composition in practice.

[0185] The present invention uses cross-sectional area to calculate the composition coefficient of Chinese fir and Phoebe bournei:

[0186] ;

[0187] In the formula: Species represents the type of tree species, BA species Indicates the estimated basal area per hectare (m 2 / hm 2 ), BA is based on the stand basal area per hectare obtained from the sample plot (m 2 / hm 2 ).

[0188] In order to reflect the different near-natural transformation modes of Cunninghamia lanceolata, it is proposed to introduce the near-natural transformation mode Mode as a dummy variable to analyze whether there are statistical differences in the tree height curves of different transformation modes. Since there are three near-natural modes of interplanting Cunninghamia lanceolata under the forest farm (see Table 2), namely M1, M2 and M3, and a control mode CK. Therefore, the dummy variables (1) Z1=1, M1, and the rest are 0; (2) Z2=1, M2, and the rest are 0; (3) Z3=1, M3, and the rest are 0; (4) Z1=Z2=Z3=0, which is the CK control plot.

[0189] 3. Screening of the basic model for the growth of breast height diameter of Chinese fir.

[0190] The eight basic models proposed in Table 1 were used to construct the DBH growth models of Chinese fir, and the determination coefficient R 2 , root mean square error RMSE and mean absolute error MAE (Mean Absolute Error), and the statistical indicator results in Table 3 and Figure 5 Comparison of the fitting curves of single tree breast diameter growth based on 8 basic models.

[0191] Table 3 Parameter estimation and evaluation indexes of alternative basic models for tree DBH growth (all DBH ~ t data)

[0192]

[0193] From Table 3 and Figure 5 , it can be seen that all the models have good fitting effects, R 2 The evaluation indicators of all the alternative DBH growth models varied within a small range. The best performing one was the Korf model, with R 2 The value of Korf is 0.8321, the RMSE is 2.5762, and the MAE is 2.1215. Therefore, Korf is selected as the basic model BDG for the DBH growth of Chinese fir.

[0194] 4. Construct a breast diameter growth model for Chinese fir based on multivariate and dummy variables.

[0195] In order to more accurately evaluate the impact of the near-natural transformation model of Cunninghamia lanceolata on the growth of DBH of retained Cunninghamia lanceolata trees, this example only used the DBH growth data after the transformation (i.e., after 2015).

[0196] According to the variance inflation factor VIF and correlation analysis results mentioned above, four types of variable factors were considered at the same time: stand factor, competition factor, pattern factor and site factor. Finally, the site index SI, the number of plants per hectare N, the fir composition coefficient fir% and the pattern Mode were selected as variables to construct the fixed effect model GDG and the dummy variable model DGDG.

[0197] Because the values ​​of the mode and site index are integers and vary within a small range (SI is 10-20, with a unit of 2; M is set to 1, 2, 3, and 4, representing mode 1, mode 2, mode 3, and control, respectively), they can be considered as dummy variables and added to the model.

[0198] Specifically, the site index has a significant positive correlation with both DBH and age ( P < 0.05), the number of plants per hectare was significantly positively correlated with age ( P < 0.05), and the composition coefficient of Chinese fir was also significantly negatively correlated with both DBH and age ( P < 0.05), the pattern was significantly negatively correlated with DBH and age ( P < 0.05).

[0199] Therefore, taking Korf as the basic model (BDG), SI, N, fir%, and Mode were introduced to construct a dummy variable model. After multiple screenings, two groups of considerations were considered: (1) N and fir% were selected as fixed effects, and SI and mode were selected as dummy variables; (2) N, fir% and SI were added as fixed variables, and mode was added as a dummy variable.

[0200] And by comparing the model fitting abilities of different combinations, the optimal dummy variable model is screened.

[0201] Based on consideration (1), since SI has 6 values, namely 10, 12, 14, 16, 18, and 20, according to the definition of dummy variables, the dummy variables are set to A. When SI is 10, set S1=1, and the rest are 0; B. When SI is 12, set S2=1, and the rest are 0; C. When SI is 14, set S3=1, and the rest are 0; D. When SI is 16, set S4=1, and the rest are 0; E. When SI is 18, set S5=1, and the rest are 0; F. When SI is 20, set S1=S2=S3=S4=S5=0.

[0202] Therefore, the DGDG model based on two groups of considerations is in the form of DGDG1 and DGDG2, and its specific form is:

[0203] DGDG1:

[0204] ;

[0205] DGDG2:

[0206] ;

[0207] Where: DBH ij For the i The DBH value of the jth tree in a plot at age t (cm); N i is the number of trees per hectare in the ith plot (trees / hm2); fir% i is the fir composition coefficient of the i-th plot; SI i is the site index of the ith plot; ɑ0, ɑ1, ɑ2, ɑ3, b, c1, c2, c3, d1, d2, d3, d4, d5 are the parameters to be estimated for the model; D m is the limit of the diameter at breast height (asymptotic diameter at breast height).

[0208] As shown in Table 4 below, the evaluation indicators of the DGDG1 and DGDG2 models obtained based on the fitting data are given.

[0209] Table 4 Evaluation indexes of DBH growth model of Chinese fir based on dummy variables

[0210]

[0211] The results showed that after introducing fixed effects and dummy variables, the fitting effect of the DBH growth model of Chinese fir was quite considerable, with AIC and BIC values ​​ranging from 3005 to 3070. There was little difference in the fitting results of the two models, so both models were selected as alternative models for constructing the mixed effect DBH growth model.

[0212] 5. The breast diameter growth model of Chinese fir under the near-natural transformation of Chinese fir and Phoebe fujianensis based on nonlinear mixed effects.

[0213] DGDG1 and DGDG2 were used separately, and the nlme function was used to construct a mixed effect model with plot as a random variable, and the autocorrelation and heteroskedasticity of the data were taken into account.

[0214] Considering the random effect combination of different parameters: The model constructed based on DGDG1 has There are 9 combinations, 9 of which converge; the models built based on DGDG2 have There are 11 combinations of which converge.

[0215] The best performing model was selected based on the model evaluation index. The fitting results of the mixed effect model RDGDG1 constructed based on DGDG1 are shown in Table 5. It can be seen that the fitting effects of random effects added to different parameter positions are very small, with AIC values ​​ranging from 2540.4 to 2573.4, BIC values ​​ranging from 2605.4 to 2661.6, and R 2 The fluctuation range of RMSE values ​​is smaller, only 0.003 and 0.01; the model with random effect added to b (model RDGDG1r4) has the best fitting effect, R 2 It increased to 0.896, and the AIC and BIC values ​​were the lowest among the 9 mixed effect models.

[0216] The respective optimal hybrid models are RDGDG1r4 and RDGDG2r1.5, as shown below:

[0217] RDGDG1r4:

[0218] ;

[0219] RDGDG2r1.5:

[0220] ;

[0221] Where β0 is the random effect parameter, and other symbols have the same meanings as above.

[0222] Table 5 Comparison of model fitting results of RDGDG1 based on each variable considering different random effect parameters

[0223]

[0224] The fitting results of the mixed effect model RDGDG2 constructed based on DGDG2 are shown in Table 6. The results are similar to those of the RDGDG1 model. The fitting effects of the models are very different. The AIC values ​​are between 2551.4 and 2562.8, the BIC values ​​are between 2597.9 and 2632.5, and the R 2 The values ​​are both 0.893; the fluctuation range of RMSE and MAE values ​​is only 0.002 and 0.003. The results are the same as those of RDGDG1, and the best fitting effect is obtained when the random effect b is added (model RDGDG2r1.5). 2 It increased from 0.747 to 0.893, and the AIC and BIC values ​​also decreased from 3070.8 and 3112.6 to 2551.4 and 2597.9, respectively.

[0225] Table 6 Comparison of model fitting results of RDGDG2 based on each variable considering different random effect parameters

[0226]

[0227] In summary: Based on models RDGDG1-4 and RDGDG2-1.5, the first-order autoregression method is used to eliminate the autocorrelation of the model, and the exponential function, power function and constant plus power function are used to eliminate the heteroskedasticity between data. The fitting results are shown in Table 7.

[0228] Table 7 Comparison of model fitting results based on each model considering autocorrelation and heteroskedasticity structure

[0229]

[0230] It can be seen that the model RDGDG1r4 built with the power function as the heteroskedastic structure has the best effect (AIC and BIC are the smallest, Loglik is the largest); the LRT test shows that there is a significant difference between using the power function as the heteroskedastic structure and not considering the heteroskedastic structure ( P < 0.0001), the difference was not significant when the constant plus power function was used as the heteroscedastic structure ( P = 0.9893); Finally, the model RDGDG1r4_ah1 with the sample plot as the random effect factor, the random effect added to the parameter b, the first-order autoregression to eliminate autocorrelation, and the power function as the heteroskedastic structure was selected as the optimal mixed effect model.

[0231] The fitting results of model RDGDG2r1.5 are similar. Considering the first-order autoregression and using the power function as the heteroskedastic structure, the model has the best effect. The AIC is reduced from the original 2551.4 to 1913.8, and the BIC is reduced from 2597.9 to 1969.5. The optimal mixed effect model is RDGDG2r1.5_ah1.

[0232] Comparing the evaluation indicators of BDG, DGDG, and RDGDG models, the results show that the mixed models with fixed effects, dummy variables, and random effects are better than the basic models, with smaller AIC and R 2 and Loglik are larger than those of the basic model, indicating that the introduction of random effects improves the fitting accuracy of the model. However, by comparing the fitting indices of the mixed effect models, it can be found that the fitting effect of random effects on one parameter and multiple parameters is very small; and no matter which group of models is selected, the random effect on the parameters b The fitting effects are the best in comparison. Finally, the optimal nonlinear mixed effect model RDGDG2r1.5_ah1 is obtained through comprehensive comparison of model tests.

[0233] 6. Verification of the nonlinear mixed-effect growth model of the breast diameter of Chinese fir under the near-natural transformation model of Chinese fir and Phoebe burgdorferi.

[0234] 384 pairs of test data sets were used to test the prediction effect of the DBH growth model of Chinese fir and evaluate the prediction accuracy of the model. The results are shown in Table 8.

[0235] The smaller the TRE and MAE values ​​are, the higher the prediction accuracy is. The model RDGDG2-1.5 can accurately predict the growth of breast diameter of Chinese fir under the near-natural transformation mode of Chinese fir and Phoebe fujian.

[0236] Table 8 Summary of fitting and test indicators of DBH growth model of Chinese fir

[0237]

[0238] From the above specific embodiments, it can be seen that the present application uses a single survey data combined with the annual ring width data to establish a model, which can complete the construction of the breast diameter growth model of Chinese fir under the near-natural transformation mode of Chinese fir and Phoebe chinensis in the absence of consecutive years of re-measurement data, thereby providing a reference basis for the subsequent prediction of the yield of Chinese fir species under the near-natural transformation mode of Chinese fir and the screening of suitable near-natural transformation modes of Chinese fir and Phoebe chinensis.

[0239] Example 3

[0240] This embodiment provides a system for predicting the growth of diameter at breast height of Chinese fir. For details not disclosed in the system for predicting the growth of diameter at breast height of Chinese fir in this embodiment, please refer to the specific implementation contents of the method for predicting the growth of diameter at breast height of Chinese fir in other embodiments.

[0241] Figure 6 Schematic diagram of the structure of the Chinese fir breast diameter growth prediction system according to an embodiment of the present application is shown in FIG.

[0242] like Figure 6 As shown, the Chinese fir DBH growth prediction system includes a data set module 10 , a basic model module 20 , a mixed effect model module 30 , a random effect adaptation module 40 and a DBH prediction module 50 .

[0243] Specifically described,

[0244] The data set module 10 is used to obtain the breast diameter data of Cunninghamia lanceolata under near-natural transformation as a model data set;

[0245] The basic model module 20 is used to construct a plurality of candidate growth models for the diameter at breast height of Chinese fir, and to perform model evaluation to determine the basic growth model for the diameter at breast height;

[0246] Specifically, the following steps are included:

[0247] Multiple growth equations were selected as alternative equations for the breast height growth model of Chinese fir, and multiple alternative breast height growth models of Chinese fir were constructed; the alternative breast height growth models of Chinese fir were fitted separately and the model fitting effects were compared; the first evaluation formula consisting of three statistical indicators, namely the determination coefficient R2, root mean square error RMSE and mean absolute error MAE, was used to evaluate the model and determine the basic breast height growth model.

[0248] The specific calculation formula of the first evaluation formula is as follows:

[0249] ;

[0250] ;

[0251] ;

[0252] Among them, n is the number of samples, is the measured value of the DBH of the jth tree in the i-th plot, is the estimated value of DBH of the jth tree in the i-th plot, is the average breast diameter.

[0253] The mixed effect model module 30 is used to screen the multi-category independent variables of the model and determine the effective independent variables; and add the effective independent variables to the basic model of DBH growth to construct a DBH growth model based on multiple variables and dummy variables;

[0254] The random effect adaptation module 40 is used to construct an optimal DBH mixed effect growth model based on the DBH growth model and considering the random effects of different parameters;

[0255] The DBH prediction module 50 is used to predict the DBH growth of Chinese fir by using the DBH mixed effect growth model of Chinese fir.

[0256] In summary, the Chinese fir breast diameter growth prediction system of the present application is adopted, and multiple alternative Chinese fir breast diameter growth models are constructed through the basic model module 20, and model evaluation is performed to determine the breast diameter growth basic model; the mixed effect model module 30 screens the multi-category independent variables of the model to determine the effective independent variables; and the effective independent variables are added to the breast diameter growth basic model to construct a breast diameter growth model based on multiple variables and dummy variables; the random effect adaptation module 40 constructs the optimal breast diameter mixed effect growth model according to the breast diameter growth model, considering the random effects of different parameters.

[0257] This application goes through the process of acquiring the breast diameter data of Chinese fir under the near-natural transformation of Chinese fir, introducing the model after evaluating and calculating various variables that affect the breast diameter growth, and constructing the breast diameter growth model of Chinese fir using a nonlinear mixed effects model, and constructing the optimal breast diameter mixed effects growth model considering the random effects of different parameters, which greatly improves the accuracy of the breast diameter growth model estimation, improves the accuracy and efficiency of the breast diameter prediction, and is of great significance to improving the quality and yield of Chinese fir plantations.

[0258] In addition, this application uses a single survey data combined with annual ring width data to establish a model, which can complete the construction of the breast diameter growth model of Chinese fir under the near-natural transformation mode of Chinese fir and Phoebe chinensis in the absence of consecutive years of re-measurement data, thereby providing a reference for the subsequent prediction of the yield of Chinese fir species under the near-natural transformation mode of Chinese fir and the screening of suitable near-natural transformation modes of Chinese fir and Phoebe chinensis.

[0259] Example 4

[0260] This embodiment provides a device for predicting the growth of Chinese fir's diameter at breast height. For details not disclosed in the device for predicting the growth of Chinese fir's diameter at breast height in this embodiment, please refer to the specific implementation contents of the method or system for predicting the growth of Chinese fir's diameter at breast height in other embodiments.

[0261] Figure 7 Schematic diagram of the structure of the Chinese fir breast diameter growth prediction device 400 according to an embodiment of the present application is shown in FIG.

[0262] like Figure 7 As shown, the Chinese fir diameter at breast height growth prediction device 400 includes: a storage unit 402: used to store executable instructions; and a processing unit 401: used to connect to the storage unit 402 to execute the executable instructions to complete the Chinese fir diameter at breast height growth prediction method or image recognition method.

[0263] Those skilled in the art will understand that Figure 7 It is only an example of the growth prediction device 400 for the breast height of Chinese fir and does not constitute a limitation of the growth prediction device 400 for the breast height of Chinese fir. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the growth prediction device 400 for the breast height of Chinese fir may also include input and output devices, network access devices, buses, etc.

[0264] The processing unit 401 (Central Processing Unit, CPU) may also be other general processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general processor may be a microprocessor or the processing unit 401 may also be any conventional processor, etc. The processing unit 401 is the control center of the Chinese fir diameter at breast height growth prediction device 400, and uses various interfaces and lines to connect various parts of the entire Chinese fir diameter at breast height growth prediction device 400.

[0265] The storage unit 402 can be used to store computer-readable instructions. The processing unit 401 realizes various functions of the Chinese fir diameter at breast height growth prediction device 400 by running or executing the computer-readable instructions or modules stored in the storage unit 402 and calling the data stored in the storage unit 402. The storage unit 402 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the Chinese fir diameter at breast height growth prediction device 400, etc. In addition, the storage unit 402 can include a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM) or other non-volatile / volatile storage devices.

[0266] If the modules integrated in the device 400 for predicting the growth of Chinese fir diameter at breast height are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also instruct the relevant hardware to complete through computer-readable instructions. The computer-readable instructions can be stored in a computer-readable storage medium, and when the computer-readable instructions are executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented.

[0267] Example 5

[0268] This embodiment provides a computer-readable storage medium on which a computer program is stored; the computer program is executed by a processor to implement the method for predicting the growth of the diameter at breast height of Chinese fir in other embodiments.

[0269] In summary, the Chinese fir breast diameter growth prediction device and storage medium of the present application include constructing multiple alternative Chinese fir breast diameter growth models, and performing model evaluation to determine the breast diameter growth basic model; screening the multi-category independent variables of the model to determine the effective independent variables; and adding the effective independent variables to the breast diameter growth basic model to construct a breast diameter growth model based on multiple variables and dummy variables; according to the breast diameter growth model, considering the random effects of different parameters to construct the optimal breast diameter mixed effect growth model.

[0270] This application goes through the process of acquiring the breast diameter data of Chinese fir under the near-natural transformation of Chinese fir, introducing the model after evaluating and calculating various variables that affect the breast diameter growth, and constructing the breast diameter growth model of Chinese fir using a nonlinear mixed effects model, and constructing the optimal breast diameter mixed effects growth model considering the random effects of different parameters, which greatly improves the accuracy of the breast diameter growth model estimation, improves the accuracy and efficiency of the breast diameter prediction, and is of great significance to improving the quality and yield of Chinese fir plantations.

[0271] In addition, this application uses a single survey data combined with annual ring width data to establish a model, which can complete the construction of the breast diameter growth model of Chinese fir under the near-natural transformation mode of Chinese fir and Phoebe chinensis in the absence of consecutive years of re-measurement data, thereby providing a reference for the subsequent prediction of the yield of Chinese fir species under the near-natural transformation mode of Chinese fir and the screening of suitable near-natural transformation modes of Chinese fir and Phoebe chinensis.

[0272] Those skilled in the art will appreciate that the terms used in the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms of "a", "said" and "the" used in the present invention and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0273] It should be understood that although the terms first, second, third, etc. may be used in the present invention to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present invention, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0274] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0275] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for predicting the growth of breast diameter of Chinese fir based on a near-natural transformation of Chinese fir and Phoebe fujianensis stands, characterized in that: include: The diameter at breast height data of Cunninghamia lanceolata under near-natural transformation are obtained as a model data set; the diameter at breast height data of Cunninghamia lanceolata includes single tree diameter at breast height data, annual ring width data, and historical diameter at breast height growth data calculated according to the annual ring width data; multiple candidate Cunninghamia lanceolata diameter at breast height growth models are constructed in combination with the growth law of trees and the characteristics of the model, and a model evaluation is performed based on the model data set to determine a basic model for diameter at breast height growth; Among them, the first evaluation formula consisting of three statistical indicators, namely, the coefficient of determination R2, the root mean square error RMSE and the mean absolute error MAE, is used to evaluate the model and determine the basic model of breast diameter growth; the first evaluation formula is calculated as follows: ； ； ； Among them, n is the number of samples, is the measured value of the DBH of the jth tree in the i-th plot, is the estimated value of DBH of the jth tree in the i-th plot, is the average breast diameter; R 2 The closer it is to 1, the smaller the values ​​of RMSE and MAE are, indicating that the prediction effect of the selected model is more accurate; Screen the multi-category independent variables of the model to determine the effective independent variables; add the effective independent variables to the basic model of DBH growth to construct a DBH growth model based on multiple variables and dummy variables; specifically, use a nonlinear mixed effect model to construct a DBH growth model, the general form of which is: ； Among them, D ij is the number of trees in the i-th plot t The diameter at breast height at the age; Among them, f (β, u i , t) describes D ij The relationship between μ and t, that is, the function between DBH and forest age; i is the standard random effect parameter estimate; D is the random effect variance-covariance matrix; R ij is the variance-covariance matrix within the plot; μ i ~N(0, t) means that the random parameter has an expectation of 0 and a variance of D Normal distribution of ; ε ij is the error term, with an expected value of 0 and a variance of R ij The normal distribution of ij With μ i Independent of each other; According to the DBH growth model, after considering the random effects of different parameters, an optimal DBH mixed-effect growth model is constructed; the DBH growth of Chinese fir is predicted by the DBH mixed-effect growth model; The optimal DBH mixed effect growth model considering the random effects of different parameters includes: 1) Based on the DBH growth model of Chinese fir, the nlme function was used and the plots were used as random variables to construct multiple random effect models configured with multiple parameters; and the optimal random effect model was determined based on the model evaluation indicators; 2) According to the optimal random effects model, the first-order autoregression method is used to eliminate the autocorrelation of the model, and the exponential function, power function and constant plus power function are used to eliminate the heteroscedasticity between data. The formula is as follows: ； ； ； ； Where: The exponential function is VarExp (ε ij ), the power function is VarPower(ε ij ), the constant plus power function is VarConstPower (ε ij ); represents the model error variance; G i is a diagonal matrix used to capture the heteroskedasticity of the random effects; is a variance matrix describing the autocorrelation of random effects; x ij is the predicted value of DBH of the j-th plant in the i-th plot based on the fixed effect parameters; λ, λ1, λ2 are the parameters to be estimated; And judge by comparing the AIC (Akaike Information Criterion) and BIC (Bayesian Information Criterion) values ​​of the model. The smaller the AIC and BIC values ​​of these two indicators, the better. At the same time, we should also consider the Loglik (log likelihood value), which is better when the larger the value is. Among them, the formulas for AIC and BIC are as follows: ； ； In the formula, p is the number of model parameters, n is the number of sample plants; finally, the optimal DBH mixed effect growth model is determined.

2. The method for predicting the growth of diameter at breast height of Chinese fir based on the near-natural transformation of Chinese fir and Phoebe fujianensis stands according to claim 1, characterized in that: The method of obtaining the breast diameter data of Chinese fir under near-natural transformation of Phoebe fujianensis comprises the following steps: Based on the survey data of each tree, standard trees of various natural transformation sites of Cunninghamia lanceolata and Phoebe fujianensis were determined, and several trees with diameters closest to the standard trees were selected as tree samples; Using Swedish growth cones Haglof Sweden, cores were drilled from the tree samples at breast height along the east-west and north-south directions, with a total of two core samples from each tree; The tree core samples were dried and polished for dating analysis, and the tree ring width was accurately measured using the LinTAB 6 tree ring analyzer and TASP-Win software; TSAP software was used to cross-date all samples.

3. The method for predicting the growth of diameter at breast height of Chinese fir based on the near-natural transformation of Chinese fir and Phoebe fujianensis stands according to claim 1, characterized in that: The method of constructing a plurality of candidate growth models for the diameter at breast height of Chinese fir and performing model evaluation to determine a basic growth model for the diameter at breast height comprises the following steps: Selecting multiple growth equations as candidate equations for the growth model of the diameter at breast height of Chinese fir, and constructing multiple candidate growth models of the diameter at breast height of Chinese fir; fitting the candidate growth models of the diameter at breast height of Chinese fir respectively and comparing the model fitting effects; The first evaluation formula consisting of three statistical indicators, namely, the coefficient of determination R2, the root mean square error RMSE and the mean absolute error MAE, was used to evaluate the model and determine the basic model of breast diameter growth.

4. The method for predicting the growth of diameter at breast height of Chinese fir based on the near-natural transformation of Chinese fir and Phoebe fujianensis stands according to claim 1, characterized in that: The multi-category independent variables of the screening model are used to determine the effective independent variables, including: The independent variables of the model are determined from four aspects: forest stand factor, competition factor, pattern factor and site factor; During the modeling process, independent variables with weak correlation with DBH and insignificant parameters were removed, and independent variables with collinearity with other variables were screened to determine the effective independent variables.

5. The method for predicting the growth of diameter at breast height of Chinese fir based on the near-natural transformation of Chinese fir and Phoebe fujianensis stands according to claim 4, characterized in that: The filtering out of independent variables with weak correlation with DBH and insignificant parameters during the modeling process, screening independent variables with collinearity with other variables, and determining effective independent variables comprises the following steps: Use linear regression method to screen insignificant factors; Calculate the variance inflation factor (VIF) value of each variable and remove factors with VIF values ​​greater than 5; Use the correlation matrix to filter out factors with absolute correlation values ​​less than 0.1; Calculate the VIF value again, and obtain the final selectable factors as valid independent variables through secondary screening.

6. The method for predicting the growth of diameter at breast height of Chinese fir based on the near-natural transformation of Chinese fir and Phoebe fujianensis stands according to claim 1, characterized in that: According to the DBH growth model, the optimal DBH mixed effect growth model is constructed by considering the random effects of different parameters, including: According to the growth model of the diameter at breast height of Chinese fir, using the nlme function and taking the sample plot as a random variable, multiple random effect models configured with multiple parameters are constructed; and the optimal random effect model is determined according to the model evaluation index; According to the optimal random effects model, the first-order autoregression method is used to eliminate the autocorrelation of the model, and the exponential function, power function and constant plus power function are used to eliminate the heteroscedasticity between data respectively; and the second evaluation formula is used to evaluate the goodness of fit of the model, and the model with the highest convergence and fitting accuracy is selected as the optimal model based on a certain variable; at the same time, a likelihood ratio test is performed on multiple random effects models, and models with significant differences are selected; finally, the mixed effect growth model of the breast diameter of Chinese fir is determined.

7. A system for predicting the growth of diameter at breast height of Chinese fir based on a near-naturally transformed forest stand of Chinese fir and Phoebe fujianensis using the method according to any one of claims 1 to 6, characterized in that: include: The data set module is used to obtain the breast diameter data of Cunninghamia lanceolata under near-natural transformation as a model data set; The basic model module is used to construct multiple candidate growth models for the diameter at breast height of Chinese fir and to conduct model evaluation to determine the basic growth model for the diameter at breast height; A mixed effect model module is used to screen the multi-category independent variables of the model and determine the effective independent variables; and add the effective independent variables to the basic model of DBH growth to construct a DBH growth model based on multiple variables and dummy variables; A random effect adaptation module, used for constructing an optimal DBH mixed effect growth model based on the DBH growth model and considering the random effects of different parameters; The prediction module is used to predict the growth of the diameter at breast height of Chinese fir by using the mixed effect growth model of the diameter at breast height of Chinese fir.

8. A device for predicting the growth of breast diameter of Chinese fir based on a near-natural transformation of Chinese fir and Phoebe fujianensis stands, characterized in that: include: A storage unit for storing executable instructions; as well as A processing unit, configured to be connected to the memory to execute executable instructions to complete the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon; the computer program is executed by a processor to implement the method according to any one of claims 1 to 6.

Images