Cunninghamia bournei tree height curve prediction method based on near-natural transformed forest stand of phoebe bournei

Abstract

The invention provides a cunninghamia lanceolata tree height curve prediction method based on a cunninghamia lanceolata phoebe bournei near-natural transformation forest stand, and the method comprises the steps: completing the construction of a cunninghamia lanceolata tree height curve model in a cunninghamia lanceolata phoebe bournei near-natural transformation mode by using a multivariate and dumb variable combined and a basic method support is provided for estimating the height of the cedarwood in the near-natural transformation mode of the cedarwood phoebe bournei, subsequently estimating the cedarwood storage amount and the wood outturn amount of the wood species, and screening a proper near-natural transformation mode of the cedarwood phoebe bournei.

Description

Technical Field

[0001] The present application belongs to the technical field of forest stand growth and prediction models, and in particular, relates to a method for predicting a tree height curve of Chinese fir based on a nearly natural transformation of a forest stand of Chinese fir and Phoebe fujian. 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] However, due to the long time interval of forest stand survey data, inaccurate growth variables and low precision of the extrapolation process, the current prediction accuracy and efficiency of Chinese fir tree height are low.

[0004] 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

[0005] The invention proposes a method for predicting the height curve of a Chinese fir tree. This application introduces various variables that affect the growth of diameter at breast height into the model to improve the accuracy of tree height curve model estimation, thereby greatly improving the accuracy and efficiency of tree height prediction, which is of great significance to improving the quality and yield of Chinese fir plantations.

[0006] According to a first aspect of an embodiment of the present application, a method for predicting a tree height curve of Chinese fir based on a near-naturally transformed forest stand of Chinese fir Phoebe fujianensis is provided, comprising: Obtain survey data of near-natural transformation of Chinese fir and Phoebe fujianensis stands to obtain model building data and test data; Combining the growth law of trees and the characteristics of the model, multiple tree height curve models were determined as candidate equations for the Chinese fir tree height curve model; multiple tree height curve models were fitted and the model fitting effects were compared, and the first evaluation of the model was performed to determine the basic tree height curve model; Determine the effective variables that reflect environmental factors and the effect of DBH growth under the near-natural transformation model, add the effective variables to the basic tree height curve model, and obtain a multivariate tree height curve model; According to the multivariate tree height curve model, a dummy variable tree height curve model is constructed based on the dummy variable method; the optimal Chinese fir tree height curve model is determined through the second evaluation formula of the model; and the Chinese fir tree height is predicted through the optimal Chinese fir tree height curve model.

[0007] In some embodiments of the present application, the first evaluation of the model includes using 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. The specific calculation formula is: ; ; ; in, y ij and Respectively i The first j Observed and predicted tree heights for each plant; m is the number of sample plots, n i It is i The number of trees in a plot.

[0008] In some embodiments of the present application, determining effective variables that reflect environmental factors and influence breast diameter growth under near-natural transformation models includes the following steps: Select multiple variables including stand factors and tree species coefficients; stand factors include at least one of the number of trees per hectare N, stand average breast diameter Dq, stand basal area BA and stand density index SDI; tree species composition coefficient includes Chinese fir composition coefficient fir%; In statistical analysis, independent variables with variance inflation factors (VIF) greater than 5 were eliminated, and variables that were significant in statistical tests were retained to obtain the screened variables; The correlation between tree height and the screened variables was analyzed, and on the basis of the basic tree height curve model, at least one variable with a higher correlation was selected as a valid variable.

[0009] In some embodiments of the present application, effective variables are added to the basic tree height curve model to obtain a multivariate tree height curve model, including: The basic tree height curve model is the Wykoff model, and the expression of the Wykoff model is: ; Where H and D are tree height (m) and diameter at breast height (cm), respectively; a and b are model parameters to be estimated; The effective variables are site index, stand condition and tree species composition; the general expression of the multivariate tree height curve model GHD1 is as follows: ; Among them, H ij For the i Sample site j Tree height, DBH ij For the i Sample site j Diameter at breast height of a tree; SI i For the i Site index of each plot; For the i The stand density index of each plot; fir% i It is i The fir composition coefficient of each plot; ɑ0, ɑ1, ɑ2, b1, b2 are the parameters to be estimated in the model; And considering that the three effective variables all act on the parameter ɑ0, the expression of the model form GHD2 is: ; in, ɑ 0 、ɑ 1 、ɑ 2 、ɑ 3 、b 1 are the parameters to be estimated for the model; Finally, the optimal model GHD is determined for subsequent fitting based on whether the model parameters are significant and the model fitting effect.

[0010] In some embodiments of the present application, a dummy variable tree height curve model is constructed based on a dummy variable method according to a multivariate tree height curve model, including: Based on the optimal model GHD, dummy variables Z are added to different parameters. i ( i =1, 2, 3), we get the dummy variable model DGHD1 acting on parameter a and the dummy variable model DGHD2 acting on parameter b; The expression of DGHD1 is: ; The expression of DGHD2 is: ; in, ɑ 0 、ɑ 1 、ɑ 2、b 1 、b 2 、K 1 、k 2 、k 3 are the parameters to be estimated for the model.

[0011] In some embodiments of the present application, the optimal fir tree height curve model is determined by the second evaluation formula of the model, including: Compare the Akaike Information Criterion AIC value and the Bayesian Information Criterion BIC value of the model. The smaller the AIC value and BIC value, the better. At the same time, consider the log-likelihood value Loglik, the larger the value, the better. ; ; in, p is the number of model parameters, n is the number of sample plants, logn is the logarithm of n, and logLik is the logarithm of likelihood.

[0012] In some embodiments of the present application, the second evaluation formula of the model also includes the prediction ability of the evaluation model, specifically including: The model test index is calculated using the observed values ​​and predicted values ​​of a certain number of sample strains to evaluate the prediction ability of the model; the model test index includes the absolute coefficient R 2 , total relative error (TRE) and mean absolute error (MAE). The specific calculation formula is: ; ; ; in, y ij and Respectively i The first j Observed and predicted tree heights for each plant; m is the number of sample plots, n i It is i The number of trees in a plot.

[0013] According to a second aspect of an embodiment of the present application, a system for predicting a height curve of a Chinese fir tree is provided, comprising: The data set module is used to obtain the survey data of the near-natural transformation of Cunninghamia lanceolata and Phoebe fujianensis stands to obtain model building data and test data; The basic model module is used to combine the growth law of trees and the characteristics of the model to determine multiple tree height curve models as candidate equations for the Chinese fir tree height curve model; and to fit multiple tree height curve models respectively and compare the model fitting effects, and at the same time perform the first evaluation of the model to determine the basic tree height curve model; The multivariate tree height curve model is used to determine the effective variables that reflect environmental factors and the growth of DBH under the near-natural transformation model. The effective variables are added to the basic tree height curve model to obtain the multivariate tree height curve model. The optimal fir tree height curve model module is used to construct a dummy variable tree height curve model based on the multivariate tree height curve model based on the dummy variable method; and determine the optimal fir tree height curve model through the second evaluation formula of the model; The prediction module is used to predict the height of Chinese fir trees by using the optimal Chinese fir tree height curve model.

[0014] According to a third aspect of an embodiment of the present application, a device for predicting a height curve of a Chinese fir tree 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 a height curve of a Chinese fir tree.

[0015] 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 a Chinese fir tree height curve.

[0016] A method for predicting the height curve of Chinese fir based on the near-natural transformation forest stand of Chinese fir and Phoebe fujianensis is adopted in the present application, including combining the growth law of trees and the characteristics of the model, determining multiple tree height curve models as candidate equations for the height curve model of Chinese fir; fitting multiple tree height curve models respectively and comparing the model fitting effects, and performing the first evaluation of the model at the same time to determine the basic tree height curve model; determining the effective variables that reflect the environmental factors and the influence of breast diameter growth under the near-natural transformation mode, adding the effective variables to the basic tree height curve model, and obtaining a multivariate tree height curve model; constructing a dummy variable tree height curve model based on the dummy variable method according to the multivariate tree height curve model; determining the optimal Chinese fir tree height curve model through the second evaluation formula of the model; and predicting the height of Chinese fir through the optimal Chinese fir tree height curve model. The present application uses a combination of multivariate and dummy variables to complete the construction of the height curve model of Chinese fir under the near-natural transformation mode of Chinese fir and Phoebe fujianensis, providing a basic method support for estimating the height of Chinese fir under the near-natural transformation mode of Chinese fir and the subsequent estimation of the volume of Chinese fir and the yield of timber species, and screening suitable near-natural transformation mode of Chinese fir and Phoebe fujianensis. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] 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: Figure 1 FIG. 2 shows a schematic diagram of steps of a method for predicting a height curve of a Chinese fir tree according to an embodiment of the present application; Figure 2 A schematic diagram of the steps of determining effective variables according to an embodiment of the present application is shown in FIG. Figure 3 A schematic diagram of the steps of performing the second evaluation of the model according to an embodiment of the present application is shown in FIG. Figure 4 The tree height-DBH fitting curve based on the Wykoff equation (M5) is shown in; Figure 5 The distribution diagram of the observed values ​​(test data) and predicted values ​​of each model is shown in ; Figure 6 : shows a schematic diagram of the structure of a Chinese fir tree height curve prediction system according to an embodiment of the present application; Figure 7 Schematic diagram of the structure of the device for predicting the height curve of a Chinese fir tree according to an embodiment of the present application is shown in FIG. DETAILED DESCRIPTION

[0018] Regarding this application, considering that tree height is a key indicator for evaluating forest productivity, and is also an important indicator for calculating tree volume and even biomass. However, in forest surveys, tree height measurement is relatively time-consuming and labor-intensive, and the survey error is relatively large. Therefore, forestry workers often estimate tree height by constructing a tree height-DBH curve model.

[0019] Because the DBH is easy to measure and has a small error, the height of Chinese fir is estimated through the relationship model between tree height and DBH, stand density and near-natural transformation mode, which provides an important method for accurately estimating the height of Chinese fir under the currently commonly used near-natural transformation mode of Chinese fir and Phoebe fujianensis.

[0020] However, due to the long time interval of forest stand survey data, inaccurate growth variables and low precision of the extrapolation process, the current prediction accuracy and efficiency of Chinese fir tree height are low.

[0021] To solve the above problems, the method for estimating the breast diameter growth model of Chinese fir based on a near-naturally transformed forest stand of Chinese fir and improving the estimation accuracy of the breast diameter growth model provided by the present invention mainly consists of the following steps: Obtain survey data of near-natural transformation of Chinese fir and Phoebe fujianensis stands to obtain model building data and test data; Combining the growth law of trees and the characteristics of the model, multiple tree height curve models were determined as candidate equations for the Chinese fir tree height curve model; multiple tree height curve models were fitted and the model fitting effects were compared, and the first evaluation of the model was performed to determine the basic tree height curve model; Determine the effective variables that reflect environmental factors and the effect of DBH growth under the near-natural transformation model, add the effective variables to the basic tree height curve model, and obtain a multivariate tree height curve model; According to the multivariate tree height curve model, a dummy variable tree height curve model is constructed based on the dummy variable method; the optimal Chinese fir tree height curve model is determined through the second evaluation formula of the model; and the Chinese fir tree height is predicted through the optimal Chinese fir tree height curve model.

[0022] In addition, the present application also provides a system and device for predicting the height curve of a Chinese fir tree.

[0023] 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 height-diameter curve model 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 estimating the height of Cunninghamia lanceolata under the forest stand of near-natural transformation mode of Cunninghamia lanceolata, and accurately improving the quality and yield of Cunninghamia lanceolata plantations.

[0024] Specifically, the present invention uses a combination of multivariate and dummy variable methods to complete the construction of a tree height curve model of Chinese fir under a near-natural transformation mode of Chinese fir, providing a basic method support for estimating the tree height of Chinese fir under a near-natural transformation mode of Chinese fir and subsequent estimation of Chinese fir stock and timber output, and screening suitable near-natural transformation modes of Chinese fir.

[0025] 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.

[0026] Example 1 Figure 1 Schematic diagram of the steps of the method for predicting the height curve of a Chinese fir tree according to an embodiment of the present application is shown in FIG.

[0027] like Figure 1 As shown, the method for predicting the height curve of Chinese fir tree includes the following steps: S1: Obtain survey data of near-natural transformation of Chinese fir and Phoebe fujianensis stands to obtain model building data and test data; S2: Combining the growth law of trees and the characteristics of the model, multiple tree height curve models are determined as candidate equations for the Chinese fir tree height curve model; multiple tree height curve models are fitted and the model fitting effects are compared, and the first evaluation of the model is performed to determine the basic tree height curve model; S3: Determine the effective variables that reflect environmental factors and the effect of DBH growth under the near-natural transformation model, add the effective variables to the basic tree height curve model, and obtain a multivariate tree height curve model; S4: Construct a dummy variable tree height curve model based on the multivariate tree height curve model and the dummy variable method; determine the optimal Chinese fir tree height curve model through the second evaluation of the model; S5: Predict Chinese fir tree height using the optimal Chinese fir tree height curve model.

[0028] This application improves the accuracy of tree height curve model estimation by introducing various variables that affect the growth of breast diameter into the model, greatly improving the accuracy and efficiency of tree height prediction, which is of great significance to improving the quality and yield of Chinese fir plantations.

[0029] In the specific implementation, S2 combines the growth law of trees and the characteristics of the model, and selects 13 commonly used growth equations as shown in Table 1 as alternative equations for the breast diameter growth model of Chinese fir, thereby forming thirteen alternative models as shown in Table 1: Stoffels, Naslund, Curtis, Schumachter, Wykoff, Bates and Watts, Von Bertalanffy, Meyer, Strand, Logistic, Weibull, Prodan and Richards.

[0030] Table 1 Candidate tree height-DBH basic model

[0031] In Table 1, H and D are tree height (m) and breast diameter (cm), respectively; a, b, and c are model parameters to be estimated.

[0032] Then, the thirteen alternative models were fitted separately and the model fitting effects were compared, and the model with the best fitting effect was selected as the basic tree height curve model.

[0033] The 13 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.

[0034] 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, thereby obtaining the most suitable basic model of the Chinese fir tree height curve.

[0035] In the specific implementation, the first evaluation of the model in S2 includes using the determination coefficient R 2 The model is evaluated using three statistical indicators: root mean square error (RMSE) and mean absolute error (MAE). The specific calculation formula is: ; ; ; in, y ij and Respectively i The first j Observed and predicted tree heights for each plant; m is the number of sample plots, n i It is i The number of trees in a plot.

[0036] Figure 2 A schematic diagram of the steps of determining effective variables according to an embodiment of the present application is shown in FIG.

[0037] like Figure 2 As shown in S3, determining the effective variables that affect DBH growth under environmental factors and near-natural transformation models includes the following steps: S31: Select multiple variables including stand factor and tree species coefficient; stand factor includes at least one of the number of trees per hectare N, average breast diameter Dq, basal area BA and stand density index SDI; tree species composition coefficient includes Chinese fir composition coefficient fir; S32: In statistical analysis, the independent variables with variance inflation factor (VIF) greater than 5 were eliminated, and the variables that were significant in the statistical test were retained to obtain the screened variables; S33: Perform correlation analysis on the tree height and the screened variables, and select at least one variable with a higher correlation as a valid variable based on the basic tree height curve model.

[0038] In order to reflect the influence of environmental factors and near-natural transformation mode on the Chinese fir tree height curve model, the present invention introduces indicators such as stand density and tree species composition coefficient when constructing the Chinese fir tree height curve model. The specific steps are: (1) selecting stand factors such as the number of trees per hectare N, the average breast diameter of the stand Dq, the basal area of ​​the stand BA, the stand density index SDI, etc., and the tree species composition coefficient such as the Chinese fir composition coefficient fir%; when performing statistical analysis, the independent variables with variance inflation factor (VIF) greater than 5 should be eliminated, and those variables that are significant in the statistical test (P < 0.05) should be retained.

[0039] (2) Perform correlation analysis between tree height and the variables selected in (1). Based on the determined basic model of Chinese fir tree height curve, select variables with higher correlation (1-3). Then add models with different combinations to test the effects of different factors on the tree height-DBH relationship.

[0040] In the following implementation, valid variables are added to the basic tree height curve model to obtain a multivariate tree height curve model, including: Assume that the basic tree height curve model is the Wykoff model, and the expression of the Wykoff model is: ; Where H and D are tree height (m) and diameter at breast height (cm), respectively; a and b are model parameters to be estimated; Moreover, the effective variables are site index, stand condition and tree species composition; the general expression of the multivariate tree height curve model GHD1 is as follows: ; Among them, H ij For the i Sample site j Tree height, DBH ij For the i Sample site j Diameter at breast height of a tree; SI i For the i Site index of each plot; For the i The stand density index of each plot; fir% i It is i The fir composition coefficient of each plot; ɑ0, ɑ1, ɑ2, b1, b2 are the parameters to be estimated in the model; And considering that the three effective variables all act on the parameter ɑ0, the expression of the model form GHD2 is: ; Among them, ɑ0, ɑ1, ɑ2, ɑ3, and b1 are the parameters to be estimated for the model; Finally, the optimal model GHD is determined for subsequent fitting based on whether the model parameters are significant and the model fitting effect.

[0041] Next, enter step S4 to construct a dummy variable tree height curve model based on the multivariate tree height curve model based on the dummy variable method, including: Based on the optimal model GHD, dummy variables Z are added to different parameters. i ( i =1, 2, 3), we get the dummy variable model DGHD1 acting on parameter a and the dummy variable model DGHD2 acting on parameter b; The expression of DGHD1 is: ; The expression of DGHD2 is: ; Among them, ɑ0, ɑ1, ɑ2, b1, b2, K1, k2, and k3 are the model parameters to be estimated.

[0042] To further illustrate, the effective variables screened by S3 are added to the optimal basic model obtained by S2 to construct a Chinese fir tree height curve model including dummy variables.

[0043] In order to reflect the afforestation measures of near-natural transformation, 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 between different transformation modes.

[0044] Therefore, in order to evaluate the growth changes of DBH of Cunninghamia lanceolata under different near-natural transformation modes, three types of models were established in the present invention: (1) basic model BHD with only tree height H~DBH; (2) model GHD including stand, site and diversity index; (3) model with mode Mode as dummy variable (DGHD).

[0045] Then, the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and Log-Likelihood (LogLik; LogLikelihood) were used to evaluate and compare the models. In the process of determining the best tree height curve model, priority was given to those models that could achieve model convergence and have high prediction accuracy.

[0046] Figure 3 A schematic diagram of the steps for performing the second evaluation of the model according to an embodiment of the present application is shown in FIG.

[0047] like Figure 3 As shown, in further implementation, in S4, the optimal fir tree height curve model is determined through the second model evaluation, including: S41: Compare the Akaike information criterion AIC value and the Bayesian information criterion BIC value of the model. The smaller the AIC value and BIC value, the better. At the same time, consider the log-likelihood value Loglik, the larger the value, the better. ; ; in, p is the number of model parameters, n is the number of sample plants, logn is the logarithm of n, and logLik is the logarithm of likelihood.

[0048] In further implementation, the second evaluation formula of the model in S4 also includes the prediction ability of the evaluation model, specifically including: S42: Use the observed values ​​and predicted values ​​of a certain number of sample strains to calculate the model test index and evaluate the prediction ability of the model; the model test index includes the absolute coefficient R 2 , total relative error (TRE) and mean absolute error (MAE). The specific calculation formula is: ; ; ; in, y ij and Respectively i The first j Observed and predicted tree heights for each plant; m is the number of sample plots, n i It is i The number of trees in a plot.

[0049] In summary, a method for predicting the height curve of Chinese fir based on a near-naturally transformed forest stand of Chinese fir and Phoebe fujianensis according to an embodiment of the present application includes determining multiple tree height curve models as alternative equations for the Chinese fir tree height curve model in combination with the growth patterns of trees and the characteristics of the model; fitting the multiple tree height curve models respectively and comparing the model fitting effects, while performing a first evaluation of the model to determine a basic tree height curve model; determining effective variables that reflect environmental factors and the effects on breast diameter growth under the near-natural transformation model, adding the effective variables to the basic tree height curve model to obtain a multivariate tree height curve model; constructing a dummy variable tree height curve model based on the dummy variable method according to the multivariate tree height curve model; determining the optimal Chinese fir tree height curve model through the second evaluation formula of the model; and predicting the height of Chinese fir through the optimal Chinese fir tree height curve model.

[0050] This application uses a combination of multivariate and dummy variable methods to complete the construction of a tree height curve model for Chinese fir under the near-natural transformation model of Chinese fir and Fujian Phoebe, providing basic methodological support for estimating the tree height of Chinese fir under the near-natural transformation model of Chinese fir and the subsequent estimation of Chinese fir stock and timber output, and screening suitable near-natural transformation models of Chinese fir and Fujian Phoebe.

[0051] This application proposes a modeling method based on the current commonly used near-natural transformation model of Cunninghamia lanceolata in my country, which can not only improve the estimation accuracy of the height-diameter curve model of Cunninghamia lanceolata, but also comprehensively consider the forest stand, site, competition index and near-natural transformation model. This method is of great significance for accurately estimating the height of Cunninghamia lanceolata under the near-natural transformation model of Cunninghamia lanceolata, and accurately improving the quality and yield of Cunninghamia lanceolata plantations.

[0052] Example 2 This embodiment further illustrates Embodiment 1 in conjunction with a specific implementation scenario.

[0053] The research object of this embodiment is a Chinese fir plantation after near-natural transformation of Chinese fir from Phoebe fujianensis in Xishan Forest Farm of a certain city.

[0054] 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 m 2 (0.06 hm 2 ), of which there were 27 near-natural transformation plots and 3 pure forest control plots.

[0055] The retention density of Chinese fir and the planting density of Phoebe bournei under each mode are shown in Table 2.

[0056] Table 2 The near-natural transformation model of Chinese fir and Phoebe chinensis in Xishan Forest Farm

[0057] (I) First, obtain survey data on near-naturally transformed stands of Cunninghamia lanceolata and Phoebe chinensis, and obtain data on tree height and breast diameter of Cunninghamia lanceolata.

[0058] In 2022, each tree in the 30 near-natural transformation plots of Xishan Forest Farm will be inspected to measure the breast diameter (1.3m), 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) of each fir tree.

[0059] 2 / 3 H-DBH data pairs were randomly selected to form the modeling data set (862 pairs), and the remaining data pairs constituted the test data set (429 pairs), as shown in Table 3.

[0060] Table 3 Statistics of tree height curve modeling data and test data

[0061] In order to reflect the near-natural transformation model of Cunninghamia lanceolata and Phoebe fujianensis 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 composition coefficients of each tree species include the composition coefficient of Chinese fir (fir%) and the composition coefficient of Phoebe bournei (nm%).

[0062] The tree species composition coefficient reflects the proportion of a specific tree species in the total forest volume or total cross-sectional area. Since the cross-sectional area at breast height is easier to measure than the volume, the tree species composition is often determined by the cross-sectional area in practice.

[0063] In particular, this application uses cross-sectional area to calculate the composition coefficients of Chinese fir and Phoebe bournei: ; Among them, 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 ).

[0064] In order to reflect the afforestation measures of near-natural transformation, 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 between different transformation modes.

[0065] Since there are three near-natural patterns of interplanting Phoebe bournei under the fir forest in Xishan Forest Farm (see Table 2), namely M1, M2 and M3, and a control pattern CK, we set 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.

[0066] (ii) Next, based on the growth patterns of trees and the characteristics of the model, 13 commonly used growth equations as shown in Table 1 were selected as candidate equations for the height curve model of Chinese fir.

[0067] Then, thirteen alternative breast diameter growth models of Chinese fir were screened to determine the basic tree height curve model of Chinese fir.

[0068] First, the 13 basic models proposed in Table 1 were used to construct the height curve models of Chinese fir respectively; and the absolute coefficient R 2 The three indicators of evaluation are total relative error (TRE) and mean absolute error (MAE), and the following results are obtained as shown in Table 4: Table 4 Parameter estimation and evaluation indicators of alternative basic models

[0069] The 13 candidate models in Table 1 were fitted separately, and the model 13 with insignificant parameters was discarded. Finally, the fitting results of the last 12 models converged, and the results are shown in Table 4.

[0070] See Table 4, the R 2 The evaluation indicators of all candidate tree height curve models vary within a small range. Although increasing the number of model parameters can improve the accuracy of the model, in practical applications, the differences between basic models are negligible, and models with more parameters are prone to convergence problems. The inherent simplicity of the two-parameter model makes it a good match in various situations.

[0071] Therefore, after comprehensive consideration, the best performing model 5 among the two-parameter models was selected as the Wykoff equation, which was selected as the basic tree height curve model BHD of the Chinese fir tree height curve. 2 , RMSE and MAE are 0.4439, 2.3981 and 1.9763 respectively.

[0072] Figure 4 The tree height-DBH fitting curve based on the Wykoff equation (M5) is shown in FIG.

[0073] like Figure 4 The tree height-DBH fitting curve corresponding to the Wykoff curve is shown in Figure 4 As shown in Figure 2, the model fitting results converge.

[0074] (iii) Next, we determined the effective variables that affect DBH growth under environmental factors and near-natural transformation models.

[0075] According to the VIF screening results and correlation analysis results, it was found that tree height, DBH and site index were significantly positively correlated ( P < 0.05), and had a significant negative correlation with the composition coefficient of Chinese fir ( P < 0.05). Tree height also showed a significant negative correlation with the stand density index ( P < 0.05). All three variables have significant effects on tree height, which is statistically significant.

[0076] Therefore, based on the Wykoff basic model, three variables, namely site index SI, stand density index SDI and fir composition coefficient fir%, were introduced to construct the tree height curve model, and the fitting ability of models with different combinations was compared to screen the model.

[0077] The results show that after adding variables, the model accuracy is significantly improved. 2 The value is 0.7574; the RMSE and MAE values ​​are reduced to 1.5838 and 1.2410 respectively; it shows that considering the site index, stand conditions and tree species composition can improve the fitting ability of the tree height-DBH model, and the general expression of the model GHD1 is as follows: ; Where: H ij For the i Sample site j Tree height, DBH ij For the i Sample site j Diameter at breast height of a tree; SI i For the i Site index of each plot; For the i The stand density index of each plot; fir% i It is i The fir composition coefficient of each plot; ɑ0, ɑ1, ɑ2, b1, b2 are the model parameters to be estimated.

[0078] The model form GHD2 in which all three variables act on the parameter ɑ0 is also considered: ; Where: ɑ 0 、ɑ 1 、ɑ 2 、ɑ 3 、b 1 are the parameters to be estimated for the model; It was found that the parameters of the GHD2 model were ɑ 2 (Parameters acting on fir%) were not significant (see Table 5), and the fitting effect was slightly smaller than that of GHD1, so GHD1 was selected as the optimal model for subsequent fitting.

[0079] Table 5 Estimated values ​​of model parameters

[0080] (iv) Finally, a dummy variable tree height curve model based on dummy variables is constructed.

[0081] This application uses dummy variables to reflect the impact of different near-natural transformation modes on the tree height-DBH growth curve.

[0082] Based on the model GHD1 selected in (III) above, try adding dummy variables Z in different parameters i ( i =1,2, 3), and use R 2 , AIC, BIC and logLik and the statistical significance of parameters were used to evaluate the fitting effect.

[0083] Finally, we obtain the dummy variable model DGHD1 acting on parameter a and the dummy variable model DGHD2 acting on parameter b.

[0084] The results showed that the fitting effect of the Chinese fir tree height curve model was significantly improved after the introduction of dummy variables (Table 5). 2 It increased from 0.7574 to 0.7825 and 0.7824 respectively; RMSE also showed a downward trend, indicating that the transformation pattern affected the variation of the tree height curve of Chinese fir.

[0085] Among them, the dummy variable model DGHD1 acting on parameter a and the dummy variable model DGHD2 acting on parameter b, the final form of the two models is as follows: DGHD1: ; DGHD2: ; In the formula: ɑ0, ɑ1, ɑ2, b1, b2, K1, k2, k3 are the model parameters to be estimated.

[0086] Finally, a second evaluation of the model was conducted to determine the optimal fir tree height curve model.

[0087] Based on Table 5 and Table 6 shown below, the parameter estimates and evaluation indicators of each model obtained based on the fitted data are given respectively.

[0088] The results show that the tree height curve model performs better than the basic model after considering the stand density index, tree species composition coefficient and transformation mode. The AIC is smaller than that of the basic model, and R 2 The and logLik values ​​are larger than those of the basic model, indicating that the introduction of stand factors and transformation patterns improves the fitting accuracy of the model.

[0089] After comprehensive comparison, we get the effect on the parameters a The tree height curve model DGHD1 has the best effect (the smallest AIC is 3161.77; the largest logLik is -1571.89).

[0090] Table 6 Evaluation indicators based on each model

[0091] Finally, the tree height curve model of Chinese fir was tested based on the near-natural transformation model of Chinese fir and Phoebe fujianensis.

[0092] By analyzing 429 sets of tree height-DBH measurement data, the prediction effects of various tree height curve models of Chinese fir were tested, and the performance of these models in predicting Chinese fir of different sizes was compared. It was found that the DGHD1 model had the largest R2 and the smallest TRE. Figure 5 The predicted values ​​and observed values ​​of the DGHD1 model are more concentrated near the straight line with a slope of 1, which also shows that the prediction is more accurate.

[0093] Figure 5 The distribution diagram of observed values ​​(test data) and predicted values ​​of each model is shown in .

[0094] Therefore, the present invention uses a combination of multivariate and dummy variable methods to complete the construction of a tree height curve model of Chinese fir under the near-natural transformation mode of Chinese fir and Fujian Phoebe, providing a basic method support for estimating the tree height of Chinese fir under the near-natural transformation mode of Chinese fir and the subsequent estimation of Chinese fir stock and timber output, and screening suitable near-natural transformation modes of Chinese fir and Fujian Phoebe.

[0095] Example 3 This embodiment provides a Chinese fir tree height curve prediction system. For details not disclosed in the Chinese fir tree height curve prediction system of this embodiment, please refer to the specific implementation content of the Chinese fir tree height curve prediction method in other embodiments.

[0096] Figure 6 Schematic diagram of the structure of the Chinese fir tree height curve prediction system according to an embodiment of the present application is shown in FIG.

[0097] like Figure 6 As shown, the Chinese fir tree height curve prediction system includes a data set module 10 , a basic model module 20 , a multivariate tree height curve model 30 , an optimal Chinese fir tree height curve model module 40 and a tree height prediction module 50 .

[0098] Specifically described, Data set module 10, used to obtain survey data of near-natural transformation forest stands of Cunninghamia lanceolata and Phoebe fujianensis to obtain model building data and test data; The basic model module 20 is used to determine multiple tree height curve models as candidate equations for the Chinese fir tree height curve model based on the growth law of trees and the characteristics of the model; and to fit the multiple tree height curve models respectively and compare the model fitting effects, and to perform a first evaluation of the model to determine the basic tree height curve model; A multivariate tree height curve model 30 is used to determine effective variables that reflect environmental factors and influence DBH growth under a near-natural transformation model, and add the effective variables to the basic tree height curve model to obtain a multivariate tree height curve model; The optimal Chinese fir tree height curve model module 40 is used to construct a dummy variable tree height curve model based on the multivariate tree height curve model based on the dummy variable method; and determine the optimal Chinese fir tree height curve model through the second evaluation formula of the model; The prediction module 50 is used to predict the height of Chinese fir trees by using the optimal Chinese fir tree height curve model.

[0099] The fir tree height curve prediction system of the present application is adopted, and a basic model module 20 is used to combine the growth law of trees and the characteristics of the model to determine multiple tree height curve models as alternative equations for the fir tree height curve model; and multiple tree height curve models are fitted respectively and the model fitting effects are compared, and the first model evaluation is performed at the same time to determine the basic tree height curve model; the multivariate tree height curve model 30 determines the effective variables that reflect environmental factors and the influence of breast diameter growth under the near-natural transformation model, and the effective variables are added to the basic tree height curve model to obtain the multivariate tree height curve model; the optimal fir tree height curve model module 40 constructs a dummy variable tree height curve model based on the dummy variable method according to the multivariate tree height curve model; the optimal fir tree height curve model is determined through the second model evaluation formula; finally, the tree height prediction module 50 predicts the fir tree height through the optimal fir tree height curve model.

[0100] This application uses a combination of multivariate and dummy variable methods to complete the construction of a tree height curve model for Chinese fir under the near-natural transformation model of Chinese fir and Fujian Phoebe, providing basic methodological support for estimating the tree height of Chinese fir under the near-natural transformation model of Chinese fir and the subsequent estimation of Chinese fir stock and timber output, and screening suitable near-natural transformation models of Chinese fir and Fujian Phoebe.

[0101] Example 4 This embodiment provides a device for predicting a height curve of a Chinese fir tree. For details not disclosed in the device for predicting a height curve of a Chinese fir tree in this embodiment, please refer to the specific implementation contents of the method or system for predicting a height curve of a Chinese fir tree in other embodiments.

[0102] Figure 7 Schematic diagram of the structure of a Chinese fir tree height curve prediction device 400 according to an embodiment of the present application is shown in FIG.

[0103] like Figure 7 As shown, the Chinese fir tree height curve 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 tree height curve prediction method or image recognition method.

[0104] Those skilled in the art will understand that Figure 7It is merely an example of the Chinese fir tree height curve prediction device 400 and does not constitute a limitation of the Chinese fir tree height curve prediction device 400. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the Chinese fir tree height curve prediction device 400 may also include input and output devices, network access devices, buses, etc.

[0105] 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 gate 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 tree height curve prediction device 400, and uses various interfaces and lines to connect various parts of the entire Chinese fir tree height curve prediction device 400.

[0106] The storage unit 402 can be used to store computer-readable instructions. The processing unit 401 implements various functions of the Chinese fir tree height curve 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 tree height curve 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.

[0107] If the modules integrated in the Chinese fir tree height curve prediction device 400 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.

[0108] Example 5 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 a Chinese fir tree height curve in other embodiments.

[0109] In summary, the use of the Chinese fir tree height curve prediction device and storage medium of the present application includes 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.

[0110] 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.

[0111] 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.

[0112] 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.

[0113] 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".

[0114] 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.

[0115] 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 height curve of Chinese fir based on a near-natural transformation of Chinese fir and Phoebe fujianensis stands, characterized in that: include: Obtain survey data of near-natural transformation of Chinese fir and Phoebe fujianensis stands to obtain model building data and test data; Combining the growth law of trees and the characteristics of the model, multiple tree height curve models are determined as candidate equations for the Chinese fir tree height curve model; and the multiple tree height curve models are fitted respectively and the model fitting effects are compared, and a first evaluation of the model is performed at the same time to determine the basic tree height curve model; Determine effective variables that reflect environmental factors and influence DBH growth under the near-natural transformation model, add the effective variables to the basic tree height curve model, and obtain a multivariate tree height curve model; According to the multivariate tree height curve model, a dummy variable tree height curve model is constructed based on the dummy variable method; through the second evaluation of the model, the optimal Chinese fir tree height curve model is determined; and the Chinese fir tree height is predicted by the optimal Chinese fir tree height curve model.

2. The method for predicting the height curve of Chinese fir according to claim 1, characterized in that: The first evaluation of the model includes using 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. The specific calculation formula is: ； ； ； in, y ij and Respectively i The first j Observed and predicted tree heights for each plant; m is the number of sample plots, n i It is i The number of trees in a plot.

3. The method for predicting the height curve of Chinese fir according to claim 1, characterized in that: Determining the effective variables that reflect environmental factors and the effect on DBH growth under the near-natural transformation model comprises the following steps: Select multiple variables including stand factors and tree species coefficients; the stand factors include at least one of the number of trees per hectare N, the average breast diameter of the stand Dq, the basal area BA and the stand density index SDI; the tree species composition coefficient includes the fir composition coefficient fir; In statistical analysis, independent variables with variance inflation factors (VIF) greater than 5 were eliminated, and variables that were significant in statistical tests were retained to obtain the screened variables; A correlation analysis is performed between the tree height and the screened variables, and based on the basic tree height curve model, at least one variable with a higher correlation is selected as a valid variable.

4. The method for predicting the height curve of Chinese fir according to claim 1, characterized in that: The adding of the effective variables into the basic tree height curve model to obtain a multivariate tree height curve model includes: The basic tree height curve model is the Wykoff model, and the expression of the Wykoff model is: ； Where H and D are tree height (m) and diameter at breast height (cm), respectively; a and b are model parameters to be estimated; The effective variables are site index, stand conditions and tree species composition; the general expression of the multivariate tree height curve model GHD1 is as follows: ； Among them, H ij For the i Sample site j Tree height, DBH ij For the i Sample site j Diameter at breast height of a tree; SI i For the i Site index of each plot; For the i The stand density index of each plot; fir% i It is i The fir composition coefficient of each plot; ɑ0, ɑ1, ɑ2, b1, b2 are the parameters to be estimated in the model; And considering that the three effective variables all act on the parameter ɑ0, the expression of the model form GHD2 is: ； Among them, ɑ0, ɑ1, ɑ2, ɑ3, and b1 are the parameters to be estimated for the model; Finally, the optimal model GHD is determined for subsequent fitting based on whether the model parameters are significant and the model fitting effect.

5. The method for predicting the height curve of Chinese fir according to claim 4, characterized in that: The method of constructing a dummy variable tree height curve model based on the multivariate tree height curve model and the dummy variable method comprises: Based on the optimal model GHD, dummy variables Z are added to different parameters. i ( i =1, 2, 3), we get the dummy variable model DGHD1 acting on parameter a and the dummy variable model DGHD2 acting on parameter b; The expression of DGHD1 is: ； The expression of DGHD2 is: ； Among them, ɑ0, ɑ1, ɑ2, b1, b2, K1, k2, and k3 are the model parameters to be estimated.

6. The method for predicting the height curve of Chinese fir according to claim 1, characterized in that: The method of determining the optimal fir tree height curve model through the second model evaluation includes: Compare the Akaike information criterion AIC value and the Bayesian information criterion BIC value of the model. The smaller the AIC value and BIC value, the better. At the same time, consider the log-likelihood value Loglik. The larger the value, the better. ； ； in, p is the number of model parameters, n is the number of sample plants, logn is the logarithm of n; logLik is the logarithm of likelihood.

7. The method for predicting the height curve of Chinese fir according to claim 1 or 6, characterized in that: The second evaluation of the model also includes evaluating the predictive ability of the model, specifically including: The model test index is calculated using the observed values ​​and predicted values ​​of a certain number of sample strains to evaluate the prediction ability of the model; the model test index includes the absolute coefficient R 2 , total relative error (TRE) and mean absolute error (MAE). The specific calculation formula is: ； ； ； in, y ij and Respectively i The first j Observed and predicted tree heights for each plant; m is the number of sample plots, n i It is i The number of trees in a plot.

8. A tree height curve prediction system for Chinese fir based on a near-natural transformation of Chinese fir and Phoebe fujianensis stands, characterized in that: include: The data set module is used to obtain the survey data of the near-natural transformation of Cunninghamia lanceolata and Phoebe fujianensis stands to obtain model building data and test data; The basic model module is used to determine multiple tree height curve models as candidate equations for the Chinese fir tree height curve model based on the growth law of trees and the characteristics of the model; and to fit the multiple tree height curve models respectively and compare the model fitting effects, and at the same time perform a first evaluation of the model to determine the basic tree height curve model; A multivariate tree height curve model is used to determine effective variables that reflect environmental factors and influence breast diameter growth under a near-natural transformation model, and the effective variables are added to the basic tree height curve model to obtain a multivariate tree height curve model; The optimal Chinese fir tree height curve model module is used to construct a dummy variable tree height curve model based on the multivariate tree height curve model based on the dummy variable method; and determine the optimal Chinese fir tree height curve model through the second evaluation formula of the model; A prediction module is used to predict the height of Chinese fir trees using the optimal Chinese fir tree height curve model.

9. A device for predicting the height curve of a Chinese fir tree, 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 7.

10. 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 7.

Images