A method for predicting carbon accumulation of individual growth of forest trees in post-harvest secondary forest restoration of moso bamboo

By constructing a multiple linear regression prediction model, the problem of predicting carbon accumulation in individual tree growth during the restoration of secondary forests after bamboo felling was solved, achieving high-precision carbon accumulation prediction and supporting the evaluation of carbon sequestration efficiency and the achievement of carbon neutrality goals in secondary forest restoration.

CN117689068BActive Publication Date: 2026-11-10ZHEJIANG FORESTRY ACAD +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202311702152.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-12
Publication Date
2026-11-10
Estimated Expiration
2043-12-12

AI Technical Summary

Technical Problem

There is no existing technology for predicting the carbon accumulation of individual trees during the restoration of secondary forests after bamboo felling, which affects the evaluation of the carbon sequestration efficiency of secondary forests and the achievement of carbon neutrality goals.

Method used

By constructing a multiple linear regression prediction model, using the initial data and carbon accumulation rate of surviving trees, and combining the standard deviation standardization method, stepwise regression method, variance expansion factor, Akaike information criterion, and Bayesian information criterion, the optimal carbon accumulation prediction model is determined to predict the carbon accumulation of individual trees during growth.

Benefits of technology

It achieves high-precision prediction of carbon accumulation in individual tree growth, avoids the problem of multicollinearity among independent variables, and provides a scientific basis for evaluating carbon sequestration efficiency and formulating carbon sequestration management technologies in secondary forest restoration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure BDA0004602362170000061
    Figure BDA0004602362170000061
  • Figure BDA0004602362170000071
    Figure BDA0004602362170000071
  • Figure BDA0004602362170000081
    Figure BDA0004602362170000081
Patent Text Reader

Abstract

The application discloses a kind of bamboo felled secondary forest recovery in the prediction method of individual growth carbon accumulation of forest tree. A plurality of variables are selected to establish a multiple linear regression model, first, the variables are dimensionless by standard deviation standardization method, then the optimal carbon accumulation prediction model is determined by stepwise regression method combined with VIF, AIC and BIC: the prediction model determined when VIF is less than 10 and AIC and Bayesian information standard BIC are minimum is the optimal carbon accumulation prediction model; the value of the parameter of the optimal carbon accumulation prediction model is estimated by maximum likelihood method, and the prediction model of individual growth carbon accumulation of forest tree is obtained. Avoid the multicollinearity and over-parameterization problem between independent variables, and obtain the prediction model of individual growth carbon accumulation of forest tree which objectively reflects the key influencing factors of individual growth carbon accumulation of forest tree in bamboo felled secondary forest recovery. The prediction model has high precision, small error and practical application value.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of forest ecology, specifically to a method for predicting carbon accumulation in individual trees during the restoration of secondary forests after bamboo felling. Background Technology

[0002] The ecological impacts of bamboo expansion into adjacent forest communities have attracted significant attention from ecologists, foresters, and government departments both domestically and internationally. To curb bamboo expansion, promote the healthy and stable development of zonal forest communities, and ensure their continued role in carbon sequestration and biodiversity maintenance, secondary forest communities in severely affected forest areas are restored through methods such as felling bamboo to promote natural regeneration or artificially induced regeneration. Secondary forests, as an important component of my country's forest resources, possess strong carbon sequestration capabilities. For example, research by Dong Lili et al. found that the carbon storage of Mongolian oak (Quercus mongolica) secondary forest vegetation was 106.95-126.38 t·hm². -2 The carbon storage of secondary forests is approximately 2-2.4 times that of temperate and northern forest vegetation, and 1.2-2.7 times that of forest vegetation in Northeast China (Dong Lili, Liu Hongmin, Wang Chengcheng, et al. Short-term and long-term effects of thinning on carbon storage in secondary forest ecosystems of Mongolian oak [J]. Journal of Shenyang Agricultural University, 2019, 50(5):614-620). During the restoration of secondary forests, the growth, death, and relocation of individual trees are the main ecological processes affecting the carbon sequestration efficiency of secondary forests. Among them, the growth of individual trees is the direct source of carbon accumulation, which determines the contribution of secondary forests to addressing climate change and contributing to the achievement of carbon neutrality goals.

[0003] Individual tree growth is affected by stand density, stand age, site, and stand competition. For example, Jiang Li et al. studied the effect of stand density index on the diameter growth of individual Chinese fir trees in Shaowu, Fujian (Jiang Li, Zhang Xiongqing, Duan Aiguo et al. Application of different stand density indices in the annual diameter growth model of individual Chinese fir trees [J]. Forestry Science Research, 2022, 35(4):123-129.); Gao Dongqi et al. predicted the sectional area and volume growth of individual Pinus tabuliformis trees based on the Richards equation and found that site, age, stand density and competition have a significant impact on their growth (Gao Dongqi, Deng Huafeng, Jiang Yi et al. Method of coupling individual tree growth model with stand growth model - taking Pinus tabuliformis as an example [J]. Journal of Jiangxi Agricultural University, 2014, 36(6):1298-1304). However, there are no reports on predicting carbon accumulation from individual tree growth. Prediction is a crucial basis for evaluating carbon sequestration efficiency, predicting carbon sequestration potential, and developing carbon sequestration management techniques in secondary forest restoration, directly determining the contribution of secondary forests to climate change mitigation and carbon neutrality. To date, there are no reports on methods for predicting carbon accumulation from individual tree growth in secondary forest restoration after bamboo felling. Summary of the Invention

[0004] To address the technological gaps in the existing technology, this invention provides a method for predicting the carbon accumulation of individual trees during the restoration of secondary forests after bamboo felling. This method has small errors, high accuracy, and practical application value.

[0005] A method for predicting individual tree growth carbon accumulation during the restoration of secondary forests after bamboo felling includes:

[0006] S1. Determine surviving trees: Trees with a diameter at breast height (DBH) ≥ 5 cm at the beginning of the period and still surviving at the end of the period are determined to be surviving trees;

[0007] S2. Using surviving individuals of each tree as the target trees, determine the species, initial data, and final data for each target tree;

[0008] S3. The measured carbon accumulation rate R of the target wood. c Take the logarithm Ln(R) c Using the initial data of the object as the model dependent variable and the initial data of the object as the model independent variable, a multiple linear regression prediction model is constructed according to equation (1):

[0009] Ln(R c )=a0+a1x1+a2x2++……+ a m x m (1);

[0010] In equation (1), a0 is the intercept, a1, a2, ..., a m The parameters are for their respective variables; x1, x2, ..., x m Both are independent variables, namely the initial diameter at breast height (D0) of the target tree and the total number of individuals of the same species within a certain radius r of the target tree (N). c-T The total number of different species within a certain radius r neighborhood of the target tree (N) h-T ), the total chest height cross-sectional area of ​​individuals of the same species within a certain radius r of the target tree (BA) c-T ), the total chest height cross-sectional area (BA) of heterogeneous individuals within a certain radius r neighborhood of the target tree. h-T The total number of individuals of the same species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r neighborhood (N). c The total number of individuals of different species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r neighborhood (N). h The diameter at breast height (DBH) of a certain radius r of the target tree is greater than the total breast height of the same species of individual trees (BA). c ), the cross-sectional area of ​​individuals of different species within a certain radius r of the target tree whose diameter at breast height (DBH) is greater than the total DBH of the target tree (BA). h ), the number of species (S) in a certain radius r neighborhood of the target tree;

[0011] First, the variables were standardized using the standard deviation method to make them dimensionless. Then, the optimal carbon accumulation prediction model was determined by stepwise regression combined with the variance inflation factor (VIF), the Akaike information criterion (AIC), and the Bayesian information criterion (BIC). The prediction model determined when the variance inflation factor is less than 10 and both the Akaike information criterion (AIC) and the Bayesian information criterion (BIC) are minimized is the optimal carbon accumulation prediction model.

[0012] S4. Estimate the parameters a1, a2, ..., a of the optimal carbon accumulation prediction model using the maximum likelihood method. m The value of is used to obtain a predictive model for carbon accumulation in individual tree growth;

[0013] S5. The predicted value of the carbon accumulation rate of individual trees is obtained by using a prediction model of individual tree growth carbon accumulation.

[0014] This invention has discovered that 10 factors are correlated with the growth carbon accumulation of surviving trees: initial diameter at breast height (DBH), total number of individuals of the same species within a certain radius r of the target tree, total number of individuals of different species within a certain radius r of the target tree, total basal area of ​​DBH of individuals of the same species within a certain radius r of the target tree, total basal area of ​​DBH of individuals of different species within a certain radius r of the target tree, total number of individuals of the same species with a DBH greater than the target tree within a certain radius r of the target tree, total number of individuals of different species with a DBH greater than the target tree within a certain radius r of the target tree, total basal area of ​​DBH of individuals of the same species with a DBH greater than the target tree within a certain radius r of the target tree, total basal area of ​​DBH of individuals of different species with a DBH greater than the target tree within a certain radius r of the target tree, and number of species within a certain radius r of the target tree. Therefore, a predictive model for growth carbon accumulation of surviving trees is constructed based on these 10 factors as variables.

[0015] To achieve better invention results, the following optimizations are made:

[0016] In step S3, the neighborhood of the target wood with a certain radius r is the region with the target wood as the center and r as the radius. Among them, r = 2.5m-20m is more suitable, r = 10m-20m is further preferred, and r = 10m is the most preferred.

[0017] In step S3, the measured value of carbon accumulation rate R c Calculate according to formula (2):

[0018] R c =(C ti -C t-1i ) / C t-1i×100% (2);

[0019] In the formula, R c C represents the measured carbon accumulation rate generated by the growth of surviving individual tree i. ti C represents the total carbon sequestration of surviving individual i at the end of period i. t-1i The total carbon sequestration of surviving individual i at the beginning of period.

[0020] The total carbon sequestration C of the surviving individual i of the forest trees at the end of period ti Total carbon sequestration C of surviving individual tree i at the beginning of the period t-1i The calculation formulas are shown in equations (3) and (4) respectively:

[0021] C ti =W ti ×C i (3);

[0022] C t-1i =W t-1i ×C i (4);

[0023] ΔC survivali =C ti -C t-1i (5);

[0024] In the formula, ΔC survivali W represents the measured carbon accumulation produced by the growth of surviving individual tree i. ti W represents the total biomass of surviving individual i at the end of the period. t-1i C represents the total biomass of surviving individual i at the beginning of the period. i denoted as the carbon content coefficient of surviving individual i of the forest tree.

[0025] The total biomass W of the surviving individual i of the forest trees at the end of period i ti The total biomass W of the surviving individual i at the beginning of the period can be calculated using existing methods, such as data on the species of the target tree and its final diameter at breast height (DBH), final tree height, and final crown length, based on existing biomass models for pines, firs, hardwood I, hardwood II, and softwoods. t-1i The calculation is performed using existing methods, such as using data on the species of the target tree and its initial diameter at breast height (DBH), initial tree height, and initial crown length, based on existing biomass models for pines, firs, hardwoods I, hardwoods II, and softwoods. The carbon content coefficient is an existing carbon content coefficient.

[0026] The following data can be obtained by surveying the forest land: diameter at breast height (DBH), tree height, crown length, total number of individuals of the same species within a certain radius r of the target tree, total number of individuals of different species within a certain radius r of the target tree, total number of individuals of the same species with a DBH greater than the target tree within a certain radius r of the target tree, total number of individuals of different species with a DBH greater than the target tree within a certain radius r of the target tree, and number of species within a certain radius r of the target tree.

[0027] The total basal area of ​​the same species within a certain radius r of the target tree, the total basal area of ​​the same species within a certain radius r of the target tree, the total basal area of ​​the same species within a certain radius r of the target tree with a diameter at breast height greater than that of the target tree, and the total basal area of ​​the same species within a certain radius r of the target tree with a diameter at breast height greater than that of the target tree can be calculated using existing calculation methods based on the results of forest land surveys.

[0028] This invention can use all surviving trees as target trees to construct a multiple linear regression prediction model, thereby obtaining a prediction model for the growth carbon accumulation of individual trees applicable to all tree species.

[0029] This invention can also first divide all surviving trees into two groups according to tree species: a coniferous tree group and a broadleaf tree group. Surviving trees in the coniferous tree group are used as target trees, and a multiple linear regression prediction model is constructed to obtain a prediction model applicable to the carbon accumulation of growth in coniferous trees. Similarly, surviving trees in the broadleaf tree group are used as target trees, and a multiple linear regression prediction model is constructed to obtain a prediction model applicable to the carbon accumulation of growth in broadleaf trees.

[0030] The formula for calculating the variance inflation factor (VIF) is shown in equation (6):

[0031] VIF = 1 / (1-R) i 2 (6);

[0032] In the formula, VIF is the variance inflation factor, and R... i Let x be the i-th independent variable i The coefficient of determination between the variable and the other independent variables.

[0033] Both AIC and BIC are calculated using existing calculation methods.

[0034] The model of this invention is validated using the total relative error (Rs), average relative error (E1), absolute value of average relative error (E2), and prediction accuracy (P) to assess its prediction accuracy and performance. These can be calculated using the following formula:

[0035] Total relative error:

[0036] Mean relative error:

[0037] Mean relative absolute value:

[0038] Prediction accuracy:

[0039] In the formula: yi is the measured value, gi is the predicted value, n is the sample size, and t is the measured value. α This represents the t-distribution value at a confidence level of α = 0.05, where T is the number of parameters in the regression model. The average of the predicted values ​​is

[0040] This invention can establish a prediction model based on the carbon accumulation cycle of individual trees that you want to predict. The interval between the beginning and the end of the period is the carbon accumulation cycle of individual trees that you want to predict, which can be used flexibly.

[0041] The present invention has the following beneficial effects:

[0042] This invention establishes a multiple linear regression model using selected variables. First, the variables are standardized using the standard deviation method to make them dimensionless. Then, a stepwise regression method combined with VIF, AIC, and BIC is used to determine the optimal carbon accumulation prediction model: the prediction model determined when VIF is less than 10 and both AIC and the Bayesian information standard (BIC) are minimized is considered the optimal carbon accumulation prediction model. The maximum likelihood method is used to estimate the parameters of the optimal carbon accumulation prediction model, resulting in a prediction model for individual tree growth carbon accumulation. This avoids multicollinearity and overparameterization problems among independent variables, and yields a prediction model for individual tree growth carbon accumulation that objectively reflects the key influencing factors of individual tree growth carbon accumulation in secondary forest restoration after bamboo felling. This provides an important basis for predicting and analyzing carbon sequestration efficiency evaluation, carbon sequestration potential prediction, and the formulation of carbon sequestration management technologies in secondary forest restoration.

[0043] The prediction method of this invention is particularly suitable for predicting the carbon accumulation of individual trees during the restoration of secondary forests after bamboo felling. The prediction model has high accuracy and small error, and can objectively reflect the correlation between dependent and independent variables. It has practical application value and fills the technical gap in the prior art.

[0044] The prediction method of this invention is also applicable to the prediction of carbon accumulation in the growth of individual trees in other forests. Detailed Implementation

[0045] The present invention will be further described in detail below with reference to specific embodiments, so that those skilled in the art can better understand the technical solution of the present invention, but it should not be construed as a limitation of the present invention.

[0046] Example 1

[0047] This invention presents a method for predicting carbon accumulation in individual tree growth during the restoration of secondary forests after bamboo felling in the Huangtan Forest Farm of Pan'an County, Zhejiang Province. The research process is as follows:

[0048] 1. Research Objective

[0049] This study aims to develop a method for predicting carbon accumulation in individual tree growth during the restoration of secondary forests after bamboo felling. This research provides a scientific basis for revealing the influencing factors, dynamic changes, and carbon accumulation potential of individual tree growth during the restoration process, and offers a reference for developing carbon sequestration management techniques for secondary forests.

[0050] 2. Overview of the Study Area

[0051] The study area is located in Huangtan Forest Farm, Pan'an County, Zhejiang Province, at 120°16′–120°26′ east longitude and 28°42′–28°52′ north latitude. The average annual temperature is 14℃–16℃, with the coldest month (January) averaging -0.6℃ to 0.9℃ and the hottest month (July) averaging 23℃–24.4℃. The accumulated temperature ≥10℃ is 2500℃–5100℃, and the frost-free period is 185–200 days. The annual precipitation is 1500 mm–1800 mm, and the annual relative humidity is 79%. It belongs to the mid-subtropical monsoon climate zone, with four distinct seasons, warm climate, abundant rainfall, and suitable sunshine. The soil is mainly yellow soil: yellow soil is generally deep and fertile below 600m altitude; above 600m altitude, it is mostly ash soil, with a soil layer thickness of generally 40cm–80cm. Its forest vegetation belongs to the mid-subtropical evergreen broad-leaved forest, and the vertical zone is the Zhejiang-Fujian hilly area of ​​sweet oak and Michelia champaca vegetation zone. The main forest types include mixed coniferous and broad-leaved forests, evergreen broad-leaved forests, moso bamboo forests and Chinese fir forests. The dominant tree species are mainly Fagaceae species such as Quercus glauca and Castanopsis sclerophylla, Michelia champaca, Pinus huangshanensis, Chinese fir, and moso bamboo. The main plantations are Chinese fir, Pinus huangshanensis, Cryptomeria japonica and moso bamboo.

[0052] 3. Plot survey and data processing

[0053] 3.1 Plot setup and survey

[0054] From April to June 2022, a typical secondary forest restoration area relying on natural regeneration after bamboo harvesting was selected in Huangtan Forest Farm, Pan'an County. A large plot of 100m×100m was set up (altitude 879m, slope 5°, south-facing). The southwest corner of the plot was set as the origin, with coordinates of 30°19'17"N, 119°26'30"E. The east-west direction was the horizontal axis (x), and the north-south direction was the vertical axis (y). The plot was divided into 25 plots of 20m×20m using the adjacent grid method, and each plot was further divided into 5m×5m quadrats. All woody plants (including trees, shrubs, and woody vines) with a diameter at breast height (DBH) ≥ 5cm within the quadrats were investigated and recorded. The tree species name, DBH (D), tree height (H), crown length (L), and relative coordinates (x, y) within the quadrats were investigated and recorded, and each tree was attached with an aluminum numbered tag. In 2023, the sample plots set up and surveyed in 2022 were re-examined. All woody plants (including trees, shrubs and woody vines) with a diameter at breast height (DBH) ≥ 5 cm within the sample plots were surveyed. The tree species name, DBH (D), tree height (H), crown length (L), and relative coordinates within the sample plot were recorded. The importance value of each tree species was calculated according to formula (11), and the average DBH of each tree species was calculated according to formula (12). The forest stand overview of the sample plots is shown in Table 1.

[0055] Importance value (IV) = (relative frequency + relative abundance + relative significance) / 3 (11);

[0056]

[0057] In the formula, D mean The value represents the average diameter at breast height (DBH); Di is the DBH of an individual tree, and n is the total number of trees.

[0058] Table 1. Overview of forest stands in the sample plots

[0059]

[0060] 3.2 Determination of surviving individual trees

[0061] Trees with a diameter at breast height (DBH) ≥ 5 cm at the beginning of the period (2022 survey) and still alive at the end of the period (2023 survey) were defined as surviving trees.

[0062] 3.3 Calculation method for carbon accumulation in individual trees

[0063] Based on data from two phases of surveys, the species (i.e., tree species name), number of trees, diameter at breast height (D), tree height (H), and crown length (L) of surviving trees were determined. This data was used for pine species (including Masson pine, slash pine, loblolly pine, black pine, Huangshan pine, golden larch, etc.), fir species (including Chinese fir, yew, dawn redwood, Japanese cedar, pond cypress, bald cypress, cypress, etc.), and hardwood species (mainly including *Symplocos edulis*, *Zelkova spp.*, *Machilus thunbergii*, *Machilus chinensis*, *Machilus chinensis*, *Cinnamomum camphora*, *Elaeocarpus decipiens*, *Toona sinensis*, *Prunus camphora*, etc.), whose trunk wood density is generally less than 0.7 g / cm³. 3Hardwood II (mainly including species such as Castanopsis sclerophylla, Quercus glauca, Castanopsis sclerophylla, Castanopsis sclerophylla, Ilex chinensis, and Quercus variabilis, whose trunk wood density is generally greater than or equal to 0.7 g / cm³) 3 The biomass models of individual trees (alder, willow, maple, sycamore, sweetgum, sassafras, etc.) and carbon content coefficients (see Table 3) were calculated using formula (2) to obtain the measured values ​​of carbon accumulation rate of individual tree growth. (Jiang Bo. Research on the Theory and Technology of Rapid Construction of Forest Ecosystems [M]. China Forestry Publishing House, 2010)

[0064] Carbon accumulation rate R c =(C ti -C t-1i ) / C t-1i ×100% (2);

[0065] In the formula, C ti C represents the total carbon sequestration of surviving individual tree i at the end of period i (year t). t-1i This represents the total carbon sequestration of surviving individual i at the beginning of period (year t-1). Ci ti C t-1i The calculation formulas are as follows:

[0066] C ti =W ti ×C i (3);

[0067] C t-1i =W t-1i ×C i (4);

[0068] In the formula, W ti W represents the total biomass of surviving individual i trees at the end of period i (year t). t-1i C represents the total biomass of surviving individual i at the beginning of period (year t-1). i denoted as the carbon content coefficient of surviving individual i of the forest tree.

[0069] Calculate the measured values ​​of individual tree growth carbon accumulation using formula (5):

[0070] ΔC survivali =C ti -C t-1i (5);

[0071] In the formula, ΔC survivali The measured value of carbon accumulation generated by the growth of surviving individual tree i.

[0072] Table 2. Individual tree biomass models for each tree species group

[0073]

[0074] Table 3 Carbon content coefficients of various tree species

[0075]

[0076]

[0077] 3.4 Carbon Accumulation Prediction Methods

[0078] i) Predictive model construction

[0079] The target trees were surviving individuals with an initial diameter at breast height (D0) ≥ 5 cm. The measured carbon accumulation rate R of the target trees was used as the target trees. c Take the logarithm Ln(R) c Using the initial data of the object as the model dependent variable and the initial data of the object as the model independent variable, a multiple linear regression prediction model is constructed according to equation (1):

[0080] Ln(R c )=a0+ a1x1+ a2x2+……+ a m x m (1);

[0081] In the formula, a0 is the intercept, a1, a2, ..., a m For parameters; x1, x2, ..., x m All are independent variables, namely the initial diameter at breast height (D0) of the target tree, the total number of individuals of the same species in the neighborhood of a certain radius r (10m) of the target tree (N). c-T The total number of different species within a certain radius r (10m) of the target tree (N) h-T ), the total chest height cross-sectional area (BA) of individuals of the same species within a certain radius r (10m) of the target tree. c-T ), the total chest height cross-sectional area (BA) of different species within a certain radius r (10m) of the target tree. h-T The total number (N) of individuals of the same species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree. c The total number (N) of individuals of different species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree. h ), the cross-sectional area (BA) of individuals of the same species within a certain radius r (10m) of the target tree whose diameter at breast height (DBH) is greater than the total DBH of the target tree. c ), the total basal area (BA) of heterogeneous individuals with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) neighborhood. h ) and the number of species (S) in a certain radius r (10m) of the target tree.

[0082] First, the variables are standardized using the standard deviation method to make them dimensionless. Then, a stepwise regression method is used in conjunction with the variance inflation factor (VIF), the Akaike information criterion, and the Bayesian information criterion to determine the optimal carbon accumulation prediction model: the prediction model determined when the VIF is less than 10 and both the Akaike information criterion and the Bayesian information criterion are minimized is considered the optimal carbon accumulation prediction model. To avoid multicollinearity and overparameterization among the independent variables, the VIF is used to test for multicollinearity among the model's independent variables. Generally, a VIF < 10 indicates relatively low multicollinearity among the independent variables.

[0083] ii) Parameter solving and model verification

[0084] The parameters a1, a2, ..., a of the optimal carbon accumulation prediction model are estimated using the maximum likelihood method. m The model prediction accuracy and effectiveness were verified using the total relative error (Rs), average relative error (E1), absolute value of average relative error (E2), and prediction accuracy (P), thus obtaining a prediction model for carbon accumulation in individual forest trees.

[0085] The predicted value of the carbon accumulation rate of individual trees was obtained by using a prediction model of individual tree growth carbon accumulation.

[0086] The formula for calculating the variance inflation factor is shown in equation (6).

[0087] VIF = 1 / (1-R) i 2 (6);

[0088] In the formula, VIF is the variance inflation factor, and R... i Let x be the i-th independent variable i The coefficient of determination between the variable and the other independent variables.

[0089] The formulas for calculating the total relative error (Rs), average relative error (E1), absolute value of average relative error (E2), and prediction accuracy (P) are as follows:

[0090] Total relative error:

[0091] Mean relative error:

[0092] Mean relative absolute value:

[0093] Prediction accuracy:

[0094] In the formula: yi is the measured value, gi is the predicted value, n is the sample size, and t is the measured value. α This represents the t-distribution value at a confidence level of α = 0.05, where T is the number of parameters in the regression model. The average of the predicted values ​​is

[0095] 3.5 Data Analysis

[0096] The calculation of carbon accumulation in individual tree growth was completed in Excel, the prediction of carbon accumulation in individual tree growth was completed in DPS software, and the analysis of factors affecting carbon accumulation in individual tree growth was completed in R software, in order to improve work efficiency and avoid human error.

[0097] 4 Results Analysis

[0098] 4.1 Distribution characteristics of tree species with large diameter growth

[0099] The survey and analysis of the secondary forests following the felling of moso bamboo in the Huangtan Forest Farm in 2022-2023 yielded the following results (see Table 4): 1974 surviving trees belonging to 32 species were recorded in 2022-2023. The average diameter at breast height (DBH) of the surviving individuals for each species is recorded as the average diameter. Among the surviving individuals, the average diameter of Japanese cedar was the largest in 2022, reaching 38.9 cm, followed by cypress (27.5 cm), dawn redwood (23.8 cm), Chinese fir (17.2 cm), sassafras (15.5 cm), and winter cherry (11.4 cm). The average diameter of the remaining species ranged from 5.1 to 9.8 cm. In 2023, Japanese cedar still had the largest average diameter, reaching 39.9 cm, followed by cypress (28.0 cm), dawn redwood (24.5 cm), Chinese fir (17.9 cm), sassafras (16.5 cm), winter cherry (12.3 cm), and sweetgum (10.7 cm). The average diameter of the remaining species ranged from 5.6 to 7.9 cm. The average diameter of all surviving individuals increased from 15.4 cm to 16.1 cm in 2022-2023, with a diameter growth rate of 4.5%. Among the dominant tree species, Castanopsis fargesii had the highest diameter growth rate at 9.8%, followed by Zelkova serrata (9.3%), Liquidambar formosana (9.2%), Toona sinensis var. chinensis (8.6%), Prunus camphora (7.9%), Cinnamomum camphora (6.8%), and Sassafras japonica (6.5%). The average diameter growth rate of the remaining tree species was less than 6%, and the average diameter growth rate of the other 20 tree species was 5.5%.

[0100] Table 4. Diameter growth rate of surviving individuals

[0101]

[0102] Tree species distribution characteristics of carbon accumulation rate

[0103] Based on the survey and carbon sequestration calculation of the secondary forests after bamboo felling in Huangtan Forest Farm in 2022-2023 (see Table 5), among the surviving individuals, Chinese fir had the largest carbon sequestration in 2022, reaching 39671.49 kg, followed by Japanese cedar (7194.58 kg), cypress (4728.19 kg), dawn redwood (3256.45 kg), sassafras (2470.74 kg), and liquidambar (1969 kg). The carbon sequestration of the top 20 non-dominant tree species in 2023 was 1519.86 kg, with the highest being 46 kg for *Cephalotaxus fortunei* (43294.9 kg) and 1581.28 kg for *Prunus cerasifera* (1581.28 kg). The ranking of carbon sequestration in 2023 remained consistent with 2022, with *Cephalotaxus fortunei* having the highest at 43294.9 kg, followed by *Taxus chinensis* (130.5 kg) and the lowest at 130.5 kg. The total carbon sequestration of the remaining 20 non-dominant tree species was 1745.6 kg. The carbon accumulation rate of all surviving individuals in 2022-2023 was 9.4%. Among the dominant tree species, *Cephalotaxus fortunei* had the highest carbon accumulation rate at 21.9%, followed by *Liquidambar formosana* (21.4%), *Toona sinensis* (21.2%), *Zelkova serrata* (20%), *Prunus cerasifera* (19%), *Cinnamomum zaengense* (16.5%), and *Sassafras chinensis* (14.1%). *Cypressus cypress* had the lowest rate at only 3.9%, and the remaining 20 non-dominant tree species had a carbon accumulation rate of 14.9%.

[0104] Table 5 Carbon accumulation rate of surviving individuals

[0105]

[0106] 4.3 Predictive Model and Accuracy for Carbon Accumulation in Surviving Individuals

[0107] To ensure the predictive effectiveness of the model, based on the biological characteristics of surviving individual tree species, three carbon accumulation prediction models were established according to the "3.4 Carbon Accumulation Prediction Method" for coniferous, broadleaf, and all tree species. The independent variable parameters and test indicators of the models are shown in Table 6.

[0108] The independent variables in the prediction models for individual growth carbon accumulation of coniferous, broadleaf, and all tree species include the initial diameter at breast height (D0) of the target tree, the total number of individuals of the same species in the neighborhood of a certain radius r (10m) of the target tree (N). c-T The total number of different species within a certain radius r (10m) of the target tree (N) h-T ), the total chest height cross-sectional area (BA) of individuals of the same species within a certain radius r (10m) of the target tree. c-T ), the total chest height cross-sectional area (BA) of different species within a certain radius r (10m) of the target tree. h-T The total number (N) of individuals of the same species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree. c The total number (N) of individuals of different species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree. h), the cross-sectional area (BA) of individuals of the same species within a certain radius r (10m) of the target tree whose diameter at breast height (DBH) is greater than the total DBH of the target tree. c ), the total basal area (BA) of heterogeneous individuals with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) neighborhood. h ) and the number of species (S) in a certain radius r (10m) of the target tree.

[0109] Referring to Table 6, the total relative error (Rs), average relative error (E1), and absolute value of average relative error (E2) of the carbon accumulation prediction models for coniferous, broadleaf, and all tree species are all low, while the prediction accuracy (P) reaches 81.2%, 85.1%, and 79.5%, respectively. This indicates that the carbon accumulation prediction method for individual tree growth of the present invention has good predictive effect and can reflect the influence level of various factors on carbon accumulation. In particular, after dividing all tree species into coniferous and broadleaf categories, the carbon accumulation prediction models established for coniferous and broadleaf tree species respectively have higher accuracy than the carbon accumulation prediction model established directly for all tree species.

[0110] 4.4 Analysis of factors influencing carbon accumulation in surviving individuals

[0111] Table 6 Model parameters a0-a 10 The size and sign of D0 and N reflect the effect and degree of influence of carbon accumulation in the growth of individual coniferous, broad-leaved, and all tree species. c-T N h-T The larger the absolute value of the parameters of the 10 independent variables, the greater the influence of the independent variable on carbon accumulation during growth. Negative values ​​indicate that the larger the value of the independent variable, the lower the carbon accumulation rate and the lower the carbon fixation capacity; conversely, positive values ​​indicate that the larger the value of the independent variable, the higher the carbon accumulation rate and the stronger the carbon fixation capacity.

[0112] Among the main factors influencing carbon accumulation in surviving individuals of coniferous trees, the total number of individuals of the same species (N) within a certain radius r (10m) of the target tree is one of the key factors. c-T ), the total chest height cross-sectional area (BA) of individuals of the same species within a certain radius r (10m) of the target tree. c-T The total number (N) of individuals of the same species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree. c ) and the total cross-sectional area (BA) of individuals of the same species within a certain radius r (10m) of the target tree whose diameter at breast height is greater than that of the target tree. c ) exhibits an inhibitory effect, meaning that the larger the value, the lower the carbon accumulation rate; while the total number of heterogeneous individuals (N) within a certain radius r (10m) neighborhood of the target tree h-T ), the total chest height cross-sectional area (BA) of different species within a certain radius r (10m) of the target tree. h-T The total number (N) of individuals of different species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree.h ), the total basal area (BA) of heterogeneous individuals with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) neighborhood. h The parameters corresponding to the number of species (S) in a certain radius r (10m) of the target tree are positive, indicating a positive impact. However, the corresponding parameters are relatively low. The main reason is that coniferous species such as cypress, fir, Japanese cedar, and dawn redwood are mostly tall trees and are in the dominant layer. The broad-leaved species under the forest are mostly regenerated species and are in the sub-dominant layer. They have no significant impact on the growth of coniferous species and do not show any inhibitory effect. Moreover, they are conducive to maintaining soil and water in the forest and promoting the growth of coniferous species in the upper layer.

[0113] For broadleaf tree species, among the factors influencing carbon accumulation in surviving individuals, the total number of individuals of the same species (N) within a certain radius r (10m) of the target tree is one of the factors. c-T ), the total chest height cross-sectional area (BA) of individuals of the same species within a certain radius r (10m) of the target tree. c-T The total number of different species within a certain radius r (10m) of the target tree (N) h-T ), the total chest height cross-sectional area (BA) of different species within a certain radius r (10m) of the target tree. h-T The total number (N) of individuals of the same species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree. c ), the cross-sectional area (BA) of individuals of the same species within a certain radius r (10m) of the target tree whose diameter at breast height (DBH) is greater than the total DBH of the target tree. c The total number (N) of individuals of different species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree. h ), the total basal area (BA) of heterogeneous individuals with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) neighborhood. h The parameters corresponding to the number of species (S) within a certain radius r (10m) of the target tree are both negative, indicating an inhibitory effect. This suggests that not only do dominant coniferous species inhibit the growth of surviving broadleaf species, but also broadleaf species in the sub-dominant layer, due to competition for nutrients such as light and water with growth space, significantly inhibit the carbon accumulation of surviving broadleaf species. Furthermore, the number of individuals of the same species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree (N) is also negative. c The total number (N) of individuals of different species with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) of the target tree. h ), the total basal area (BA) of heterogeneous individuals with a diameter at breast height (DBH) greater than that of the target tree within a certain radius r (10m) neighborhood. hThe parameters corresponding to the negative values ​​and relatively large values ​​indicate that the inhibitory effect on the growth carbon accumulation of surviving individuals of broad-leaved trees is stronger than the inhibitory effect of all surviving individuals within the neighborhood radius on the growth carbon accumulation of surviving individuals of broad-leaved trees. In addition, the more neighboring species (S) a surviving individual has, the more intense the competition and the slower the growth carbon accumulation.

[0114] Among the factors influencing carbon accumulation in surviving individuals of all tree species, the inhibitory and promoting factors did not exhibit a clear classification pattern, primarily related to the different biological characteristics of the surviving individuals. However, regardless of coniferous, broadleaf, or all tree species, the initial diameter at breast height (DBH) D0 of the target tree consistently showed an inhibitory effect on carbon accumulation, and this parameter value was the largest among all 10 influencing factors. This indicates that after the initial DBH reaches 5 cm, the forest develops into a middle-aged or young-aged forest, with rapid growth. In the near-mature, mature, or over-mature forests, the larger the DBH, the slower the growth and the lower the carbon accumulation rate. Furthermore, the constant term a0 indicates that carbon accumulation in surviving individuals is influenced not only by D0 and N... c-T N h-T In addition to the influence of 10 factors, it may also be related to site conditions, climate or soil, and the combined effect will be positive during 2022-2023.

[0115] In summary, based on the prediction method of carbon accumulation by surviving individuals in the restoration of secondary forests after bamboo felling, we can scientifically carry out secondary forest carbon sequestration management practices for different forest types, tree species structures, and developmental stages to promote the healthy and stable development and growth of secondary forests, enhance their carbon sequestration capacity, and help achieve carbon neutrality as soon as possible.

[0116] Table 6. Parameters and test indices for each variable in the carbon accumulation prediction model for surviving individuals.

[0117]

[0118]

Claims

1. A method for predicting carbon accumulation in individual tree growth during the restoration of secondary forests after bamboo felling, characterized in that, include: S1. Determine surviving trees: Trees with a diameter at breast height (DBH) ≥ 5 cm at the beginning of the period and still surviving at the end of the period are determined to be surviving trees; S2. Using surviving individuals of each tree as the target trees, determine the species, initial data, and final data for each target tree; S3. The measured carbon accumulation rate R of the target wood. c Take the logarithm Ln(R) c Using the initial data of the object as the model dependent variable and the initial data of the object as the model independent variable, a multiple linear regression prediction model is constructed according to equation (1): Ln(R c )=a0+a1x1+a2x2+……+ a m x m (1); In equation (1), a0 is the intercept, a1, a2, ..., a m The parameters are for their respective variables; x1, x2, ..., x m All are independent variables, namely: initial diameter at breast height (DBH) of the target tree, total number of individuals of the same species in a certain radius r neighborhood of the target tree, total number of individuals of different species in a certain radius r neighborhood of the target tree, total basal area of ​​DBH of individuals of the same species in a certain radius r neighborhood of the target tree, total basal area of ​​DBH of individuals of different species in a certain radius r neighborhood of the target tree, total number of individuals of the same species with a DBH greater than that of the target tree in a certain radius r neighborhood of the target tree, total number of individuals of different species with a DBH greater than that of the target tree in a certain radius r neighborhood of the target tree, total basal area of ​​DBH of individuals of the same species with a DBH greater than that of the target tree in a certain radius r neighborhood of the target tree, total basal area of ​​DBH of individuals of different species with a DBH greater than that of the target tree in a certain radius r neighborhood of the target tree, and number of species in a certain radius r neighborhood of the target tree. First, the standard deviation standardization method was used to make the variables dimensionless. Then, the stepwise regression method was used in combination with the variance inflation factor, the Akaike information criterion and the Bayesian information criterion to determine the optimal carbon accumulation prediction model: the prediction model determined when the variance inflation factor is less than 10 and both the Akaike information criterion and the Bayesian information criterion are minimized is the optimal carbon accumulation prediction model. S4. Estimate the parameters a1, a2, ..., a of the optimal carbon accumulation prediction model using the maximum likelihood method. m The value of is used to obtain a predictive model for carbon accumulation in individual tree growth; S5. The predicted value of the carbon accumulation rate of individual trees is obtained by using the prediction model of individual tree growth carbon accumulation; In step S3, the measured value R of the carbon accumulation rate c Calculate according to formula (2): R c =(C ti -C t-1i ) / C t-1i ×100% (2); In the formula, R c C represents the measured carbon accumulation rate generated by the growth of surviving individual tree i. ti C represents the total carbon sequestration of surviving individual i at the end of period i. t-1i The total carbon sequestration of surviving individual i of the forest tree at the beginning of period; The total carbon sequestration C of the surviving individual i of the forest trees at the end of period ti Total carbon sequestration C of surviving individual tree i at the beginning of the period t-1i The calculation formulas are shown in equations (3) and (4) respectively: C ti =W ti ×C i (3); C t-1i =W t-1i ×C i (4); In the formula, W ti W represents the total biomass of surviving individual i at the end of the period. t-1i C represents the total biomass of surviving individual i at the beginning of the period. i denoted as the carbon content coefficient of surviving individual i of the forest tree.

2. The prediction method according to claim 1, characterized in that, In step S3, the neighborhood of the target wood with a certain radius r is the region with the target wood as the center and r as the radius, where r = 2.5m-20m.

3. The prediction method according to claim 1 or 2, characterized in that, In step S3, r = 10m - 20m.

4. The prediction method according to claim 1, characterized in that, The total biomass W of the surviving individual i of the forest trees at the end of period i ti The biomass of the target trees and the data on the diameter at breast height, tree height, and crown length at the end of the period were calculated based on the biomass models of pines, firs, hardwoods I, hardwoods II, and softwoods. The total biomass W of the surviving individual i of the forest trees at the beginning of period t-1i The biomass of the target trees, along with initial diameter at breast height (DBH), initial tree height, and initial crown length data, were calculated using biomass models for pines, firs, hardwoods I, hardwoods II, and softwoods.

5. The prediction method according to claim 1, characterized in that, By treating all surviving trees as target trees, a multiple linear regression prediction model was constructed to obtain a prediction model for the growth carbon accumulation of individual trees applicable to all tree species.

6. The prediction method according to claim 1, characterized in that, All surviving trees were first divided into two groups according to tree species: coniferous tree species group and broadleaf tree species group. All surviving trees in the coniferous tree group were used as target trees. A multiple linear regression prediction model was constructed to obtain a prediction model for the growth carbon accumulation of individual trees in coniferous tree species. By using surviving individuals of broadleaf tree species as target trees, a multiple linear regression prediction model was constructed to obtain a prediction model applicable to the growth carbon accumulation of individual broadleaf tree species.

7. The prediction method according to claim 1, characterized in that, The model is validated using total relative error, average relative error, absolute value of average relative error, and prediction accuracy to assess its prediction accuracy and effectiveness.

Citation Information

Patent Citations

  • Country ecosystem carbon reserve prediction method

    CN116151413A

  • Method and system for wood harvest and storage, carbon sequestration and carbon management

    US20220374912A1