A method for evaluating forest carbon storage potential based on site factors

Abstract

This invention provides a method for evaluating forest carbon storage potential based on site factors. It employs the Mitscherlich equation to construct the relationship between average diameter at breast height (DBH) and corresponding survey time intervals, thus inversely determining stand age. Based on the growth of average stand carbon storage, site quality grades are classified, and a graded growth model for carbon storage is established. After differentiation, stand age is used as a predictor variable to determine the annual growth rate of carbon storage. Based on a general linear regression model, a backward elimination method is used to screen out factors affecting carbon sequestration rates, quantifying the direction and extent of the influence of stand factors and environmental drivers on forest carbon storage potential. Furthermore, the differences in the influence of tree species on carbon sequestration potential are explored based on site factors. This invention establishes a linear relationship between annual carbon storage growth and stand factors and environmental drivers, providing a new method for quantitatively assessing the carbon sequestration potential of suitable afforestation sites at a future time under climate change, and providing technical support for formulating appropriate afforestation plans and carbon reduction policies.

Description

Technical Field

[0001] This invention belongs to the field of forestry and specifically proposes a method for evaluating the potential of forest carbon storage based on site factors. Background Technology

[0002] Forests, as the largest carbon sink in terrestrial ecosystems, possess complex ecological functions and engage in frequent carbon exchange with the atmosphere. They play an irreplaceable role in mitigating global climate change and maintaining regional, national, and even global carbon balance. Against the backdrop of current environmental degradation and global warming, forests have garnered significant attention due to their enormous carbon sequestration capacity. Carbon absorption by forest vegetation is currently a crucial pathway to reducing atmospheric CO2 concentrations. The Kyoto Protocol, the Paris Agreement, and various IPCC reports have all explicitly identified afforestation and the proper management and tending of existing forest stands as important means to increase forest carbon storage and enhance forest carbon sequestration capacity. Forest carbon storage not only reflects the structure and function of forest ecosystems but is also a vital indicator for assessing forest quality and a crucial parameter for measuring the carbon balance of forest ecosystems. Accurately estimating the spatiotemporal distribution and dynamic changes of forest carbon storage under climate change is of substantial significance for formulating China's emission reduction policies, achieving afforestation and carbon sequestration targets, and participating in international climate negotiations.

[0003] Forest carbon storage is the result of continuous material cycling and energy flow between trees and their environment during tree growth and development, and is closely related to site quality. Site quality is the sum of tree species characteristics and all abiotic environmental factors. The better the site quality, the more suitable the forest stand is for growth, the more frequent the material cycling with the outside world, and the greater the forest carbon storage density. For forested land, its carbon storage potential can be predicted by constructing a carbon storage growth model that combines factors such as temperature, precipitation, and stand age. However, for suitable afforestation land, since there is no vegetation growth and no tree measurement factors such as diameter at breast height (DBH) and stand age, it can be regarded as any stand. The carbon sequestration potential of all tree species types on a fixed forest site can only be assessed based on environmental factors on that site when they grow to a certain condition (requiring a certain stand age or DBH). In recent years, in order to adapt to global climate change and achieve carbon sequestration targets, my country has intensified its afforestation efforts. Estimating the carbon sequestration potential of newly afforested biomass carbon pools and its contribution to future forest carbon sinks in China will help formulate appropriate afforestation plans and carbon emission reduction policies, and provide data support for the implementation of my country's forest resource management, ecological and environmental protection policies. Summary of the Invention

[0004] To address the shortcomings of existing methods, this invention provides a method for evaluating forest carbon storage potential based on site factors. This method uses only diameter at breast height (DBH) data from multiple consecutive surveys to estimate stand age and classify site quality. It then establishes a cluster of carbon storage growth curves reflecting site quality differences, thereby estimating carbon sequestration rates under different site qualities and constructing a linear relationship between these rates and stand factors and environmental drivers. This provides a new method for quantitatively assessing the carbon sequestration potential of suitable afforestation sites at a future time under climate change, contributing to a comprehensive evaluation of the carbon sequestration potential of China's forest ecosystems and providing technical support for formulating appropriate afforestation plans and carbon reduction policies.

[0005] A method for evaluating forest carbon storage potential based on site factors includes the following steps: The Mitscherlich equation is used to construct the relationship between the average diameter at breast height (DBH) of the stand and the corresponding survey time interval, and the stand age is determined by inverse calculation; based on the growth of the average wood carbon storage of the stand, site quality grades are classified, a carbon storage graded growth model is established, and after differentiation, the stand age is used as a predictor variable to determine the annual growth of carbon storage; based on a general linear regression model, a backward elimination method is used to screen out factors that affect the carbon sequestration rate, quantify the direction and degree of influence of stand factors and environmental driving forces on forest carbon storage potential, and explore the differences in the influence of tree species on carbon sequestration potential based on site factors.

[0006] The specific technical solution is as follows:

[0007] Step 1: Select sample trees from multiple consecutive surveys within the sample plot, reorganize the sample plot, and match abiotic environmental factors;

[0008] Step 2: Based on the constrained Mitscherlich growth equation, establish the relationship between the average diameter at breast height (DBH) of the stand and the corresponding survey time interval, and then determine the stand age by inverse calculation based on the average DBH of the stand.

[0009] Step 3: Classify site quality grades based on the average stand carbon storage.

[0010] Step 4: Based on the Richards growth equation, establish a carbon storage hierarchical growth model, determine the annual carbon storage growth equation, use forest age as the predictor variable, and calculate the carbon sequestration rate at a certain time point for each sample plot, i.e., the annual carbon storage growth.

[0011] Step 5: Use the backward elimination method to screen out the factors that affect the carbon fixation rate;

[0012] Step 6: Based on a general linear regression model, quantify the direction and extent of the influence of stand factors and environmental driving forces on forest carbon storage potential, and explore the differences in the influence of tree species on carbon sequestration potential based on site factors.

[0013] Specifically:

[0014] Step one involves selecting sample trees from multiple consecutive surveys within the sample plot, including the following steps:

[0015] The basic data consists of continuous survey data at fixed intervals, starting from the first survey period and ending with the last survey period, for a total of k survey periods. For each sample plot, corresponding geographical factors, topographic factors, soil factors, and climatic factors (collectively referred to as environmental drivers) must be obtained. Within each survey period, the species type and measured diameter at breast height (DBH) of all individual trees within the sample plot are recorded. k surviving trees from all survey periods are selected, resulting in k measured DBH data points. It is required that at least 5 trees of the same species are retained. When multiple constituent tree species exist within a sample plot, they are classified into different sample plots according to species type, and a unique number is generated by combining the species number and the sample plot number to reorganize the sample plot. Individual tree carbon storage (aboveground carbon storage + belowground carbon storage) is calculated using the standing biomass model and carbon storage measurement parameters according to forestry industry standards. The average DBH of the reorganized sample plot is the square mean DBH of all retained trees, and the sample plot carbon storage is the arithmetic mean of the carbon storage of all retained trees, i.e., the average stand carbon storage, calculated using the following formula:

[0016]

[0017]

[0018] In equations (1) and (2): n is the number of trees retained in the reconstituted plot. To preserve the diameter at breast height of the timber, To preserve charcoal reserves.

[0019] Step two, which involves reverse-engineering the age of the forest stand, includes the following steps:

[0020] Step 2-1: Calculate the annual growth rate S based on the average diameter at breast height of the stand in the first and last two periods, arrange them in descending order, select an appropriate number of classes according to the sample size, use ordered sample clustering for initial classification, and obtain the number of growth levels M. The ordered sample clustering algorithm uses the optimal segmentation method for classification, and its minimum error function recursive formula is as shown in equation (4).

[0021]

[0022]

[0023] In equations (3) to (4): This represents the average diameter at breast height (DBH) of the initial sample plot. The average diameter at breast height (DBH) of the sample plot in the kth period (the last period) is... The time interval between the first and last survey periods. The loss function representing classification. This represents the classification scheme, where M is the number of classification levels, N is the number of sample plots, and X is the number of sample plots under a certain classification level. It represents the sum of squares of deviations for a certain type of sample.

[0024] Step 2-2: The Mitscherlich growth equation with starting point constraint is simplified as shown in equation (5). The parameter b is constructed with tree species as dummy variable. The parameter a is constructed with the initial growth level as dummy variable through the equal interval construction dummy variable method. The relationship between the average diameter at breast height of the stand and the corresponding time interval is fitted as shown in equation (6). The final growth level of each sample plot is determined through the double iteration algorithm.

[0025]

[0026]

[0027] In equations (5) to (6): Let be the average diameter at breast height (DBH) of the stand in the i-th sample plot during the j-th period. Let be the average diameter at breast height (DBH) of the stand in the initial stage of the i-th sample plot, 'a' be a parameter reflecting the forest land's productive potential, and 'b' be a parameter related to the growth rate of the tree species. The time interval between period j and the initial stage of the survey. , For model parameters, It is a dummy variable for the initial growth level (i.e., its value is 1 when the growth level belongs to the m-th level, and 0 otherwise). This represents the difference in parameter 'a' between adjacent growth grades in the diameter at breast height (DBH) growth model. It is a dummy variable for the tree species (i.e., its value is 1 when the tree species is z, and 0 otherwise).

[0028] Steps 2-3: Based on the final growth level of each plot, reconstruct parameter a as shown in equation (7), and obtain the model parameters after fitting;

[0029]

[0030] In the formula: This is a dummy variable for the final growth level (i.e., its value is 1 when the growth level belongs to level m, and 0 otherwise).

[0031] Steps 2-4: Using the diameter at breast height (DBH) growth model from the starting point, as shown in Equation (8), the formula for calculating the initial stand age is derived, as shown in Equation (9). Based on the tree species type and growth level of each sample plot, the model parameters are determined and substituted into the average DBH of the initial stand to estimate the initial stand age of each sample plot. Based on the time interval of different survey periods, the age of the sample plots in other periods can be obtained, as shown in Equation (10).

[0032] )

[0033]

[0034]

[0035] In equations (8) to (10): Let be the initial forest age of the i-th sample plot. Let be the forest age of the i-th sample plot in the j-th period.

[0036] Step three, classifying site quality grades, includes the following steps:

[0037] Step 3-1: The average charcoal reserves of the same plot over multiple periods are marked with a unique number (tree species number + plot number). The unique number of the plot is used to represent the site quality of each plot, and the plots are arranged in ascending order of plot number.

[0038] Step 3-2: Based on the number of sample plots and the difference in average wood carbon storage among sample plots, determine the site quality grade number H. Based on the Richards growth equation, with the average wood carbon storage of the stand as the dependent variable and the stand age as the independent variable, fit a cluster of carbon storage growth curves in stages, as shown in equation (11).

[0039] )

[0040] In equation (11): Let be the average carbon storage of the forest stand in the i-th sample plot during the j-th period. , , For model parameters, This is a dummy variable for the initial site quality grade (i.e., its value is 1 when the site quality grade belongs to level h, and 0 otherwise). This represents the difference in parameter 'a' between adjacent curves in the carbon reserve growth curve cluster. This represents the difference in parameter c between adjacent curves in the carbon reserve growth curve cluster.

[0041] Step 3-3: Calculate the estimated average charcoal reserves of each sample plot under H site quality grades, select the curve with the smallest sum of squared deviations as the site quality grade to which the sample plot belongs, and iterate repeatedly until the site quality grade of each sample plot no longer changes. The classification result at this point is determined as the final site quality grade.

[0042] Step four, establishing a carbon storage staged growth model and calculating the carbon sequestration rate, includes the following steps:

[0043] Step 4-1: Based on the final site quality grade, the site quality grade is reflected in parameters a and c by constructing dummy variables at equal intervals. Parameter b is used to distinguish tree species. The Richards growth equation is used to fit the relationship between the average wood carbon storage of the stand and the stand age to establish a carbon storage graded growth model, as shown in equation (12).

[0044] )

[0045] In equation (12): This is a dummy variable for the final site quality grade.

[0046] Step 4-2: Differentiate the stand age in equation (12) to obtain the carbon storage annual growth classification equation, that is, the average carbon sequestration rate of the stand in a certain year, as shown in equation (13).

[0047]

[0048] In equation (13): , , , denoted as the annual growth of average wood carbon storage in the i-th sample plot during the j-th period.

[0049] Step 4-3: Using forest age as the predictor variable, substitute it into equation (13) to calculate the carbon sequestration rate at a certain time point for each sample plot.

[0050] Step five involves screening factors that affect the carbon fixation rate, including the following steps:

[0051] To analyze the impacts of stand factors, climate factors, topographic factors, geographical factors, and soil factors on forest carbon storage potential and to clarify the driving forces of forest carbon sequestration capacity, a generalized analysis of variance (ANOVA) was used. The average carbon sequestration rate of trees in each sample plot was used as the dependent variable, and all factors were used as independent variables. A backward elimination method was employed, and the model was adjusted using the coefficient of determination (COP). Based on the principle of maximizing the interaction between factors, all first-order qualitative factors and first- and second-order quantitative factors were selected for interaction. A termination condition was set for the interaction terms, where the number of model parameters was less than 20% of the total sample size. This process identified factors that influenced the carbon fixation rate at a specific time point. The calculation formula is as shown in (14);

[0052]

[0053] In equation (14): N is the sample size, and p is the number of parameters. These are observed values. It is a predicted value. This represents the average of the observed values.

[0054] Step six includes the following steps:

[0055] Step 6-1: To clarify the contribution of environmental driving forces and tree species diversity to carbon sequestration potential, a linear regression model is established based on a general linear regression model containing qualitative factors. The annual growth of average carbon storage of trees in the forest stand at a certain time point is used as the dependent variable, and the factors that affect the carbon sequestration rate are used as independent variables. The model is shown in Equation (15) to quantify the direction and degree of influence of forest stand factors and environmental driving forces on forest carbon storage potential.

[0056]

[0057] In equation (15): For constant terms, This is the first quantitative factor affecting the carbon fixation rate. yes The regression coefficients, This is the nth quantitative factor affecting the carbon fixation rate. yes The regression coefficients, This is the first qualitative factor (out of R levels) that affects the carbon fixation rate. express The effect at the r-th level, express dummy variables (i.e., when) Its value is 1 when it belongs to the r-th level, and 0 otherwise. This is the nth qualitative factor that affects the carbon fixation rate (out of a total of T levels). express The effect at the t-th level, express dummy variables (i.e., when) Its value is 1 when it belongs to the t-th level, and 0 otherwise. This is the error term.

[0058] Step 6-2: Using a linear model that includes site factors and tree species type, assuming that the tree species growing in the sample plot is any one of all tree species, predict the carbon sequestration potential of each tree species at a certain time point based on the site factors of the sample plot, and compare the differences in the impact of tree species on carbon sequestration potential. Attached Figure Description

[0059] Figure 1 This is a flowchart illustrating the implementation of the present invention;

[0060] Figure 2a This represents the limit value of the annual growth of average charcoal reserves for different tree species in Example 1 of the present invention;

[0061] Figure 2bThis refers to the age at which the average annual growth of charcoal reserves of different tree species reaches its maximum value in Example 1 of this invention;

[0062] Figure 3a This is a graph showing the relationship between the actual and predicted values ​​of the linear regression model for the annual growth of the average carbon storage of the sample plot in the 10th year of Example 1 of the present invention.

[0063] Figure 3b This is a graph showing the relationship between the actual and predicted values ​​of the linear regression model for the annual growth of average wood carbon storage in the sample plot in the 20th year of Example 1 of the present invention.

[0064] Figure 3c This is a graph showing the relationship between the actual and predicted values ​​of the linear regression model for the annual growth of average wood carbon storage in the sample plot at year 30 in Example 1 of the present invention.

[0065] Figure 4 This is a comparison chart of the average annual growth of charcoal reserves in all sample plots under the original tree species and the eight tree species in Example 1 of the present invention.

[0066] Figure 5 This is a comparison chart showing the difference in carbon sequestration potential between the original tree species and the most suitable tree species in Embodiment 1 of the present invention;

[0067] Figure 6a The difference in annual growth of average carbon storage of Masson pine forest in 322 sample plots in the 30th year of Example 1 of the present invention (I).

[0068] Figure 6b The difference in annual growth of average carbon storage of Masson pine forest in 322 sample plots in the 30th year of Example 1 of the present invention (II).

[0069] Figure 6c This is a comparison of the annual growth of average wood carbon storage in 322 sample plots of wetland pine forest in the 30th year of Example 1 of the present invention.

[0070] Figure 6d This is a comparison of the annual growth of the average carbon storage of Chinese fir trees in 322 sample plots in the 30th year of Example 1 of the present invention.

[0071] Figure 6e The difference in annual growth of average carbon storage of oak trees in 322 sample plots in Example 1 of the present invention (I).

[0072] Figure 6f The difference in annual growth of average carbon storage of oak trees in 322 sample plots in Example 1 of the present invention (II).

[0073] Figure 6gThis is a summary of the annual differences in carbon storage and growth of the average wood in 322 sample plots of Phoebe zhennan forest in the 30th year of Example 1 of the present invention.

[0074] Figure 6h This is a comparison of the annual growth of the average carbon storage of the *Symplocos edulis* forest in 322 sample plots in the 30th year of Example 1 of the present invention.

[0075] Figure 6i This is a comparison of the annual growth of carbon storage of other hardwood trees on 322 sample plots in the 30th year of Example 1 of the present invention.

[0076] Figure 6j This shows the annual differences in carbon storage and growth of other soft broad-leaved forests on 322 sample plots in the 30th year of Example 1 of the present invention.

[0077] Note: The original tree species (i.e., the tree species in the parentheses in the upper left corner) represents the actual value. Masson pine, slash pine, Chinese fir, oak, nanmu, Michelia champaca, other hardwoods and other softwoods represent the predicted growth values ​​that the original tree species can achieve when the original tree species is replaced. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and examples.

[0079] It should be noted that the formula numbers in the claims and in Example 1 are all self-contained systems and there is no shared numbering.

[0080] Example 1

[0081] The basic data consisted of five forest resource inventories conducted in Guangdong Province from 1997 to 2017, with a five-year interval between adjacent survey periods and a fixed sample plot area of ​​0.0667 hm². 2 The abiotic environmental factors and stand factors matched for each fixed sample plot are as follows: 3 geographical factors (longitude, latitude, and watershed code); 7 topographic factors (aspect, slope class, and altitude); 19 climatic factors (annual minimum temperature, annual precipitation, and hottest season precipitation); 25 soil factors (organic matter, parent rock, and water content); and 1 stand factor (dominant tree species). Among these, there are 45 quantitative factors and 10 qualitative factors. A method for evaluating forest carbon storage potential based on site factors is presented, with the following specific steps:

[0082] Step 1: Select sample trees from multiple consecutive surveys within the sample plot, reorganize the sample plot, and match abiotic environmental factors;

[0083] Using 1997 as the starting point and 2017 as the ending point, the continuous survey period consisted of five phases. For each sample plot, corresponding geographical, topographical, soil, and climatic factors were collected. During each phase, the species type and measured diameter at breast height (DBH) of all individual trees within the sample plot were recorded. Five surviving trees from the survey were selected, meaning five sets of DBH data were available, and at least five trees of the same species were required to be retained. When multiple constituent tree species existed within a sample plot, they were classified into different sample plots according to species type. A unique number was assigned to each sample plot using the species number plus the plot number, resulting in a total of 322 sample plots: 86 Masson pine forests, 15 slash pine forests, 45 Chinese fir forests, 81 oak forests, 23 Phoebe zhennan forests, 27 Schima superba forests, 28 other hardwood forests, and 17 other softwood forests. Individual tree carbon storage (aboveground carbon storage + belowground carbon storage) was calculated using the standing tree biomass model and carbon storage measurement parameters according to forestry industry standards. The average diameter at breast height (DBH) of the reconstituted plots was the square mean DBH of all retained trees, and the plot carbon storage was the arithmetic mean of the carbon storage of all retained trees, i.e., the average stand carbon storage. The calculation formula is as follows:

[0084]

[0085]

[0086] In equations (1) and (2): n is the number of trees retained in the reconstituted plot. To preserve the diameter at breast height of the timber, To preserve charcoal reserves.

[0087] Step 2: Based on the constrained Mitscherlich growth equation, establish the relationship between the average diameter at breast height (DBH) of the stand and the corresponding survey time interval, and then determine the stand age by inverse calculation based on the average DBH of the stand.

[0088] Step 2-1: Calculate the annual growth rate S based on the average diameter at breast height of the stand in the first and last two periods, arrange them in descending order, select an appropriate number of classes according to the sample size, and use ordered sample clustering for initial classification to obtain the number of growth levels M (M = 9 in Example 1). The ordered sample clustering algorithm uses the optimal segmentation method for classification, and its minimum error function recursive formula is as shown in equation (4).

[0089]

[0090]

[0091] In equations (3) to (4): This represents the average diameter at breast height (DBH) of the initial sample plot. The average diameter at breast height (DBH) of the last sample plot. The time interval between the first and last survey periods (in Example 1) = 20), The loss function representing classification. This represents the classification scheme, where M is the number of classification levels, N is the number of sample plots, and X is the number of sample plots under a certain classification level. It represents the sum of squares of deviations for a certain type of sample.

[0092] Step 2-2: Using the Mitscherlich growth equation with starting point constraints, as shown in Equation (5), construct parameter b with tree species as dummy variable, construct parameter a with initial growth level as dummy variable through the equal interval dummy variable construction method, fit the relationship between the average diameter at breast height of the stand and the corresponding time interval, as shown in Equation (6), and determine the final growth level of each sample plot through the double iteration algorithm.

[0093]

[0094]

[0095] In equations (5) to (6): Let be the average diameter at breast height (DBH) of the stand in the i-th sample plot during the j-th period. Let be the average diameter at breast height (DBH) of the stand in the initial stage of the i-th sample plot, 'a' be a parameter reflecting the forest land's productive potential, and 'b' be a parameter related to the growth rate of the tree species. The time interval between period j and the initial stage of the survey. , For model parameters, It is a dummy variable for the initial growth level (i.e., its value is 1 when the growth level belongs to the m-th level, and 0 otherwise). This represents the difference in parameter 'a' between adjacent growth grades in the diameter at breast height (DBH) growth model. It is a dummy variable for the tree species (i.e., its value is 1 when the tree species is z, and 0 otherwise).

[0096] Steps 2-3: Based on the final growth level of each plot, reconstruct parameter a as shown in equation (7). After fitting, the model parameters can be obtained, as shown in Table 1.

[0097]

[0098] In the formula: This is a dummy variable for the final growth level (i.e., its value is 1 when the growth level belongs to level m, and 0 otherwise).

[0099] Steps 2-4: Using the diameter at breast height (DBH) growth model from the starting point, as shown in Equation (8), the initial stand age calculation formula is derived, as shown in Equation (9). Based on the tree species type and growth level of each sample plot, the model parameters are determined and substituted into the average DBH of the initial stand to estimate the initial stand age of each sample plot. Based on the time interval of different survey periods, the age of the sample plots in other periods can be obtained, as shown in Equation (10). The stand age statistics for each tree species are shown in Table 2.

[0100] )

[0101]

[0102]

[0103] In equations (8) to (10): Let be the initial forest age of the i-th sample plot. Let be the forest age of the i-th sample plot in the j-th period.

[0104] Step 3: Classify site quality grades based on the average stand carbon storage.

[0105] Step 3-1: The average charcoal reserves of the same plot over multiple periods are marked with a unique number (tree species number + plot number). The unique number of the plot is used to represent the site quality of each plot, and the plots are arranged in ascending order of plot number.

[0106] Step 3-2: Based on the number of sample plots and the difference in average wood carbon storage among sample plots, determine the site quality grade number H (H = 9 in Example 1). Based on the Richards growth equation, with the average wood carbon storage of the stand as the dependent variable and the stand age as the independent variable, fit a cluster of carbon storage growth curves in stages, as shown in Equation (11).

[0107] )

[0108] In equation (11): Let be the average carbon storage of the forest stand in the i-th sample plot during the j-th period. , , For model parameters, This is a dummy variable for the initial site quality grade (i.e., its value is 1 when the site quality grade belongs to level h, and 0 otherwise). This represents the difference in parameter 'a' between adjacent curves in the carbon reserve growth curve cluster. This represents the difference in parameter c between adjacent curves in the carbon reserve growth curve cluster.

[0109] Step 3-3: Calculate the average charcoal reserves of each sample plot under H site quality grades, select the curve with the smallest sum of squared deviations as the site quality grade to which the sample plot belongs, and iterate repeatedly until the site quality grade of each sample plot no longer changes. The classification result at this time is determined as the final site quality grade. The specific classification results are shown in Table 3.

[0110] Step 4: Based on the Richards growth equation, establish a carbon storage graded growth model, determine the annual carbon storage growth equation, and calculate the carbon sequestration rate at different time points for each sample plot using forest age as the predictor variable.

[0111] Step 4-1: Based on the final site quality grade, the site quality grade is reflected in parameters a and c by constructing dummy variables at equal intervals. Parameter b is used to distinguish tree species. The Richards growth equation is used to fit the relationship between the average wood carbon storage of the stand and the stand age to establish a carbon storage graded growth model, as shown in equation (12). The model parameters can be obtained after fitting with the least squares method, as shown in Table 4.

[0112] )

[0113] In equation (12): This is a dummy variable for the final site quality grade.

[0114] Step 4-2: Differentiate the stand age in equation (12) to obtain the graded equation for annual carbon storage growth, i.e., the actual carbon sequestration rate of the average stand in a certain year, as shown in equation (13). The limit values ​​of annual average carbon storage growth for each tree species and the age at which the limit values ​​are reached are shown in [reference needed]. Figures 2a to 2b ;

[0115]

[0116] In equation (13): , , , denoted as the annual growth of average wood carbon storage in the i-th sample plot during the j-th period.

[0117] Step 4-3: Using forest age as the predictor variable, substitute it into equation (13) to calculate the carbon sequestration rate of each sample plot in the 10th, 20th and 30th years.

[0118] Step 5: Use the backward elimination method to screen out the factors that affect the carbon fixation rate;

[0119] To analyze the impacts of stand factors, climate factors, topographic factors, geographical factors, and soil factors on forest carbon storage potential and to clarify the driving forces of forest carbon sequestration capacity, a generalized analysis of variance (ANOVA) was used. The average carbon sequestration rate of trees in each sample plot was used as the dependent variable, and all factors were used as independent variables. A backward elimination method was employed, and the model was adjusted using the coefficient of determination (COP). Based on the principle of maximizing the interaction between factors, all first-order qualitative factors and first- and second-order quantitative factors were selected for interaction. The termination condition for the interaction terms was set to be less than 20% of the total sample size. Factors affecting carbon fixation rate at three different time points were screened, and the screening results are shown in Table 5. The calculation formula is as shown in (14);

[0120]

[0121] In equation (14): N is the sample size, and p is the number of parameters. These are observed values. It is a predicted value. This represents the average of the observed values.

[0122] Step 6: Based on a general linear regression model, quantify the direction and extent of the influence of stand factors and environmental driving forces on forest carbon storage potential, and compare the differences in carbon sequestration potential of tree species within the same site at different time points.

[0123] Step 6-1: To clarify the contributions of environmental driving forces and tree species diversity to carbon sequestration potential, a general linear regression model containing qualitative factors was established, with the annual growth of average wood carbon storage in the 10th, 20th, and 30th years as the dependent variable and factors affecting the carbon sequestration rate as independent variables, as shown in Equation (15). This quantifies the direction and degree of influence of stand factors and environmental driving forces on forest carbon storage potential. The regression coefficients of each influencing factor in the 30th year are shown in Table 6. The relationship between the actual values ​​and predicted values ​​of the linear regression model at the three time points is shown in Figure 3a to 3b. Figure 3c .

[0124]

[0125] In equation (15): For constant terms, This is the first quantitative factor affecting the carbon fixation rate. yes The regression coefficients, This is the nth quantitative factor affecting the carbon fixation rate. yes The regression coefficients, This is the first qualitative factor (out of R levels) that affects the carbon fixation rate. express The effect at the r-th level, express dummy variables (i.e., when) Its value is 1 when it belongs to the r-th level, and 0 otherwise. This is the nth qualitative factor that affects the carbon fixation rate (out of a total of T levels). express The effect at the t-th level, express dummy variables (i.e., when) Its value is 1 when it belongs to the t-th level, and 0 otherwise. This is the error term.

[0126] Step 6-2: Using a linear model incorporating site factors and tree species type, assuming the tree species growing in the sample plot are any one of eight species, predict the carbon sequestration potential of each tree species at 10, 20, and 30 years based on the site factors of the sample plot, and compare the differences in the impact of tree species on carbon sequestration potential. Figure 4 , Figure 5 and Figures 6a to 6j .

[0127] Table 1 lists the fitting evaluation indices and parameter values ​​of the diameter at breast height (DBH) growth model including growth grades. As shown in Table 1, the DBH growth model exhibits a relatively good fit, with a coefficient of determination R0. 2 The mean squared diameter (DBD) can reach 0.9836, the standard deviation of the estimated value (SEE) is less than 0.8 cm, the mean prediction error (MPE) and total relative error (TRE) tend to be 0, and the mean percentage standard error (MPSE) is less than 3.5%. In addition, the nine growth grades reflected in parameter a represent different growth potentials, with growth grade 1 having the greatest growth potential and growth grade 9 having the least. Parameter b, which is related to the growth rate of tree species, also shows that the DBD growth rates of the eight tree species are not completely equal, with slash pine growing the fastest and Michelia macclurei growing the slowest.

[0128] Table 1 Fitting parameters and evaluation indices for the diameter at breast height (DBH) growth model with growth gradations

[0129]

[0130] Table 2 lists the average age of each tree species plots calculated using the diameter at breast height (DBH) growth model. Table 2 shows that other hardwood forests and oak forests are older, with an average initial age exceeding 20 years, while the slash pine forest is younger, less than 10 years old. This indicates significant differences in species characteristics among different tree species; even with the same DBH, the required growth time is not entirely consistent. For the age estimation of the 27 *Symplocos macrantha* plots, the age differences among plots are greater than those for other tree species, possibly due to the larger differences in average DBH among the *Symplocos macrantha* plots.

[0131] Table 2. Statistics on stand age estimates for each tree species in sample plots.

[0132]

[0133] Table 3 lists the site quality classification results based on the average carbon storage of the sample plots and the average carbon storage of each tree species under different site quality grades over 20 years. As shown in Table 3, not every tree species has nine site quality grades; only Masson pine forests, oak forests, and other hardwood forests have nine grades. Grade 1 represents the best site quality. However, regardless of the tree species, the number of sample plots reaching Grade 1, Grade 2, and Grade 3 is relatively small, with most sample plots having a medium site quality. Furthermore, when within the same site quality grade, other softwood forests have the highest average carbon storage over 20 years, while Chinese fir forests have the lowest. For example, at Grade 5, the average carbon storage of other softwood forests can reach a maximum of 62.38 kg over 20 years, while that of Chinese fir forests is only 21.61 kg.

[0134] Table 3. Average carbon storage of trees of various species in sample plots under different site quality grades at 20 years.

[0135]

[0136] Note: The values ​​in the table are the average charcoal reserves of each sample plot. The values ​​in parentheses are the number of sample plots that fall within that site quality grade. The higher the grade, the worse the site quality, i.e., grade 1 has the best site quality.

[0137] Table 4 lists the fitting parameters and evaluation indices of the carbon reserve tiered growth model. As shown in Table 4, the carbon reserve tiered growth model achieved a good fit, with a coefficient of determination R0. 2 The mean squared error (MSE) can reach 0.9595, the standard deviation (SEE) of the estimated value is less than 13.5 kg, the mean prediction error (MPE) and total relative error (TRE) tend to be 0, and the mean percentage standard error (MPSE) is less than 17.0%. The better the site quality, the larger the value of parameter 'a', and the greater the limit that the average carbon storage of the sample plot can reach; the better the site quality, the smaller the value of parameter 'c', and the shorter the number of years required for the annual growth of carbon storage to reach its maximum.

[0138] Table 4 Fitting parameters and model evaluation indices for the carbon storage tiered growth model

[0139]

[0140] Note: a 1，…， a9 represents the parameter 'a' value under nine site quality grades, and c 1，…， c9 represents the parameter c value for the nine site quality grades. Since the site quality grades are constructed using dummy variables at equal intervals, the data in the table are calculated data; the original parameters are as follows: =638.41, = -66.33, = 2.15, = 0.19.

[0141] Figure 2a and Figure 2b The maximum annual growth rate of average carbon storage for each tree species and the age at which the maximum value is reached are given. As shown in the figure, the influence of site quality on the annual growth rate of carbon storage is consistent across the eight tree species: the better the site quality, the larger the maximum value and the shorter the time required to reach it. However, under the same site quality, there are certain differences among the tree species; that is, the maximum value that each species can reach and the age at which the maximum value is reached are not the same. The maximum growth rate for each tree species is: Chinese fir < Masson pine < oak < *Schima superba* < other hardwoods < *Phoebe zhennan* < *Pinus slashii* < other softwoods. The age at which the maximum value is reached is: other softwoods < *Pinus slashii* < *Phoebe zhennan* < other hardwoods < *Schima superba* < oak < *Pinus massoniana* < Chinese fir.

[0142] Table 5 lists the screening results of factors influencing carbon sequestration rate (annual growth of carbon storage) at the 10th, 20th, and 30th years. The table shows that the selected influencing factors do not differ significantly at different time points. Based on the combined results of the three screenings and the significance values, it can be determined that among the single factors, isothermal conditions, annual minimum air temperature, monthly average diurnal temperature range, precipitation in the hottest season, and tree species have extremely significant effects on carbon sequestration rate, while the interaction terms all have extremely significant effects. Furthermore, among geographical factors, longitude has the greatest impact; among climatic factors, isothermal conditions have the greatest impact; and among soil factors, non-capillary porosity has a relatively large impact.

[0143] Table 5. Screening results of factors influencing carbon fixation rate at three time points.

[0144]

[0145]

[0146] Note:

[0147] (1) When screening factors that affect the carbon sequestration rate of the average wood in the sample plot in the 10th year, a total of 38 screenings were conducted. Single factors were eliminated in turn and interaction terms between factors were introduced: aspect, slope grade, watershed code, soil texture, slope position, altitude grade, parent rock, annual maximum temperature, total nitrogen, precipitation in the coldest season, P2O5 composition, bulk density, hydrolyzed nitrogen, annual temperature variation range, gravel, water content, soil thickness, average temperature in the coldest season, precipitation in the driest month, precipitation in the wettest season, hydrolyzed phosphorus, seasonal variation of temperature, slope, clay, coarse silt, altitude, tree species * slope * clay, precipitation in the wettest month, landform, sand, hydrolyzed potassium, K2O composition * humus layer thickness * aspect, humus layer thickness, K2O composition, aspect * hydrolyzed phosphorus * hydrolyzed phosphorus, maximum water holding capacity, capillary water holding capacity, tree species * capillary porosity * annual minimum temperature.

[0148] (2) When screening factors that affect the average carbon sequestration rate of trees in the sample plot in the 20th year, a total of 37 screenings were conducted. Single factors were eliminated in turn and interaction terms between factors were introduced: aspect, soil texture, watershed code, slope position, altitude class, slope class, parent rock, annual maximum temperature, precipitation in the driest month, average temperature in the coldest season, precipitation in the coldest season, total nitrogen, gravel, hydrolyzed phosphorus, annual temperature variation range, hydrolyzed nitrogen, P2O5 composition, water content, bulk density, precipitation in the wettest season, soil thickness, slope, seasonal variation of temperature, clay, hydrolyzed potassium, precipitation in the wettest month, tree species * slope * clay, coarse silt, sand, landform, K2O composition * humus layer thickness * aspect, humus layer thickness, K2O composition, aspect * hydrolyzed phosphorus * hydrolyzed phosphorus, maximum water holding capacity, capillary water holding capacity, tree species * capillary porosity * annual minimum temperature.

[0149] (3) When screening factors that affect the average carbon sequestration rate of trees in the sample plot in the 30th year, a total of 36 screenings were conducted. Single factors were eliminated in turn and interaction terms between factors were introduced: aspect, soil texture, altitude class, slope position, watershed code, annual maximum temperature, hydrolytic P, gravel, average temperature in the coldest season, precipitation in the driest month, total N, annual temperature variation range, precipitation in the coldest season, parent rock, slope class, composition P2O5, slope, water content, precipitation in the wettest season, hydrolytic N, bulk density, clay, soil thickness, hydrolytic K, average temperature in the hottest season, seasonal variation of temperature, average temperature in the wettest season, tree species * slope * clay, coarse silt, sand, landform, precipitation in the wettest month, humus layer thickness * soil structure * latitude, aspect * hydrolytic N * hydrolytic N, tree species * capillary porosity * annual minimum temperature, slope * aspect * hydrolytic P.

[0150] Table 6 lists the regression coefficients of factors influencing the annual growth of carbon storage in the average timber of the sample plots at year 30. As shown in Table 6, the coefficients can be positive or negative. For quantitative factors, a positive coefficient indicates a positive correlation between the factor and the annual growth of carbon storage. For example, higher values ​​for factors such as annual precipitation, annual minimum temperature, and organic matter lead to greater annual growth in carbon storage, indicating a promoting effect. A negative coefficient indicates a negative correlation, indicating an inhibitory effect. For qualitative factors, the positive or negative value of a coefficient at a certain level represents the direction of influence on the annual growth of carbon storage relative to a reference level. For example, among tree species, using other softwoods as a reference level, the coefficients for species such as Masson pine, slash pine, and Chinese fir are negative, indicating that the annual growth of carbon storage in these species is smaller than that of other softwoods. The magnitude of the coefficient also reflects the size of the difference. For instance, the coefficient for Masson pine is -6.9036, indicating that its difference from softwoods is the greatest.

[0151] Table 6. Regression coefficients of factors influencing the average carbon storage of timber in the sample plot at year 30 and annual growth.

[0152]

[0153]

[0154]

[0155] Note: The following are the evaluation metrics for the linear regression model: coefficient of determination R0 2 = 0.6420, Corrected coefficient of determination = 0.5385.

[0156] Figures 3a to 3c The relationship between the actual and predicted values ​​of the linear regression model for the annual growth of average carbon storage in sample plots at the 10th, 20th, and 30th years is presented. As shown in the figure, the coordinate points of the predicted and actual values ​​fitted by the model are relatively evenly distributed on both sides of the diagonal. However, there are sample plots with overestimated annual growth of carbon storage and underestimated annual growth of carbon storage.

[0157] When the same tree species grows on all sample plots, the carbon storage that the eight tree species can achieve varies from year to year. Figure 4The average annual growth of carbon storage in all plots under eight tree species is presented, and the growth under the original tree species is compared. As shown in the figure, compared to the original tree species, the annual growth is smaller when only Chinese fir or other hardwoods grow in the plots, while it is larger when only slash pine, oak, or other softwoods grow. The differences among tree species are not entirely consistent at different time points. For example, for slash pine, oak, and other softwoods, the annual growth of carbon storage is greatest in the 10th year when only other softwoods are grown in the plots. However, in the 20th and 30th years, the annual growth of carbon storage is greatest when only slash pine is grown in the plots. This is related to the growth characteristics of the tree species; the age at which the annual growth of carbon storage reaches its maximum value is earlier than that of slash pine and oak.

[0158] In the 10th, 20th, and 30th years, the tree species with the largest annual increase in carbon storage among eight tree species were selected from each sample plot. This species was considered the most suitable for growth on that sample plot and was assumed to be the tree species to be grown on it. The differences in carbon sequestration potential between the original tree species and the most suitable tree species were then compared. Figure 5 A comparative graph of the annual growth of carbon storage between the original tree species and the most suitable tree species at three time points is presented. The graph shows that when the tree species most suitable for the site conditions of the sample plot are selected for planting, the average annual growth of carbon storage achievable by all sample plots is higher than that of the original tree species, exceeding it by more than 2 kg. This indicates that selecting suitable afforestation tree species for a given site can improve the carbon sequestration potential of the vegetation.

[0159] Figures 6a to 6j The study presents the differences in annual carbon storage growth of various tree species across 322 sample plots at year 30. As shown in the figure, regardless of the original tree species in the sample plots, the tree species with the highest predicted annual carbon storage growth among the eight species in the 322 sample plots are mainly concentrated in three species: slash pine, oak, and other softwood broadleaf trees. Specifically, there are 121 slash pine plots, 93 oak plots, and 98 other softwood broadleaf trees. The tree species with the lowest predicted annual growth is mainly Chinese fir, with 272 plots. Due to length constraints, the differences in annual carbon storage growth for each tree species across the 322 sample plots in years 10 and 20 are not presented. In year 10, the tree species with the highest predicted annual carbon storage growth were mainly slash pine (103 plots), oak (83 plots), and other softwood broadleaf trees (136 plots), while the species with the lowest predicted annual carbon storage growth was mainly Chinese fir (271 plots). In year 20, the tree species with the highest predicted annual carbon storage growth were mainly slash pine (134 plots), oak (71 plots), and other softwood broadleaf trees (113 plots), while the species with the lowest predicted annual carbon storage growth was mainly Chinese fir (298 plots). In conclusion, slash pine, oak, and other softwood broadleaf trees are suitable carbon sink tree species for afforestation in Guangdong Province. Figure 4It can be seen that other soft broad-leaved trees have the highest carbon sequestration potential in the short term, while in the long term, around 30 years, slash pine and oak have higher carbon sequestration potential than other soft broad-leaved trees.

[0160] The parameter values ​​listed in Embodiment 1 of this invention are only applicable to the distribution range of the modeling samples used. However, the modeling method and process provided by this invention can provide reference and guidance for model construction in other regions.

[0161] Those skilled in the art will recognize that the embodiments and descriptions of this invention are merely preferred embodiments to illustrate the basic principles and main features of the invention. The embodiments described herein are intended to help readers understand the implementation methods of the invention and should be understood as not limiting the scope of protection of the invention to such specific statements and embodiments. Without departing from the definitions of the appended claims, the invention can have various modifications, such as differences in the number of survey periods, the structural parameters of site quality grades, the basic growth equation for modeling, and the type of interaction terms between screening factors. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the essence of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A method for evaluating forest carbon storage potential based on site factors, characterized in that, Includes the following steps: Step 1: Select sample trees from multiple consecutive surveys within the sample plot, reorganize the sample plot, and match abiotic environmental factors; The basic data consists of continuous survey data at fixed intervals, with the first survey period as the starting point and the last survey period as the ending point, for a total of k survey periods. During each sample plot survey, corresponding geographical factors, topographic factors, soil factors, and climatic factors need to be obtained, collectively referred to as environmental driving forces. Within each survey period, the species type and measured diameter at breast height (DBH) of all individual trees in the sample plot are recorded, and sample trees that have survived for all k survey periods are selected, i.e., there are k measured DBH data, and the number of retained trees of the same species is required to be no less than 5. When there are multiple constituent tree species in a sample plot, they are classified into different sample plots according to tree species type, and a unique number is formed by combining the tree species number and the sample plot number to reorganize the sample plot. The carbon storage per tree is the sum of the aboveground carbon storage and the underground carbon storage, calculated using the standing biomass model and carbon storage measurement parameters standard in the forestry industry. The average diameter at breast height (DBH) of the reconstituted plots is the square mean DBH of all retained trees, and the plot carbon storage is the arithmetic mean of the carbon storage of all retained trees, i.e., the average stand carbon storage, calculated using the following formula: ， ， In equations (1) and (2): n is the number of trees retained in the reconstituted plot. To preserve the diameter at breast height of the timber, To preserve charcoal reserves; Step 2: Based on the constrained Mitscherlich growth equation, establish the relationship between the average diameter at breast height (DBH) of the stand and the corresponding survey time interval, and then determine the stand age by inverse calculation based on the average DBH of the stand. Determining the age of a forest stand by reverse calculation involves the following steps: Step 2-1: Calculate the annual growth rate S based on the average diameter at breast height of the stand in the first and last two periods, arrange them in descending order, select an appropriate number of classes according to the sample size, use ordered sample clustering for initial classification, and obtain the number of growth levels M. The ordered sample clustering algorithm uses the optimal segmentation method for classification, and its minimum error function recursive formula is as shown in equation (4). ， ， In equations (3) to (4): This represents the average diameter at breast height (DBH) of the initial sample plot. The average diameter at breast height (DBH) of the sample plot in the kth period (the last period) is... The time interval between the first and last survey periods. The loss function representing classification. This represents the classification scheme, where M is the number of classification levels, N is the number of sample plots, and X is the number of sample plots under a certain classification level. Represents the sum of squares of deviations for a certain type of sample; Step 2-2: The Mitscherlich growth equation with starting point constraint is simplified as shown in equation (5). The parameter b is constructed with tree species as dummy variable. The parameter a is constructed with the initial growth level as dummy variable through the equal interval construction dummy variable method. The relationship between the average diameter at breast height of the stand and the corresponding time interval is fitted as shown in equation (6). The final growth level of each sample plot is determined through the double iteration algorithm. ， ， In equations (5) to (6): Let be the average diameter at breast height (DBH) of the stand in the i-th sample plot during the j-th period. Let be the average diameter at breast height (DBH) of the stand in the initial stage of the i-th sample plot, 'a' be a parameter reflecting the forest land's productive potential, and 'b' be a parameter related to the growth rate of the tree species. The time interval between period j and the initial stage of the survey. , For model parameters, This is a dummy variable for the initial growth level; that is, its value is 1 when the growth level belongs to the m-th level, and 0 otherwise. This represents the difference in parameter 'a' between adjacent growth grades in the diameter at breast height (DBH) growth model. This is a dummy variable for the tree species; that is, its value is 1 when the tree species is z, and 0 otherwise. Steps 2-3: Based on the final growth level of each plot, reconstruct parameter a as shown in equation (7), and obtain the model parameters after fitting; ， In the formula: This is a dummy variable for the final growth level; that is, its value is 1 when the growth level belongs to the m-th level, and 0 otherwise. Steps 2-4: Using the diameter at breast height (DBH) growth model from the starting point, as shown in Equation (8), the formula for calculating the initial stand age is derived, as shown in Equation (9). Based on the tree species type and growth level of each sample plot, the model parameters are determined and substituted into the average DBH of the initial stand to estimate the initial stand age of each sample plot. Based on the time interval of different survey periods, the age of the sample plots in other periods can be obtained, as shown in Equation (10). ）， ， ， In equations (8) to (10): Let be the initial forest age of the i-th sample plot. The forest age of the i-th sample plot in the j-th period; Step 3: Classify site quality grades based on the average stand carbon storage. The process of classifying site quality grades includes the following steps: Step 3-1: The average charcoal reserves of the same plot over multiple periods are all marked with a unique number. The unique number of the plot is used to replace the site quality of each plot, and the plots are arranged in ascending order by plot number. Step 3-2: Based on the number of sample plots and the difference in average wood carbon storage among sample plots, determine the site quality grade number H. Based on the Richards growth equation, with the average wood carbon storage of the stand as the dependent variable and the stand age as the independent variable, fit a cluster of carbon storage growth curves in stages, as shown in equation (11). ）， In equation (11): Let be the average carbon storage of the forest stand in the i-th sample plot during the j-th period. , , For model parameters, This is a dummy variable for the initial site quality grade, meaning its value is 1 when the site quality grade belongs to level h, and 0 otherwise; This represents the difference in parameter 'a' between adjacent curves in the carbon reserve growth curve cluster. This represents the difference in parameter c between adjacent curves in the carbon reserve growth curve cluster. Step 3-3: Calculate the average charcoal reserves of each sample plot under H site quality grades, select the curve with the smallest sum of squared deviations as the site quality grade to which the sample plot belongs, and iterate repeatedly until the site quality grade of each sample plot no longer changes. The classification result at this time is determined as the final site quality grade. Step 4: Based on the Richards growth equation, establish a carbon storage hierarchical growth model, determine the annual carbon storage growth equation, use forest age as the predictor variable, and calculate the carbon sequestration rate at a certain time point for each sample plot, i.e., the annual carbon storage growth. Step 5: Use the backward elimination method to screen out the factors that affect the carbon fixation rate; Establishing a carbon storage staged growth model and calculating the carbon sequestration rate includes the following steps: Step 4-1: Based on the final site quality grade, the site quality grade is reflected in parameters a and c by constructing dummy variables at equal intervals. Parameter b is used to distinguish tree species. The Richards growth equation is used to fit the relationship between the average wood carbon storage of the stand and the stand age to establish a carbon storage graded growth model, as shown in equation (12). ）， In equation (12): This is a dummy variable for the final site quality grade; Step 4-2: Differentiate the stand age in equation (12) to obtain the carbon storage annual growth classification equation, that is, the average carbon sequestration rate of the stand in a certain year, as shown in equation (13). ， In equation (13): , , , The annual growth of average wood carbon storage in the i-th sample plot during the j-th period; Step 4-3: Using forest age as the predictor variable, substitute it into equation (13) to calculate the carbon sequestration rate at a certain time point for each sample plot; Step 6: Based on a general linear regression model, quantify the direction and extent of the influence of stand factors and environmental driving forces on forest carbon storage potential, and explore the differences in the influence of tree species on carbon sequestration potential based on site factors.

2. The method for evaluating forest carbon storage potential based on site factors according to claim 1, characterized in that, Step five includes the following steps: To analyze the impacts of stand factors, climate factors, topographic factors, geographical factors, and soil factors on forest carbon storage potential and to clarify the driving forces of forest carbon sequestration capacity, a generalized analysis of variance (ANOVA) was used. The average carbon sequestration rate of trees in each sample plot was used as the dependent variable, and all factors were used as independent variables. A backward elimination method was employed, and the model was adjusted using the coefficient of determination (COP). Based on the principle of maximizing the interaction between factors, all first-order qualitative factors and first- and second-order quantitative factors were selected for interaction. A termination condition was set for the interaction terms, where the number of model parameters was less than 20% of the total sample size. This process identified factors that influenced the carbon fixation rate at a specific time point. The calculation formula is as shown in (14); ， In equation (14): N is the sample size, and p is the number of parameters. These are observed values. It is a predicted value. This represents the average of the observed values.

3. The method for evaluating forest carbon storage potential based on site factors according to claim 1, characterized in that, Step six includes the following steps: Step 6-1: To clarify the contribution of environmental driving forces and tree species diversity to carbon sequestration potential, a linear regression model with qualitative factors is established based on the annual growth of average carbon storage of trees in the forest at a certain time point as the dependent variable and factors that affect the carbon sequestration rate as independent variables, as shown in Equation (15), to quantify the direction and degree of influence of forest factors and environmental driving forces on forest carbon storage potential. ， In equation (15): For constant terms, This is the first quantitative factor affecting the carbon fixation rate. yes The regression coefficients, This is the nth quantitative factor affecting the carbon fixation rate. yes The regression coefficients, This is the first qualitative factor affecting the carbon fixation rate, with a total of R levels; express The effect at the r-th level, express dummy variables, that is, when Its value is 1 when it belongs to the r-th level, and 0 otherwise; This is the nth qualitative factor that affects the carbon fixation rate, with a total of T levels; express The effect at the t-th level, express dummy variables, that is, when Its value is 1 when it belongs to the t-th level, and 0 otherwise; This is the error term; Step 6-2: Using a linear model that includes site factors and tree species type, assuming that the tree species growing in the sample plot is any one of all tree species, predict the carbon sequestration potential of each tree species at a certain time point based on the site factors of the sample plot, and compare the differences in the impact of tree species on carbon sequestration potential.

Images