Biomass model construction and evaluation of dominant species Alnus Jiangnan in subtropical mountain marsh

Abstract

The invention discloses a biomass model construction and evaluation method for subtropical mountain marsh dominant species Alnus Jiangnan. The method comprises the following specific steps: S1, a field investigation method; s2, performing an indoor experiment; s3, data analysis: performing statistical analysis on the data, and constructing a relative growth model, a compatibility biomass model and an optimal biomass model; s4, a model precision evaluation method: evaluating the model precision by using a root-mean-square error, a mean absolute error and a decision coefficient; and S5, obtaining a result and analyzing. According to the method, the relative growth model and the compatibility biomass model for the biomass of the single Alnus Jiangnan and the biomass of each component of the alnus Jiangnan are constructed, the optimal parameters are fitted through the multivariate nonlinear least square method, and it is ensured that the models do not have the heterovariance problem and have high prediction precision; important theoretical basis and practical guidance are provided for evaluation of the biomass of the alnus Jiangnan in the subtropical mountain marsh, and the carbon sequestration capacity of the ecological system of the mountain marsh under the climate change background can be better understood and predicted.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of forestry, in particular to the biomass model construction and evaluation of Alnus trabeculosa Hand.-Mazz., a dominant species in subtropical mountain swamp. BACKGROUND

[0002] Alnus trabeculosa Hand.-Mazz. is one of the dominant tree species in subtropical mountain swamp ecosystems, with significant ecological and economic value. It has a fast growth rate, high economic benefits, and strong carbon sequestration capacity, making it one of the preferred tree species for carbon sequestration afforestation. In addition, Alnus trabeculosa prefers light and warmth and is tolerant to water, occupying a significant ecological niche in subtropical mountain swamp habitats. Biomass is an important indicator of ecosystem productivity and carbon sequestration capacity. Therefore, accurately evaluating the biomass of Alnus trabeculosa is crucial for understanding and predicting the carbon sequestration capacity of mountain swamp ecosystems under climate change.

[0003] Tree biomass models are an important method for estimating forest biomass, usually based on the relationship between tree biomass and other growth parameters such as diameter at breast height and tree height. Currently, commonly used biomass models include relative growth models, compatible biomass models, and volume-biomass conversion models. Among them, relative growth models and compatible biomass models are the most commonly used. Relative growth models are a simple and effective method for biomass estimation, with high fitting accuracy for each component. However, in this modeling approach, each component is estimated independently, and there is a problem of incompatibility between the component models and the total model, i.e., the sum of the estimated values of each component model does not completely equal the estimated value of the total model. On the contrary, compatible biomass models decompose the total biomass of a tree into the biomass of each component, ensuring that the total biomass equals the sum of the biomass of each component, thereby effectively solving the incompatibility problem and reducing the prediction error between components.

[0004] Currently, research on the biomass of Alnus trabeculosa has mainly focused on plantations in northern Fujian and hilly areas of central Sichuan, with insufficient evaluation of the biomass of Alnus trabeculosa in mountain swamps. Therefore, this study takes Alnus trabeculosa in mountain swamps in Fengshuanyangfeng and Huangdan areas of Zhejiang Province as an example, uses relative growth models and compatible biomass models to simulate the total biomass of individual Alnus trabeculosa and the biomass of each component, and evaluates their accuracy, providing a theoretical basis for the evaluation of Alnus trabeculosa biomass in mountain swamps. SUMMARY

[0005] The purpose of the present application is to solve the shortcomings in the prior art and to provide a biomass model construction and evaluation of Alnus trabeculosa, a dominant species in subtropical mountain swamps.

[0006] To achieve the above-mentioned purpose, the present application adopts the following technical solutions: The biomass model construction and evaluation of the subtropical mountain swamp dominant species Alnus cremastogyne include the following specific steps: S1: Field investigation method: the field investigation time is September 2018, 9 and 11 20m*20m Alnus cremastogyne forest sample plots are set up in two study areas (Wangdongyuan and Huangtang forest farm) according to the altitude gradient, the center longitude and latitude of each sample plot is recorded (recorded by GPS), the terrain (slope, slope direction, slope position), soil type and vegetation characteristics (coverage, canopy density) are recorded, the diameter at breast height and tree height of all trees in the sample plot are measured, so as to fully understand the growth of Alnus cremastogyne in the study area (Table 1); in order to ensure the high precision simulation and prediction of biomass, 34 standard trees are selected according to different diameter at breast height (1.9-21.9cm) and tree height (2.95-13.6m) for felling, the roots are carefully dug and taken back to the laboratory together with the trunk (bark + wood) and crown (branches + leaves); S2: Indoor experiment: the trunk, crown and root are treated respectively, and then the biomass is measured; S3: Data analysis: statistical analysis is carried out on the data, and the relative growth model, compatible biomass model and optimal biomass model are constructed; S4: Model precision evaluation method: the root mean square error, mean absolute error and determination coefficient are used to evaluate the precision of the model; S5: Results and analysis: the construction and precision evaluation of single Alnus cremastogyne relative growth model, the construction and precision evaluation of single Alnus cremastogyne compatible biomass model, and the selection of single Alnus cremastogyne biomass and its components optimal biomass model are carried out.

[0007] As a further technical solution of the present application, S2 specifically includes: S21: The trunk is cut by layering cutting method, and the fresh weight of bark and wood is measured respectively; S22: The crown is divided into upper, middle and lower three layers, the fresh weight of each leafy branch is measured, the average fresh weight of leafy branches in each layer is calculated, 3 standard branches are selected according to the average fresh weight of leafy branches in each layer, the leaf quality and branch quality of the standard branches are measured after the leaves are removed, the fresh weight of each layer of branches and leaves is calculated according to the fresh weight of each layer of standard branches, and then the fresh weight of branches and leaves in the whole crown is calculated; S23: The soil on the surface of the root is washed, and the fresh weight is measured, then 500g or so of fresh samples of each component are placed in an oven at 85℃ and dried to a constant weight, the moisture content of the sample is measured, and then the dry weight (i.e. biomass) of each component is calculated.

[0008] As a further technical solution of the present application, in S2, the biomass of the Jiangnan alder standard tree and its components in the study area is shown in Table 2, and the total biomass of the Jiangnan alder standard tree in the study area is 58.87 kg / plant, wherein the stem accounts for 61% of the total biomass (bark 22%, wood 78%), the crown accounts for 27% of the total biomass (branches 73%, leaves 23%), and the roots account for 12% of the total biomass.

[0009] As a further technical solution of the present application, S3 specifically includes: S31: statistical analysis: using the ANOVA single factor analysis method of SPSS 22.0 software, the correlation between the height and diameter at breast height of the Jiangnan alder standard tree and the biomass of the Jiangnan alder and its components is analyzed (Table 3), and it is concluded that the diameter at breast height and the height have extremely significant correlation with all other variables (total weight, stem, bark, crown, leaves, branches, wood, and roots) (P<0.01). p <0.01); S32: relative growth model: the relative growth model biomass equation is usually in the form of power function, logarithmic function, polynomial, and exponential function, and according to the research of Yin Huiyan et al., in the form of the relative growth model of arbor species, the equation in the form of power function accounts for about 51.96% of the total number of arbor biomass equations, and the binary function form considering the diameter at breast height and the height and the unary form using only the diameter at breast height are the most common, that is, the commonly used allometric growth equation form includes three kinds: , , , wherein: W is the total biomass and the biomass of each component (kg), D is the diameter at breast height (cm), H is the height (m), a , b , c is the estimated parameter; S33: compatible biomass model: the common methods for establishing a compatible model include nonlinear likelihood independent regression method, nonlinear proportional adjustment method, and nonlinear simultaneous equation method, but in recent years, the component addition method has excellent fitting effect on multiple tree species, the model precision is high, and it is favored by many scholars; based on the component addition method, three power function model formulas are selected: , , As the basis of the compatible biomass model, the component addition method is used to establish the compatible biomass model of the Jiangnan alder, and then the parameters of the equation set are fitted through joint estimation. S34: Biomass model construction: The Cvpartition function of Matlab software is used for cross-validation partitioning of the measured data set of Alnus cremastogyne in Jiangnan, 80% of the data is used as the training set, and the most suitable initial parameters are screened out by using multiple cycles to construct the biomass model suitable for the study area of Alnus cremastogyne; the remaining 20% of the data is used as the test set, the prediction accuracy of the biomass model is evaluated, and the best biomass model of single Alnus cremastogyne and each component in the study area is screened out, the specific information of the training set and the test set is shown in Table 4, and the White Test is used to test whether the variance of the model residual is constant or not, and the significance level is 0.05.

[0010] As a further technical solution of the application, in S33, The compatible biomass equation group formula 4 established is: The compatible biomass equation group formula 5 established is: The compatible biomass equation group formula 6 established is: In the formula, W is the compatible total biomass and the biomass of each component (kg), respectively: tree roots, wood, bark, branches, leaves, tree trunks, tree crowns, and total biomass; respectively: power function model of tree roots, wood, bark, branches, leaves, tree trunks, tree crowns, and total biomass, in formula 4, in formula 5, in formula 6, D is the diameter at breast height (cm); H is the tree height (m); a and b are estimated parameters; is the error term of each model.

[0011] As a further technical solution of the application, S4 specifically comprises: The root mean square error (RMSE) is the square root of the average value of the square of the difference between the predicted value and the actual value, and the mean absolute error (MAE) is the average value of the absolute value of the difference between the predicted value and the actual value, both of which can measure the accuracy of model prediction; the determination coefficient (R2) is the percentage of the model explaining the variation of the target variable, which can measure the goodness of model fitting; the specific calculation formula is as follows: In the formula, MSE is the mean square error, is the number of observation values,​​​​ is the actual value, is the predicted value, is the average value of the actual value; the smaller the values of RMSE and MAE, the higher the accuracy of the model prediction; the closer the value is to 1, the better the prediction effect of the model.

[0012] As a further technical solution of the present application, S5 specifically comprises: S51: selecting (Formula 1), (Formula 2), and (Formula 3) as the basic model, respectively using the three equations to establish the total biomass model of single Jiangnan alder and the model of each component of the tree trunk, bark, crown, leaves, branches, wood, and roots (Tables 5-7), and evaluating the modeling accuracy of each component biomass of Jiangnan alder; the results show that the White test significance levels of all relative growth models are greater than 0.01, indicating that there is no heteroscedasticity problem, and by comparing the average values of the determination coefficients of the three equations, it is concluded that the monomial function model =0.923) has higher overall modeling accuracy for the biomass of single Jiangnan alder and each component thereof than the binomial function ( =0.910), and ( =0.850); Among the biomass models of each component of single Jiangnan alder, the monomial function model (Formula 1) has higher modeling accuracy for the biomass of single Jiangnan alder, wood, and roots than the binomial function models (Formula 2 and Formula 3), the binomial function model (Formula 2) has higher modeling accuracy for the biomass of the tree trunk and crown than the monomial function model, and the binomial function model (Formula 3) has higher modeling accuracy for the biomass of leaves and branches than the monomial function model; S52: using , , as the component basic model, three compatible biomass models of Jiangnan alder are constructed by using the component addition method, with the constraint conditions of tree trunk = bark + wood, crown = branch + leaf, and total biomass = root + crown + crown, and the modeling accuracy of the total biomass of single Jiangnan alder is evaluated (Tables 8-10); the results show that the White test significance levels of the three compatible biomass models are all greater than 0.01, indicating that there is no heteroscedasticity problem; the monomial function compatible biomass model (Formula 4) has higher modeling accuracy for the total biomass of single Jiangnan alder =0.909) than the binomial function compatible biomass model (Formula 5, =0.885) and ( Formula 6, =0.886) are high; by comparing the determination coefficients of the relative growth model (Formula 1-3) and the compatible growth model (Formula 4-6) in the biomass modeling of single Jiangnan alder, it is concluded that compared with the relative growth model ( =0.968), the modeling accuracy of the compatible biomass model for the total biomass of single Jiangnan alder ( =0.898) decreases ( p <0.05); In the modeling of the biomass of each component of single Jiangnan alder, the modeling accuracy of the unary compatible biomass model (Formula 4) for the biomass of branches, bark and crown of Jiangnan alder is higher than that of the unary relative growth model (Formula 1), and the modeling accuracy of the rest of the components of Jiangnan alder is opposite; the modeling accuracy of the binary compatible biomass model (Formula 5) for the biomass of bark and stem of single Jiangnan alder is higher than that of the binary relative growth model (Formula 2); the modeling accuracy of the binary compatible biomass model (Formula 6) for the biomass of bark, crown, leaves and branches of single Jiangnan alder is higher than that of the binary relative growth model (Formula 3); S53: The prediction accuracy of the relative growth model and the compatible biomass model for the biomass of single Jiangnan alder and each component thereof is evaluated using the test set, and the results are shown in Tables 11 and 12. By comparing and analyzing the modeling accuracy and prediction accuracy of the relative growth model in Tables 11 and 5-7, it is concluded that Formula 1 has the highest simulation accuracy for the biomass of leaves of single Jiangnan alder, Formula 2 has the highest simulation accuracy for the biomass of single Jiangnan alder and its roots, branches, leaves and crown, and Formula 3 has the highest simulation accuracy for the biomass of wood and bark of single Jiangnan alder, thereby screening out the best relative growth model for the biomass of each component of single Jiangnan alder as follows: In the formula, W is the total biomass and the biomass of each component of single Jiangnan alder (kg), are respectively: roots, wood, bark, branches, leaves, stem, crown; D is the diameter at breast height (cm); H is the tree height (m); Table 12 is the prediction accuracy of the three compatible biomass models. By comparing and analyzing Table 12 and Tables 8-10, it is concluded that Formula 4 has the highest simulation accuracy for the total biomass of single Jiangnan alder (test set =0.869, training set =0.869); the results show that the compatible biomass model of Jiangnan alder constructed by the unary power function is the best prediction model: In the formula, WFor the compatibility of total biomass and each component biomass (kg), Respectively: tree roots, wood, bark, branches, leaves, tree trunks, tree crowns, total biomass; D For the diameter at breast height (cm).

[0013] The beneficial effects of the present application are: 1. The relative growth model and the compatible biomass model for single Jiangnan alder and each component biomass are constructed, and the optimal parameters are fitted by multivariate nonlinear least squares method, so as to ensure that the model has no heteroscedasticity problem and has high prediction accuracy.

[0014] 2. The present application provides an important theoretical basis and practical guidance for the evaluation of Jiangnan alder biomass in subtropical mountain swamp, and helps to better understand and predict the carbon sink capacity of mountain swamp ecosystem under the background of climate change. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 Schematic diagram of the study area; Figure 2 The flow chart of the biomass model construction and evaluation of the subtropical mountain swamp dominant species Jiangnan alder put forward in the present application. DETAILED DESCRIPTION

[0016] In order to make the technical means, creative features, purposes and effects realized by the present application easy to understand, the present application is further described below in combination with specific embodiments.

[0017] Please refer to the attached Figure 2 The biomass model construction and evaluation of the subtropical mountain swamp dominant species Jiangnan alder, including the following specific steps: S1: Field investigation method: the field investigation time is September 2018, 9 and 11 20m x 20m Jiangnan alder forest sample plots are set up according to the altitude gradient in two study areas (Wangdongyuan and Huangtanglinchang), the center longitude and latitude (recorded by GPS), terrain (slope, aspect, slope position), soil type and vegetation characteristics (coverage, canopy density) of each sample plot are investigated and recorded, the diameter at breast height and tree height of all trees in the sample plot are measured to fully understand the growth of Jiangnan alder in the study area (Table 1); in order to ensure high precision simulation and prediction of biomass, 34 standard trees are selected according to different diameter at breast height (1.9-21.9cm) and tree height (2.95-13.6m) for felling, the tree roots are carefully dug and taken back to the laboratory together with the tree trunk (bark + wood), tree crown (branches + leaves); Study area profile: the study area (such as Figure 1The study area is located in the Wangedongyuan Mountain Wetland Nature Reserve (27°40'00"N-27°44'19"N, 119°34'28"E-119°38'54"E) in Jingning She Autonomous County, Lishui City, Zhejiang Province and Huangdan Forest Farm (28°53'25"N-28°49'57"N, 120°28'51"E-120°25'20"E) in Pan'an County, Jinhua City, Zhejiang Province, with an elevation of 930-1117 m. Both areas are typical of subtropical monsoon climate, with distinct seasons, mild climate, and abundant rainfall. The average annual temperature is 16.5°C, the annual precipitation is above 1445.8 mm, and the average annual relative humidity is above 77%. The soil type is mountain meadow soil. Alnus cremastogyne grows on the lower slope with a west or northwest aspect. Through the study of Alnus cremastogyne forest plots at different altitudes, the spatial distribution and variation of the biomass of Alnus cremastogyne in mountain marshes in Zhejiang Province can be more comprehensively reflected, providing a scientific basis for understanding and predicting the carbon sink capacity of the ecological system under regional climate change.

[0018] Table 1 Stand conditions of the sample plots in Wangedongyuan of Jingning and Huangdan Forest Farm of Pan'an S2: Indoor experiment: the trunk, crown, and roots are treated separately, and then the biomass is measured; S2 specifically includes: S21: The trunk is cut using the stratified cutting method, and the fresh weight of the bark and wood is measured; S22: The crown is divided into three layers: upper, middle, and lower. The fresh weight of each leafy branch is measured, and the average fresh weight of the leafy branches in each layer is calculated. Three standard branches are selected according to the average fresh weight of the leafy branches in each layer. After removing the leaves from the standard branches, the branch mass and leaf mass are measured. The fresh weight of the branches and leaves in each layer is calculated based on the fresh weight of the standard branches, and then the fresh weight of the branches and leaves in the entire crown is calculated; S23: The soil on the surface of the roots is washed off, and the fresh weight is measured. Then, about 500 g of fresh samples of each component are placed in an oven at 85°C and dried to a constant weight. The moisture content of the samples is measured, and then the dry weight (i.e., biomass) of each component is calculated; According to the measurement, the biomass of the Alnus cremastogyne standard tree and its components in the study area is shown in Table 2. The total biomass of the Alnus cremastogyne standard tree in the study area is 58.87 kg per plant, of which the trunk accounts for 61% of the total biomass (22% for bark and 78% for wood), the crown accounts for 27% of the total biomass (73% for branches and 23% for leaves), and the roots account for 12% of the total biomass.

[0019] Table 2 Statistics of the biomass of the Alnus cremastogyne standard tree and its components in the study area S3: Data analysis: statistical analysis of data, and construction of relative growth model, compatibility biomass model and optimal biomass model; S3 specifically includes: S31: statistical analysis: using ANOVA single factor analysis method of SPSS 22.0 software, analyzing the correlation between Jiangnan alder standard tree height, diameter at breast height and Jiangnan alder and its component biomass (Table 3), and obtaining that there is extremely significant correlation between diameter at breast height and tree height and all other variables (total weight, stem, bark, crown, leaves, branches, wood, roots) (P<0.01); p Table 3 Correlation between Jiangnan alder standard tree and its component biomass and tree height, diameter at breast height in the study area Correlation coefficient Note: ** indicates extremely significant correlation (P<0.01) p S32: relative growth model: the relative growth model biomass equation is usually in the form of power function, logarithmic function, polynomial and exponential function, according to the research of Yin Huiyan et al., among the forms of relative growth model of arbor species, the equation in the form of power function accounts for about 51.96% of the total number of arbor biomass equations, and the binary function form considering diameter at breast height and tree height and the unary form using only diameter at breast height are the most common, that is, the commonly used allometric growth equation form includes three kinds: , , , in which: W is total biomass and component biomass (kg), D is diameter at breast height (cm), H is tree height (m), a , b , c is the estimated parameter; S33: compatibility biomass model: the common methods for establishing compatibility model include nonlinear likelihood independent regression method, nonlinear proportional adjustment method and nonlinear simultaneous equation method, but in recent years, the component addition method has excellent fitting effect on multiple tree species, the model precision is high, and it is favored by many scholars; based on the component addition method, three power function model formulas are selected as the basis of compatibility biomass model: , , as the basis of compatibility biomass model, and the component addition method is used to establish the compatibility biomass model of Jiangnan alder, and then the parameters of the equation set are fitted through joint estimation; In S33, the established compatibility biomass equation set formula 4 is: ​​ The established compatible biomass equation set Formula 6 is: The established compatible biomass equation set Formula 6 is: In the formula, W is the total biomass and the biomass of each component (kg), respectively: tree roots, wood, bark, branches, leaves, trunks, crowns, total biomass; respectively: power function model of tree roots, wood, bark, branches, leaves, trunks, crowns, total biomass, in Formula 4, in Formula 5, and in Formula 6, D is the diameter at breast height (cm); H is the tree height (m); a, b are estimated parameters; is the error term of each model; S34: Biomass model construction: cross-validation partitioning of the measured data set of Alnus cremastogyne in southern China was performed using the Cvpartition function of Matlab software, 80% of the data was used as the training set, and the most suitable initial parameters were screened out using multiple cycles to construct the biomass model of Alnus cremastogyne suitable for the study area; the remaining 20% of the data was used as the test set, the prediction accuracy of the biomass model was evaluated, and the best biomass model of Alnus cremastogyne and its components in the study area was screened out, the specific information of the training set and the test set is shown in Table 4, and White Test was used to test whether the variance of the model residual was constant or not, with a significance level of 0.05.

[0020] Table 4 Training set and test set data of Alnus cremastogyne standard trees in the study area S4: Model accuracy evaluation method: root mean square error, mean absolute error and determination coefficient are used to evaluate the model accuracy; S4 specifically includes: Root mean square error (RMSE) is the square root of the average of the square of the difference between the predicted value and the actual value, mean absolute error (MAE) is the average of the absolute value of the difference between the predicted value and the actual value, both of which can measure the accuracy of model prediction; determination coefficient (R2) is the percentage of model explaining the variation of target variable, which can measure the goodness of model fitting; the specific calculation formula is as follows: , , , , , in which: MSE is mean square error, is the number of observed values, is the actual value, is the predicted value, is the average value of the actual value; the smaller the values of RMSE and MAE, the higher the accuracy of the model prediction; the closer the value of R2 to 1, the better the prediction effect of the model.

[0021] S5: Results and analysis: the construction and precision evaluation of the single Alnus cremastogyne relative growth model, the construction and precision evaluation of the single Alnus cremastogyne compatibility biomass model, and the selection of the best biomass model of single Alnus cremastogyne and its components were carried out respectively; S5 specifically includes: S51: Selecting (Equation 1), (Equation 2), and (Equation 3) as the basic model, the models of total biomass and each component of single Alnus cremastogyne (Table 5-7) were established using the three equations respectively, and the modeling accuracy of each component of Alnus cremastogyne was evaluated; the results showed that the White test significance level of all relative growth models was greater than 0.01, indicating that there was no heteroscedasticity problem, by comparing the average values of the determination coefficients of the three equations, it was concluded that the monomial power function model =0.923) had higher overall modeling accuracy for single Alnus cremastogyne and its components than the binomial function ( =0.910) and ( =0.850); Table 5 Monomial function relative growth model Table 6 Binomial function relative growth model Table 7 Binomial function relative growth model Among the single Alnus cremastogyne component biomass models, the monomial power function model (Equation 1) had higher modeling accuracy than the binomial power function models (Equations 2 and 3) for single Alnus cremastogyne, wood and root biomass, the binomial power function model (Equation 2) had higher modeling accuracy than the monomial power function model for stem and crown biomass, and the binomial power function model (Equation 3) had higher modeling accuracy than the monomial power function model for leaf and branch biomass; S52: Taking 、 、 As a component-based model, the compatible biomass models of three Jiangnan Alnus trees were constructed using the component addition method. The modeling accuracy of the total biomass of a single Jiangnan Alnus tree was evaluated with the constraints of trunk = bark + wood, crown = branches + leaves, and total biomass = roots + crown + crown (Tables 8-10). The results showed that the White test significance levels of the three compatible biomass models were all greater than 0.01, indicating that there was no heteroscedasticity problem. The univariate function compatible biomass model The modeling accuracy of the total biomass of a single Alnus jiangnanensis tree ( =0.909), compared with the binary function compatible biomass model (Formula 5, =0.885) and (Formula 6, =0.886) has high accuracy; by comparing the determination coefficients of the relative growth model (Formula 1-3) and the compatible growth model (Formula 4-6) in modeling the biomass of a single Alnus chinensis, it is concluded that: =0.968), the compatibility biomass model has a significant impact on the total biomass of individual Alnus jiangnanensis ( =0.898) decreases in modeling accuracy ( p <0.05); Table 8 Unary functions Compatible biomass model Table 9 Binary functions Compatible biomass model Table 10 Binary functions Compatible biomass model In the modeling of the biomass of individual Jiangnan Alnus trees, the univariate compatible biomass model (Equation 4) had higher modeling accuracy for the biomass of branches, bark, and crowns than the univariate relative growth model (Equation 1), while the opposite results were found for the modeling accuracy of the biomass of the other components of Jiangnan Alnus. The binary compatible biomass model (Equation 5) had higher modeling accuracy for the biomass of bark and trunks of individual Jiangnan Alnus than the binary relative growth model (Equation 2). The binary compatible biomass model (Equation 6) had higher modeling accuracy for the biomass of bark, crown, leaves, and branches of individual Jiangnan Alnus than the binary relative growth model (Equation 3). S53: The relative growth model and the compatible biomass model were evaluated for the prediction accuracy of the biomass of individual Jiangnan alder and its components using the test set, and the results are shown in Tables 11 and 12. By comparing and analyzing the modeling accuracy and prediction accuracy of the relative growth model in Tables 11 and 5-7, it is concluded that formula 1 has the highest simulation accuracy for the leaf biomass of individual Jiangnan alder, formula 2 has the highest simulation accuracy for the biomass of individual Jiangnan alder and its roots, branches, leaves, and crown, and formula 3 has the highest simulation accuracy for the wood and bark of individual Jiangnan alder. Thus, the best relative growth model for the biomass of individual Jiangnan alder is as follows: In the formula, W is the total biomass and the biomass of each component of individual Jiangnan alder (kg), respectively: roots, wood, bark, branches, leaves, tree trunk, crown; D is the diameter at breast height (cm); H is the tree height (m); Table 11: Prediction accuracy of the relative growth model Table 12 shows the prediction accuracy of the three compatible biomass models. By comparing and analyzing Table 12 with Tables 8-10, it is concluded that formula 4 has the highest simulation accuracy for the total biomass of individual Jiangnan alder (test set = 0.869, training set = 0.869); and the results show that the compatible biomass model of Jiangnan alder constructed by the monomial function is the best prediction model: In the formula, W is the compatible total biomass and the biomass of each component (kg), respectively: roots, wood, bark, branches, leaves, tree trunk, crown, total biomass; D is the diameter at breast height (cm).

[0022] Table 12: Accuracy evaluation of the compatible biomass model In this study, we used two methods, relative growth model and compatible biomass model, to model and evaluate the accuracy of the biomass of single Jiangnan alder (Alnus petenensis) tree in Zhejiang Province. The results showed that the best model for the total biomass of single Jiangnan alder tree was the one-dimensional power function compatible biomass model, which was consistent with previous research results. On the one hand, under the conditions of mountain marsh, the soil water content is higher, and trees can more easily absorb water and nutrients, which makes the root system not have to be too developed, so the underground biomass of Jiangnan alder accounts for a lower proportion than other trees, and this also leads to the aboveground biomass of Jiangnan alder accounting for a higher proportion than other trees. The compatible biomass model can fully consider the internal correlation and additivity between component biomass, and can more comprehensively and objectively reflect the distribution relationship of biomass between different components, thereby improving the simulation accuracy of the biomass of single Jiangnan alder tree. Secondly, by using correlation analysis, we found that the correlation between diameter at breast height and the biomass of each component of Jiangnan alder was higher than that between tree height and the biomass of each component (Table 3), so the simulation accuracy of the one-dimensional power function compatible biomass model for the biomass of single Jiangnan alder tree was higher.

[0023] The biomass of different components of Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz. was modeled and the precision was analyzed, and the results showed that the best model form of the biomass of different components of single Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz. was different. The best model of the biomass of leaves was a unary power function model, the best model of the biomass of roots, branches, trunks and crowns was a binary power function model, and the best model of the biomass of wood and bark was a binary power function model. The results showed that, except for leaves, the biomass of other components of single Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz. was better simulated by using diameter at breast height (D) and tree height (H) as double variables in a binary power function model. The results were consistent with the results of previous studies, which showed that the fitting effect and prediction accuracy of the biomass of each component were improved by adding the variable of tree height. As a key variable in the field of forest biomass modeling, the addition of tree height to the biomass model usually improves the simulation accuracy of the biomass of trees. However, in the leaf biomass model of Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz., the addition of tree height reduced the simulation accuracy. The reason may be that the influence of tree height on the growth rate and biomass of leaves of Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz. was smaller than that of other variables. At the same time, by comparing the correlation analysis results of diameter at breast height, tree height and the biomass of each component of single Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz.) (Table 3), it can be concluded that the correlation coefficient of tree height and the biomass of crown of Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz. was relatively low. Some studies have shown that crown length (L) is an important factor affecting the biomass of crown. It can reflect the horizontal expansion of the crown. Trees with larger crown width generally have more complex and extensive crown structure, can obtain more light energy, and promote their growth and biomass accumulation. At the same time, the crown morphology of different tree species is different, which leads to the difference in the best crown prediction variable of different tree species. The introduction of crown width variable helps to increase the accuracy of the crown biomass model. In addition, tree species with higher branch height usually have straight trunks. These tree species have better growth environment, no excessive branch differentiation or pest infestation, and their crown is at a higher height, which is easier to obtain sufficient sunlight and can more effectively utilize resources to increase biomass. Therefore, in the selection of independent variables for the biomass model of each component of single Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz., diameter at breast height (D) can be used as the basic independent variable, and then crown width and crown length, branch height and other variables can be selected as the covariates of the model according to the characteristics of different components to model the biomass of trees.

[0024] From the above description, it can be seen that the above-mentioned embodiments of the present application achieve the following technical effects: The present study constructs the relative growth model and compatible biomass model for the biomass of single Alnus cremastogyne Buch. var. sinuata (Burfh.) Hand.-Mazz. and each component thereof, and the optimal parameters are fitted by multivariate nonlinear least squares method, which ensures that the model has no heteroscedasticity problem and has high prediction accuracy.

[0025] The results showed that the biomass of different components was suitable for different model forms. The best model for leaves was the one-dimensional power function relative growth model, while the best models for roots, branches, trunks, and crowns were two-dimensional power function relative growth models. The one-dimensional power function compatible biomass model showed higher precision and reliability in estimating the total biomass of individual Jiangnan alder.

[0026] The results of this study provide important theoretical basis and practical guidance for the biomass assessment of Jiangnan alder in subtropical mountain mire, and help better understand and predict the carbon sink capacity of mountain mire ecosystems under the background of climate change. Future studies can further optimize the model structure and introduce more variables to improve the universality and prediction accuracy of the model.

[0027] Those skilled in the art will understand that the above discussion of any of the embodiments is merely exemplary and is not intended to be limiting of the scope of the present application; the above embodiments or technical features among different embodiments can also be combined, and the steps can be implemented in any order, and there are many other changes of different aspects of the present application as described above, which are not provided in details for the sake of brevity.

[0028] The present application is intended to cover all such alternatives, modifications, and variations as fall within the broad scope of the description. Accordingly, any and all such modifications, variations, and equivalents that fall within the spirit and scope of the present application are intended to be included within the scope of the present application.

Claims

1. Construction and evaluation of biomass model of Alnus jiangnanensis, a dominant species in subtropical mountain swamps, characterized by: The specific steps include: S1: Field survey methods: Nine and eleven 20m × 20m Alnus chinensis forest plots were established in the two study areas, respectively, along the elevation gradient. The central latitude and longitude, topography, soil type, and vegetation characteristics of each plot were recorded. The diameter at breast height (DBH) and tree height of all trees within the plots were measured. Thirty-four standard trees were evenly selected based on their DBH and height, and their roots were dug up and brought back to the laboratory along with the trunks and crowns. S2: Indoor experiment: the trunk, crown and root were treated separately and then the biomass was measured; S3: Data analysis: Statistical analysis of the data was performed, and relative growth model, compatible biomass model and optimal biomass model were constructed; S4: Model accuracy evaluation method: Root mean square error, mean absolute error and coefficient of determination are used to evaluate model accuracy; S5: Results and analysis: The relative growth model of a single Alnus jiangnanensis tree and its accuracy evaluation, the compatible biomass model of a single Alnus jiangnanensis tree and its accuracy evaluation, and the optimal biomass model screening of the biomass of a single Alnus jiangnanensis tree and its components were carried out respectively.

2. The biomass model construction and evaluation of Alnus chinensis, a dominant species in subtropical mountain swamps, according to claim 1, is characterized in that: In S1, the field survey time was September 2018; the central longitude and latitude of each sample plot were recorded using GPS; the topography of each sample plot included slope, aspect, and position, and the vegetation characteristics included coverage and canopy density.

3. The biomass model construction and evaluation of Alnus chinensis, a dominant species in subtropical mountain swamps, according to claim 1, is characterized in that: The S2 specifically includes: S21: The trunk was cut using the layered cutting method, and the fresh weights of the bark and wood were measured separately; S22: Divide the crown into three layers: upper, middle, and lower. Measure the fresh weight of each leaf-bearing branch. Calculate the average fresh weight of leaf-bearing branches in each layer. Select three standard branches based on the average fresh weight of leaf-bearing branches in each layer. Remove leaves from the standard branches and measure the branch and leaf weights. Calculate the fresh weight of branches and leaves in each layer based on the fresh weight of standard branches in each layer. Calculate the fresh weight of branches and leaves in the entire crown. S23: Clean the soil on the surface of the tree roots and weigh its fresh weight. Then, place about 500 g of fresh samples of each component in an oven at 85°C and dry them to a constant weight. Measure the moisture content of the sample and then calculate the dry weight of each component.

4. The biomass model construction and evaluation of Alnus chinensis, a dominant species in subtropical mountain swamps, according to claim 3, is characterized in that: In S2, the total biomass of standard Alnus jiangnanensis trees in the study area was 58.87 kg / tree, of which the trunk accounted for 61% of the total biomass, the crown accounted for 27% of the total biomass, and the root accounted for 12% of the total biomass.

5. The biomass model construction and evaluation of Alnus chinensis, a dominant species in subtropical mountain swamps, according to claim 1, is characterized in that: The S3 specifically includes: S31: Statistical analysis: SPSS 22.0 software was used to analyze the correlation between the height and diameter at breast height of standard trees of Alnus jiangnanensis and the biomass of Alnus jiangnanensis and its components. The results showed that the diameter at breast height and tree height had extremely significant correlations with all other variables. S32: Relative growth model: Among the relative growth models of tree species, the most common are the power function form that considers both DBH and tree height, and the unary form that uses only DBH. Commonly used allometric growth equations include three forms: , , , where: W is the total biomass and the biomass of each component, D is the diameter at breast height, H For the height of the tree, a 、 b 、 c is the estimated parameter; S33: Compatible biomass model: Based on the component addition method, three power function models are selected: 、 、 As the basis of the compatible biomass model, the compatible biomass model of Alnus jiangnanensis was established by component addition method, and then the parameters of the equation group were fitted by joint estimation. S34: Biomass model construction: The Cvpartition function of Matlab software was used to perform cross-validation partitioning on the measured dataset of Alnus jiangnanensis. 80% of the data was used as the training set, and the most appropriate initial parameters were screened through multiple cycles to construct a biomass model of Alnus jiangnanensis suitable for the study area. The remaining 20% ​​of the data was used as the test set. By evaluating the prediction accuracy of the biomass model, the optimal biomass model of single Alnus jiangnanensis trees and their components in the study area was screened. At the same time, the White test was used to test whether the variance of the model residuals was not constant, with a significance level of 0.

05.

6. The biomass model construction and evaluation of Alnus chinensis, a dominant species in subtropical mountain swamps, according to claim 5, is characterized in that: In the S33, The established compatible biomass equation 4 is: The established compatible biomass equation 5 is: The established compatible biomass equation (6) is: Where W is the compatible total biomass and the biomass of each component, They are: roots, wood, bark, branches, leaves, trunk, crown, and total biomass; are the power function models of roots, wood, bark, branches, leaves, trunks, crowns, and total biomass, respectively. Formula 4 is , in Formula 5, , where Equation 6 is ; D is the diameter at breast height; H is the tree height; a and b are estimated parameters; is the error term of each model.

7. The biomass model construction and evaluation of Alnus chinensis, a dominant species in subtropical mountain swamps, according to claim 6, is characterized in that: The S4 specifically includes: ， ， ， , where: MSE is the mean square error, is the number of observations, is the actual value, is the predicted value, It is the average value of the actual value; the smaller the value of RMSE and MAE, the higher the accuracy of the model prediction; The closer the value is to 1, the better the prediction effect of the model.

8. The biomass model construction and evaluation of Alnus chinensis, a dominant species in subtropical mountain swamps, according to claim 7, is characterized in that: The S5 specifically includes: S51: The significance level of the White test for all relative growth models is greater than 0.01, indicating that there is no heteroscedasticity problem. By comparing the average values ​​of the coefficients of determination of the three equations, it is concluded that the overall modeling accuracy of the power function model for the biomass of a single Alnus jiangnanensis tree and its components is higher than that of the binary function model. and High precision; In the biomass models of individual Alnus jiangnanensis components, the modeling accuracy of the univariate power function model in the biomass of individual Alnus jiangnanensis, wood, and roots is higher than that of the bivariate power function model. The modeling accuracy of trunk and crown biomass is higher than that of the univariate power function model, and the bivariate power function model is more accurate. The modeling accuracy of leaf and branch biomass is higher than that of the univariate power function model; S52: The White test significance levels of the three compatible biomass models are all greater than 0.01, indicating that there is no heteroscedasticity problem; the univariate function compatible biomass model The modeling accuracy of the total biomass of a single Alnus jiangnanensis tree is better than that of the binary function compatible biomass model. and The accuracy is high; by comparing the determination coefficients of the relative growth model and the compatible growth model in modeling the biomass of a single Alnus jiangnanensis, it was found that the modeling accuracy of the compatible biomass model for the total biomass of a single Alnus jiangnanensis tree is lower than that of the relative growth model; In the modeling of the biomass of each component of a single Jiangnan Alnus tree, the univariate compatible biomass model had higher modeling accuracy for the biomass of branches, bark, and crown than the univariate relative growth model, while the modeling accuracy of the other components of the Jiangnan Alnus was the opposite. The binary compatible biomass model had higher modeling accuracy for the biomass of the bark and trunk of a single Jiangnan Alnus tree than the binary relative growth model. The binary compatible biomass model had higher modeling accuracy for the biomass of the bark, crown, leaves, and branches of a single Jiangnan Alnus tree than the binary relative growth model. S53: Equation 1 has the highest simulation accuracy for the leaf biomass of a single Alnus chinensis tree, Equation 2 has the highest simulation accuracy for the biomass of a single Alnus chinensis tree and its roots, branches, leaves, and crown, and Equation 3 has the highest simulation accuracy for the wood and bark of a single Alnus chinensis tree. The optimal relative growth model for each component biomass of a single Alnus chinensis tree is screened out as follows: Where, W is the total biomass and biomass of each component of a single Alnus jiangnanensis tree, They are: roots, wood, bark, branches, leaves, trunk, and crown; D is the diameter at breast height; H is the tree height; the results show that the compatible biomass model of Alnus jiangnanensis constructed by the univariate power function is the best choice of prediction model: Where, W is the compatible total biomass and the biomass of each component, They are: roots, wood, bark, branches, leaves, trunk, crown, and total biomass; D is the diameter at breast height.

Images