Forest ecosystem carbon reserve determination method based on multi-source remote sensing data

Abstract

The invention discloses a forest ecosystem carbon reserve determination method based on multi-source remote sensing data, and the method comprises the steps: constructing and analyzing a dominant tree species multi-temporal multi-feature classification data set according to Sentinel-1 and Sentinel-2 multi-source remote sensing data and research area forest resource checking vector data extraction; classifying the dominant tree species according to the multi-feature multi-temporal remote sensing data; and estimating the forest carbon reserves of different dominant tree species according to the optimal classification data set and the forest stock. According to the method, forest multi-feature and multi-temporal feature data are obtained by means of multi-source remote sensing data, and combined dimension reduction and separability analysis are performed, so that advantageous tree species spatial distribution is extracted, a forest stock and biomass linear model is fitted, and carbon reserve spatial distribution and features of a research area are estimated. And basic data and a practical basis are provided for further strengthening forest ecological system protection and implementing forest carbon reserve accurate improvement measures.

Description

Technical Field

[0001] The present invention relates to a forest ecosystem carbon storage estimation technology, and in particular to a forest ecosystem carbon storage determination method based on multi-source remote sensing data. Background Art

[0002] With the rapid development of remote sensing technology, the rapid, large-scale, high-precision assessment of forest carbon stocks using multi-source, high-temporal-spatial remote sensing imagery combined with ground-based survey data warrants further investigation. Currently, most studies estimating forest carbon stocks fail to consider the heterogeneity of tree species composition, focusing on the overall or secondary vegetation subtypes such as broad-leaved and coniferous forests. Detailed carbon stock estimation studies at the level of different dominant tree species are lacking. In summary, considering the urgent challenges and practical needs of modern remote sensing surveys of forest resources under the "dual carbon" framework, it is necessary to further clarify the current status of regional forest carbon storage and calculate the biomass parameters and spatial distribution characteristics of different dominant tree species.

[0003] At present, the carbon storage estimation technologies of forest ecosystems mainly include integration methods based on sample plot surveys and remote sensing data, geostatistics and spatial interpolation methods based on the spatial correlation thinking of the first law of geography, the application of lidar technology, machine learning and deep learning methods, ecological models and process simulation methods, etc.

[0004] Technical solutions of existing technology:

[0005] (1) Integration of plot surveys and remote sensing data: This method measures and estimates the carbon content of vegetation and soil by setting up plots in different types of forest ecosystems and combining them with high-resolution remote sensing imagery. By analyzing the field data obtained from the plot surveys and the remote sensing data, it is possible to estimate the carbon storage of large-scale forest ecosystems.

[0006] (2) Geostatistics and spatial interpolation methods that apply the first law of geography, spatial correlation thinking: The closer the distance between geographical elements, the stronger their correlation. Geostatistics and spatial interpolation techniques are widely used in estimating carbon storage in forest ecosystems. These methods can effectively use limited sample data to infer the distribution of carbon storage in the entire study area. Spatial interpolation methods such as kriging and pan-kriging can be used to estimate and predict carbon parameters in unmeasured areas.

[0007] (3) Application of LiDAR technology: LiDAR technology can provide high-precision three-dimensional structural information of the ground and vegetation, and is widely used in estimating carbon storage in forest ecosystems. Using LiDAR data, the height, density, and volume of forest vegetation can be accurately measured, thereby inferring the biomass and carbon storage of vegetation.

[0008] (4) Machine learning and deep learning methods: Machine learning and deep learning technologies play an important role in estimating carbon storage in forest ecosystems. By utilizing a large amount of sample plot data and remote sensing data, establishing a carbon storage estimation model, and using machine learning and deep learning algorithms for training and optimization, accurate carbon storage estimation can be achieved.

[0009] (5) Ecological model and process simulation methods: Ecological model and process simulation methods can use the theories of forest ecology and the laws of ecological processes to establish mathematical models to simulate forest growth and carbon cycle processes, thereby inferring the carbon storage of forest ecosystems. This method can take into account the dynamic changes and complexity of ecosystems and has certain advantages in estimating forest carbon storage. In addition, there are also methods that rely entirely on forest resource survey data to estimate carbon storage, but these methods require high comprehensiveness, detail, and accuracy of data acquisition.

[0010] Disadvantages of existing technology:

[0011] (1) Currently, few studies have estimated the carbon storage of multiple dominant tree species, and most studies focus on a single tree species;

[0012] (2) Most studies did not conduct further research after obtaining data on dominant tree species, which lacked applicability;

[0013] (3) Using forest resource inventory data and high temporal and spatial resolution remote sensing data to more accurately estimate the carbon storage of different dominant tree species is an important direction for future research. Summary of the Invention

[0014] The main purpose of this invention is to provide a method for measuring forest ecosystem carbon reserves based on multi-source remote sensing data. This method addresses the difficulties of existing ground forest resource inventories, the lack of precise forest species classification, and the inaccurate measurement of carbon reserves. This method uses multi-source remote sensing data to obtain multi-feature and multi-temporal forest characteristic data and conducts combined dimensionality reduction and separability analysis. This method is used to extract the spatial distribution of dominant tree species, fit linear models of forest volume and biomass, and estimate the spatial distribution and characteristics of carbon reserves in the study area. This method provides basic data and a practical basis for further strengthening forest ecosystem protection and implementing precise measures to increase forest carbon reserves.

[0015] The technical solution adopted by the present invention is: a method for measuring forest ecosystem carbon storage based on multi-source remote sensing data, comprising:

[0016] Extract, construct and analyze a multi-temporal and multi-feature classification dataset of dominant tree species based on Sentinel-1 and Sentinel-2 multi-source remote sensing data and forest resource inventory vector data of the study area;

[0017] Classify dominant tree species based on multi-feature and multi-temporal remote sensing data;

[0018] Estimate the forest carbon storage of different dominant tree species based on the optimal classification dataset and forest stock volume.

[0019] Furthermore, the extraction, construction and analysis of the multi-temporal and multi-feature classification dataset of dominant tree species based on Sentinel-1 and Sentinel-2 multi-source remote sensing data and the vector data of forest resources inventory in the study area include:

[0020] With the help of the ESTARFM model, fuse and supplement to generate long-time series high-resolution image data; through two sets of low-resolution remote sensing images and high-resolution remote sensing images at times t1 and t2, and a low-resolution remote sensing image at time t p (1 < p < 2), to predict the high-resolution remote sensing image at time t p (1 < p < 2);

[0021] Using the vector data of forest resources inventory, obtain the sample points of dominant tree species;

[0022] Based on the eigenvalue of each band of Sentinel-2 and Sentinel-1, obtain the spatial texture features, vegetation indices, and VV and VH scattering coefficients respectively; analyze and compare the differences in each eigenvalue and the time variation law of different dominant tree species, especially the comparison between the growing season and the non-growing season.

[0023] Even further, the spatial texture features include: [[ID=I]]

[0024] Mean: Variance: Contrast: Entropy: Correlation: Variance: / / 这里原文重复了“方差”,译文保留 Angular second moment: Homogeneity: Even further, the vegetation indices include: [[ID=I]]

[0025] Kernel normalized difference vegetation index, specifically expressed as:

[0026]

[0027] NIR and RED are the reflectance values of the near-infrared, red, and blue bands of the image respectively;

[0028]

[0029] Where σ is the length scale parameter specified in each specific application, indicating the sensitivity of the index to sparse or dense vegetation areas; a reasonable choice is to take the average value σ = 0.5(NIR + RED), and the simplified formula is expressed as kNDVI = tanh(NDVI2 );

[0030] Enhanced Vegetation Index, specifically expressed as:

[0031]

[0032] Among them, BLUE is the blue band;

[0033] Soil Adjusted Vegetation Index, specifically expressed as:

[0034]

[0035] The green chlorophyll vegetation index is expressed as:

[0036]

[0037] Ratio vegetation index, specifically expressed as:

[0038]

[0039] The weighted difference vegetation index is specifically expressed as:

[0040] WDVI=NIR-g×RED.

[0041] Furthermore, the classification of dominant tree species based on multi-feature and multi-temporal remote sensing data includes:

[0042] The obtained single-phase multi-feature and multi-phase multi-feature data sets were optimized by principal component analysis to obtain the principal components that meet the contribution rate; the separability of the optimized feature sets was calculated using the JM distance to complete the sample and dataset quality evaluation; finally, the dominant tree species classification and classification accuracy evaluation of the 13 optimized feature sets were carried out based on the SVM and RF models to select the optimal classification option.

[0043] Furthermore, the estimation of forest carbon storage of different dominant tree species based on the optimal classification data set and forest stock volume includes:

[0044] Based on the correlation analysis between the optimal classification data set and forest stock volume, a stock volume regression model was constructed by selecting strong correlation factors to calculate the stock volume of each dominant tree species. At the same time, a conversion equation for the stock volume-biomass-carbon storage of dominant tree species was constructed to evaluate and analyze the carbon density, storage and spatial distribution of dominant tree species.

[0045] The linear regression model in the empirical parameter model was used to achieve biomass prediction through the existing regression model; the stock-biomass conversion model was used to estimate the biomass of dominant tree species, and the average biomass model was used to estimate the biomass of some dominant tree species that lacked biomass fitting models;

[0046] Biomass was converted into carbon storage based on the carbon content coefficients of different dominant tree species, and above-ground and underground carbon storage were calculated based on the ratios:

[0047] (1) Stock-biomass conversion model:

[0048] Bio Q =a×V+b

[0049] In the formula, Bio Q It represents the forest biomass per unit area, in t / ha; V represents the forest stock per unit area, in m 3 / ha; a and b are regression parameters, which are obtained through regression simulation of existing sample points;

[0050] (2) Average biomass model:

[0051] Bio J =A×S

[0052] Where Bio J is the biomass of a certain type of dominant tree species, t, A is the unit biomass of this type of dominant tree species, t / ha, S is the area of ​​this type of dominant tree species, unit ha;

[0053] (3) Carbon storage conversion:

[0054] CS Above =Bio*Factor

[0055] CS Under =CS Above *R

[0056] CS Total =CS Above +CS Under

[0057] CS Above Aboveground carbon storage of dominant tree species, CS Under Belowground carbon storage of dominant tree species, CS Total The total carbon storage of dominant tree species, Factor is the carbon content of dominant tree species, and R is the ratio of above-ground to underground carbon storage.

[0058] Advantages of the present invention:

[0059] The present invention uses forest resource inventory data and high temporal and spatial resolution remote sensing data to analyze and compare the differences in characteristic values ​​of different dominant tree species, thereby achieving the estimation of carbon storage of multiple dominant tree species.

[0060] A multi-temporal and multi-feature classification dataset was constructed and optimized, and SVM and RF were used to classify the dominant tree species in the classification set, which improved the classification accuracy of dominant tree species.

[0061] With the help of forest resource inventory data and high temporal and spatial resolution remote sensing data, the uncertainty of single data is avoided, and the influence of other factors on single data is avoided, thereby improving the accuracy of the data and obtaining more accurate and reliable carbon storage of different dominant tree species.

[0062] In addition to the above-described objects, features and advantages, the present invention has other objects, features and advantages. The present invention will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The illustrative embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0064] Figure 1 This is a flow chart of the method for determining forest ecosystem carbon reserves based on multi-source remote sensing data of the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0066] Figure 1 The flowchart of the method for determining the carbon storage of forest ecosystems based on multi-source remote sensing data of the present invention is shown.

[0067] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0068] The present invention is based on multi-source remote sensing data such as Sentinel-1 and Sentinel-2 and forest resource inventory vector data, and obtains surface reflectance, vegetation characteristic index, spatial texture characteristics and backscattering coefficient rate in multiple bands of the study area. It adopts principal component dimensionality reduction and performs separability calculation of different dominant tree species. It uses two classifiers, random forest model and support vector machine, to classify dominant tree species, and selects out-of-bag data for accuracy evaluation. In this way, a detailed distribution map of forest types in the study area is obtained. Using the preferred characteristic variables, the mathematical modeling method of accumulation and biomass is introduced to estimate the carbon storage-related parameters of the study area and analyze the spatial distribution characteristics. The steps are as follows:

[0069] 1. Extract, construct, and analyze a multi-temporal and multi-feature classification dataset of dominant tree species based on multi-source remote sensing data such as Sentinel-1 and Sentinel-2 and vector data from the forest inventory of the study area.

[0070] The specific steps include:

[0071] (1) With the help of the ESTARFM model, fuse and supplement to generate long - time - series high - resolution image data. Through two sets of low - resolution remote sensing images at times t1 and t2, high - resolution remote sensing images, and a low - resolution remote sensing image at time t p (1 < p < 2), predict the high - resolution remote sensing image at time t p (1 < p < 2).

[0072] (2) Use the forest resources inventory vector data to obtain the sample points of dominant tree species.

[0073] (3) Based on the characteristic values of each band of Sentinel - 2 and Sentinel - 1, obtain spatial texture features, vegetation indices, and VV and VH scattering coefficients respectively. Analyze and compare the differences in characteristic values and the time - varying rules of different dominant tree species, especially the comparison between the growing season and the non - growing season.

[0074] ① Spatial texture features:

[0075] The Gray Level Co - occurrence Matrix (GLCM) is very effective in interpreting and processing remote sensing images and is currently the most commonly used method for calculating spatial texture. In this study, 8 spatial texture features were extracted using GLCM based on the ENVI platform, namely Mean, Variance, Homogeneity, Contrast, Dissimilarity, Entropy, Second Moment, and Correlation (Table 1). Specifically, they are expressed as:

[0076] Table 1 Calculation formulas for spatial texture features

[0077]

[0078]

[0079] ② Vegetation indices:

[0080] The kernel Normalized Difference Vegetation Index (kNDVI) can effectively address the saturation and mixed - pixel problems encountered by traditional indices and has strong applicability. Specifically, it is expressed as:

[0081]

[0082] NIR and RED are the reflectance values ​​of the near-infrared, red, and blue bands of the image, respectively.

[0083]

[0084] in σ is a length scale parameter specified in each specific application, which represents the sensitivity of the index to sparse or dense vegetation areas. A reasonable choice is to take the average value σ = 0.5 (NIR + RED), and the simplified formula is kNDVI = tanh (NDVI 2 ).

[0085] The Enhanced Vegetation Index (EVI) is a vegetation index widely used in remote sensing to monitor and assess the state, changes, and health of vegetation growth. It incorporates an atmospheric correction step in its calculations. Compared to the classic NDVI, the EVI incorporates the sensitivity of infrared bands and the spectral characteristics of vegetation cover, overcoming the shortcomings of the classic vegetation index in areas with high vegetation cover. This helps to more accurately estimate vegetation biomass. Specifically, it is expressed as:

[0086]

[0087] Among them, BLUE is the blue band

[0088] The Soil Adjusted Vegetation Index (SAVI) is a new index that reduces the sensitivity of the vegetation index to soil shading by introducing a soil adjustment parameter L (0>L>1, L=0.5). It is suitable for areas where high vegetation cover and bare soil coexist. It is specifically expressed as:

[0089]

[0090] The Green Chlorophyll Vegetation Index (GCVI) is calculated using information from the green band, making it more sensitive to the content of green chlorophyll, because green chlorophyll primarily absorbs light in the blue and red bands and relatively less in the green band. GCVI can provide a relative measure of the chlorophyll content of vegetation. Specifically expressed as:

[0091]

[0092] The Ratio Vegetation Index (RVI) uses spectral information from different bands to provide quantitative information about vegetation by calculating the ratio of the reflectance values ​​of these bands. This is used to monitor the chlorophyll synthesis status of the vegetation canopy. Higher RVI values ​​generally indicate richer and healthier vegetation. Specifically, it is expressed as:

[0093]

[0094] The Weighted Difference Vegetation Index (WDVI) is an optimized index that introduces a soil adjustment weight index g (g is generally set to 0.5) based on the difference vegetation index to eliminate the influence of soil background. It is specifically expressed as:

[0095] WDVI=NIR-g×RED

[0096] 2. Classify dominant tree species based on multi-feature and multi-temporal remote sensing data.

[0097] One of the embodiments of the present invention is to perform principal component analysis optimization processing on the obtained single-phase multi-feature (12) and multi-phase multi-feature (1) data sets to obtain the principal components that meet the contribution rate; use the JM distance to perform separability calculation on the optimized feature set to complete the sample and data set quality evaluation work; finally, based on the SVM and RF models, the 13 optimized feature sets are classified into dominant tree species and the classification accuracy is evaluated to select the optimal classification option.

[0098] 3. Estimate forest carbon storage of different dominant tree species based on the optimal classification dataset and forest stock.

[0099] Correlation analysis was conducted between the optimal classification dataset and forest stock volume. Strongly correlated factors were selected to construct a stock volume regression model to calculate the stock volume of each dominant tree species. A conversion equation for the stock volume-biomass-carbon storage of dominant tree species was also constructed to assess and analyze the carbon density, storage, and spatial distribution of dominant tree species.

[0100] The forest stock per unit area (abbreviated as unit stock) is the premise for calculating biomass. By constructing a regression model between the unit stock of the sample point and the preferred variable, the visualization of the stock in the study area can be achieved. The present invention uses the linear regression model in the empirical parameter model to realize biomass prediction through the existing regression model. The stock-biomass conversion model is used to estimate the biomass of the dominant tree species of the arbor type. The average biomass model is used to estimate the dominant tree species that lack a biomass fitting model, such as other shrubs, fruit trees, peach trees, etc. Finally, the biomass is converted into carbon storage based on the carbon content coefficient of different dominant tree species, and the above-ground and underground carbon storage are respectively calculated based on the ratio.

[0101] (1) Stock-biomass conversion model

[0102] Bio Q =a×V+b

[0103] In the formula, Bio Q It represents the forest biomass per unit area, in t / ha; V represents the forest stock per unit area, in m 3 / ha; a and b are regression parameters, which are obtained through regression simulation of existing sample points.

[0104] (2) Average biomass model

[0105] Bio J =A×S

[0106] Where Bio J is the biomass of a certain type of dominant tree species, t, A is the unit biomass of this type of dominant tree species, t / ha, and S is the area of ​​this type of dominant tree species, unit ha.

[0107] (3) Carbon storage conversion

[0108] CS Above =Bio*Factor

[0109] CS Under =CS Above *R

[0110] CS Total =CS Above +CS Under

[0111] CS Above Aboveground carbon storage of dominant tree species, CS Under Belowground carbon storage of dominant tree species, CS Total The total carbon storage of dominant tree species, Factor is the carbon content of dominant tree species, and R is the ratio of above-ground to underground carbon storage.

[0112] This paper proposes the basic concepts and specific implementation paths for mapping carbon density of dominant tree species, estimating carbon reserves, and analyzing spatial characteristics. This provides data support and a theoretical foundation for evaluating forest ecosystem services and assessing the economic value of carbon offsets. It also provides decision-making and action guidance for assessing the effectiveness of forest resource conservation and utilization and further optimizing management measures.

[0113] The present invention is based on the Sentinel-1 and Sentinel-2 high temporal and spatial resolution multispectral image data sets, and obtains multiple vegetation indices, VV, VH polarization scattering coefficients, and homogeneous spatial texture features to preliminarily form a classification band set. The principal component method is used to compress the band set and the JM distance method is used to evaluate the separability of sample points and classification features. The random forest model (RF) and support vector machine (SVM) are used to classify the multi-phase and multi-feature band sets and determine the optimal classification combination. Finally, the accumulation-optimal feature set and biomass-accumulation fitting models are constructed in sequence, which are helpful to obtain relevant parameters such as the accumulation and carbon storage of dominant tree species and analyze the spatial distribution characteristics of carbon storage. The present invention intends to protect this process-based step for calculating carbon storage.

[0114] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for determining forest ecosystem carbon storage based on multi-source remote sensing data, characterized in that: include: Extract, construct and analyze a multi-temporal and multi-feature classification dataset of dominant tree species based on Sentinel-1 and Sentinel-2 multi-source remote sensing data and forest resource inventory vector data of the study area; Classify dominant tree species based on multi-feature and multi-temporal remote sensing data; The forest carbon storage of different dominant tree species was estimated based on the optimal classification data set and forest stock volume.

2. The method for determining forest ecosystem carbon storage based on multi-source remote sensing data according to claim 1, characterized in that: The extraction, construction and analysis of the multi-temporal and multi-feature classification dataset of dominant tree species based on Sentinel-1 and Sentinel-2 multi-source remote sensing data and forest resource inventory vector data of the study area include: With the ESTARFM model, fuse and supplement to generate long-time series high-resolution image data; through two sets of low-resolution remote sensing images and high-resolution remote sensing images at t1 and t2 moments, and a low-resolution remote sensing image at t p (1 < p < 2) moment, to predict the high-resolution remote sensing image at t p (1 < p < 2) moment; Use forest resource inventory vector data to obtain sample points of dominant tree species; Based on the characteristic values ​​of each band of Sentinel-2 and Sentinel-1, spatial texture characteristics, vegetation index, and VV and VH scattering coefficients were obtained respectively; the differences in the characteristic values ​​of different dominant tree species and the temporal variation patterns were analyzed and compared, especially the comparison between the growing season and the non-growing season.

3. The method for determining forest ecosystem carbon storage based on multi-source remote sensing data according to claim 2, characterized in that: The spatial texture features include: Mean: Differences: Contrast ratio: entropy: Dependencies: variance: Angular second moment: Homogeneity: N is the size of the gray-level co-occurrence matrix N×N, i and j are gray values, P i,j is the gray-level co-occurrence matrix, μ i is the row mean, μ j is the mean of the column, σ i is the standard deviation of the row, σ j is the standard deviation of the column.

4. The method for determining forest ecosystem carbon storage based on multi-source remote sensing data according to claim 2, characterized in that: The vegetation indices include: The kernel normalized difference vegetation index is expressed as: NIR and RED are the reflectance values ​​of the near-infrared, red, and blue bands of the image, respectively; where σ is a length scale parameter specified in each specific application, representing the sensitivity of the index to sparse or dense vegetation areas; a reasonable choice is to take the average value σ = 0.5 (NIR + RED), and the simplified formula is kNDVI = tanh (NDVI 2 ); Enhanced Vegetation Index, specifically expressed as: Among them, BLUE is the blue band; Soil Adjusted Vegetation Index, specifically expressed as: Green chlorophyll vegetation index, specifically expressed as: Ratio vegetation index, specifically expressed as: The weighted difference vegetation index is specifically expressed as: WDVI=NIR-g×RED.

5. The method for determining forest ecosystem carbon storage based on multi-source remote sensing data according to claim 1, characterized in that: The classification of dominant tree species based on multi-feature and multi-temporal remote sensing data includes: The obtained single-phase multi-feature and multi-phase multi-feature data sets were optimized by principal component analysis to obtain the principal components that meet the contribution rate; the separability of the optimized feature sets was calculated using the JM distance to complete the sample and dataset quality evaluation; finally, the dominant tree species classification and classification accuracy evaluation of the 13 optimized feature sets were carried out based on the SVM and RF models to select the optimal classification option.

6. The method for determining forest ecosystem carbon storage based on multi-source remote sensing data according to claim 1, characterized in that: The estimation of forest carbon stocks of different dominant tree species based on the optimal classification data set and forest stock volume includes: Based on the correlation analysis between the optimal classification data set and forest stock volume, a stock volume regression model was constructed by selecting strong correlation factors to calculate the stock volume of each dominant tree species. At the same time, a conversion equation for the stock volume-biomass-carbon storage of dominant tree species was constructed to evaluate and analyze the carbon density, storage and spatial distribution of dominant tree species. The linear regression model in the empirical parameter model was used to achieve biomass prediction through the existing regression model; the stock-biomass conversion model was used to estimate the biomass of dominant tree species, and the average biomass model was used to estimate the biomass of some dominant tree species that lacked biomass fitting models; Biomass was converted into carbon storage based on the carbon content coefficients of different dominant tree species, and above-ground and underground carbon storage were calculated based on the ratios: (1) Stock-biomass conversion model: Cinema Q =a×V+b In the formula, Bio Q It represents the forest biomass per unit area, in t / ha; V represents the forest stock per unit area, in m 3 / ha; a and b are regression parameters, which are obtained through regression simulation of existing sample points; (2) Average biomass model: Was J =A×S Where Bio J is the biomass of a certain type of dominant tree species, t, A is the unit biomass of this type of dominant tree species, t / ha, S is the area of ​​this type of dominant tree species, unit ha; (3) Carbon storage conversion: CS Above =Bio*Factor CS Under =CS Above *R CS Total =CS Above +CS Under CS Above Aboveground carbon storage of dominant tree species, CS Under Belowground carbon storage of dominant tree species, CS Total The total carbon storage of dominant tree species, Factor is the carbon content of dominant tree species, and R is the ratio of above-ground to underground carbon storage.

Images