Carbon reserve estimation method for ecological system

Abstract

The invention provides a carbon reserve estimation method for an ecological system, and the method comprises the steps: S11, recognizing tree species, herbaceous plants and soil in an obtained remote sensing satellite image, and generating a land utilization / land planting and covering current situation map; s12, determining the accuracy degree of the land utilization / land planting and covering current situation map according to a preset grading and classification rule, carrying out grid division, obtaining a plurality of estimation regions, and obtaining sample data in the estimation regions; s13, estimating and analyzing the carbon density by taking the individual trees in the estimation area as basic units, and generating a carbon density attribute table; s14, carrying out spatial position attribute connection on the carbon density attribute table and the land utilization / land planting and covering current situation map, and then constructing an ecological carbon density database; and S15, linking the ecological carbon density database with the InVEST model to estimate the carbon reserve. According to the method, the remote sensing image, the ground survey data and the InVEST model are combined, the carbon reserve of the research area is accurately estimated, and the precision and the reliability of a carbon reserve estimation result are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon reserve estimation, and in particular to a carbon reserve estimation method for an ecosystem. Background Art

[0002] Currently, the main methods for estimating carbon storage in terrestrial ecosystems include plot inventory, model estimation, and remote sensing estimation. The plot inventory method obtains vegetation information from typical plots and calculates vegetation carbon storage with the help of relevant models. This method is technically simple and highly reliable. The model estimation method, with the InVEST model as a typical example, has the advantages of easy data acquisition, simple operation, and strong visualization. However, due to large errors in parameters such as carbon density and carbon storage coefficient, the reliability of the estimation results is relatively low. Remote sensing estimation is often used in carbon sequestration research in medium and large-scale areas. However, remote sensing methods such as satellites and airborne lidar have problems such as low accuracy in individual tree segmentation and difficulty in obtaining breast diameter and accurate tree height. Summary of the Invention

[0003] The purpose of this invention is to provide a carbon storage estimation method for ecosystems. By combining remote sensing images, ground survey data and the InVEST model, the carbon storage of the study area can be accurately estimated, thereby improving the accuracy and reliability of the carbon storage estimation results.

[0004] To achieve the above object, the present invention provides a method for estimating carbon storage in an ecosystem, the method comprising: S11. Obtain remote sensing satellite images of the field investigation study area, identify tree species, herbs, and soil in the remote sensing satellite images, and generate a land use / land cover status map; S12. Determine the accuracy of the land use / land cover status map according to preset classification rules and perform grid division to obtain a number of estimated areas, and acquire data on above-ground samples, underground samples, and soil samples in the estimated areas; S13. Using individual trees in the estimation area as the basic unit, estimate and analyze the aboveground vegetation carbon density, belowground vegetation carbon density, soil carbon density, and dead organic matter carbon density to generate a carbon density attribute table; S14, after connecting the carbon density attribute table with the land use / land cover status map based on spatial location attributes, an ecological carbon density database is constructed; S15. Link the ecological carbon density database with the InVEST model to estimate carbon storage, and obtain the current estimated regional total carbon storage, aboveground vegetation carbon storage, belowground vegetation carbon storage, soil carbon storage, and dead organic matter carbon storage.

[0005] Furthermore, in step S11, tree species, herbs and soil in the remote sensing satellite image are identified, specifically including: The remote sensing satellite images of the field investigation study area are sorted according to the acquisition time to obtain an image sequence; Extracting image data at several different acquisition moments from the image sequence according to preset rules and marking them as target image data; Extract image features from target image data and divide the image features into tree species features, herb features and soil features; Boundary identification is performed on tree species, herbaceous features, and soil features, and the identification results are marked with color values ​​to generate a current land use / land cover map.

[0006] Furthermore, in step S12, the accuracy of the field investigation study area is determined using the national land cover level III classification system.

[0007] Furthermore, in step S12, the land use / land cover status map is gridded, specifically including: Collect historical data sets that grid land use / land cover maps based on accuracy, and use the historical data sets to pre-train a land gridding model to determine the relationship between accuracy and land use / land cover map division; The currently determined accuracy and the land use / land cover status map are gridded using a land grid division model to obtain a number of estimated areas.

[0008] Furthermore, the land grid division model is pre-trained using historical data sets to determine the relationship between the degree of accuracy and the division of the land use / land cover status map, including: Collect gridded land use / land cover maps for different years, different types of regions, and different vegetation cover conditions, as well as the corresponding field survey accuracy data; For each grid cell in the land use / land cover map, extract the type information and the corresponding accuracy data, perform matching coding, and construct a sample data set; The sample data set is input into the spatial structure extraction branch and the accuracy extraction branch of the land grid division model to obtain spatial data and grid data respectively; The spatial data and grid data are integrated to determine the relationship between the accuracy and the division of the land use / land cover status map.

[0009] Furthermore, data on above-ground samples, underground samples, and soil samples in the estimated area are obtained, specifically including: Obtain the diameter at breast height (DBH) and height of the tree species in the above-ground samples within the estimated area. The DBH is measured starting at 5.0 cm, 1.3 m from the root collar of the upslope root of the tree trunk. If a tree has an irregular trunk shape, measure the diameter in two vertical directions and take the average value. The tree height is measured using a stadiometer. The tree density is estimated using the plot method combined with the spacing between trees to obtain the above-ground sample data. Randomly enclose a 2m×2m to 5m×5m box within the estimated area corresponding to the underground sample. After recording the herb and litter species and their coverage, all above-ground herbs and litter within the box are harvested. 300g of herbs and litter are collected from each box. The collected herbs and litter are dried at a preset temperature and weighed to obtain underground sample data. The soil drilling method and the profile method were used to sample, analyze and determine the estimated soil organic carbon in the area to obtain soil data.

[0010] Furthermore, in step S13, the generation of the carbon density attribute table specifically includes: Based on the aboveground sample data, the biomass of individual trees is calculated. Based on the biomass of individual trees, herbaceous data, and the land use / land cover status map, the total aboveground vegetation biomass is estimated using the biomass equation, thereby calculating the total aboveground vegetation carbon density. The litter biomass was estimated based on the litter data, and thus the dead organic matter carbon density was calculated; The underground biomass is estimated based on the aboveground biomass, thereby calculating the underground vegetation carbon density; The soil carbon density is calculated by estimating the soil layer carbon density based on the acquired soil data; Based on the design tables of aboveground vegetation carbon density, dead organic matter carbon density, underground vegetation carbon density and soil carbon density, a carbon density attribute table is generated.

[0011] Furthermore, in step S14, the construction of the ecological carbon density database specifically includes: Based on the carbon density attribute table and the land use / land cover status map, a database structured carbon density attribute table and land use / land cover status map association table are constructed; Digitize and geocode the vector or raster data in the database structure land use / land vegetation map association table, and assign a unique identifier to each land unit / vegetation unit; The carbon density attributes corresponding to each land unit / vegetation unit identifier in the database structure carbon density attribute table are linked through smart contracts to construct an ecological carbon density database.

[0012] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a method for estimating carbon storage in ecosystems, which adopts a gridded, high-precision ground field survey method. It can more accurately grasp the growth status of ground vegetation and record key data such as tree species, diameter at breast height and tree height in detail, so as to more accurately estimate the carbon storage in the region. By assigning carbon density values ​​to each patch of the land use map, each patch will be connected to all attributes of the sampling point closest to it, including above-ground carbon density, underground carbon density, soil carbon density and dead organic matter carbon density, which significantly improves the coverage of carbon density values ​​and more intuitively presents the spatial differences in carbon storage in different regions. Combining the InVEST model for carbon storage estimation not only improves the efficiency and quality of data processing, but also generates a raster map of carbon storage distribution, which helps to more comprehensively and scientifically reflect the carbon storage status of regional ecosystems. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work. Figure 1 A schematic flow chart of a method for estimating carbon reserves in an ecosystem provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0014] The present invention will be further described in detail below with reference to the accompanying drawings and examples. It will be understood that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present invention, not all structures.

[0015] Reference Figure 1 This embodiment provides a method for estimating carbon storage in an ecosystem, the method comprising: S11. Obtain remote sensing satellite images of the field investigation study area, identify tree species, herbs and soil in the remote sensing satellite images, and generate a land use / land cover status map.

[0016] S12. Determine the accuracy of the land use / land cover status map according to preset classification rules and perform grid division to obtain several estimated areas, and acquire data on above-ground samples, underground samples, and soil samples in the estimated areas.

[0017] S13. Taking individual trees in the estimation area as the basic unit, estimate and analyze the carbon density of aboveground vegetation, belowground vegetation, soil carbon density, and dead organic matter carbon density to generate a carbon density attribute table.

[0018] S14. After connecting the carbon density attribute table with the land use / land cover status map based on spatial location attributes, an ecological carbon density database is constructed.

[0019] S15. Link the ecological carbon density database with the InVEST model to estimate carbon storage, and obtain the current estimated regional total carbon storage, aboveground vegetation carbon storage, belowground vegetation carbon storage, soil carbon storage, and dead organic matter carbon storage.

[0020] In this embodiment, the remote sensing satellite images obtained from the field investigation study area are used to identify and distinguish tree species, herbs and soil through classification algorithms, and are assigned color values ​​to represent them, thereby generating a land use / land cover status map. To determine the accuracy of the ground survey, a grid-based, high-density field survey is conducted based on the land use / land cover status map on the basis of fully collecting the latest land surveys, forest inventories and related research literature. The main data of the survey include above-ground samples, underground samples and soil sample data in the estimated area in the land use / land cover status map. Taking a single tree as the basic unit, based on the biomass model of each tree species, the carbon density of a single tree is explored, and through plot surveys and soil sampling analysis, the carbon density of different vegetation types and soils is estimated and analyzed, thereby obtaining the carbon density of the tree layer, shrub layer, herb layer and soil layer, and then calculating the carbon density of above-ground vegetation, underground vegetation, soil carbon density and dead organic matter carbon density, thereby generating a carbon density attribute table.

[0021] The carbon density attribute table estimated through field surveys is assigned its attributes to the patches of the current land use / land cover map (each patch will be assigned all the attributes of the point closest to it, namely aboveground carbon density, belowground carbon density, soil carbon density, and dead organic matter carbon density). For example, the attributes in the carbon density attribute table are assigned to the current land use / land cover map through ArcGIS software. To ensure accuracy, each land type and the points that fall into that land type are extracted and connected separately to construct an ecological carbon density database. In the InVEST model, the land use map within the region is a raster file. After rasterizing the data in the ecological carbon density database, it is imported into the InVEST model to obtain the current estimated total regional carbon storage, aboveground vegetation carbon storage, belowground vegetation carbon storage, soil carbon storage, and dead organic matter carbon storage. For example, the data in the ecological carbon density database is rasterized using ArcGIS software.

[0022] Specifically, the attributes in the carbon density attribute table were assigned to the land use / land cover status map using Arcgis software. The main steps are as follows: ①Divide remote sensing satellite images into patches according to tree species, herbs and soil Right-click the data element in the list on the left - open the attribute table - table options - select by attribute, divide the map into tree species, herbs and soil, and place the vector files in the specified folder after classification.

[0023] Sample operation steps (taking exporting grassland land class vector files as an example): Enter "dlmc" = 'Grassland' in the tool's expression and select Apply to filter out the target data.

[0024] Finally, the vector file Agrassland.shp of the grassland land class is obtained (the other land classes use the same naming rule, Ai, where i represents the land class).

[0025] ② Derive carbon density attribute points falling into different land types Import the carbon density attribute table C of all survey points into the regional grassland land class map A grassland.shp. According to Select - Select by Location, select the carbon density attribute points that fall into different land classes and export them.

[0026] Example steps: Select Selection by Location in the menu bar. Selection method: Select features from the following layers Target layer: Carbon density attribute table C Source layer: the land class file A (grassland.shp) selected in step ① Spatial selection method for target layer features: Intersect with source layer features After the required data is selected, right-click the point file - data - export data, and make sure to set the export to the selected elements. The point data C grassland.shp (carbon density points that fall into the grassland land class) is saved in the specified folder.

[0027] ③ Image matching carbon pool The carbon density attributes of the survey points where each land type falls, namely the carbon pool (aboveground vegetation carbon density, belowground vegetation carbon density, soil carbon density and dead organic matter carbon density) are connected to the land type map unit.

[0028] Example steps: Import the land class file A Grassland.shp and its divided point data C Grassland.shp into the workspace Right click on the land class file - Connect and Associate - Connect Select another layer of data based on spatial location Select C Grassland.shp data in the connection layer Connection rules: Each polygon will be assigned all the attributes of the point closest to its border, as well as a distance field showing the proximity to the point (in the units of the target layer). Points that fall inside the polygon are considered to be the closest line to the polygon (i.e., the distance is 0). Name the layer output data A'Grassland.shp and save it to the specified folder. Click OK to perform the spatial connection.

[0029] ④Merge spots and set identification codes The above steps ①②③ are carried out for each land type, and finally different land type files A'i.shp are obtained; there are two ways to merge land type files. The first is to use the merge tool, and the second is to use the append tool. The merge tool is prone to crash when processing large amounts of data. The append tool can add small amounts in sequence to reduce the data volume and computing pressure. It is recommended to use the append tool first.

[0030] Example steps: Select Toolbox - Data Management Tools - General - Append After all data are appended using the append tool, a file is generated with a land use / land cover status map M' containing carbon density data; Right-click to open the attribute table - Add field - Field name (can be named identification code) - Field attribute select double precision The last column of the attribute table is the latest added field, right-click and click Field calculator a. Select the scripting language: the default is VB, which needs to be manually changed to Python; b. Double-click the field name of a feature or table, and the code for the field name will be automatically generated in the calculator box; c. By default, [Show code blocks] is not checked. If the calculation script is complex and requires multiple lines, check it; d. After completing the field calculation code, click OK to generate the field value. Note: The primary purpose of this step is to establish a spatial association between the land use / vegetation map raster data and the carbon density attribute table to meet the input requirements of the InVEST model. Unique identifiers can be set using a sequential numbering scheme from 1 to n (ensuring no duplication with other field values ​​in the attribute table) to facilitate subsequent spatial analysis and calculations.

[0031] ⑤Carbon density data export By exporting and processing the attribute table of the land use / land cover status map M', an ecological carbon density database C" with the land use / land cover status map patches as units will be obtained.

[0032] In the InVEST model, the land use map of the region is a raster file. Therefore, it is necessary to convert the land use map M.shp file using ArcGIS software and rasterize the land use map. Example steps: Toolbox - Conversion Tools - Convert to Raster - Polygon to Raster File M".

[0033] Carbon stock estimation: Open the InVEST model, select the carbon storage and sequestration module, import the ecological carbon density database, and run the model for calculation.

[0034] Example steps: a. Convert the ecological carbon density database C" into csv format; b. Create a new folder E, which only contains carbon pool data and raster files; c. Open InVEST, select the Carbon Storage and Sequestration module, and set the workspace to the specific folder E; d. The current LULC is a raster file M"; e. Select the carbon density data table C in the carbon pool. csv file; f. Operation; Finally, the total regional carbon storage, aboveground vegetation carbon storage, belowground vegetation carbon storage, soil carbon storage and dead organic matter carbon storage were obtained.

[0035] As a preferred embodiment, in step S11, identifying tree species, herbs and soil in remote sensing satellite images specifically includes: The remote sensing satellite images of the field investigation study area are sorted according to the acquisition time to obtain an image sequence.

[0036] Image data at several different acquisition times are extracted from the image sequence according to preset rules and marked as target image data.

[0037] Extract image features from target image data and divide the image features into tree species features, herb features and soil features.

[0038] Boundary identification is performed on tree species, herbaceous features, and soil features, and the identification results are marked with color values ​​to generate a current land use / land cover map.

[0039] In this embodiment, carbon density changes over time. For example, trees absorb carbon dioxide as they grow, increasing their carbon storage. By sorting remote sensing satellite images by acquisition time to form an image sequence, carbon absorption by vegetation at different growth stages can be observed. Target image data at different acquisition times is extracted from the image sequence to capture the dynamic changes in carbon density. For example, in a forest ecosystem, trees begin to sprout in spring, gradually increasing their carbon absorption capacity; in summer, trees grow vigorously, reaching a high carbon density; and in autumn, some trees shed their leaves, decreasing their carbon absorption capacity.

[0040] Remote sensing satellite images are sorted by acquisition time to form an image sequence, allowing for clear observation of changes in ground features over time in the same area. This provides a foundation for subsequent time series analysis. For example, this can be used to analyze vegetation growth cycles, such as tree budding in spring, lush growth in summer, and leaf drop in autumn, or to monitor soil moisture and erosion in different seasons.

[0041] Tree species characteristics, herb characteristics, and soil characteristics contribute differently to carbon density. Tree species are one of the main carriers of carbon storage in terrestrial ecosystems. Their trunks, branches, and leaves all contain large amounts of carbon. Different tree species have different carbon storage capacities. For example, coniferous forests and broad-leaved forests differ in carbon absorption and storage. Although herbaceous plants have smaller individual carbon reserves, they have short growth cycles and can absorb a certain amount of carbon in a short period of time. The organic carbon content in the soil is affected by a variety of factors, including vegetation type and soil texture. By dividing features and performing boundary identification and color value marking, a basis can be provided for subsequent estimation of tree species, herbaceous plants, and soil carbon density, thereby obtaining more accurate carbon density estimation results.

[0042] Specifically, by processing the remote sensing satellite images collected in real time in the same field investigation and study area and combining it with subsequent steps, the carbon storage change results of the field investigation and study area throughout the year can be obtained.

[0043] As a preferred embodiment, in step S12, the accuracy of the field survey study area is determined using the National Land Cover Classification System (Level III) for Ecological Carbon Sequestration Research (based on Level I and II classifications, with further subdivision of vegetation community types down to tree species) to determine the accuracy of the field survey study area.

[0044] As a preferred embodiment, in step S12, the land use / land cover status map is gridded, specifically including: Collect historical data sets for gridding land use / land cover maps based on the degree of accuracy, and pre-train the land gridding model with the historical data sets to determine the relationship between the degree of accuracy and the division of land use / land cover maps.

[0045] The currently determined accuracy and the land use / land cover status map are gridded using a land grid division model to obtain a number of estimated areas.

[0046] In this embodiment, traditional land map gridding often requires extensive manual effort. Due to differences in operator experience, understanding, and work habits, it is difficult to ensure consistent gridding results. Machine learning extracts insights from a large amount of historical land map gridding data, comprehensively considering multiple land attributes to achieve precise gridding based on the actual characteristics of the existing land map. Land use and vegetation cover types vary significantly across regions. Gridding allows carbon storage estimates to be performed for each grid area, taking into account specific characteristics such as soil type, vegetation cover, and topography. For example, in areas where forestland and grassland are adjacent, gridding allows separate estimates to be made for the two areas, improving estimation accuracy. Gridded land maps are used for long-term dynamic monitoring. Carbon storage estimates for each grid area can be updated using remote sensing imagery or other data acquired over time. For example, in an area undergoing forest restoration, gridding allows accurate monitoring of forest vegetation growth and carbon storage increases in each grid area.

[0047] As a preferred embodiment, the relationship between the accuracy and the division of the land use / land cover status map is determined by pre-training the land grid division model with the historical data set, specifically including: Collect gridded land use / land cover maps of different years, different types of areas, and different vegetation coverage conditions, as well as the corresponding field survey accuracy data.

[0048] For each grid cell in the land use / land cover map, the type information and the corresponding accuracy data are extracted, matched and coded, and a sample dataset is constructed.

[0049] The sample data sets are input into the spatial structure extraction branch and the accuracy extraction branch of the land grid division model to obtain spatial data and grid data respectively.

[0050] The spatial data and grid data are integrated to determine the relationship between the accuracy and the division of the land use / land cover status map.

[0051] In this embodiment, land maps from different years, different types of regions, and different vegetation cover conditions, along with field survey accuracy data, are collected, exposing the land gridding model to a wide variety of land use and vegetation cover scenarios. Through matching coding, the type information of each grid cell in the land map is combined with the accuracy data to accurately extract the characteristics of each grid cell, thereby learning how to divide different types of land and vegetation at different levels of accuracy. For example, in forest areas, the distribution of different tree species requires different levels of accuracy. Through matching coding, the model can learn that coniferous forest areas, due to their relatively dense distribution and consistent growth patterns, require slightly larger grids for division; whereas mixed forest areas, due to the diverse and complex distribution of vegetation types, require smaller grids for accurate division. The sample dataset is input into the spatial structure extraction branch and the accuracy extraction branch of the land gridding model to obtain spatial data and grid data, respectively. The spatial structure extraction branch focuses on the spatial distribution characteristics of land use and vegetation cover, such as shape, size, and adjacency; the accuracy extraction branch focuses on understanding the scale and details of grid divisions under different accuracy requirements.

[0052] Specifically, land maps of different years, different types of regions, and different vegetation cover conditions, as well as corresponding field survey accuracy data, are collected. For each grid cell in the land map, its type information and corresponding accuracy data are extracted and matched and coded. For example, land use types are categorized as farmland, woodland, grassland, etc., and each is assigned a different code, such as 1 for farmland, 2 for woodland, and 3 for grassland. Accuracy data is categorized into three levels: high, medium, and low, represented by codes H, M, and L, respectively. Then, for each grid cell, its land use type code and accuracy code are combined into a composite code. For example, a grid cell with forestland and high accuracy is coded as 3H.

[0053] The constructed sample dataset was fed into the spatial structure extraction branch and the precision extraction branch of a land gridding model constructed using a deep learning framework. The spatial structure extraction branch used a convolutional neural network (CNN) to extract spatial distribution features of land use and vegetation cover, such as texture, shape, and edges. The precision extraction branch employed a multi-layer perceptron (MLP) to analyze the relationship between different types of information and precision. A loss function was used to optimize the land gridding model parameters, enabling the model to accurately determine the relationship between precision and land map divisions. The feature vectors obtained from the spatial structure extraction branch and the precision extraction branch were concatenated and fed into the final division judgment layer as fused features. An attention mechanism was used to emphasize the feature dimensions that determine division relationships. Finally, the model was tested on a validation set, and metrics such as division accuracy, recall, and F1 score were calculated to evaluate the performance of the land gridding model.

[0054] As a preferred embodiment, acquiring data of above-ground samples, underground samples, and soil samples in the estimation area specifically includes: Obtain the diameter at breast height and tree height of the tree species in the above-ground samples within the estimation area. The starting point for the diameter at breast height is 5.0 cm, and the measurement is carried out at a distance of 1.3 m from the root neck of the upslope of the trunk. If there are trees with irregular trunk shapes, the diameters in two directions are measured vertically and the average value is taken. The tree height is measured using a height meter. The forest density is estimated using the plot method combined with the spacing between trees to obtain the above-ground sample data.

[0055] Randomly select boxes of 2m×2m~5m×5m in the estimated area corresponding to the underground sample, record the types of herbs and litter and their coverage, and then harvest all the above-ground herbs and litter in the boxes. Based on the divided boxes, collect 300g of herbs and litter respectively, dry them at a preset temperature, weigh them and record them to obtain the underground sample data.

[0056] The soil drilling method and the profile method were used to sample, analyze and determine the estimated soil organic carbon in the area to obtain soil data.

[0057] As a preferred embodiment, in step S13, generating the carbon density attribute table specifically includes: Based on the above-ground sample data, the biomass of individual trees is calculated. Based on the biomass of individual trees, herbaceous data and the land use / land cover status map, the total above-ground vegetation biomass is estimated through the biomass equation, thereby calculating the total above-ground vegetation carbon density.

[0058] Litter biomass was estimated based on litter data, and thus the dead organic matter carbon density was calculated.

[0059] The belowground biomass is estimated based on the aboveground biomass, thereby calculating the belowground vegetation carbon density.

[0060] The soil carbon density was calculated by estimating the soil layer carbon density based on the obtained soil data.

[0061] Based on the design tables of aboveground vegetation carbon density, dead organic matter carbon density, underground vegetation carbon density and soil carbon density, a carbon density attribute table is generated.

[0062] In this example, the carbon density of each carbon pool is calculated and integrated into a table to provide a comprehensive understanding of the carbon storage situation within a region. This can also serve as a basis for dynamic monitoring. By generating carbon density attribute tables at different time periods, changes in carbon storage at different points in time can be compared, thereby assessing the changing trends of the ecosystem.

[0063] As a preferred embodiment, in step S14, the construction of the ecological carbon density database specifically includes: Based on the carbon density attribute table and the land use / land cover status map, a database structured carbon density attribute table and land use / land cover status map association table are constructed.

[0064] The vector or raster data in the database structure land use / land vegetation cover status map association table is digitized and geocoded, and a unique identifier is assigned to each land unit / vegetation unit.

[0065] The carbon density attributes corresponding to each land unit / vegetation unit identifier in the database structure carbon density attribute table are linked through smart contracts to construct an ecological carbon density database.

[0066] In this embodiment, the carbon density attribute table stores the carbon density data corresponding to different land use types and vegetation types, and is the core data source for estimating ecosystem carbon storage. By constructing a carbon density attribute table with a database structure, the carbon density information of various land units can be systematically managed and queried. The land use / land vegetation cover status map is spatial data, which reflects the spatial distribution of land use types and vegetation coverage. By constructing a land use / land vegetation cover status map association table with a database structure and digitizing and geocoding vector or raster data, the spatial data is effectively associated with the carbon density data to achieve data integration. Assigning a unique identifier to each land unit / vegetation unit can ensure the uniqueness and traceability of each unit in the database, and the carbon density information of each land unit can be accurately located during the query and analysis process.

[0067] The immutable nature of blockchain ensures the authenticity and reliability of carbon density data and land use / land cover maps. By uploading data to the blockchain, any modifications are detected by other nodes in the network, enhancing the data's credibility. Smart contracts enable automatic data updates and verification. When land use / land cover status changes, smart contracts automatically trigger updates to the carbon density attribute table, ensuring real-time and accurate data and reducing errors and delays caused by manual intervention.

[0068] 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 estimating carbon storage in an ecosystem, characterized in that: The method comprises: S11. Obtain remote sensing satellite images of the field investigation study area, identify tree species, herbs, and soil in the remote sensing satellite images, and generate a land use / land cover status map; S12. Determine the accuracy of the land use / land cover status map according to preset classification rules and perform grid division to obtain a number of estimated areas, and acquire data on above-ground samples, underground samples, and soil samples in the estimated areas; S13. Using individual trees in the estimation area as the basic unit, estimate and analyze the aboveground vegetation carbon density, belowground vegetation carbon density, soil carbon density, and dead organic matter carbon density to generate a carbon density attribute table; S14, after connecting the carbon density attribute table with the land use / land cover status map based on spatial location attributes, an ecological carbon density database is constructed; S15. Link the ecological carbon density database with the InVEST model to estimate carbon storage, and obtain the current estimated regional total carbon storage, aboveground vegetation carbon storage, belowground vegetation carbon storage, soil carbon storage, and dead organic matter carbon storage.

2. The method for estimating carbon storage in an ecosystem according to claim 1, wherein: In step S11, tree species, herbs and soil in the remote sensing satellite image are identified, specifically including: The remote sensing satellite images of the field investigation study area are sorted according to the acquisition time to obtain an image sequence; Extracting image data at several different acquisition moments from the image sequence according to preset rules and marking them as target image data; Extract image features from target image data and divide the image features into tree species features, herb features and soil features; Boundary identification is performed on tree species, herbaceous features, and soil features, and the identification results are marked with color values ​​to generate a current land use / land cover map.

3. The method for estimating carbon storage in an ecosystem according to claim 1, wherein: In step S12, the accuracy of the field investigation study area is determined using the national land cover level III classification system.

4. The method for estimating carbon storage in an ecosystem according to claim 1, wherein: In step S12, the land use / land cover status map is gridded, specifically including: Collect historical data sets that grid land use / land cover maps based on accuracy, and use the historical data sets to pre-train a land gridding model to determine the relationship between accuracy and land use / land cover map division; The currently determined accuracy and the land use / land cover status map are gridded using a land grid division model to obtain a number of estimated areas.

5. The method for estimating carbon storage in an ecosystem according to claim 4, characterized in that: The land grid division model is pre-trained using historical datasets to determine the relationship between the accuracy and the division of land use / land cover status maps, including: Collect gridded land use / land cover maps for different years, different types of regions, and different vegetation cover conditions, as well as the corresponding field survey accuracy data; For each grid cell in the land use / land cover map, extract the type information and the corresponding accuracy data, perform matching coding, and construct a sample data set; The sample data set is input into the spatial structure extraction branch and the accuracy extraction branch of the land grid division model to obtain spatial data and grid data respectively; The spatial data and grid data are integrated to determine the relationship between the accuracy and the division of the land use / land cover status map.

6. The method for estimating carbon storage in an ecosystem according to claim 1, wherein: In step S13, data of above-ground samples, underground samples, and soil samples in the estimation area are acquired, specifically including: Obtain the diameter at breast height (DBH) and height of the tree species in the above-ground samples within the estimated area. The DBH is measured starting at 5.0 cm, 1.3 m from the root collar of the upslope root of the tree trunk. If a tree has an irregular trunk shape, measure the diameter in two vertical directions and take the average value. The tree height is measured using a stadiometer. The tree density is estimated using the plot method combined with the spacing between trees to obtain the above-ground sample data. Randomly enclose a 2m×2m to 5m×5m box within the estimated area corresponding to the underground sample. After recording the herb and litter species and their coverage, all above-ground herbs and litter within the box are harvested. 300g of herbs and litter are collected from each box. The collected herbs and litter are dried at a preset temperature and weighed to obtain underground sample data. The soil drilling method and the profile method were used to sample, analyze and determine the estimated soil organic carbon in the area to obtain soil data.

7. The method for estimating carbon storage in an ecosystem according to claim 6, characterized in that: In step S13, the generation of the carbon density attribute table specifically includes: Based on the aboveground sample data, the biomass of individual trees is calculated. Based on the biomass of individual trees, herbaceous data, and the land use / land cover status map, the total aboveground vegetation biomass is estimated using the biomass equation, thereby calculating the total aboveground vegetation carbon density. The litter biomass was estimated based on the litter data, and thus the dead organic matter carbon density was calculated; The underground biomass is estimated based on the aboveground biomass, thereby calculating the underground vegetation carbon density; The soil carbon density is calculated by estimating the soil layer carbon density based on the acquired soil data; Based on the design tables of aboveground vegetation carbon density, dead organic matter carbon density, underground vegetation carbon density and soil carbon density, a carbon density attribute table is generated.

8. The method for estimating carbon storage in an ecosystem according to claim 1, wherein: In step S14, the construction of the ecological carbon density database specifically includes: Based on the carbon density attribute table and the land use / land cover status map, a database structured carbon density attribute table and land use / land cover status map association table are constructed; Digitize and geocode the vector or raster data in the database structure land use / land vegetation map association table, and assign a unique identifier to each land unit / vegetation unit; The carbon density attributes corresponding to each land unit / vegetation unit identifier in the database structure carbon density attribute table are linked through smart contracts to construct an ecological carbon density database.

Images