Ecological system carbon sink assessment method

Abstract

The invention relates to an ecological system carbon sink evaluation method, which comprises the steps of generating a continuous climate distribution grid based on a global ecological system type and climate element data, extracting multi-dimensional characteristic parameters of temperature, rainfall and a carbon sink value through an improved weighted K-means clustering algorithm, quantifying climate transition zone boundary fuzziness in combination with a covariance matrix, and evaluating the ecological system carbon sink evaluation result. Dynamically adjusting the weight of the spatial variation coefficient; a simulated annealing algorithm is utilized to optimize correlation between the division scale and the carbon sink response, and a multi-scale regression model is constructed to generate a high-precision carbon sink distribution diagram; further fusing the remote sensing vegetation index and the ground actual measurement data, and eliminating the evaluation error caused by the conflict between the boundary fuzziness and heterogeneity of the transition zone; according to the method, the problems of insufficient climate-ecology interaction dynamic response modeling and low boundary division precision in the prior art can be solved, the reliability of carbon sink evaluation in complex areas such as forest-grassland interlaced areas and coastal wetlands is remarkably improved, and scientific data support is provided for ecological restoration project site selection and carbon trading markets.

Description

Technical Field

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

[0002] Assessing the carbon sequestration capacity of ecosystems has become a key scientific issue in balancing carbon emissions and ecological protection. Carbon sequestration assessments quantify the ability of ecosystem components, such as vegetation and soil, to absorb and fix carbon dioxide, providing core data support for the formulation of ecological restoration strategies and carbon trading policies. However, the spatial distribution of terrestrial ecosystem carbon sequestration capacity is highly complex, especially in transitional regions where climatic conditions and vegetation types intersect. The carbon sequestration response patterns are susceptible to microclimate fluctuations and ecological interactions, making it difficult for traditional assessment models to accurately capture their dynamic characteristics.

[0003] In existing technologies, carbon sink assessments often rely on single ecological type assumptions or static climate parameter interpolation, which makes it difficult to reflect the synergistic effects of climate gradients and ecosystem heterogeneity at the regional scale. For example, in interlaced areas such as forest-grassland transition zones and coastal wetlands, slight changes in climate factors (such as precipitation variability and temperature thresholds) may trigger nonlinear responses in carbon sink capacity. However, existing methods often mask the dynamic characteristics of such sensitive areas due to the use of fixed boundary divisions or global unified parameters. In addition, carbon sink distribution maps generated based on low-resolution data or simplified models are prone to jumps or distortions in assessment results in areas with fuzzy boundaries, making it difficult to support the needs of refined ecological management. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide an ecosystem carbon sink assessment method that can integrate multi-source data, adaptively divide climate-ecological response units, and dynamically optimize boundary accuracy.

[0005] The purpose of the present invention is achieved by the following scheme:

[0006] In a first aspect, the present invention provides a method for assessing ecosystem carbon sequestration, comprising the following steps:

[0007] S1: Based on the global ecosystem type distribution database and climate element database, temperature and precipitation data of the target area are obtained. The temperature and precipitation data are processed based on the kriging spatial interpolation method to generate a continuous climate distribution grid. The continuous climate distribution grid is used to indicate the spatial continuous distribution characteristics of temperature and precipitation.

[0008] S2: Processing carbon sink data and extracting features from the continuous climate distribution grid based on the K-means clustering algorithm to generate a first characteristic parameter set containing temperature change rate, precipitation change rate, and carbon sink per unit area;

[0009] S3: Calculating and processing the first characteristic parameter set based on a covariance matrix algorithm to generate a first fuzzy degree distribution, where the first fuzzy degree distribution is used to indicate the uncertainty intensity of the boundary demarcation within the climate transition zone;

[0010] S4: Identify whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold; if so, perform boundary adjustment and coefficient optimization on the first fuzzy degree distribution to generate a first distribution map of regionalized carbon sink distribution per unit area;

[0011] S5: Identify whether the square value of the regression equation of the first distribution map is lower than the preset square threshold. If so, perform data fusion processing on the carbon sink data of the first distribution map and the continuous climate distribution grid to generate a second distribution map. The second distribution map is used to eliminate the carbon sink assessment error caused by the conflict between the ambiguity of the climate transition zone boundary and the spatial heterogeneity.

[0012] In one embodiment, S2 of an ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0013] S21: Perform Z-score normalization on the temperature, precipitation and carbon sink values ​​in the continuous climate distribution grid to generate a three-dimensional normalized feature vector;

[0014] S22: Clustering the three-dimensional standardized feature vectors based on the weighted K-means clustering algorithm to generate an initial geographic unit division result. The calculation formula for the initial geographic unit division result is:

[0015]

[0016] w j =α*CV T +β*CV P +γ*CV C

[0017] Among them, Cluster i is the initial geographic unit division result, that is, the set of all grids belonging to the i-th geographic unit, V j is the three-dimensional normalized eigenvector of the j-th grid, μ i is the i-th cluster center, CV T 、CV P 、CV C are the spatial variation coefficients of temperature, precipitation, and carbon sink, respectively; α, β, and γ are weight coefficients and satisfy α+β+γ=1;

[0018] S23: Calculating the boundary fuzzy index of the initial geographic unit division result based on the fuzzy membership function, determining the climate transition zone area, and generating a first geographic unit set containing the transition zone mark;

[0019] S24: Based on the initial boundaries of the first geographic unit set, extract the temperature change rate, precipitation change rate, and carbon sequestration per unit area to generate a first characteristic parameter set.

[0020] In one embodiment, S21 of an ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0021] S211: normalizing the temperature data, precipitation data, and carbon sink values ​​in the continuous climate distribution grid to generate three-dimensional standardized data, where the three-dimensional standardized data includes standardized temperature data, standardized precipitation data, and standardized carbon sink values;

[0022] S212: Perform dimensionality reduction processing on the standardized temperature data, standardized precipitation data, and standardized carbon sink value of the three-dimensional standardized data using a principal component analysis algorithm to generate a three-dimensional principal component eigenvector. The expression of the three-dimensional principal component eigenvector is:

[0023] V = [PC1, PC2, PC3]

[0024]

[0025] Among them, V is the three-dimensional principal component eigenvector, Z T is the normalized temperature data, Z P is the standardized precipitation data, Z C is the standardized carbon sink value, a i 、b i 、c i (i=1, 2, 3) are principal component analysis coefficients used to determine Z T 、Z P 、Z C For PC i the extent of contribution;

[0026] S213: Based on the three-dimensional principal component eigenvector, construct a standardized input data set for K-means clustering and generate a three-dimensional standardized eigenvector.

[0027] In one embodiment, S3 of an ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0028] S31: performing spatial derivative calculations on the temperature change rate and precipitation change rate in the first characteristic parameter set to generate the temperature gradient modulus and the partial derivative of the main wind direction of precipitation;

[0029] S32: Calculate the joint gradient difference variance of the temperature gradient mode length and the precipitation main wind direction partial derivative through the covariance matrix to generate the transition zone fuzzy quantification index;

[0030] S33: Based on the joint gradient difference variance, a probability distribution map reflecting the uncertainty of the climate transition zone boundary is constructed to generate the first fuzzy degree distribution.

[0031] In one embodiment, S4 of the ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0032] S41: identifying whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold; if so, processing the first fuzzy degree distribution based on a weighted Voronoi diagram algorithm, recalculating the cell boundary distance function using the spatial variation coefficient as a weight, and generating a revised geographic cell boundary;

[0033] S42: performing Gaussian convolution filtering on the corrected geographic unit boundary to generate a smoothed geographic unit boundary;

[0034] S43: Based on the smoothed geographic unit boundary, updating the weighted average of the carbon sequestration per unit area within the geographic unit boundary to generate a second geographic unit set;

[0035] S44: performing coefficient optimization processing on the second geographic unit set to generate a first distribution map of regionalized carbon sink distribution per unit area.

[0036] In one embodiment, S44 of an ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0037] S441: performing dynamic correlation analysis on the division scale and spatial variation coefficient of the second geographic unit set based on a correlation coefficient algorithm to generate a dynamic correlation coefficient;

[0038] S442: performing iterative optimization processing on the dynamic correlation coefficient based on a simulated annealing algorithm to generate an optimal partitioning scale parameter;

[0039] S443: Model the carbon sink per unit area and climate factors under the optimal division scale parameters based on linear regression technology to generate a first distribution map of the regionalized carbon sink per unit area.

[0040] In one embodiment, S5 of an ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0041] S51: identifying whether the square value of the regression equation of the first distribution map is lower than a preset square threshold; if so, performing vegetation index extraction processing on the remote sensing image of the coverage area of ​​the first distribution map based on remote sensing vegetation index inversion technology to generate NDVI raster data;

[0042] S52: Perform spatial weighted fusion processing on the NDVI raster data and the carbon sink data of the continuous climate distribution grid to generate an updated carbon sink distribution dataset;

[0043] S53: Based on the multi-scale clustering optimization technology, the updated carbon sink distribution dataset is subjected to transition zone boundary correction and regression modeling processing to generate a second distribution map.

[0044] In a second aspect, the present invention provides an ecosystem carbon sink assessment system, comprising

[0045] The data acquisition and processing module is used to obtain temperature and precipitation data of the target area based on the global ecosystem type distribution database and the climate element database, and process the temperature and precipitation data based on the Kriging spatial interpolation method to generate a continuous climate distribution grid. The continuous climate distribution grid is used to indicate the spatial continuous distribution characteristics of temperature and precipitation;

[0046] A feature clustering extraction module is used to process carbon sink data and extract features from the continuous climate distribution grid based on the K-means clustering algorithm, and generate a first feature parameter set containing the temperature change rate, precipitation change rate, and carbon sink per unit area;

[0047] a fuzzy distribution calculation module, configured to calculate and process the first characteristic parameter set based on a covariance matrix algorithm to generate a first fuzzy degree distribution, wherein the first fuzzy degree distribution is used to indicate the uncertainty intensity of the boundary demarcation within the climate transition zone;

[0048] a distribution map generation module, configured to identify whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold, and if so, perform boundary adjustment processing and coefficient optimization processing on the first fuzzy degree distribution to generate a first distribution map of the regionalized unit area carbon sink distribution;

[0049] The distribution map fusion optimization module is used to identify whether the square value of the regression equation of the first distribution map is lower than the preset square threshold. If so, the carbon sink data of the first distribution map and the continuous climate distribution grid are fused to generate a second distribution map. The second distribution map is used to eliminate the carbon sink assessment error caused by the conflict between the ambiguity of the climate transition zone boundary and spatial heterogeneity.

[0050] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements any of the above-mentioned ecosystem carbon sink assessment methods.

[0051] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements any of the above-mentioned ecosystem carbon sink assessment methods when executed by a processor.

[0052] In summary, the ecosystem carbon sink assessment method provided by the present invention can effectively assess the carbon sink capacity of an ecosystem by combining multiple statistical and data processing techniques. From data acquisition to interpolation processing, and then to cluster analysis and covariance calculation, each step provides important information and basis for the final carbon sink assessment. Through continuous optimization and adjustment, the accuracy and reliability of the assessment can be improved, providing strong support for ecological management and policy formulation. This comprehensive assessment method can not only capture the complex ecological dynamics within the climate transition zone, but also play a key role in ecological protection and climate change response.

[0053] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A flow chart of an ecosystem carbon sequestration assessment method provided in an embodiment of the present application;

[0055] Figure 2 A schematic diagram of a process for generating a first characteristic parameter set provided in an embodiment of the present application;

[0056] Figure 3 A schematic diagram of a process for generating a first distribution map according to an embodiment of the present application;

[0057] Figure 4 This is a structural diagram of an ecosystem carbon sink assessment system provided in another embodiment of the present application. DETAILED DESCRIPTION

[0058] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate preferred embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the disclosure.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0060] In one embodiment, Figure 1As shown, a method for assessing carbon sequestration in an ecosystem is provided. This embodiment uses the method applied to a terminal as an example. It is understood that the method can also be applied to a server, or to a system including a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0061] S1: Based on the global ecosystem type distribution database and climate element database, temperature and precipitation data of the target area are obtained, and the temperature and precipitation data are processed based on the Kriging spatial interpolation method to generate a continuous climate distribution grid. The continuous climate distribution grid is used to indicate the spatial continuous distribution characteristics of temperature and precipitation.

[0062] Specifically, the system uses the API interfaces of the global ecosystem type distribution database (such as WWF TerrestrialEcoregions and UNEP-WCMC database) and the climate factor database (such as CRU TS and WorldClim) to obtain the original observation data of temperature (Ta) and precipitation (P) in the target area through the OGC WFS protocol. Preferably, the data resolution of the acquired data is not less than 0.5°×0.5°. The global ecosystem type distribution database provides spatial information such as vegetation cover type and land use classification in the target area, which is used to determine the spatial heterogeneity of the ecosystem. The climate factor database contains observation data of climate parameters such as historical temperature and precipitation, and is combined with remote sensing data to supplement areas with insufficient spatial coverage.

[0063] Specifically, the system uses an improved Kriging spatial interpolation method (OK+EBK combined model), introduces terrain elevation (Digital Elevation Model, DEM) and vegetation cover (Normalized Difference Vegetation Index, NDVI) as covariates, constructs a three-dimensional spatial interpolation model, and generates a continuous climate distribution grid with a resolution of 30m×30m. The grid calibrates the boundary fuzzy area through the Thiessen polygon method to ensure that the spatial continuity index (SCI) is ≥0.92. Kriging spatial interpolation is a statistically based method that uses regionalized variables and variation functions to estimate the values ​​of unknown sample points. This method can provide an unbiased optimal estimate and can give the variance of the estimated value, thereby reflecting the reliability of the estimate. Through Kriging interpolation, the system can generate a continuous climate distribution grid, which can more accurately indicate the spatial continuous distribution characteristics of temperature and precipitation.

[0064] S2: Based on the K-means clustering algorithm, carbon sink data processing and feature extraction are performed on the continuous climate distribution grid to generate a first characteristic parameter set containing temperature change rate, precipitation change rate and carbon sink per unit area.

[0065] Specifically, the K-means clustering algorithm divides data points into K clusters, maximizing the similarity of data points within a cluster and minimizing the similarity of data points between clusters. During the clustering process, appropriate initial cluster centers and iteration times are selected to ensure the stability and accuracy of the clustering results. In this embodiment, the system uses temperature and precipitation data from a continuous climate distribution grid as input features, combined with carbon sink data (such as vegetation carbon storage, soil carbon storage, etc.) in the target area, to form a multidimensional data set. An appropriate K value (number of clusters) is selected and the K-means clustering algorithm is used to perform cluster analysis on the data set. Preferably, the K value can be determined by methods such as the elbow method or the silhouette coefficient. After completing the cluster analysis, the system extracts key feature parameters from the clustering results and generates a first feature parameter set, which includes the temperature change rate, precipitation change rate, and carbon sink per unit area. The temperature change rate and precipitation change rate reflect the spatial variation trend of climate elements, while the carbon sink per unit area directly indicates the distribution characteristics of carbon sink capacity.

[0066] S3: Calculate and process the first characteristic parameter set based on a covariance matrix algorithm to generate a first fuzzy degree distribution, where the first fuzzy degree distribution is used to indicate the uncertainty intensity of the boundary division within the climate transition zone.

[0067] Specifically, the covariance matrix is ​​a matrix used to describe the correlation between variables in a multidimensional data set. In the present invention, the covariance matrix algorithm is used to calculate the correlation between the parameters in the first characteristic parameter set and generate a first fuzzy degree distribution. The system standardizes the temperature change rate, precipitation change rate and carbon sequestration per unit area in the first characteristic parameter set to eliminate the influence of dimension and magnitude differences, and calculates the covariance matrix between the standardized characteristic parameters. The diagonal elements of the covariance matrix represent the variance of each parameter, and the non-diagonal elements represent the covariance between the parameters; and by analyzing the eigenvalues ​​and eigenvectors of the covariance matrix, the first fuzzy degree distribution is generated. The fuzzy degree distribution reflects the uncertainty intensity of the boundary division within the climate transition zone, that is, the degree of fuzziness of the boundary area. The first fuzzy degree distribution can effectively indicate the uncertainty of the boundary within the climate transition zone, providing a basis for subsequent boundary adjustment and coefficient optimization.

[0068] S4: Identify whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold; if so, perform boundary adjustment processing and coefficient optimization processing on the first fuzzy degree distribution to generate a first distribution map of regionalized unit area carbon sink distribution.

[0069] Specifically, the gradient difference variance reflects the severity of spatial variations in the fuzzy degree distribution. If the gradient difference variance exceeds the preset variance threshold, it indicates that there is a large uncertainty in the boundary demarcation of the climate transition zone, and boundary adjustment and coefficient optimization are required. Specifically, the system performs boundary adjustment processing and coefficient optimization processing on the first fuzzy degree distribution. Boundary adjustment adjusts the boundary position of the fuzzy degree distribution to make the boundary more consistent with the actual distribution characteristics of the climate transition zone; coefficient optimization improves the accuracy and reliability of the fuzzy degree distribution by adjusting the parameters of the clustering algorithm and the interpolation algorithm. After processing, a first distribution map of the regionalized unit area carbon sink distribution is generated. This distribution map eliminates the uncertainty of the boundary demarcation within the climate transition zone to a certain extent, and improves the accuracy of carbon sink assessment.

[0070] S5: Identify whether the square value of the regression equation of the first distribution map is lower than the preset square threshold. If so, perform data fusion processing on the carbon sink data of the first distribution map and the continuous climate distribution grid to generate a second distribution map. The second distribution map is used to eliminate the carbon sink assessment error caused by the conflict between the ambiguity of the climate transition zone boundary and the spatial heterogeneity.

[0071] Specifically, the system compares the square value of the regression equation with the preset square threshold. If the square value is lower than the preset square threshold, it means that the current model does not fit the data well and data fusion processing is required. At this time, the system fuses the first distribution map with the carbon sink data in the continuous climate distribution grid. Data fusion can be achieved through methods such as weighted averaging and Kalman filtering to improve the consistency and reliability of the data. Through data fusion processing, the system generates a second distribution map. On the basis of eliminating the conflict between the ambiguity of the boundaries of the climate transition zone and the conflict between spatial heterogeneity, the second distribution map further improves the accuracy and reliability of carbon sink assessment, can accurately reflect the carbon sink distribution characteristics of the target area, and can effectively eliminate the assessment errors caused by the conflict between boundary ambiguity and spatial heterogeneity, providing more scientific and reliable data support for the formulation of ecological restoration strategies and carbon trading policies.

[0072] In summary, the ecosystem carbon sink assessment method provided by the present invention can effectively assess the carbon sink capacity of an ecosystem by combining a variety of statistical and data processing techniques. From data acquisition to interpolation processing, to cluster analysis and covariance calculation, each step provides important information and basis for the final carbon sink assessment. Through continuous optimization and adjustment, the accuracy and reliability of the assessment can be improved, providing strong support for ecological management and policy formulation. This comprehensive assessment method can not only capture the complex ecological dynamics within the climate transition zone, but also provide a scientific basis for achieving the goal of carbon neutrality, thereby playing a key role in ecological protection and climate change response.

[0073] In one embodiment, Figure 2As shown, S2 of the ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0074] S21: Perform Z-score normalization on the temperature, precipitation, and carbon sink values ​​in the continuous climate distribution grid to generate a three-dimensional normalized feature vector.

[0075] Preferably, the three-dimensional normalized feature vector is generated by the following steps:

[0076] S211: Normalizing the temperature data, precipitation data, and carbon sink values ​​in the continuous climate distribution grid to generate three-dimensional standardized data, where the three-dimensional standardized data includes standardized temperature data, standardized precipitation data, and standardized carbon sink values.

[0077] Specifically, Z-score standardization is a common data normalization method that eliminates the impact of different dimensions and dimension ranges on data processing by converting data into a standard normal distribution with a mean of 0 and a standard deviation of 1. For the temperature, precipitation and carbon sink values ​​in the continuous climate distribution grid, their mean and standard deviation are calculated respectively. Specifically, the system performs Z-score standardization on the temperature data, precipitation data and carbon sink values ​​respectively to generate three-dimensional standardized data, including standardized temperature data Z T , Standardized precipitation data Z P and the standardized carbon sink value Z C These standardized data can reflect the relative variation characteristics of each variable in space and provide a basis for subsequent feature extraction and cluster analysis.

[0078] S212: Performing dimensionality reduction processing on the standardized temperature data, standardized precipitation data, and standardized carbon sink value of the three-dimensional standardized data using a principal component analysis algorithm to generate a three-dimensional principal component eigenvector.

[0079] Specifically, Principal Component Analysis (PCA) is a statistical dimensionality reduction technique that transforms the original data into a new coordinate system through orthogonal transformation, maximizing the variance of the data on the new coordinate axes. In three-dimensional standardized data, PCA can extract the main characteristic components, reducing the data dimension while retaining most of the information. The expression of the three-dimensional principal component eigenvector is:

[0080] V = [PC1, PC2, PC3]

[0081]

[0082] Among them, V is the three-dimensional principal component eigenvector, Z T is the normalized temperature data, Z P is the standardized precipitation data, ZC is the standardized carbon sink value, a i 、b i 、c i (i=1, 2, 3) are principal component analysis coefficients used to determine Z T 、Z P 、Z C For PC i By selecting the principal components with higher cumulative contribution rates (usually the first two or three), a three-dimensional principal component eigenvector is generated. This vector can effectively retain the main characteristic information of the original data, while reducing the data dimension and improving the efficiency and accuracy of subsequent clustering analysis.

[0083] S213: Based on the three-dimensional principal component eigenvector, construct a standardized input data set for K-means clustering and generate a three-dimensional standardized eigenvector.

[0084] Specifically, the system constructs a standardized input data set for K-means clustering based on the generated three-dimensional principal component eigenvectors. The three-dimensional principal component eigenvector of each grid point is used as a sample of the data set to form a data matrix containing all grid point samples. Each row of the data matrix represents the eigenvector of a grid point, and each column corresponds to a principal component feature. It should be noted that the constructed standardized input data set has a uniform data scale and dimension, and can effectively reflect the main changing trends of the climate and carbon sink characteristics of the target area. In the K-means clustering algorithm, this data set is used as input to make the clustering process more stable and accurate, avoid clustering bias caused by differences in data dimensions and dimensions, and provide a reliable data basis for generating high-quality initial geographic unit division results.

[0085] S22: Clustering the three-dimensional standardized feature vectors based on the weighted K-means clustering algorithm to generate an initial geographic unit division result.

[0086] Specifically, the weighted K-means clustering algorithm is an improvement on the traditional K-means algorithm. By introducing weight coefficients, the importance of different features in the clustering process is differentiated. In the ecosystem carbon sink assessment, the degree of influence of temperature, precipitation and carbon sink values ​​on the division of geographical units may be different. By setting weight coefficients, the clustering results can be made more in line with the actual situation. Specifically, the system first initializes the cluster center and randomly selects K samples as the initial cluster center; then calculates the distance from each sample to each cluster center according to the weighted Euclidean distance formula, and assigns the sample to the cluster to which the nearest cluster center belongs; updates the cluster center and calculates the new cluster center of each cluster by weighted average; repeats the above steps until the cluster center no longer changes or the maximum number of iterations is reached, and the stabilized result is defined as the initial geographical unit division result. Among them, the calculation formula for the initial geographical unit division result is:

[0087]

[0088] w j =α*CV T +β*CV P +γ*CV C

[0089] Among them, Cluster i is the initial geographic unit division result, that is, the set of all grids belonging to the i-th geographic unit, V j is the three-dimensional normalized eigenvector of the j-th grid, μ i is the i-th cluster center, CV T 、CV P 、CV C where α, β, and γ are the spatial coefficients of variation of temperature, precipitation, and carbon sink, respectively. These coefficients satisfy the weighting requirements of α + β + γ = 1. By iteratively optimizing cluster centers and sample allocation, a stable initial geographic unit division is generated. This result reflects the spatial distribution pattern of climate and carbon sink characteristics in the target region, providing a basis for subsequent boundary fuzzy index calculations and climate transition zone determination.

[0090] S23: Calculating the boundary fuzzy index of the initial geographic unit division result based on the fuzzy membership function, determining the climate transition zone area, and generating a first geographic unit set containing transition zone marks.

[0091] Specifically, in the division of geographic units, the grid points in the boundary area may belong to multiple geographic units at the same time, which has a certain degree of fuzziness. The fuzzy membership function is used to describe the degree to which the grid points belong to different geographic units. The value range is between 0 and 1. The closer the value is to 1, the higher the degree of membership. For the grid point V j , which goes to the i-th cluster center μ i The fuzzy membership u j，i It can be defined as:

[0092]

[0093] Among them, K is the number of clusters, m is the fuzzy coefficient (usually between 1.5 and 2), ||V j -μ i || is the grid point V j To the cluster center μ i The Euclidean distance of the grid point is quantified by considering the relative distance between the grid point and each cluster center. The boundary fuzziness index is used to measure the fuzziness of the boundary area of ​​the geographical unit, reflecting the uncertainty intensity of the boundary division within the climate transition zone. For each grid point V j , its boundary fuzzy index F jIt can be determined by calculating the difference in fuzzy membership to adjacent cluster centers:

[0094] F j =1-max(u j，1 ,u j，2 ,…,u j，k )

[0095] Among them, the boundary fuzzy index F j The value range is between 0 and 1. The closer the value is to 1, the higher the degree of boundary fuzziness is, and the closer the grid point is to the core area of ​​the climate transition zone. j Climate transition zones are identified based on the distribution of carbon sinks. A fuzzy index threshold is typically set, and grid points with boundary fuzziness indexes greater than the threshold are designated as climate transition zones. These areas represent complex variations in climate conditions and ecosystem types, making them difficult to accurately characterize using traditional clear-cut boundary delineations. Therefore, they require special attention and consideration in carbon sink assessments.

[0096] Based on the results of the boundary fuzzy index calculation, the system adds special markings to the grid points in the climate transition zone, building on the initial geographic unit division results to generate a first geographic unit set containing transition zone markings. This set not only includes the traditional clear geographic unit divisions but also incorporates information about the climate transition zone, enabling a more comprehensive and accurate reflection of the spatial distribution pattern of climate and ecological characteristics in the target area.

[0097] The first geographic unit set provides a more refined and accurate spatial unit division foundation for subsequent carbon sink characteristic parameter extraction and distribution map generation. In climate transition zones, due to their complex climatic and ecological characteristics, traditional geographic unit divisions can lead to inaccurate and discontinuous carbon sink assessment results. By identifying transition zones in the first geographic unit set, more refined assessment methods and models can be applied to these areas in subsequent analyses, improving the overall accuracy and reliability of carbon sink assessments and providing more scientific and rational data support for ecological management and policy formulation.

[0098] S24: Based on the initial boundaries of the first geographic unit set, extract the temperature change rate, precipitation change rate, and carbon sequestration per unit area to generate a first characteristic parameter set.

[0099] Specifically, after determining the first set of geographic units and their transition zone markers, key characteristic parameters need to be extracted from these geographic units, including the temperature change rate, precipitation change rate, and carbon sink per unit area, to generate the first characteristic parameter set. Specifically, within the initial boundaries of the first set of geographic units, the temperature change rate is calculated by calculating the ratio of the difference in temperature values ​​between adjacent grid points to the distance. Similarly, the precipitation change rate can be obtained by calculating the ratio of the difference in precipitation values ​​between adjacent grid points to the distance. The carbon sink per unit area is directly derived from the carbon sink value data in the continuous climate distribution grid. After standardization and clustering, it can more accurately reflect the carbon absorption and fixation capacity of each geographic unit per unit area. In the climate transition zone, due to the complexity and sensitivity of the ecosystem, the carbon sink per unit area may show large spatial differences and dynamic changes, requiring a comprehensive analysis in combination with other characteristic parameters.

[0100] After data extraction, the system integrates the extracted temperature and precipitation change rates, as well as the carbon sink per unit area, to generate a first set of characteristic parameters. Each element in this set corresponds to a geographic unit or grid point and contains the values ​​of the three key characteristic parameters, comprehensively describing the spatial variation of the target region's climate and carbon sink characteristics.

[0101] This ecosystem carbon sink assessment method, through statistical analysis, spatial modeling, and visualization of the characteristic parameters in this set, can provide a deeper understanding of the relationship between regional climate and carbon sinks, revealing the mechanisms by which climate transition zones influence carbon sink capacity, and providing a scientific basis for developing targeted ecological restoration strategies and carbon trading policies. Furthermore, in subsequent data fusion and error elimination processes, the first characteristic parameter set will serve as an important input for integration with other data sources, further improving the accuracy and reliability of carbon sink assessments and ensuring that the assessment results truly reflect the ecosystem carbon sink status and dynamic trends in the target region.

[0102] In one embodiment, S3 of an ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0103] S31: Perform spatial derivative calculations on the temperature change rate and precipitation change rate in the first characteristic parameter set to generate the temperature gradient modulus and the partial derivative of the main wind direction of precipitation.

[0104] Specifically, spatial derivative calculation is a method used to quantify the rate of change of spatial field variables, and is widely used in geographic information science and ecological modeling. By calculating the spatial derivatives of the temperature and precipitation rates, we can reveal the spatial trends and gradient characteristics of these variables, providing basic data support for subsequent fuzzy quantification and boundary uncertainty analysis.

[0105] First, the system extracts temperature rate of change data from the first characteristic parameter set. This data reflects the spatial rate of temperature change. The spatial resolution of the data should be consistent with the resolution of the entire evaluation system to ensure the accuracy and comparability of the calculation results. The spatial derivative of the temperature rate of change data is calculated using either the finite difference method or the Gaussian filter derivative algorithm. The finite difference method calculates the difference in temperature change rate between adjacent grid points to approximate the temperature gradient. The Gaussian filter derivative algorithm calculates the derivative while smoothing the data, effectively reducing the impact of noise on the gradient calculation.

[0106] After completing the derivative calculation of the temperature change rate data, the modulus of the temperature gradient vector is calculated for the derivative to obtain the temperature gradient modulus data. The temperature gradient modulus reflects the overall intensity of temperature changes in space. A higher modulus value indicates a more drastic temperature change, which may correspond to the core area of ​​the climate transition zone or the sensitive response area of ​​the ecosystem.

[0107] Secondly, the system integrates precipitation rate data with prevailing wind direction data provided by regional meteorological stations. Predominant wind direction data reflects the primary direction of precipitation transport and is crucial for understanding the spatial distribution and changing trends of precipitation. During this data integration process, the wind direction data must be converted into a spatial grid format that matches the precipitation rate data. Based on the prevailing wind direction, the directional derivative of the precipitation rate in that direction is calculated. This directional derivative reflects the rate of change of the precipitation rate in the direction of the prevailing wind, revealing the spatial variation characteristics of precipitation transport.

[0108] After calculating the directional derivative of the precipitation rate of change in the main direction, the system then calculates the directional derivative to generate the partial derivative of the main wind direction of precipitation. This data not only contains information on the spatial variation of the precipitation rate of change but also incorporates the influence of wind direction on precipitation distribution. It can more comprehensively reflect the dynamic spatial variation characteristics of precipitation and provide key input for the subsequent calculation of the joint gradient difference variance.

[0109] S32: The joint gradient difference variance of the temperature gradient modulus and the partial derivative of the main wind direction of precipitation is calculated through the covariance matrix to generate the transition zone fuzzy quantification index.

[0110] Specifically, the system pairs the temperature gradient modulus data with the precipitation main wind direction partial derivative data to form a two-dimensional data set. Each data pair contains the temperature gradient modulus value and the precipitation main wind direction partial derivative value at the corresponding grid point. Based on this two-dimensional data set, a covariance matrix is ​​constructed to quantify the joint variation characteristics and correlation between the two variables. The calculation formula of the covariance matrix is:

[0111]

[0112] Cov(T, T) and Cov(P, P) are the variances of the temperature gradient mode and the partial derivative of the main wind direction of precipitation, respectively, reflecting the degree of spatial variation of each variable; Cov(T, P) and Cov(P, T) are the covariances between the temperature gradient mode and the partial derivative of the main wind direction of precipitation, reflecting the strength and direction of the linear correlation between the two variables.

[0113] The system extracts the joint gradient difference variance through eigenvalue decomposition of the covariance matrix. Eigenvalue decomposition can decompose the covariance matrix into eigenvalues ​​and eigenvectors. The eigenvalues ​​represent the variance of the data in the corresponding characteristic direction. The characteristic direction corresponding to the larger eigenvalue represents the main trend of data change, while the smaller eigenvalue reflects the secondary change direction and difference variance of the data. The specific calculation steps are:

[0114] 1. Calculate the eigenvalues ​​λ1 and λ2 of the covariance matrix, and the corresponding eigenvectors v1 and v2.

[0115] 2. The joint gradient difference variance is defined as the smaller eigenvalue λ2, which reflects the degree of data difference in the secondary change direction, that is, the degree of inconsistency between the temperature gradient mode and the partial derivative of the main wind direction of precipitation.

[0116] The joint gradient difference variance λ² is used as a quantification metric for the transition zone fuzziness. Larger values ​​of the joint gradient difference variance indicate greater spatial inconsistency in the temperature and precipitation gradients, leading to a more ambiguous boundary demarcation within the climate transition zone. Smaller values ​​indicate more consistent trends between the two, resulting in a clearer boundary. This method can quantitatively characterize the degree of uncertainty in the boundary demarcation within the climate transition zone, providing a scientific basis for the subsequent construction of probability distribution maps.

[0117] S33: Based on the joint gradient difference variance, a probability distribution map reflecting the uncertainty of the climate transition zone boundary is constructed to generate the first fuzzy degree distribution.

[0118] Specifically, a probability distribution map can be constructed by selecting an appropriate probability distribution model (such as the normal distribution or t-distribution) and using the variance of the joint gradient difference as an input parameter. This yields a probability distribution map that reflects boundary uncertainty. This probability distribution map can intuitively demonstrate the degree of ambiguity in the climate transition zone boundary, providing a basis for subsequent boundary adjustment and optimization.

[0119] In one embodiment, Figure 3 As shown, S4 of the ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0120] S41: Identify whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold. If so, process the first fuzzy degree distribution based on a weighted Voronoi diagram algorithm, recalculate the cell boundary distance function using the spatial variation coefficient as a weight, and generate a corrected geographic cell boundary.

[0121] Specifically, the system calculates the gradient difference variance of the first fuzzy degree distribution and evaluates whether it exceeds the preset variance threshold. The gradient difference variance reflects the degree of spatial variation of the fuzzy degree. A larger variance value indicates that there is greater uncertainty in the boundary division of the climate transition zone and boundary adjustment is required. The weighted Voronoi diagram improves the traditional Voronoi diagram by introducing a weight factor, which can better reflect the attribute differences and spatial relationships of spatial units. In this step, the spatial variation coefficient is used as the weight to recalculate the unit boundary distance function and generate a revised geographic unit boundary. The calculation formula for the revised geographic unit boundary is:

[0122]

[0123] Among them, d i，j represents the boundary distance between the i-th and j-th geographic units, w i and w j are the spatial variation coefficients of the two units, p i and p j In this way, the boundaries of the revised geographic units are generated to make them more consistent with the actual distribution characteristics of the climate transition zone.

[0124] S42: Performing Gaussian convolution filtering on the corrected geographic unit boundary to generate a smoothed geographic unit boundary.

[0125] Specifically, Gaussian convolution filtering is a commonly used image processing and data smoothing method. By applying a weighted average to data using a Gaussian kernel function, it effectively reduces noise and irregularities while preserving the primary characteristic structure. In geographic information processing, Gaussian convolution filtering can be used to smooth geographic unit boundaries, improving the accuracy and aesthetics of boundary delineation.

[0126] Specifically, the system selects an appropriate Gaussian kernel function, whose standard deviation and size should be adjusted according to the resolution and complexity of the geographic unit boundary. A larger standard deviation is suitable for smoothing large-scale boundaries, while a smaller standard deviation is suitable for smoothing fine-scale boundaries. The Gaussian kernel function is applied to each point on the geographic unit boundary, and the new boundary point position is calculated through convolution operation. The convolution operation formula is:

[0127]

[0128] Where G(i) is the weight of the Gaussian kernel function at position i, Boundary(x) is the coordinate value of the original boundary at position x, and New Boundary(x) is the coordinate value of the smoothed boundary at position x. After Gaussian convolution filtering, a smoothed geographic unit boundary is generated. The smoothed boundary is more regular and continuous, reducing jagged edges and fluctuations caused by data noise and irregularities, improving the quality and readability of geographic unit delineation, and providing a more stable foundation for subsequent carbon sink updates and distribution map generation.

[0129] S43: Based on the smoothed geographic unit boundaries, update the weighted average of the carbon sequestration per unit area within the geographic unit boundaries to generate a second geographic unit set.

[0130] Specifically, the weighted average calculation formula for carbon sequestration per unit area is:

[0131]

[0132] Among them, C i is the carbon sink per unit area of ​​the i-th grid unit, A i is the area of ​​the i-th grid cell, and N is the total number of grid cells in the geographic unit. Through weighted average calculation, a second geographic unit set is generated, which contains the updated unit area carbon sink information and can more accurately reflect the carbon sink characteristics within the geographic unit.

[0133] S44: performing coefficient optimization processing on the second geographic unit set to generate a first distribution map of regionalized carbon sink distribution per unit area.

[0134] Preferably, the first distribution map is generated by the following steps:

[0135] S441: Performing dynamic correlation analysis on the division scale and spatial variation coefficient of the second geographic unit set based on a correlation coefficient algorithm to generate a dynamic correlation coefficient.

[0136] Specifically, the correlation coefficient algorithm evaluates the strength and direction of the relationship between the division scale and the spatial variation coefficient by calculating the linear correlation between the two variables. The dynamic correlation coefficient can reflect the changes in the correlation between the two at different spatial scales, providing a basis for subsequent coefficient optimization. Specifically, the system extracts the division scale parameter and spatial variation coefficient data from the second geographic unit set. The division scale parameter can be expressed as the size of the geographic unit or the resolution of the grid. The spatial variation coefficient reflects the degree of spatial variation of the carbon sink, and the Pearson correlation coefficient formula is used to calculate the correlation between the division scale and the spatial variation coefficient:

[0137]

[0138] Among them, x i and y i are the sample values ​​of the division scale and spatial variation coefficient, and where and are the sample means, respectively, and r is the correlation coefficient, ranging from -1 to 1. Positive values ​​indicate positive correlation, negative values ​​indicate negative correlation, and the closer the absolute value is to 1, the stronger the correlation. The system dynamically calculates the correlation coefficient using sliding window technology or moving average methods to generate a dynamic correlation coefficient sequence. The dynamic correlation coefficient reflects the changing correlation between the partition scale and the spatial variation coefficient at different spatial locations and scales, providing real-time feedback for subsequent simulated annealing optimization.

[0139] S442: Perform iterative optimization processing on the dynamic correlation coefficient based on a simulated annealing algorithm to generate an optimal partitioning scale parameter.

[0140] Specifically, the simulated annealing algorithm is a stochastic optimization algorithm based on a physical annealing process that effectively avoids local optimal solutions and finds the global optimal solution. During the optimization process, the dynamic correlation coefficient is maximized by adjusting the partition scale parameter, ensuring the best match between the partition scale and the spatial variation coefficient.

[0141] Specifically, the system sets control parameters for the simulated annealing algorithm, including the initial partition scale parameter, initial temperature, cooling rate, and number of iterations. The initial partition scale parameter can be determined based on experience or preliminary analysis results. The initial temperature is high and gradually decreases as the iterations proceed. The objective function is defined as the maximization problem of the dynamic correlation coefficient, that is, finding the partition scale parameter that maximizes the dynamic correlation coefficient. The objective function can be expressed as:

[0142] Maximize f(s)=r(s)

[0143] Where s is the partition scale parameter, and r(s) is the corresponding dynamic correlation coefficient. In each iteration, a new partition scale parameter is randomly generated, and the corresponding objective function value is calculated. If the objective function value of the new parameter is greater than the current parameter value, the new parameter is accepted; otherwise, the new parameter is accepted with a certain probability, which decreases with decreasing temperature. Through multiple iterations, the global optimal solution is gradually approached, and the optimal partition scale parameter is ultimately generated. This parameter maximizes the correlation between the partition scale and the spatial coefficient of variation, ensuring the spatial rationality of the geographic unit division and the accuracy of the carbon sink assessment.

[0144] S443: Model the carbon sink per unit area and climate factors under the optimal division scale parameters based on linear regression technology to generate a first distribution map of the regionalized carbon sink per unit area.

[0145] Specifically, the linear regression model quantifies the linear relationship between carbon sinks per unit area and climate factors (such as temperature and precipitation), providing an intuitive tool for predicting carbon sink distribution. The model development process includes data preparation, model fitting, model validation, and model application, ensuring the generated distribution map is highly accurate and reliable, accurately reflecting the spatial distribution characteristics of regional carbon sinks per unit area and their relationship with climate factors.

[0146] Specifically, the system collects data on carbon sequestration per unit area and related climate factors (such as temperature, precipitation, wind speed, etc.) under the optimal partitioning scale parameters. The data is cleaned, normalized, and missing values ​​are processed. A linear regression model is constructed with carbon sequestration per unit area as the dependent variable and climate factors as the independent variables. The general form of the model is:

[0147] C=β0+β1T+β2P+β3W+∈

[0148] Among them, C is the carbon sink per unit area, T, P, and W are climate factors such as temperature, precipitation, and wind speed, respectively, β0 is the intercept, β1, β2, and β3 are regression coefficients, and ∈ is the error term. In this embodiment, the system uses historical data to train the linear regression model and estimates the regression coefficients by the least squares method. Part of the data is used to verify the model, evaluate the goodness of fit and predictive ability of the model, apply the verified linear regression model to the climate factor data of the target area, predict the carbon sink per unit area of ​​each geographical unit, and generate a first distribution map of the regionalized carbon sink per unit area based on the prediction results. This distribution map can intuitively show the spatial distribution characteristics of the carbon sink and provide a scientific basis for ecological management and policy formulation. Common evaluation indicators include the coefficient of determination (R2), mean square error (MSE), and mean absolute error (MAE).

[0149] The system will visualize the generated regionalized per-unit-area carbon sink distribution map, presenting the spatial distribution of carbon sinks using color gradients, contour lines, or three-dimensional surface maps. High-carbon sink areas can be represented by dark colors, and low-carbon sink areas by light colors, intuitively reflecting the differences in regional carbon sink capacity. This distribution map can be used to guide ecological restoration projects, formulate carbon trading policies, and optimize land use planning. In ecological restoration, low-carbon sink areas are given priority for vegetation restoration and soil improvement; in carbon trading policies, carbon emission quotas are reasonably allocated based on the distribution of carbon sinks; in land use planning, high-carbon sink areas are protected to prevent them from being destroyed by development. Through these applications, the distribution map can provide strong support for achieving carbon neutrality goals and ecological protection.

[0150] The aforementioned ecosystem carbon sink assessment method identifies the gradient difference variance of a first fuzzy degree distribution and processes it using a weighted Voronoi diagram algorithm. Using the coefficient of spatial variation as a weight, the unit boundary distance function is recalculated. This method effectively optimizes the demarcation of geographic unit boundaries and addresses the assessment errors caused by boundary ambiguity in traditional methods. Furthermore, a Gaussian convolution filter is applied to the modified geographic unit boundaries to generate smoothed geographic unit boundaries, reducing data fluctuations caused by boundary effects and improving the stability of carbon sink assessment. Based on the smoothed geographic unit boundaries, the weighted average of the carbon sink per unit area within the geographic unit boundaries is updated to generate a second set of geographic units, ensuring the accuracy and consistency of carbon sink data. Through coefficient optimization, a first distribution map of the regionalized carbon sink per unit area is generated, providing a high-precision spatial reference for ecological management and policy formulation. Preferably, a correlation coefficient algorithm is used to dynamically correlate the demarcation scale of the second set of geographic units with the coefficient of spatial variation to generate a dynamic correlation coefficient, which can adapt to carbon sink assessment requirements at different scales. The dynamic correlation coefficient is iteratively optimized using a simulated annealing algorithm to generate the optimal demarcation scale parameter, further improving the model's adaptability and accuracy. Finally, linear regression techniques were used to model the carbon sink per unit area and climatic factors under the optimal scale parameters, generating a first regional distribution map of carbon sink per unit area, providing a scientific basis for the formulation of ecological restoration strategies and carbon trading policies. This series of steps not only addresses the assessment errors caused by the fuzzy boundaries of traditional methods in complex ecosystems, but also provides more refined and dynamic data support for ecological management and the achievement of carbon neutrality goals, significantly improving the scientific and practical nature of carbon sink assessments.

[0151] In one embodiment, S5 of an ecosystem carbon sink assessment method provided by the present invention specifically includes the following steps:

[0152] S51: Identify whether the square value of the regression equation of the first distribution map is lower than a preset square threshold. If so, perform vegetation index extraction processing on the remote sensing image of the coverage area of ​​the first distribution map based on remote sensing vegetation index inversion technology to generate NDVI raster data.

[0153] Specifically, the system calculates the square value of the regression equation of the first distribution map and evaluates whether it is lower than the preset square threshold. The square value of the regression equation reflects the goodness of fit of the model to the data. A lower square value indicates that the model's explanatory power is insufficient and needs further optimization. At this time, the system performs vegetation index extraction processing on the remote sensing image of the coverage area of ​​the first distribution map based on the remote sensing vegetation index inversion technology to generate NDVI raster data. The remote sensing vegetation index inversion technology extracts information such as vegetation coverage and growth status by analyzing the spectral information in the remote sensing image. NDVI (Normalized Difference Vegetation Index) is one of the most commonly used vegetation indices and can effectively reflect the health status and coverage of vegetation.

[0154] S52: Perform spatial weighted fusion processing on the NDVI raster data and the carbon sink data of the continuous climate distribution grid to generate an updated carbon sink distribution dataset.

[0155] Specifically, spatial weighted fusion is a data integration technology that assigns different weights to different data sources, comprehensively considers the advantages and disadvantages of each data source, and generates more accurate and reliable results. In this step, the NDVI raster data is spatially weighted fused with the carbon sink data of the continuous climate distribution grid to generate an updated carbon sink distribution dataset. Specifically, the system confirms the fusion weights, and when confirming, the accuracy, resolution and correlation of each data source are considered. Generally, vegetation index data has higher accuracy in reflecting the carbon sink capacity of vegetation, while climate data provides important background information at the macro scale. Through correlation analysis and cross-validation, a reasonable weight distribution scheme is determined.

[0156] Then, the system spatially registers the NDVI raster data with the carbon sink data of the continuous climate distribution grid to ensure the consistency of the spatial coordinate system and resolution between the two, and uses the weighted average method to fuse the data. The calculation formula is:

[0157] C new =w1*C climate +w2*C NDV1

[0158] Among them, C new is the updated carbon sink value, C climate is the carbon sink value of climate data, C NDV1 is the estimated carbon sink value based on NDVI, where w1 and w2 are the weights of the climate data and NDVI data, respectively, satisfying w1 + w2 = 1. The system integrates the fused carbon sink values ​​into a new dataset, forming an updated carbon sink distribution dataset. The accuracy and reliability of the fused dataset are evaluated by comparing it with measured data or high-precision reference data to ensure that it truly reflects the regional carbon sink distribution characteristics.

[0159] S53: Based on the multi-scale clustering optimization technology, the updated carbon sink distribution dataset is subjected to transition zone boundary correction and regression modeling processing to generate a second distribution map.

[0160] Specifically, multi-scale clustering optimization technology can capture the clustering structure and variation characteristics of data at different spatial scales, improving the accuracy of boundary demarcation and the explanatory power of the model. Cluster analysis is performed on carbon sink distribution data at different spatial scales to identify potential boundary areas of the climate transition zone. By adjusting the parameters of the clustering algorithm (such as the number of clusters, distance threshold, etc.), boundary characteristics at different scales are captured. Preferably, the system uses methods such as Gaussian convolution filtering to smooth and correct the identified boundary areas to remove noise and irregularities on the boundaries. At the same time, combined with geographic information and ecological knowledge, the boundaries are reasonably adjusted to make them more consistent with the actual distribution characteristics of the ecological transition zone.

[0161] Based on the revised transition zone boundary, the system reconstructed a regression model for carbon sink distribution and selected appropriate regression methods (such as linear regression, nonlinear regression, and geographically weighted regression). Using carbon sink quantity as the dependent variable and climate factors and vegetation indices as independent variables, a quantitative model was established. The regression model was optimized and validated through cross-validation and residual analysis to ensure high goodness of fit and predictive power. The optimized regression model better explains the spatial variation of carbon sink distribution and provides reliable model support for generating the second distribution map.

[0162] Finally, the system uses the optimized regression model to predict the carbon sink distribution of the entire study area and generate a second distribution map. This distribution map comprehensively considers the influence of climate factors, vegetation coverage and spatial heterogeneity, and can more accurately reflect the regional carbon sink distribution characteristics and its dynamic change trends. Preferably, the second distribution map can be visualized, using color gradients, contour lines and other methods to present the spatial distribution of carbon sinks. High carbon sink areas can be represented by dark colors, and low carbon sink areas by light colors, which intuitively reflect the differences in regional carbon sink capacity. The second distribution map can be used to guide ecological restoration projects, formulate carbon trading policies and optimize land use planning, providing a scientific basis for achieving carbon neutrality goals and ecological protection.

[0163] The above-mentioned ecosystem carbon sink assessment method identifies whether the squared value of the regression equation of a first distribution map is below a preset squared threshold. If so, the remote sensing vegetation index is extracted and processed from the remote sensing imagery of the area covered by the first distribution map based on remote sensing vegetation index inversion technology to generate NDVI raster data. This effectively obtains vegetation cover information and provides more accurate vegetation data support for carbon sink assessment. Furthermore, the NDVI raster data is spatially weighted and fused with carbon sink data from a continuous climate distribution grid to generate an updated carbon sink distribution dataset. This not only eliminates data inconsistencies but also improves the accuracy of carbon sink assessment. Finally, based on multi-scale clustering optimization technology, the updated carbon sink distribution dataset is subjected to transition zone boundary correction and regression modeling to generate a second distribution map. This effectively corrects the transition zone boundary, improves the accuracy and reliability of carbon sink distribution data, and provides more refined and dynamic data support for the formulation of ecological restoration strategies and carbon trading policies, significantly enhancing the scientific nature and practicality of carbon sink assessment.

[0164] Preferably, if Figure 4 As shown, the present invention provides an ecosystem carbon sequestration assessment system 600, which is configured with the following modules:

[0165] Data acquisition and processing module 610 is used to obtain temperature and precipitation data of the target area based on the global ecosystem type distribution database and the climate element database, and process the temperature and precipitation data based on the Kriging spatial interpolation method to generate a continuous climate distribution grid. The continuous climate distribution grid is used to indicate the spatial continuous distribution characteristics of temperature and precipitation;

[0166] A feature clustering extraction module 620 is configured to process carbon sink data and extract features from the continuous climate distribution grid based on a K-means clustering algorithm, thereby generating a first feature parameter set including temperature change rate, precipitation change rate, and carbon sink per unit area.

[0167] A fuzzy distribution calculation module 630 is configured to calculate and process the first characteristic parameter set based on a covariance matrix algorithm to generate a first fuzzy degree distribution, where the first fuzzy degree distribution is used to indicate the intensity of uncertainty in the boundary demarcation within the climate transition zone;

[0168] A distribution map generating module 640 is configured to identify whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold, and if so, to perform boundary adjustment and coefficient optimization on the first fuzzy degree distribution to generate a first distribution map of the regionalized carbon sink per unit area;

[0169] The distribution map fusion optimization module 650 is used to identify whether the square value of the regression equation of the first distribution map is lower than a preset square threshold. If so, the first distribution map and the carbon sink data of the continuous climate distribution grid are fused to generate a second distribution map. The second distribution map is used to eliminate the carbon sink assessment error caused by the conflict between the ambiguity of the climate transition zone boundary and spatial heterogeneity.

[0170] In summary, the ecosystem carbon sink assessment system provided by the present invention can effectively assess the carbon sink capacity of an ecosystem by combining a variety of statistical and data processing techniques. From data acquisition to interpolation processing, to cluster analysis and covariance calculation, each step provides important information and basis for the final carbon sink assessment. Through continuous optimization and adjustment, the accuracy and reliability of the assessment can be improved, providing strong support for ecological management and policy formulation. This comprehensive assessment method can not only capture the complex ecological dynamics within the climate transition zone, but also provide a scientific basis for achieving the goal of carbon neutrality, thereby playing a key role in ecological protection and climate change response.

[0171] Preferably, the feature clustering extraction module 620 provided by the present invention is configured with the following units:

[0172] The feature vector generating unit 621 is used to perform Z-score normalization processing on the temperature, precipitation and carbon sink values ​​in the continuous climate distribution grid to generate a three-dimensional normalized feature vector;

[0173] Preferably, the feature vector generating unit 621 is configured with the following subunits:

[0174] The data normalization subunit 6211 is used to normalize the temperature data, precipitation data and carbon sink values ​​in the continuous climate distribution grid to generate three-dimensional normalized data. The three-dimensional normalized data includes normalized temperature data, normalized precipitation data and normalized carbon sink values.

[0175] The data dimensionality reduction subunit 6212 is used to perform dimensionality reduction processing on the standardized temperature data, standardized precipitation data, and standardized carbon sink value of the three-dimensional standardized data using a principal component analysis algorithm to generate a three-dimensional principal component feature vector;

[0176] The input data set construction subunit 6213 is used to construct a standardized input data set for K-means clustering based on the three-dimensional principal component eigenvector and generate a three-dimensional standardized eigenvector.

[0177] The geographic unit division unit 622 is used to perform clustering processing on the three-dimensional normalized feature vector based on the weighted K-means clustering algorithm to generate an initial geographic unit division result;

[0178] The transition zone determination unit 623 is configured to calculate the boundary fuzzy index of the initial geographic unit division result based on the fuzzy membership function, determine the climate transition zone area, and generate a first geographic unit set containing transition zone labels;

[0179] The characteristic parameter generating unit 624 is configured to extract the temperature change rate, precipitation change rate, and carbon sequestration per unit area based on the initial boundaries of the first geographic unit set to generate a first characteristic parameter set.

[0180] Preferably, the fuzzy distribution calculation module 630 provided by the present invention is configured with the following units:

[0181] A spatial derivative calculation unit 631 is used to perform spatial derivative calculations on the temperature change rate and precipitation change rate in the first characteristic parameter set to generate a temperature gradient modulus and a partial derivative of the main wind direction of precipitation;

[0182] The quantitative index generating unit 632 is used to calculate the joint gradient difference variance of the temperature gradient modulus and the precipitation main wind direction partial derivative through the covariance matrix to generate a transition zone fuzzy quantitative index;

[0183] The fuzzy degree distribution generating unit 633 is configured to construct a probability distribution map reflecting the uncertainty of the climate transition zone boundary based on the joint gradient difference variance, and generate a first fuzzy degree distribution.

[0184] Preferably, the distribution map generation module 640 provided by the present invention is configured with the following units:

[0185] A boundary correction unit 641 is configured to identify whether the gradient difference variance of the first blur degree distribution exceeds a preset variance threshold. If so, the first blur degree distribution is processed based on a weighted Voronoi diagram algorithm, and the cell boundary distance function is recalculated using the spatial variation coefficient as a weight to generate a corrected geographic cell boundary.

[0186] A boundary smoothing unit 642 is used to perform Gaussian convolution filtering on the corrected geographic unit boundary to generate a smoothed geographic unit boundary;

[0187] The geographic unit updating unit 643 is configured to update the weighted average carbon sequestration per unit area within the geographic unit boundary based on the smoothed geographic unit boundary to generate a second geographic unit set;

[0188] The carbon sink distribution generating unit 644 is configured to perform coefficient optimization processing on the second set of geographic units to generate a first distribution map of regionalized carbon sink distribution per unit area.

[0189] Preferably, the carbon sink distribution generation unit 644 is configured with the following subunits:

[0190] The correlation analysis unit 6441 is used to perform dynamic correlation analysis on the division scale and spatial variation coefficient of the second geographic unit set based on the correlation coefficient algorithm to generate a dynamic correlation coefficient;

[0191] The parameter optimization unit 6442 is used to perform iterative optimization processing on the dynamic correlation coefficient based on the simulated annealing algorithm to generate the optimal partitioning scale parameter;

[0192] The modeling and drawing unit 6443 is used to model the carbon sink per unit area and the climate factors under the optimal division scale parameters based on the linear regression technology, and generate a first distribution map of the regionalized carbon sink per unit area.

[0193] Preferably, the distribution graph fusion optimization module 650 provided by the present invention is configured with the following units:

[0194] The vegetation index extraction unit 651 is used to identify whether the square value of the regression equation of the first distribution map is lower than a preset square threshold. If so, the vegetation index is extracted from the remote sensing image of the coverage area of ​​the first distribution map based on the remote sensing vegetation index inversion technology to generate NDVI raster data.

[0195] The data fusion unit 652 is used to perform spatial weighted fusion processing on the NDVI raster data and the carbon sink data of the continuous climate distribution grid to generate an updated carbon sink distribution dataset;

[0196] The second distribution map generating unit 653 is configured to perform transition zone boundary correction and regression modeling processing on the updated carbon sink distribution dataset based on a multi-scale clustering optimization technique to generate a second distribution map.

[0197] In one embodiment, the present application further provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned ecosystem carbon sink assessment method when executing the computer program.

[0198] In one embodiment, the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the above-mentioned ecosystem carbon sequestration assessment method when executed by a processor.

[0199] In the description of this specification, the reference terms "one embodiment," "some embodiments," "example," "specific example," or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described may be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and integrate different embodiments or examples described in this specification, as well as features of different embodiments or examples, unless they are mutually inconsistent.

[0200] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0201] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various modifications or substitutions within the technical scope disclosed in this application, and such modifications or substitutions should be included within the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for assessing ecosystem carbon sequestration, characterized in that: The following steps are involved: S1: Based on the global ecosystem type distribution database and climate element database, temperature and precipitation data of the target area are obtained, and the temperature and precipitation data are processed based on the kriging spatial interpolation method to generate a continuous climate distribution grid, which is used to indicate the spatial continuous distribution characteristics of temperature and precipitation; S2: performing carbon sink data processing and feature extraction on the continuous climate distribution grid based on a K-means clustering algorithm to generate a first feature parameter set including a temperature change rate, a precipitation change rate, and a carbon sink per unit area; S3: Calculating and processing the first characteristic parameter set based on a covariance matrix algorithm to generate a first fuzziness degree distribution, where the first fuzziness degree distribution is used to indicate the uncertainty intensity of the boundary demarcation within the climate transition zone; S4: Identifying whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold; if so, performing boundary adjustment and coefficient optimization on the first fuzzy degree distribution to generate a first distribution map of regionalized carbon sink distribution per unit area; S5: Identify whether the square value of the regression equation of the first distribution map is lower than a preset square threshold. If so, perform data fusion processing on the carbon sink data of the first distribution map and the continuous climate distribution grid to generate a second distribution map. The second distribution map is used to eliminate the carbon sink assessment error caused by the conflict between the ambiguity of the climate transition zone boundary and the spatial heterogeneity.

2. The ecosystem carbon sink assessment method according to claim 1, characterized in that: The S2 includes: S21: performing Z-score normalization processing on the temperature, precipitation, and carbon sink values ​​in the continuous climate distribution grid to generate a three-dimensional normalized feature vector; S22: performing clustering processing on the three-dimensional normalized feature vector based on the weighted K-means clustering algorithm to generate an initial geographic unit division result. The calculation formula of the initial geographic unit division result is: In j =α*CV T +β*CV P +γ*CV C Among them, Cluster i is the initial geographic unit division result, that is, the set of all grids belonging to the i-th geographic unit, V j is the three-dimensional normalized eigenvector of the j-th grid, μ i is the i-th cluster center, CV T 、CV P 、CV C are the spatial variation coefficients of temperature, precipitation, and carbon sink, respectively; α, β, and γ are weight coefficients and satisfy α+β+γ=1; S23: Calculating the boundary fuzzy index of the initial geographic unit division result based on the fuzzy membership function, determining the climate transition zone area, and generating a first geographic unit set containing transition zone marks; S24: Based on the initial boundaries of the first geographic unit set, extract the temperature change rate, precipitation change rate, and carbon sequestration per unit area to generate a first characteristic parameter set.

3. The ecosystem carbon sink assessment method according to claim 2, characterized in that: The S21 includes: S211: normalizing the temperature data, precipitation data, and carbon sink values ​​in the continuous climate distribution grid to generate three-dimensional standardized data, wherein the three-dimensional standardized data includes standardized temperature data, standardized precipitation data, and standardized carbon sink values; S212: Performing dimensionality reduction processing on the standardized temperature data, standardized precipitation data, and standardized carbon sink value of the three-dimensional standardized data using a principal component analysis algorithm to generate a three-dimensional principal component eigenvector. The expression of the three-dimensional principal component eigenvector is: V = [PC1, PC2, PC3] Among them, V is the three-dimensional principal component eigenvector, Z T is the normalized temperature data, Z P is the standardized precipitation data, Z C is the standardized carbon sink value, a i 、b i 、c i (i=1,2,3) are principal component analysis coefficients used to determine Z T 、Z P 、Z C For PC i the extent of contribution; S213: Based on the three-dimensional principal component eigenvector, construct a standardized input data set for K-means clustering to generate a three-dimensional standardized eigenvector.

4. The ecosystem carbon sink assessment method according to claim 1, characterized in that: The S3 includes: S31: performing spatial derivative calculations on the temperature change rate and precipitation change rate in the first characteristic parameter set to generate the temperature gradient modulus and the precipitation main wind direction partial derivative; S32: Calculating the joint gradient difference variance of the temperature gradient modulus and the precipitation main wind direction partial derivative using a covariance matrix to generate a transition zone fuzzy quantization index; S33: Based on the joint gradient difference variance, construct a probability distribution map reflecting the uncertainty of the climate transition zone boundary and generate a first fuzzy degree distribution.

5. The ecosystem carbon sink assessment method according to claim 1, characterized in that: The S4 includes: S41: Identifying whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold; if so, processing the first fuzzy degree distribution based on a weighted Voronoi diagram algorithm, recalculating a cell boundary distance function using a spatial variation coefficient as a weight, and generating a revised geographic cell boundary; S42: performing Gaussian convolution filtering on the corrected geographic unit boundary to generate a smoothed geographic unit boundary; S43: Based on the smoothed geographic unit boundary, updating the weighted average of the carbon sequestration per unit area within the geographic unit boundary to generate a second geographic unit set; S44: performing coefficient optimization processing on the second geographical unit set to generate a first distribution map of regionalized carbon sequestration per unit area.

6. The ecosystem carbon sink assessment method according to claim 5, characterized in that: The S44 includes: S441: Performing dynamic correlation analysis on the division scale and spatial variation coefficient of the second geographic unit set based on a correlation coefficient algorithm to generate a dynamic correlation coefficient; S442: performing iterative optimization processing on the dynamic correlation coefficient based on a simulated annealing algorithm to generate an optimal partitioning scale parameter; S443: Modeling the carbon sink per unit area and climate factors under the optimal division scale parameters based on linear regression technology to generate a first distribution map of regionalized carbon sink per unit area.

7. The ecosystem carbon sink assessment method according to claim 1, characterized in that: The S5 includes: S51: Identify whether the square value of the regression equation of the first distribution map is lower than a preset square threshold; if so, perform vegetation index extraction processing on the remote sensing image of the coverage area of ​​the first distribution map based on remote sensing vegetation index inversion technology to generate NDVI raster data; S52: performing spatial weighted fusion processing on the NDVI raster data and the carbon sink data of the continuous climate distribution grid to generate an updated carbon sink distribution dataset; S53: Based on the multi-scale clustering optimization technology, the updated carbon sink distribution dataset is subjected to transition zone boundary correction and regression modeling processing to generate a second distribution map.

8. An ecosystem carbon sink assessment system, characterized in that: The system comprises: a data acquisition and processing module, configured to acquire temperature and precipitation data of a target area based on a global ecosystem type distribution database and a climate element database, and process the temperature and precipitation data based on a kriging spatial interpolation method to generate a continuous climate distribution grid, wherein the continuous climate distribution grid is used to indicate spatially continuous distribution characteristics of temperature and precipitation; a feature clustering extraction module for performing carbon sink data processing and feature extraction on the continuous climate distribution grid based on a K-means clustering algorithm to generate a first feature parameter set including a temperature change rate, a precipitation change rate, and a carbon sink per unit area; a fuzzy distribution calculation module, configured to calculate and process the first characteristic parameter set based on a covariance matrix algorithm to generate a first fuzzy degree distribution, wherein the first fuzzy degree distribution is used to indicate the uncertainty intensity of the boundary demarcation within the climate transition zone; a distribution map generating module, configured to identify whether the gradient difference variance of the first fuzzy degree distribution exceeds a preset variance threshold, and if so, perform boundary adjustment processing and coefficient optimization processing on the first fuzzy degree distribution to generate a first distribution map of the regionalized unit area carbon sink distribution; The distribution map fusion optimization module is used to identify whether the square value of the regression equation of the first distribution map is lower than a preset square threshold. If so, the first distribution map and the carbon sink data of the continuous climate distribution grid are fused to generate a second distribution map. The second distribution map is used to eliminate the carbon sink assessment error caused by the conflict between the ambiguity of the climate transition zone boundary and spatial heterogeneity.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images