Method and system for constructing multi-characteristic ecological security pattern of metropolitan area based on som-mcr coupling

Abstract

The application relates to a metropolitan circle multi-feature ecological security pattern construction method and system based on SOM-MCR coupling. Through various network modules, multi-source spatial data is acquired and standardized processing is performed; ecological system service importance and ecological sensitivity index calculation, ecological service cluster identification based on a self-organizing mapping neural network, ecological function differentiation driving factor analysis and multi-type ecological source site screening are performed; then, a heterogeneous ecological resistance surface is constructed, multi-type ecological corridors and key ecological nodes are identified, and ecological security pattern construction and optimization are performed. The application can effectively solve the problems of insufficient identification of ecological source site function types, homogeneous expression of corridors, insufficient internal difference description of resistance surfaces and difficulty in accurate positioning of repair nodes in the existing ecological security pattern construction.

Description

Technical Field

[0001] This application relates to the fields of landscape ecology and land spatial planning technology, specifically to a method and system for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling. Background Technology

[0002] Ecological security patterns are a crucial foundation for maintaining regional ecosystem stability, ensuring the continuity of ecological processes, and supporting ecological protection of national land space. They are widely applied in areas such as national land space planning, ecological restoration, integrated watershed management, and regional ecological security assessment. Currently, the construction of ecological security patterns generally includes steps such as identifying ecological source areas, constructing ecological resistance surfaces, extracting ecological corridors, and optimizing ecological networks.

[0003] In the identification of ecological source areas, traditional techniques mainly employ methods such as comprehensive evaluation of ecosystem services, weighted overlay of multiple indicators, and extraction of ecological red lines to determine ecological source areas. These methods typically require the manual setting of indicator weights, compressing different ecosystem services such as habitat quality, annual water production, carbon storage, and food supply into a single comprehensive evaluation result. This makes it difficult to retain the original characteristic information between various ecosystem services, and fails to effectively identify multiple types of ecological source areas with different dominant ecological functions, resulting in insufficient expression of ecological function heterogeneity.

[0004] Current methods for constructing ecological resistance surfaces mostly use land use type as the basic resistance assignment, rarely considering factors such as topographic relief and the intensity of human activities. This results in insufficient expression of the spatial heterogeneity of resistance surfaces, making it difficult to truly reflect the migration resistance characteristics in ecological processes. Meanwhile, existing methods for constructing ecological corridors typically extract the minimum cost path between ecological source areas based on the minimum cumulative resistance model. They assume that ecological corridors formed between different ecological source areas have the same ecological attributes, lacking the ability to classify ecological corridors based on the dominant ecological functions of the source areas. This makes it difficult to reveal the ecological connections and material and energy flow characteristics between different ecological functions, and also fails to provide a targeted basis for differentiated ecological protection and restoration.

[0005] Existing research on the construction of ecological security patterns mainly focuses on the spatial identification of ecological source areas and ecological corridors, lacking quantitative analysis of the formation mechanisms of different ecological function types and their key driving factors. The results of the construction of ecological security patterns are not interpretable enough and cannot support regional ecological protection and land space optimization decisions. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling, which can solve the problems of insufficient identification of ecological source area functional types, homogeneous expression of corridors, insufficient characterization of internal differences in resistance surfaces, and difficulty in accurately positioning restoration nodes in the existing construction of ecological security patterns.

[0007] The first technical solution adopted in this invention is: a method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling, comprising the following steps: S1: Multi-source spatial data acquisition and standardization processing: Acquire multi-source spatial data including land use data, digital elevation model data, normalized vegetation index data, precipitation data, actual evapotranspiration data, soil data, nighttime light data, road network data, and administrative boundary data of the study area, and unify spatial benchmarks to form an ecological security pattern analysis database. S2: Calculation of Ecosystem Service Importance and Ecological Sensitivity Indicators: Based on the aforementioned ecological security pattern analysis database, six types of ecological indicators were obtained, namely habitat quality, annual water production, carbon storage, food supply capacity, soil erosion sensitivity, and geological disaster sensitivity. S3: Ecosystem service cluster identification based on self-organizing map neural network: The six types of ecological indicators obtained in step S2 are standardized and input into the self-organizing map neural network for unsupervised clustering to obtain ecosystem service clusters with different dominant ecological functions, and form an overall clustering result that reflects the category of ecosystem service clusters to which each evaluation unit belongs and the spatial distribution of each cluster. S4: Analysis of driving factors for ecological function differentiation: Conduct interpretability analysis on the ecological service clusters and overall clustering results obtained in step S3 to determine the key driving indicators and their contribution levels for different ecological service clusters; S5: Screening of multiple types of ecological source areas: By combining the type of ecosystem service cluster, patch area threshold and landscape connectivity index, ecological source areas with different dominant ecological functions are screened and classified. S6: Construction of heterogeneous ecological resistance surface: Based on land use type as the basic resistance, the basic resistance is modified by combining topographic relief and nighttime light index to obtain a comprehensive ecological resistance surface; S7: Identification of multiple types of ecological corridors and key ecological nodes: Based on the ecological source areas obtained in step S5 and the comprehensive ecological resistance surface obtained in step S6, potential ecological corridors between ecological source areas are extracted using the minimum cumulative resistance model, and the type of ecological corridor is determined according to the dominant function of the ecological source areas at both ends of the corridor; at the same time, ecological pinch points, ecological barrier points and ecological break points are identified. S8: Construction and optimization of ecological security pattern: Based on multiple types of ecological source areas, multiple types of ecological corridors and key ecological nodes, construct a multi-characteristic ecological security pattern for urban agglomerations and form differentiated ecological protection and restoration plans.

[0008] Furthermore, the specific method for step S1 is as follows: The multi-source spatial data uniformly adopts the WGS 1984 coordinate reference system; spatial clipping, resampling and null value checking are performed on raster data, and topology checking and boundary matching are performed on vector data, so that land use, topology, vegetation, meteorology, soil, human activities and road network data can participate in the evaluation within the same spatial range.

[0009] Furthermore, in step S2, habitat quality is evaluated through habitat suitability of land use types and the spatial impact intensity of external threat factors; annual water yield is estimated through the difference in water volume between precipitation and actual evapotranspiration; carbon storage is estimated through aboveground biomass carbon pools, underground biomass carbon pools, soil carbon pools, and litter carbon pools corresponding to different land use types; food supply capacity is spatially allocated through food statistics, arable land spatial distribution, and normalized difference vegetation index; soil and water loss sensitivity is evaluated through rainfall erosivity, soil erodibility, slope length and gradient, vegetation cover, and soil and water conservation measures; and geological hazard sensitivity is evaluated through elevation, slope, stratigraphic lithology, fault distance, river distance, precipitation, and human activity factors.

[0010] Furthermore, in step S3, habitat quality, annual water production, carbon storage, food supply capacity, soil erosion sensitivity, and geological disaster sensitivity are constructed into a multidimensional ecological feature vector. A self-organizing mapping neural network is constructed using the kohonen package in R language to perform topology-preserving unsupervised clustering on the evaluation units, thereby obtaining ecological service clusters with different dominant ecological functions.

[0011] Furthermore, in step S4, the local driving factors of a single ecosystem service cluster and the overall driving factors of all ecosystem service clusters are output through SHAP interpretability analysis to determine the degree of contribution of each ecological indicator to the differentiation of ecological functions.

[0012] Furthermore, in step S5, the minimum area threshold of the ecological source area is determined based on the area-quantity stability criterion, and Conefor is used to calculate the overall connectivity index IIC and the patch importance index dPC to screen, retain, and classify the preliminary ecological source areas.

[0013] Furthermore, in step S6, the basic land use resistance includes the resistance values ​​corresponding to forest land, paddy fields, dry land, grassland, water bodies, construction land, and unused land. The basic resistance is corrected by the relative values ​​of topographic relief and nighttime light index within the same landscape type to characterize the ecological migration resistance under the combined effects of natural background and human activities.

[0014] Furthermore, the type of ecological corridor in step S7 is determined based on the dominant ecological functions of the ecological source areas at both ends of the corridor; when both ends of the corridor are food production ecological source areas, it is determined as a food production corridor; when both ends of the corridor are carbon sink ecological source areas, it is determined as a carbon sink corridor; when both ends of the corridor belong to different dominant ecological function types, it is determined as a composite ecological corridor. Ecological pinch points are identified based on the concentration of current density in ecological corridors; ecological barrier points are identified based on the degree of improvement in accessibility between ecological source areas after the reduction of local resistance; and ecological fault points are identified through the spatial superposition relationship between ecological corridors and road networks.

[0015] The second technical solution adopted by this invention is: a system for constructing a multi-feature ecological security pattern in urban agglomerations based on SOM-MCR coupling, used to implement the construction method described in the first technical solution. It includes interconnected modules for multi-source spatial data preprocessing, ecological index calculation, SOM ecological service cluster identification, SHAP driving factor analysis, ecological source area screening and grading, heterogeneous ecological resistance surface construction, MCR ecological corridor and key ecological node identification, and ecological security pattern and protection and restoration scheme output. The system also includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to realize the construction of a multi-feature ecological security pattern. The multi-source spatial data preprocessing module is used to acquire land use data, digital elevation model data, normalized vegetation index data, precipitation data, actual evapotranspiration data, soil data, nighttime light data, road network data, and administrative boundary data of the study area, and to unify the spatial benchmark. The ecological index calculation module is used to calculate habitat quality, annual water production, carbon storage, food supply capacity, soil and water loss sensitivity, and geological disaster sensitivity. The SOM ecosystem service cluster identification module is used to standardize and unsupervised cluster six types of ecological indicators, and output ecosystem service clusters with different dominant ecological functions and the overall clustering results. The SHAP driving factor analysis module is used to calculate the contribution of each ecological indicator to a single ecological service cluster and the overall clustering results. The ecological source area screening and grading module is used to screen and grade multiple types of ecological source areas by combining ecological service cluster type, patch area threshold, overall connectivity index IIC and patch importance index dPC. The heterogeneous ecological resistance surface construction module is used to correct the resistance values ​​of each evaluation unit based on the land use basic resistance, using topographic relief and nighttime light index to form a comprehensive ecological resistance surface. The MCR ecological corridor and key ecological node identification module is used to extract potential ecological corridors between ecological source areas, determine the corridor type based on the dominant functions of the ecological source areas at both ends of the corridor, and identify ecological pinch points, ecological barrier points and ecological break points. The ecological security pattern and protection and restoration scheme output module is used to overlay multiple types of ecological source areas, multiple types of ecological corridors and key ecological nodes, and output the multi-characteristic ecological security pattern and differentiated ecological protection and restoration scheme of the metropolitan area.

[0016] The beneficial effects of this invention are as follows: (1) Existing SOM-MCR studies typically use SOM to complete ecological functional zoning and then directly proceed to source area and corridor analysis, resulting in unclear basis for cluster category formation. This invention uses six ecological indicators and SOM clustering results as a basis to train a random forest surrogate model and uses SHAP to calculate the contribution of each indicator to a single evaluation unit, each ecosystem service cluster, and the overall clustering results, thereby identifying the dominant functions and driving factors. This method can clarify the formation reasons of various ecological spaces, improve the interpretability of source area classification and protection measures, and avoid relying solely on empirical judgments based on cluster mean. (2) When existing methods directly extract source sites from high-value clusters, they tend to retain small fragmented patches and cannot identify patches with important connectivity functions. This invention first determines the minimum area threshold based on the stable interval of the area-quantity relationship, eliminates candidate patches that are too small in scale and lack functional stability, and then calculates IIC and dPC to evaluate the overall connectivity of the landscape and the contribution of individual patches to connectivity, thereby completing the source site screening and classification. The source sites obtained in this way have functional representativeness, spatial stability and network importance, which can improve the reliability of subsequent corridor simulation. (3) Existing SOM-MCR methods mostly assign fixed resistance based on land use type, failing to fully reflect differences in topographic conditions and human disturbance within the same land use type. This invention uses basic land use resistance as the base value and corrects it by using the ratio of topographic relief and nighttime light index to the average value of the same land use type, thus creating differentiated resistance within the same land use type. This treatment can more accurately reflect the actual passage conditions of ecological flow and reduce the deviation of the path of least resistance passing through high disturbance areas or bypassing suitable areas. (4) Existing SOM-MCR studies focus primarily on source area identification and minimum cost path extraction, lacking a unified expression of the ecological functions of corridors and different obstruction locations. This invention identifies food production, carbon sink regulation, habitat conservation, and composite corridors based on the dominant functions of the source areas at both ends of the corridor. It also identifies pinch points, obstacle points, and break points by combining current density, the degree of accessibility improvement after local resistance reduction, and the spatial superposition relationship between corridors and roads. This allows for the simultaneous acquisition of corridor functions, connectivity bottlenecks, areas with high restoration benefits, and road cutting locations, providing a basis for categorized protection, mechanism-based restoration, and engineering site selection. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a method according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating the specific steps of steps S1 to S4 in an embodiment of the present invention. Figure 3 The following is a flowchart illustrating the specific steps of steps S5 to S8 in the embodiments of the present invention. Detailed Implementation

[0019] To better understand the above-described objects, features, and advantages of the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Many specific details are set forth in the following description to provide a thorough understanding of the invention; however, the invention may be practiced in other ways different from those described herein, and therefore, the invention is not limited to the specific embodiments disclosed below.

[0020] This invention provides a system for constructing a multi-feature ecological security pattern in urban agglomerations based on SOM-MCR coupling. The system includes interconnected modules for multi-source spatial data preprocessing, ecological indicator calculation, SOM ecological service cluster identification, SHAP driving factor analysis, ecological source area screening and grading, heterogeneous ecological resistance surface construction, MCR ecological corridor and key ecological node identification, and ecological security pattern and protection / restoration scheme output. The system also includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to construct the multi-feature ecological security pattern. The multi-source spatial data preprocessing module is used to acquire land use data, digital elevation model data, normalized vegetation index data, precipitation data, actual evapotranspiration data, soil data, nighttime light data, road network data, and administrative boundary data of the study area, and to unify the spatial benchmark. The ecological index calculation module is used to calculate habitat quality, annual water production, carbon storage, food supply capacity, soil and water loss sensitivity, and geological disaster sensitivity. The SOM ecosystem service cluster identification module is used to standardize and unsupervised cluster six types of ecological indicators, and output ecosystem service clusters with different dominant ecological functions and the overall clustering results. The SHAP driving factor analysis module is used to calculate the contribution of each ecological indicator to a single ecological service cluster and the overall clustering results. The ecological source area screening and grading module is used to screen and grade multiple types of ecological source areas by combining ecological service cluster type, patch area threshold, overall connectivity index IIC and patch importance index dPC. The heterogeneous ecological resistance surface construction module is used to correct the resistance values ​​of each evaluation unit based on the land use basic resistance, using topographic relief and nighttime light index to form a comprehensive ecological resistance surface. The MCR ecological corridor and key ecological node identification module is used to extract potential ecological corridors between ecological source areas, determine the corridor type based on the dominant functions of the ecological source areas at both ends of the corridor, and identify ecological pinch points, ecological barrier points and ecological break points. The ecological security pattern and protection and restoration scheme output module is used to overlay multiple types of ecological source areas, multiple types of ecological corridors and key ecological nodes, and output the multi-characteristic ecological security pattern and differentiated ecological protection and restoration scheme of the metropolitan area.

[0021] like Figure 1 As shown, this embodiment of the invention also provides a method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling, including the following steps: S1: Multi-source spatial data acquisition and standardization processing: Acquire multi-source spatial data including land use data, digital elevation model data, normalized vegetation index data, precipitation data, actual evapotranspiration data, soil data, nighttime light data, road network data, and administrative boundary data of the study area, and unify the spatial benchmark to form an ecological security pattern analysis database; the specific steps are as follows: S101: Acquisition of basic geographic data: Acquire land use data, digital elevation model (DEM) data, normalized difference vegetation index (NDVI) data, precipitation data, actual evapotranspiration data, soil data, nighttime light data, road network data, and administrative boundary data of the study area, and record the corresponding spatial resolution, collection time, and attribute information according to each data source; S102: Spatial Reference Unification and Data Standardization: The acquired multi-source spatial data is uniformly converted to the WGS1984 coordinate reference system to unify the raster resolution, spatial range and cell size. Data with different spatial resolutions are resampled and administrative boundaries are used to clip various raster and vector data to ensure that all data have a consistent spatial reference. S103: Multi-source data consistency processing: Spatial registration is performed on land use data, digital elevation model (DEM) data, normalized difference vegetation index (NDVI) data, precipitation data, actual evapotranspiration data, soil data, nighttime light data, road network data, and administrative boundary data of the study area to eliminate spatial offsets between different data; missing values, outliers, and null value areas are checked and corrected, and the data format is unified to form a standardized spatial database capable of pixel-level analysis.

[0022] S2: Calculation of Ecosystem Service Importance and Ecological Sensitivity Indicators: Based on the aforementioned ecological security pattern analysis database, six types of ecological indicators were obtained, namely habitat quality, annual water production, carbon storage, food supply capacity, soil erosion sensitivity, and geological disaster sensitivity.

[0023] In this embodiment of the invention, habitat quality is evaluated by the spatial influence intensity of land use type habitat suitability and external threat factors; annual water yield is estimated by the difference between precipitation and actual evapotranspiration; carbon storage is estimated by the aboveground biomass carbon pool, underground biomass carbon pool, soil carbon pool, and litter carbon pool corresponding to different land use types; food supply capacity is spatially allocated by food statistics, cultivated land spatial distribution, and normalized difference vegetation index; soil and water loss sensitivity is evaluated by rainfall erosivity, soil erodibility, slope length and gradient, vegetation cover, and soil and water conservation measures; and geological hazard sensitivity is evaluated by elevation, slope, stratigraphic lithology, fault distance, river distance, precipitation, and human activity factors. The specific calculation formulas for each indicator are as follows: The formula for calculating habitat quality is as follows: ; in, Indicates the first Habitat quality values ​​for each evaluation unit; Indicates the first Habitat suitability of each evaluation unit for its land use type; Indicates the first The comprehensive habitat degradation intensity experienced by each evaluation unit; This represents the degradation intensity adjustment constant; The exponential parameter represents the shape of the degradation response curve.

[0024] The formula for calculating annual water production is as follows: ; in, Indicates the first The annual water production capacity of each evaluation unit; Indicates the first Annual precipitation for each evaluation unit; Indicates the first The actual annual evaporation of each evaluation unit.

[0025] The formula for calculating carbon reserves is as follows: ; in, This indicates the total carbon storage in the study area; Indicates the land use type or the number of evaluation units; Indicates the first Class of land use type or the first Area of ​​each evaluation unit; Indicates the first Class of land use type or the first Aboveground biomass carbon density corresponding to each evaluation unit; Indicates the first Class of land use type or the first The underground biomass carbon density corresponding to each evaluation unit; Indicates the first Class of land use type or the first Soil carbon density corresponding to each evaluation unit; Indicates the first Class of land use type or the first The litter carbon density corresponding to each evaluation unit.

[0026] The formula for calculating grain supply capacity is as follows: ; in, Indicates the first The grain supply capacity allocated to each evaluation unit; This indicates the total grain supply in the study area; Indicates the first Normalized vegetation index for each evaluation unit; This represents the minimum value of the normalized vegetation index within the study area; Indicates the first Normalized vegetation index for each evaluation unit; This indicates the total number of evaluation units participating in the allocation.

[0027] The formula for calculating soil erosion sensitivity is as follows: ; in, Indicates the first The soil and water loss sensitivity index of each evaluation unit; This represents the rainfall erosivity factor of the i-th evaluation unit; Represents the soil erodibility factor of the i-th evaluation unit; Represents the slope length factor of the i-th evaluation unit; Represents the slope factor of the i-th evaluation unit; This represents the vegetation cover factor of the i-th evaluation unit; This represents the soil and water conservation measures factor for the i-th evaluation unit; This represents the maximum value of the calculated soil and water loss results within the study area; This represents the minimum value of the calculated soil and water loss within the study area.

[0028] The formula for calculating geological hazard sensitivity is as follows: ; in, Indicates the first The sensitivity probability of geological disasters occurring in each evaluation unit; For constant terms; Indicates the first The regression coefficients of each influencing factor; Indicates the first In the evaluation unit, the first The values ​​of each influencing factor; This indicates the number of influencing factors.

[0029] S3: Ecosystem service cluster identification based on self-organizing map neural network: The six types of ecological indicators obtained in step S2 are standardized and input into the self-organizing map neural network for unsupervised clustering to obtain ecosystem service clusters with different dominant ecological functions, and form an overall clustering result that reflects the ecosystem service cluster category of each evaluation unit and the spatial distribution of each cluster; the specific steps are as follows: S301: Ecological Feature Vector Construction: The six indicators obtained in step S2—habitat quality, annual water production, carbon storage, food supply, soil erosion sensitivity, and geological hazard sensitivity—are organized into ecological feature vectors according to evaluation units. Each evaluation unit corresponds to a six-dimensional input sample, which is used to represent the combined characteristics of ecosystem services and ecological sensitivity.

[0030] S302: SOM Network Training: A self-organizing map neural network is constructed using the kohonen package in R language for unsupervised training of ecological feature vectors. During network training, the best-matching neuron is determined through competitive learning, and the neuron weights are updated based on neighborhood relationships. The specific calculation formula is as follows: ; in, Indicates the first +1 iterations The weight vector of each neuron; Indicates the first During the nth iteration The weight vector of each neuron; Indicates the first Learning rate at the next iteration; Represents the best-matching neuron With neurons Neighborhood influence coefficient between them; Indicates the first The first input Ecological feature vectors of each evaluation unit.

[0031] S303: Ecosystem Service Cluster Division: Based on the SOM network output, evaluation units with similar ecological characteristics are divided into several ecosystem service clusters. The number of ecosystem service clusters can be determined based on network topology, inter-class differences, sample distribution rationality, and ecological interpretive significance, and is not limited to a fixed number of categories. In this embodiment of the invention, ecosystem service clusters are divided into clusters of high-quality local habitats, low-value ecosystem services in central built-up areas, dominant clusters for plain agricultural production, composite ecological conservation clusters, highly disturbed and ecologically vulnerable clusters on the periphery of built-up areas, and dominant clusters for ecological regulation in mountain forests. Simultaneously, the best-matching neurons and their ecosystem service cluster categories for each evaluation unit are summarized to form an overall clustering result that includes the category to which the evaluation unit belongs, the spatial distribution of each ecosystem service cluster, and inter-cluster differences.

[0032] S304: Extraction of candidate ecological source areas: Based on the dominant ecological functions of each ecosystem service cluster, ecosystem service clusters with high ecosystem service functions and low ecological sensitivity are selected as candidate ecological source areas. For ecosystem service clusters that undertake functions such as food production, carbon sequestration maintenance, habitat conservation, or complex ecological regulation, their candidate ecological source attributes are determined respectively.

[0033] S4: Analysis of Driving Factors for Ecological Function Differentiation: An interpretability analysis is conducted on the ecological service clusters and overall clustering results obtained in step S3 to determine the key driving indicators and their contribution levels for different ecological service clusters; the specific steps are as follows:

[0034] S401: Single Cluster Driving Factor Analysis: Using the SOM ecosystem service cluster identification results as the interpretable object, the SHAP interpretability analysis method is used to calculate the contribution of each ecological evaluation index to the formation of different ecosystem service clusters, and to identify the dominant driving factors for each type of ecosystem service cluster. The specific calculation formula is as follows: ; in, Indicates the first The ecological evaluation index is for the first The average contribution intensity of ecosystem service clusters; Indicates the first Number of evaluation units within a cluster of ecosystem services; Indicates the first The set of evaluation units corresponding to a cluster of ecosystem services; Indicates the first In the evaluation unit, the first The ecological evaluation index is for the first The SHAP contribution value of the classification results of ecosystem service clusters.

[0035] S402: Overall Driving Factor Analysis: Based on the analysis of single-cluster driving factors, this study integrates the contribution intensity of indicators from different ecosystem service clusters to identify the overall driving factors influencing the spatial differentiation of ecological functions in the study area. This analysis is used to explain the causes of different ecological function types and to provide a basis for the selection of ecological source areas and the formulation of ecological protection strategies.

[0036] S5: Screening of Multi-Type Ecological Source Areas: Combining ecosystem service cluster type, patch area threshold, and landscape connectivity index, ecological source areas with different dominant ecological functions are screened and classified; the specific steps are as follows: S501: Area threshold screening: Patch extraction is performed on the candidate ecological source areas obtained in step S304, and patches that are too small, spatially fragmented, and have insufficient ecological function stability are screened out according to the minimum area threshold to obtain the initial ecological source patches.

[0037] S502: Landscape Connectivity Assessment: The Conefor software was used to calculate the overall connectivity index and patch importance index of the initial ecological source patches to evaluate the connectivity role of each patch in the regional ecological network. The formula for calculating the overall connectivity index is: ; in, Represents the overall connectivity index; Indicates the first Area of ​​each ecological patch; Indicates the first Area of ​​each ecological patch; Indicates the number of ecological patches; Indicates the first The first patch and the first The shortest connection steps or connection distance level between patches; This indicates the total landscape area of ​​the study region.

[0038] S503: Patch Importance Assessment: By comparing changes in landscape connectivity before and after the removal of individual ecological patches, a patch importance index is calculated. The specific calculation formula is as follows:

[0039] Among them, dPC pPC represents the importance index of the p-th ecological patch; PC represents the landscape connectivity index when all ecological patches are involved; PC p ⁻ represents the potential connectivity index of the remaining landscape after removing the p-th ecological patch.

[0040] S504: Determination of multiple types of ecological source areas: The final ecological source area is determined by comprehensively considering the type of ecological service cluster, area threshold, overall connectivity index, and patch importance index; based on the dominant function of the ecological service cluster to which the ecological source area belongs, the ecological source area is divided into different types, including food production ecological source areas, carbon sink regulation ecological source areas, habitat conservation ecological source areas, or composite ecological source areas.

[0041] S6: Construction of Heterogeneous Ecological Resistance Surface: Based on land use type as the basic resistance, the basic resistance is corrected by combining topographic relief and nighttime light index to obtain a comprehensive ecological resistance surface; the specific steps are as follows: S601: Basic Resistance Assignment: The basic ecological resistance value is determined based on the land use type. Different basic resistance values ​​are assigned to different land use types, such as forest land, grassland, cultivated land, water bodies, construction land, and unused land, to characterize the degree to which different landscape matrices hinder the migration of ecological processes.

[0042] S602: Resistance Surface Correction: The basic resistance value is corrected by combining topographic relief and nighttime light index to construct a heterogeneous ecological resistance surface: The specific formula is as follows: ; in, Indicates the first The revised ecological resistance values ​​for each evaluation unit; Indicates the first The basic resistance value corresponding to the land use type of each evaluation unit; An adjustment coefficient representing the degree of terrain relief; Indicates the first The topographic relief of each evaluation unit; Indicates the first Land use type of each evaluation unit The average topographic relief; The adjustment coefficient representing the nighttime light index; Indicates the first Nighttime light index for each evaluation unit; Indicates the first Land use type of each evaluation unit The average nighttime light index.

[0043] Land use type rasters are reclassified into basic resistance rasters. Then, the average values ​​of topographic relief and nighttime light index are calculated for each land use type, and the deviation ratio of each evaluation unit relative to the average value of the same type is calculated. Then, the corrected ecological resistance value is calculated pixel by pixel according to the above correction formula to form a corrected resistance rasters. Finally, the original spatial location of each evaluation unit is written into a unified grid, outliers and null values ​​are checked, and boundary clipping and rasters are performed to obtain a continuous heterogeneous ecological resistance surface with the same spatial reference and resolution as the ecological source area.

[0044] S7: Identification of Multiple Types of Ecological Corridors and Key Ecological Nodes: Based on the ecological source areas obtained in step S5 and the comprehensive ecological resistance surface obtained in step S6, potential ecological corridors between ecological source areas are extracted using the minimum cumulative resistance model, and the type of ecological corridor is determined according to the dominant functions of the ecological source areas at both ends of the corridor; simultaneously, ecological pinch points, ecological barrier points, and ecological breakpoints are identified; the specific steps are as follows: S701: Minimum Cumulative Resistance Path Extraction: Based on the multi-type ecological source areas obtained in step S5 and the heterogeneous ecological resistance surfaces obtained in step S6, the minimum cumulative resistance model is used to simulate the potential ecological corridors between different ecological source areas. ; in, Indicates the source of the ecology With ecological source The path with the least cumulative resistance between them; Indicates connection to the ecological source area With ecological source Any candidate path; and These represent adjacent evaluation units in the candidate path. and evaluation unit The ecological resistance value; Indicates adjacent evaluation units With evaluation unit The spatial distance between them.

[0045] S702: Classification of Multi-type Ecological Corridors: The corridor type is determined based on the dominant ecological functions of the ecological source areas at both ends of the corridor. When both ends of an ecological corridor connect to the same type of ecological source area, the corridor is classified as a corresponding single-function ecological corridor. When both ends of an ecological corridor connect to different types of ecological source areas, the corridor is classified as a composite ecological corridor. Specifically, when both ends of a corridor are food production ecological source areas, it is classified as a food production corridor; when both ends of a corridor are carbon sink ecological source areas, it is classified as a carbon sink corridor; and when both ends of a corridor belong to different dominant ecological function types, it is classified as a composite ecological corridor.

[0046] S703: Ecological Pinch Identification: Based on the multi-type ecological corridors and heterogeneous ecological resistance surfaces obtained in step S6, the potential ecological flow process between ecological source areas is equivalent to a circuit conduction process, where low-resistance areas correspond to high-conductivity areas, and high-resistance areas correspond to low-conductivity areas; ecological current density is calculated for each evaluation unit, and areas with highly concentrated ecological flow and bottleneck effects on ecological network connectivity are identified as candidate ecological pinch areas. The ecological current density of each evaluation unit is calculated using the following formula: ; in, Indicates the first Ecological current density of each evaluation unit; Indicates the first The ecopotential value of each evaluation unit; Indicates the first The ecopotential value of each evaluation unit; Indicates the relationship with the first A set of adjacent evaluation units.

[0047] S704: Ecological Barrier Point Identification: Based on the concept of resistance improvement, a local restoration window is constructed for each evaluation unit and its neighborhood. The ecological resistance within the window is adjusted to a preset restoration resistance value, and the changes in access resistance between ecological source areas before and after restoration are compared. If restoration of an evaluation unit can significantly reduce the access resistance between ecological source areas, then that evaluation unit is a candidate area for ecological barrier points. The ecological barrier intensity of each evaluation unit is calculated using the following formula: ; in, Indicates the first The intensity of ecological barriers in each evaluation unit; Indicates the total number of ecological source areas; Indicates the source of the ecology With ecological source Connection weights between them; This indicates the ecological source area below the original ecological resistance surface. With ecological source The access resistance value between them; Indicates the first After resistance restoration is carried out in each evaluation unit and its surrounding area, the ecological source area With ecological source The resistance value between them.

[0048] S705: Ecological Breakpoint Identification: Spatially overlay ecological corridors and road networks to extract the intersections or cuts between roads and ecological corridors, identify ecological breakpoints, and use these breakpoints as key locations for ecological restoration, corridor connectivity, and ecological transformation of roads.

[0049] S8: Construction and Optimization of Ecological Security Pattern: Based on multiple types of ecological source areas, multiple types of ecological corridors, and key ecological nodes, construct a multi-characteristic ecological security pattern for the metropolitan area and formulate differentiated ecological protection and restoration plans; the specific steps are as follows: S801: Construction of Ecological Security Pattern: Integrating multiple types of ecological source areas, multiple types of ecological corridors, ecological junctions, ecological barriers, and ecological fault points, a multi-characteristic ecological security pattern for urban agglomerations is constructed. This multi-characteristic ecological security pattern can simultaneously express the dominant functions of ecological source areas, the connection types of ecological corridors, and the spatial locations of key ecological nodes.

[0050] S802: Ecological Protection and Restoration Strategy Output: Based on different types of ecological source areas, different types of ecological corridors, and the distribution characteristics of key ecological nodes, propose zoned and classified ecological protection and restoration strategies; for food production source areas and corridors, focus on maintaining the continuity of arable land and agricultural ecological functions; for carbon sink regulation source areas and corridors, focus on strengthening the protection of forests, wetlands, and areas with high vegetation cover; for composite ecological corridors, focus on ensuring the synergistic connection between multiple ecosystem services; for ecological pinch points, ecological barrier points, and ecological break points, propose protection, restoration, and connectivity enhancement measures respectively.

[0051] The method described in this invention constructs an ecological security pattern for the study area. Building upon the ecosystem service clusters identified by SOM (Systemic Occurrence of Mechanisms), it introduces a random forest surrogate model and SHAP (Shape-Based Analysis) to quantify the contribution of each ecological indicator to the clustering results and clarify the formation causes of different ecological functional spaces. It combines area-quantity relationships, IIC (Integrated Indicator Capacity), and dPC (Dependent Polymerization) to screen and classify ecological source areas, improving the stability and network importance of source areas. It utilizes topographic relief and nighttime light index to correct land use basic resistance, enhancing the spatial heterogeneity of resistance surfaces. Based on the dominant functions of source areas, it classifies multiple types of corridors and identifies ecological pinch points, obstacle points, and breakpoints. Compared to existing SOM-MCR methods, this invention further enables ecological function interpretation, stable source area screening, corridor classification, and diagnosis of key obstructed locations. Results show that this invention retains multidimensional ecological indicator characteristics, clarifies the formation causes of various ecological spaces, improves the interpretability of source area classification and protection measure formulation, and accurately reflects ecological network connectivity and key bottlenecks. It provides a scientific basis for ecological protection and restoration and land space optimization, demonstrating good applicability and engineering application value.

[0052] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for constructing a multi-feature ecological security pattern in urban agglomerations based on SOM-MCR coupling, characterized in that, Includes the following steps: S1: Multi-source spatial data acquisition and standardization processing: Acquire multi-source spatial data including land use data, digital elevation model data, normalized vegetation index data, precipitation data, actual evapotranspiration data, soil data, nighttime light data, road network data, and administrative boundary data of the study area, and unify spatial benchmarks to form an ecological security pattern analysis database. S2: Calculation of Ecosystem Service Importance and Ecological Sensitivity Indicators: Based on the aforementioned ecological security pattern analysis database, six types of ecological indicators were obtained, namely habitat quality, annual water production, carbon storage, food supply capacity, soil erosion sensitivity, and geological disaster sensitivity. S3: Ecosystem service cluster identification based on self-organizing map neural network: The six types of ecological indicators obtained in step S2 are standardized and input into the self-organizing map neural network for unsupervised clustering to obtain ecosystem service clusters with different dominant ecological functions, and form an overall clustering result that reflects the category of ecosystem service clusters to which each evaluation unit belongs and the spatial distribution of each cluster. S4: Analysis of driving factors for ecological function differentiation: Conduct interpretability analysis on the ecological service clusters and overall clustering results obtained in step S3 to determine the key driving indicators and their contribution levels for different ecological service clusters; S5: Screening of multiple types of ecological source areas: By combining the type of ecosystem service cluster, patch area threshold and landscape connectivity index, ecological source areas with different dominant ecological functions are screened and classified. S6: Construction of heterogeneous ecological resistance surface: Based on land use type as the basic resistance, the basic resistance is modified by combining topographic relief and nighttime light index to obtain a comprehensive ecological resistance surface; S7: Identification of multiple types of ecological corridors and key ecological nodes: Based on the ecological source areas obtained in step S5 and the comprehensive ecological resistance surface obtained in step S6, potential ecological corridors between ecological source areas are extracted using the minimum cumulative resistance model, and the type of ecological corridor is determined according to the dominant function of the ecological source areas at both ends of the corridor; at the same time, ecological pinch points, ecological barrier points and ecological break points are identified. S8: Construction and optimization of ecological security pattern: Based on multiple types of ecological source areas, multiple types of ecological corridors and key ecological nodes, construct a multi-characteristic ecological security pattern for urban agglomerations and form differentiated ecological protection and restoration plans.

2. The method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling according to claim 1, characterized in that, The specific method for step S1 is as follows: The multi-source spatial data uniformly adopts the WGS 1984 coordinate reference system; spatial clipping, resampling and null value checking are performed on raster data, and topology checking and boundary matching are performed on vector data, so that land use, topology, vegetation, meteorology, soil, human activities and road network data can participate in the evaluation within the same spatial range.

3. The method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling according to claim 2, characterized in that, In step S2, habitat quality is evaluated by the habitat suitability of land use type and the spatial impact intensity of external threat factors; annual water yield is estimated by the difference between precipitation and actual evapotranspiration; carbon storage is estimated by the aboveground biomass carbon pool, underground biomass carbon pool, soil carbon pool and litter carbon pool corresponding to different land use types. Food supply capacity is spatially allocated through food statistics, arable land spatial distribution, and normalized vegetation index; soil erosion sensitivity is evaluated through rainfall erosivity, soil erodibility, slope length and gradient, vegetation cover, and soil and water conservation measures; geological hazard sensitivity is evaluated through elevation, slope, stratigraphic lithology, fault distance, river distance, precipitation, and human activity factors.

4. The method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling according to claim 3, characterized in that, In step S3, habitat quality, annual water production, carbon storage, food supply capacity, soil erosion sensitivity, and geological disaster sensitivity are used to construct a multidimensional ecological feature vector. A self-organizing mapping neural network is constructed using the kohonen package in R language to perform topology-preserving unsupervised clustering on the evaluation units, thereby obtaining ecological service clusters with different dominant ecological functions.

5. The method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling according to claim 4, characterized in that, In step S4, SHAP interpretability analysis is used to output the local driving factors of a single ecosystem service cluster and the overall driving factors of all ecosystem service clusters, which are used to determine the degree of contribution of each ecological indicator to the differentiation of ecological functions.

6. The method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling according to claim 5, characterized in that, In step S5, the minimum area threshold of the ecological source area is determined based on the area-quantity stability criterion, and Conefor is used to calculate the overall connectivity index IIC and the patch importance index dPC to screen, retain and classify the preliminary ecological source areas.

7. The method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling according to claim 6, characterized in that, In step S6, the basic land use resistance includes the resistance values ​​corresponding to forest land, paddy fields, dry land, grassland, water bodies, construction land and unused land. The basic resistance is corrected by the relative values ​​of topographic relief and nighttime light index within the same landscape type to characterize the ecological migration resistance under the combined effects of natural background and human activities.

8. The method for constructing a multi-feature ecological security pattern of urban agglomerations based on SOM-MCR coupling according to claim 7, characterized in that, The type of ecological corridor in step S7 is determined based on the dominant ecological functions of the ecological source areas at both ends of the corridor; when both ends of the corridor are food production ecological source areas, it is determined as a food production corridor; when both ends of the corridor are carbon sink ecological source areas, it is determined as a carbon sink corridor; when both ends of the corridor belong to different dominant ecological function types, it is determined as a composite ecological corridor. Ecological pinch points are identified based on the concentration of current density in ecological corridors; ecological barrier points are identified based on the degree of improvement in accessibility between ecological source areas after the reduction of local resistance; and ecological fault points are identified through the spatial superposition relationship between ecological corridors and road networks.

9. A system for constructing a multi-feature ecological security pattern in urban agglomerations based on SOM-MCR coupling, characterized in that, The system is designed to implement the construction method according to any one of claims 1 to 8, comprising a multi-source spatial data preprocessing module, an ecological indicator calculation module, a SOM ecosystem service cluster identification module, a SHAP driving factor analysis module, an ecological source area screening and grading module, a heterogeneous ecological resistance surface construction module, an MCR ecological corridor and key ecological node identification module, and an ecological security pattern and protection and restoration scheme output module that are interconnected; the system also includes a memory and a processor, wherein the memory stores a computer program, and the processor is used to execute the computer program to realize the construction of a multi-feature ecological security pattern; The multi-source spatial data preprocessing module is used to acquire land use data, digital elevation model data, normalized vegetation index data, precipitation data, actual evapotranspiration data, soil data, nighttime light data, road network data, and administrative boundary data of the study area, and to unify the spatial benchmark. The ecological index calculation module is used to calculate habitat quality, annual water production, carbon storage, food supply capacity, soil and water loss sensitivity, and geological disaster sensitivity. The SOM ecosystem service cluster identification module is used to standardize and unsupervised cluster six types of ecological indicators, and output ecosystem service clusters with different dominant ecological functions and the overall clustering results. The SHAP driving factor analysis module is used to calculate the contribution of each ecological indicator to a single ecological service cluster and the overall clustering results. The ecological source area screening and grading module is used to screen and grade multiple types of ecological source areas by combining ecological service cluster type, patch area threshold, overall connectivity index IIC and patch importance index dPC. The heterogeneous ecological resistance surface construction module is used to correct the resistance values ​​of each evaluation unit based on the land use basic resistance, using topographic relief and nighttime light index to form a comprehensive ecological resistance surface. The MCR ecological corridor and key ecological node identification module is used to extract potential ecological corridors between ecological source areas, determine the corridor type based on the dominant functions of the ecological source areas at both ends of the corridor, and identify ecological pinch points, ecological barrier points and ecological break points. The ecological security pattern and protection and restoration scheme output module is used to overlay multiple types of ecological source areas, multiple types of ecological corridors and key ecological nodes, and output the multi-characteristic ecological security pattern and differentiated ecological protection and restoration scheme of the metropolitan area.

Images