Bird behavior frequency distribution-based urban wetland restoration effect evaluation method and system

Abstract

The application discloses a city wetland restoration effect evaluation method and system based on bird behavior frequency distribution, relates to the technical field of ecological restoration evaluation, and comprises the following steps: dividing the wetland into different habitats according to the degree of human intervention and the vegetation structure and coding; collecting bird individual behavior data in the grid; constructing a basic frequency table with the restoration type, the behavior type and the vertical level as the dimensions, extracting the restoration measure-behavior response bias characteristic value, the niche potential gradient characteristic value and the group overlap pressure characteristic value; constructing a multi-dimensional evaluation model containing the behavior function integrity balance index, the vertical structure niche support index and the community stability correction factor based on the above characteristics, and calculating the wetland restoration ecological behavior benefit comprehensive index; comparing the comprehensive index with the diagnostic threshold value and tracing back the sub-characteristic values, and outputting the restoration effect grade and the management suggestion. The application realizes multi-dimensional quantitative evaluation of the wetland restoration effect.

Description

Technical Field

[0001] This invention relates to the field of ecological restoration assessment technology, specifically to a method and system for assessing the effectiveness of urban wetland restoration based on the frequency distribution of bird behavior. Background Technology

[0002] With the acceleration of urbanization, urban wetland ecosystems have suffered varying degrees of degradation, leading to a large-scale wetland restoration project. Wetland restoration effectiveness assessment is a crucial step in evaluating the quality of restoration projects and guiding subsequent management. Birds, as important indicator species of wetland ecosystems, are highly sensitive to habitat changes in their community structure and behavioral characteristics.

[0003] Existing wetland restoration assessment methods mainly rely on water quality physicochemical indicators, vegetation cover, or species diversity indices. However, these methods have the following limitations: the species diversity index only reflects what birds are present and cannot reflect the actual functional behaviors of birds in the restored habitat, such as feeding, roosting, and movement, making it difficult to measure the true carrying capacity of the restored habitat for bird life history stages; artificial feeding or habitats with monoculture lawns often attract large flocks of birds to feed, resulting in inflated behavioral frequencies that traditional assessment methods cannot effectively identify; and there is a lack of quantitative analysis methods for the rationality of the vertical spatial structure of the restored habitat and the competitive pressure between functional groups.

[0004] Therefore, there is an urgent need for a method that can conduct a refined and quantitative assessment of the effectiveness of urban wetland restoration from multiple dimensions, including behavioral functional integrity, spatial structural heterogeneity, and community stability, in order to make up for the shortcomings of existing technologies, such as single assessment dimensions and susceptibility to data noise interference. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing a method and system for evaluating the effect of urban wetland restoration based on the frequency distribution of bird behavior.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for evaluating the effectiveness of urban wetland restoration based on the frequency distribution of bird behavior includes the following steps:

[0008] Step S1: Obtain remote sensing images and vegetation survey data of the urban wetland to be evaluated. Based on the degree of human intervention and the vertical structure characteristics of vegetation, divide the entire wetland area into natural restoration habitats, semi-natural-semi-artificial restoration habitats, and artificial restoration habitats. Generate a rule-based evaluation grid covering the entire area and assign a corresponding restoration type code to each grid.

[0009] Step S2: Deploy survey transects or sample points in grids with different repair type codes, and collect individual bird behavior data according to preset behavior state maintenance duration threshold rules and instantaneous scanning rules. The individual bird behavior data includes at least the behavior type, effective behavior frequency, vertical spatial level of the bird when the behavior occurs, and the feeding functional group category to which it belongs.

[0010] Step S3: Based on the individual bird behavior data, construct a basic frequency table with restoration type encoding, behavior type and vertical spatial hierarchy as dimensions, and extract restoration measure-behavioral response bias feature value, vertical spatial utilization niche potential energy gradient feature value and inter-group spatial niche overlap pressure feature value from the basic frequency table.

[0011] Step S4: Using the extracted feature values, construct a multidimensional composite assessment model that includes the behavioral function integrity balance index, the vertical structure niche support index, and the community stability correction factor, and calculate the comprehensive index of wetland restoration ecological behavior benefits by using a weighted geometric average method based on area contribution.

[0012] Step S5: Compare the comprehensive index of wetland restoration ecological behavior benefits with the preset multi-level diagnostic threshold, and combine the retrospective diagnosis of each sub-feature value to output the wetland restoration effectiveness level judgment result and corresponding restoration management suggestions.

[0013] Furthermore, step S1 specifically includes the following steps:

[0014] Step S1.1: Obtain remote sensing images and vegetation survey data of the urban wetland to be evaluated;

[0015] Step S1.2: Divide the acquired high-resolution remote sensing image into several habitat patches with land feature boundaries as the boundaries;

[0016] Step S1.3: For each habitat patch, if there is no direct human modification and the degree of human intervention is lower than the preset threshold, it is coded as a natural restoration habitat; if it is located at the water-land interface and has undergone native plant replanting or micro-topography modification, it is coded as a semi-natural-semi-artificial restoration habitat; if the human intervention exceeds the threshold or includes regular lawns, hard revetments and single vegetation zones, it is coded as an artificial restoration habitat.

[0017] Step S1.4: Use the Geographic Information System fishing net tool to generate a rule evaluation grid covering the entire area, and assign the restoration type code of the habitat patch where the center point of each grid is located to the grid as its habitat restoration attribute label.

[0018] Furthermore, step S2 specifically includes the following steps:

[0019] Step S2.1: Deploy survey transects or fixed radius sampling points within grids with different habitat restoration attribute labels to ensure that the total length of transects and the number of sampling points within each type of restoration habitat are not less than the preset value.

[0020] Step S2.2: Define behavior coding rules: record pecking and hunting as feeding behavior, flying, swimming and jumping as movement behavior, and resting, preening and singing as resting behavior.

[0021] Step S2.3: For trackable individuals, a behavior switch or a single behavior continuously exceeding the threshold is counted as one valid frequency; for flocking birds, the number of individuals of each behavior type is counted as a snapshot value using a fixed-interval instantaneous scanning method.

[0022] Step S2.4: Synchronously record the vertical spatial level of the bird when the behavior occurs. The vertical spatial level is divided into the ground layer, herbaceous layer, shrub layer, lower canopy layer of trees, upper canopy layer of trees, and aerial layer according to the height above the ground.

[0023] Step S2.5: Classify bird species into insectivorous, omnivorous, herbivorous, or carnivorous groups based on their diet;

[0024] Step S2.6: Cross-summarize the data according to the three dimensions of repair type coding, behavior type coding, and vertical hierarchy coding to form a basic frequency table.

[0025] Furthermore, step S3 specifically includes the following steps:

[0026] Step S3.1: Construct the remediation measure-behavioral response bias index. This index is based on standardized residuals and introduces the behavioral variance of semi-natural remediation habitats as a regularization term to quantify the induction or inhibition intensity of different remediation measures on specific behavioral types.

[0027] Step S3.2: Construct the niche potential energy gradient feature value. By calculating the sum of the square of the difference in the probability of behavior distribution between adjacent vertical layers and the information weight factor of the low utilization layer, assess the continuity of the habitat vertical structure and whether there is an ecological fault that forces the concentration of behavior.

[0028] Step S3.3: Calculate the vertical spatial overlap of different dietary functional groups in each restoration habitat, and extract the maximum overlap value as the niche overlap pressure characteristic value to reflect the habitat's ability to provide niche separation space.

[0029] Furthermore, step S4 specifically includes the following steps:

[0030] Step S4.1: Calculate the behavioral function integrity balance index, which is obtained by multiplying the weighted geometric mean of the proportions of feeding, locomotion and habitat by an exponential decay penalty factor based on the range of behavior proportions.

[0031] Step S4.2: Calculate the vertical structure niche support index. The vertical structure niche support index is obtained by multiplying the niche potential energy gradient characteristic value plus one and taking the reciprocal as the vertical continuity factor, and multiplying it with the niche separation factor calculated based on the maximum overlap.

[0032] Step S4.3: Calculate the community stability correction factor, which is obtained by nonlinear mapping with the reciprocal of the Simpson dominance index as the independent variable.

[0033] Step S4.4: Multiply the behavioral function integrity balance index, vertical structure niche support index and community stability correction factor to obtain the individual restoration efficiency score of each restoration type. Using the proportion of the square root of the area occupied by each restoration type to the square root of the total area of ​​the whole area as the weight, perform a weighted geometric average on the individual restoration efficiency scores to obtain the comprehensive index of wetland restoration ecological behavior benefits.

[0034] Furthermore, in step S4.1, the calculation of the behavioral function integrity balance index introduces a behavioral ecological function weight coefficient, wherein the weight of feeding behavior is set higher than that of movement behavior and habitat behavior.

[0035] Furthermore, step S5 specifically includes the following steps:

[0036] Step S5.1: Compare the comprehensive index of wetland restoration ecological behavior benefits with the preset excellent threshold, good threshold and general threshold level by level to determine the wetland restoration effectiveness level;

[0037] Step S5.2: Perform retrospective diagnosis on the sub-feature values ​​corresponding to each restoration type: When the behavioral function integrity balance index of the artificial restoration habitat is higher than the warning threshold of the artificial area behavioral function and the community stability correction factor is lower than the warning threshold of the artificial area community stability, output a suggestion to reduce artificial lawns and replant native shrubs; when the niche potential energy gradient feature value of the semi-natural-semi-artificial restoration habitat is higher than the warning threshold of the vertical fault in the semi-natural area, output a suggestion to check the shrub layer and replant the middle layer vegetation; when the niche overlap pressure feature value of the natural restoration habitat is higher than the warning threshold of the natural area overlap, output a suggestion to increase microhabitat diversity.

[0038] A system for evaluating the effectiveness of urban wetland restoration based on the frequency distribution of bird behavior, used to implement any of the methods for evaluating the effectiveness of urban wetland restoration based on the frequency distribution of bird behavior, including:

[0039] The data acquisition and segmentation module is used to acquire remote sensing images and vegetation survey data, segment natural restoration, semi-natural-semi-artificial restoration and artificial restoration habitats, generate rule evaluation grids and assign restoration type codes;

[0040] The behavioral data acquisition and preprocessing module is used to lay out transects and sample points within the grid, collect individual bird behavioral data according to preset rules, and then summarize the data after cleaning to form a basic frequency table.

[0041] The feature extraction module is used to extract remediation action-behavioral response bias features, niche potential gradient features, and niche overlap pressure features from the basic frequency table;

[0042] The composite index construction module is used to calculate the behavioral functional integrity balance index, vertical structure niche support index and community stability correction factor using the above feature values, and to calculate the comprehensive index of wetland restoration ecological behavior benefits by weighted geometric average.

[0043] The diagnostic output module compares the comprehensive index with preset thresholds and outputs the repair effectiveness level and management suggestions.

[0044] Furthermore, the behavioral data acquisition and preprocessing module includes:

[0045] The behavior coding submodule is used to encode observed activities as feeding, locomotion, or resting behaviors;

[0046] The frequency statistics submodule is used to count individual birds using a duration threshold method and to count flocks of birds using an instantaneous scan method.

[0047] The spatial attribution submodule is used to record the vertical spatial hierarchy at the time the behavior occurs;

[0048] The group classification submodule is used to classify bird species into corresponding groups based on their diet.

[0049] Furthermore, the composite index construction module includes:

[0050] The behavioral function assessment submodule is used to generate a behavioral function integrity balance index by using a weighted geometric mean and a range penalty factor.

[0051] The spatial structure assessment submodule is used to generate a vertical structure niche support index based on potential energy gradient and overlapping pressure.

[0052] The stability correction submodule is used to generate community stability correction factors through the hyperbolic tangent mapping of the Simpson dominance index.

[0053] The fusion calculation submodule is used to perform a weighted geometric average of the effectiveness scores of each individual unit according to the square root of the area, and generate a comprehensive index of the ecological behavior benefits of wetland restoration.

[0054] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0055] 1. This invention constructs a composite assessment index from three core dimensions: behavioral functional integrity, spatial structural heterogeneity, and community stability. This makes up for the shortcomings of traditional methods that only focus on species quantity or single behavior frequency, and can more comprehensively reflect the ecological function restoration level of wetland restoration.

[0056] 2. By introducing a behavioral range penalty factor and a community stability correction factor based on the hyperbolic tangent function, this invention can effectively identify and suppress false prosperity disturbances caused by artificial feeding or single lawns, making the assessment results closer to the true ecological restoration state.

[0057] 3. By tracing back each sub-feature value, including the bias value of restoration measures-behavioral response, the ecological niche potential gradient, and the degree of cluster overlap, this invention can accurately locate the structural defects of restoration projects, such as the absence of shrub layer and excessive cluster overlap, providing a scientific basis for subsequent precise restoration. Attached Figure Description

[0058] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0059] Figure 1 This is a flowchart illustrating an embodiment of the present invention;

[0060] Figure 2 This is a system schematic diagram according to an embodiment of the present invention;

[0061] Figure 3 This is a hierarchical structure diagram of multidimensional feature extraction according to an embodiment of the present invention. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0063] like Figure 1 As shown, the method for evaluating the effectiveness of urban wetland restoration based on the frequency distribution of bird behavior includes the following steps:

[0064] Step S1: Obtain remote sensing images and vegetation survey data of the urban wetland to be evaluated. Based on the degree of human intervention and the vertical structure characteristics of vegetation, divide the entire wetland area into natural restoration habitats, semi-natural-semi-artificial restoration habitats, and artificial restoration habitats. Generate a rule-based evaluation grid covering the entire area and assign a corresponding restoration type code to each grid.

[0065] Step S2: Deploy survey transects or sample points in grids with different repair type codes, and collect individual bird behavior data according to preset behavior state maintenance duration threshold rules and instantaneous scanning rules. The individual bird behavior data includes at least the behavior type, effective behavior frequency, vertical spatial level of the bird when the behavior occurs, and the feeding functional group category to which it belongs.

[0066] Step S3: Based on the individual bird behavior data, construct a basic frequency table with restoration type encoding, behavior type and vertical spatial hierarchy as dimensions, and extract restoration measure-behavioral response bias feature value, vertical spatial utilization niche potential energy gradient feature value and inter-group spatial niche overlap pressure feature value from the basic frequency table.

[0067] Step S4: Using the extracted feature values, construct a multidimensional composite assessment model that includes the behavioral function integrity balance index, the vertical structure niche support index, and the community stability correction factor, and calculate the comprehensive index of wetland restoration ecological behavior benefits by using a weighted geometric average method based on area contribution.

[0068] Step S5: Compare the comprehensive index of wetland restoration ecological behavior benefits with the preset multi-level diagnostic threshold, and combine the retrospective diagnosis of each sub-feature value to output the wetland restoration effectiveness level judgment result and corresponding restoration management suggestions.

[0069] Step S1 specifically includes the following steps:

[0070] Step S1.1: Obtain remote sensing images and vegetation survey data of the urban wetland to be evaluated;

[0071] Step S1.2: Divide the acquired high-resolution remote sensing image into several habitat patches with land feature boundaries as the boundaries;

[0072] Step S1.3: For each habitat patch, if there is no direct human modification and the degree of human intervention is lower than the preset threshold, it is coded as a natural restoration habitat; if it is located at the water-land interface and has undergone native plant replanting or micro-topography modification, it is coded as a semi-natural-semi-artificial restoration habitat; if the human intervention exceeds the threshold or includes regular lawns, hard revetments and single vegetation zones, it is coded as an artificial restoration habitat.

[0073] Step S1.4: Use the Geographic Information System fishing net tool to generate a rule evaluation grid covering the entire area, and assign the restoration type code of the habitat patch where the center point of each grid is located to the grid as its habitat restoration attribute label.

[0074] Acquire high-resolution remote sensing images of the urban wetlands to be evaluated, with a spatial resolution better than 1 meter and recent vegetation cover survey data. Use object-oriented image segmentation algorithms, such as multi-scale segmentation algorithms, to segment the remote sensing images into several habitat patches with high internal homogeneity. The boundaries of the patches are based on obvious ground features, such as roads, water bodies, and forest edges.

[0075] For each habitat patch, a unique restoration type code was assigned based on the degree of human intervention and the complexity of the vegetation's vertical structure as recorded in the field survey.

[0076] Naturally restored habitats must meet both of the following conditions: there must be no record of artificial planting, mowing, or terrain modification in the past five years, and the degree of human intervention must be less than 5%;

[0077] Semi-natural-semi-artificial restoration habitats meet any of the following conditions: native plant replanting, micro-topography modification, or litter management have been carried out within the past five years, with human intervention ranging from 5% to 40%; located at the junction of water and land areas, with naturally developed or artificially restored wetland vegetation zones;

[0078] Artificially restored habitats meet any of the following conditions: human intervention exceeds 40%, including formally mowed lawns, row-lined street trees, and monoculture forests; hard-walled water areas lacking continuous aquatic vegetation buffer zones;

[0079] Using the Geographic Information System (GIS) fishing net tool, a rule assessment grid covering the entire area is generated. The recommended grid size is 50m×50m. The restoration type code of the habitat patch located at the center point of each grid is assigned to the grid as the habitat restoration attribute label of the grid.

[0080] Step S2 specifically includes the following steps:

[0081] Step S2.1: Deploy survey transects or fixed radius sampling points within grids with different habitat restoration attribute labels to ensure that the total length of transects and the number of sampling points within each type of restoration habitat are not less than the preset value.

[0082] Step S2.2: Define behavior coding rules: record pecking and hunting as feeding behavior, flying, swimming and jumping as movement behavior, and resting, preening and singing as resting behavior.

[0083] Step S2.3: For trackable individuals, a behavior switch or a single behavior continuously exceeding the threshold is counted as one valid frequency; for flocking birds, the number of individuals of each behavior type is counted as a snapshot value using a fixed-interval instantaneous scanning method.

[0084] Step S2.4: Synchronously record the vertical spatial level of the bird when the behavior occurs. The vertical spatial level is divided into the ground layer, herbaceous layer, shrub layer, lower canopy layer of trees, upper canopy layer of trees, and aerial layer according to the height above the ground.

[0085] Step S2.5: Classify bird species into insectivorous, omnivorous, herbivorous, or carnivorous groups based on their diet;

[0086] Step S2.6: Cross-summarize the data according to the three dimensions of repair type coding, behavior type coding, and vertical hierarchy coding to form a basic frequency table.

[0087] Within grids with different habitat restoration attribute labels, survey transects or fixed-radius sampling points are deployed according to the feasibility principle. The total length of survey transects within each type of restoration habitat is not less than 3000 meters, and the distance between sampling points is not less than 200 meters to ensure the representativeness of the sampling.

[0088] Three core behaviors and their codes are defined: feeding behavior: actions directly related to energy intake, such as pecking, hunting, swallowing food, and tearing plant tissues, performed by individual birds; locomotion behavior: non-feeding spatial movement actions, such as flying, swimming, jumping, and climbing, performed by individual birds; and resting behavior: non-displacement and non-feeding actions, such as remaining still, preening, singing, and resting, performed by individual birds.

[0089] For a clearly trackable individual, each time a behavior type switch occurs, or when the duration of a continuous behavior type exceeds a preset duration threshold (10 seconds), it is counted as one effective behavior frequency unit.

[0090] For flocks of birds that cannot be distinguished from individuals, an instantaneous scanning method is used. At fixed time intervals, the flock is scanned and observed once every 30 seconds. The number of individuals in the identifiable state under each behavioral type at that instant is counted as the behavior frequency snapshot value at that moment.

[0091] While recording the frequency of the behavior, the vertical spatial level of the vegetation occupied by the bird at the time the behavior occurred was also recorded. The vertical spatial level was divided into the following six levels according to the height h above the ground:

[0092] L1 Surface layer: h=0, directly in contact with the ground or water surface; L2 Herbaceous layer: 0<h≤1m; L3 Shrub layer: 1m<h≤3m; L4 Undertree layer: 3m<h≤lower edge of the canopy; L5 Uppertree layer: the surface of the canopy and protruding treetops; L6 Aerial layer: the flight space significantly higher than the canopy layer;

[0093] Each observed bird species was classified into one of the following four dietary functional groups:

[0094] Insectivorous groups: their main food source is insects and other invertebrates; Omnivorous groups: their diet consists of both animal and plant matter; Herbivorous groups: their main food source is plant seeds, fruits, leaves, etc.; Carnivorous groups: their main food source is vertebrates.

[0095] After removing ambiguous records where species or behavior types could not be identified due to objective conditions such as weather and light, the cleaned data was cross-summarized according to three dimensions: restoration type coding, behavior type coding, and vertical hierarchy coding to form a basic frequency table.

[0096] Step S3 specifically includes the following steps:

[0097] Step S3.1: Construct the remediation measure-behavioral response bias index. This index is based on standardized residuals and introduces the behavioral variance of semi-natural remediation habitats as a regularization term to quantify the induction or inhibition intensity of different remediation measures on specific behavioral types.

[0098] Step S3.2: Construct the niche potential energy gradient feature value. By calculating the sum of the square of the difference in the probability of behavior distribution between adjacent vertical layers and the information weight factor of the low utilization layer, assess the continuity of the habitat vertical structure and whether there is an ecological fault that forces the concentration of behavior.

[0099] Step S3.3: Calculate the vertical spatial overlap of different dietary functional groups in each restoration habitat, and extract the maximum overlap value as the niche overlap pressure characteristic value to reflect the habitat's ability to provide niche separation space.

[0100] The specific formula for the remedial action-behavioral response bias index is as follows:

[0101]

[0102] in, This represents the response bias value of repair type i to behavior j, with positive values ​​indicating significant promotion and negative values ​​indicating inhibition or lack of support. This represents the total effective frequency of behavior j across all meshes of repair type i. This represents the theoretical expected frequency of behavior j in repair type i, under the null hypothesis that repair measures and behavior types are independent of each other. The behavior variance regularization term represents the variance of the frequency of behavior j among different grids within a semi-natural and semi-artificial restoration habitat. The semi-natural restoration area is usually considered to be the most stable state and closest to the reference target in the restoration project. Using its behavior variance as the regularization term in the denominator can avoid statistically singular values ​​caused by the small sample size of some habitats.

[0103] The specific formula for the theoretical expected frequency of behavior j in repair type i is as follows:

[0104]

[0105] in, This represents the total frequency of all behaviors of repair type i. This represents the total frequency of behavior j across all repair types. This represents the total frequency of all behaviors across the entire region. This represents the total area occupied by repair type i. This represents the average floor area occupied by the three types of restoration. This represents the habitat carrying capacity elasticity coefficient, which is usually taken as 0.43. Theoretically, the larger the habitat area, the higher the expected frequency should be. However, due to the territorial behavior and density constraints of birds, the expected frequency does not increase linearly with the area. Therefore, an exponent less than 1 is used for nonlinear correction.

[0106] The specific formula for the characteristic value of the niche potential energy gradient is as follows:

[0107]

[0108] in, This represents the vertical niche potential gradient of habitat type i. A smaller value indicates a more continuous vertical habitat structure, while a larger value indicates a drastic abrupt change in behavioral distribution between adjacent layers. Let represent the probability density of a behavior occurring at vertical level k within habitat type i. This is calculated by summing the observation frequencies of all behavior types at all vertical levels k in habitat type i, obtaining the total behavior frequency for that habitat. Then, the observation frequencies of all behavior types at vertical level k in habitat type i are summed, yielding the behavior frequency at vertical level k. Finally, the probability density is obtained by dividing the behavior frequency at vertical level k by the total behavior frequency for that habitat. This represents the probability density of behavior occurring at vertical level k+1 in habitat type i. Represents a very small positive number;

[0109] The formula for calculating the vertical spatial overlap is:

[0110]

[0111] in, This represents the vertical niche overlap between any two groups u and v in a dietary functional group. This represents the proportion of the frequency of group u's behavior at vertical level k in habitat type i to the total frequency of group u's behavior in that habitat. This represents the proportion of the frequency of group v's behavior at vertical level k in habitat type i to the total frequency of the group's behavior in that habitat;

[0112] Extracting maximum overlap The closer the maximum overlap is to 1, the more it indicates that at least two functional groups have a high degree of overlap in spatial utilization, and there is a risk of fierce interspecific competition, reflecting that the restored habitat has failed to provide sufficient niche separation space.

[0113] like Figure 3 The diagram illustrates the extraction process from the basic frequency table to core assessment features. The first layer consists of frequency statistics, calculated directly from the basic frequency table, including: the total observed frequency of behavior j in restoration type i, the frequency distribution of behavior at each vertical level k, and the proportion of behavior frequency for each dietary functional group at each vertical level. The second layer comprises intermediate variables, including: the theoretically expected frequency based on area weighting, the probability density of behavior at each vertical level, and the vertical spatial overlap between different groups. The third layer consists of core features, calculated using the intermediate variables to obtain three final features: the restoration measure-behavior response bias index, used to measure the induction intensity of restoration measures on specific behaviors; the niche potential gradient feature, used to assess the continuity of vertical structures; and the maximum overlap pressure feature, used to reflect the habitat's ability to provide niche separation space. These three core features will serve as input variables for the subsequent multidimensional composite assessment model.

[0114] Step S4 specifically includes the following steps:

[0115] Step S4.1: Calculate the behavioral function integrity balance index, which is obtained by multiplying the weighted geometric mean of the proportions of feeding, locomotion and habitat by an exponential decay penalty factor based on the range of behavior proportions.

[0116] Step S4.2: Calculate the vertical structure niche support index. The vertical structure niche support index is obtained by multiplying the niche potential energy gradient characteristic value plus one and taking the reciprocal as the vertical continuity factor, and multiplying it with the niche separation factor calculated based on the maximum overlap.

[0117] Step S4.3: Calculate the community stability correction factor, which is obtained by nonlinear mapping with the reciprocal of the Simpson dominance index as the independent variable.

[0118] Step S4.4: Multiply the behavioral function integrity balance index, vertical structure niche support index and community stability correction factor to obtain the individual restoration efficiency score of each restoration type. Using the proportion of the square root of the area occupied by each restoration type to the square root of the total area of ​​the whole area as the weight, perform a weighted geometric average on the individual restoration efficiency scores to obtain the comprehensive index of wetland restoration ecological behavior benefits.

[0119] In step S4.1, the calculation of the behavioral function integrity balance index introduces a behavioral ecological function weight coefficient, wherein the weight of feeding behavior is set higher than that of movement behavior and habitat behavior.

[0120] The specific formula for the behavioral function integrity balance index is as follows:

[0121]

[0122] in, This represents the behavioral and functional integrity balance index of habitat type i. This indicates the frequency proportion of behavior type j in repair type i. The ecological function weighting coefficients for these behaviors are: 0.5 for feeding behavior, 0.3 for locomotion behavior, and 0.2 for habitat behavior. and These represent the maximum and minimum percentages of the three behaviors, respectively. This represents the standard deviation of the proportions of the three behaviors in a semi-natural restored habitat. Indicates feeding behavior. Indicates movement behavior. Indicates habitat behavior;

[0123] The specific formula for the vertical structure niche support index is as follows:

[0124]

[0125] in, The vertical structure niche support index represents the habitat type i.

[0126] The specific formula for the community stability correction factor is as follows:

[0127]

[0128] in, This represents the community stability correction factor for habitat type i. The Simpson dominance index represents the percentage of dominance in a person. This represents a very small positive number and is used to prevent numerical overflow. This represents the sensitivity adjustment factor, which defaults to 1.5. Represents the hyperbolic tangent function;

[0129] The indices of the above three dimensions are then fused using a weighted geometric mean based on area contribution according to restoration type to obtain a comprehensive index of wetland restoration ecological behavior benefits. The specific formula is as follows:

[0130]

[0131] in, This represents the comprehensive index of ecological benefits of wetland restoration. Indicates the ecological contribution weight of the restoration type. These represent natural restoration, semi-natural / semi-artificial restoration, and artificial restoration, respectively. The specific value is the square root of the ratio of the area occupied by the restoration type to the total area of ​​the wetland assessment area. The square root is used to balance the high weighting requirements of large-area habitats with the influence of small-area key habitats, ensuring that semi-natural restoration areas with high ecological value are not ignored even if the area is small.

[0132] Step S5 specifically includes the following steps:

[0133] Step S5.1: Compare the comprehensive index of wetland restoration ecological behavior benefits with the preset excellent threshold, good threshold and general threshold level by level to determine the wetland restoration effectiveness level;

[0134] Step S5.2: Perform retrospective diagnosis on the sub-feature values ​​corresponding to each restoration type: When the behavioral function integrity balance index of the artificial restoration habitat is higher than the warning threshold of the artificial area behavioral function and the community stability correction factor is lower than the warning threshold of the artificial area community stability, output a suggestion to reduce artificial lawns and replant native shrubs; when the niche potential energy gradient feature value of the semi-natural-semi-artificial restoration habitat is higher than the warning threshold of the vertical fault in the semi-natural area, output a suggestion to check the shrub layer and replant the middle layer vegetation; when the niche overlap pressure feature value of the natural restoration habitat is higher than the warning threshold of the natural area overlap, output a suggestion to increase microhabitat diversity.

[0135] The calculated The value is compared with the preset three-level threshold:

[0136] like A value ≥0.75 indicates excellent wetland restoration results; a value ≤0.50 indicates good results. <0.75 indicates good wetland restoration effectiveness; if 0.25≤ <0.50 indicates that the wetland restoration effect is considered average; if A value <0.25 indicates poor wetland restoration effectiveness;

[0137] While determining the output level, backtesting diagnosis is performed on the sub-feature values ​​of each repair type, and corresponding repair management suggestions are generated:

[0138] like >0.7 and <0.4, Output: The artificial restoration area exhibits a single dominant species-dependent feeding aggregation; it is recommended to reduce the area of ​​artificial turf and replant native shrubs to improve habitat concealment; if >0.5, Output: There are vertical structural faults in the semi-natural restoration area. It is recommended to check the shrub cover and replant mid-layer vegetation to connect the vertical ecological corridors; if >0.85, Output: The natural restoration area has high group overlap and insufficient spatial heterogeneity. It is recommended to thin out the area or introduce fallen trees to increase microhabitat diversity.

[0139] like Figure 2 As shown, the urban wetland restoration effect evaluation system based on bird behavior frequency distribution is used to implement any of the urban wetland restoration effect evaluation methods based on bird behavior frequency distribution, including:

[0140] The data acquisition and segmentation module is used to acquire remote sensing images and vegetation survey data, segment natural restoration, semi-natural-semi-artificial restoration and artificial restoration habitats, generate rule evaluation grids and assign restoration type codes;

[0141] The behavioral data acquisition and preprocessing module is used to lay out transects and sample points within the grid, collect individual bird behavioral data according to preset rules, and then summarize the data after cleaning to form a basic frequency table.

[0142] The feature extraction module is used to extract remediation action-behavioral response bias features, niche potential gradient features, and niche overlap pressure features from the basic frequency table;

[0143] The composite index construction module is used to calculate the behavioral functional integrity balance index, vertical structure niche support index and community stability correction factor using the above feature values, and to calculate the comprehensive index of wetland restoration ecological behavior benefits by weighted geometric average.

[0144] The diagnostic output module compares the comprehensive index with preset thresholds and outputs the repair effectiveness level and management suggestions.

[0145] The behavioral data acquisition and preprocessing module includes:

[0146] The behavior coding submodule is used to encode observed activities as feeding, locomotion, or resting behaviors;

[0147] The frequency statistics submodule is used to count individual birds using a duration threshold method and to count flocks of birds using an instantaneous scan method.

[0148] The spatial attribution submodule is used to record the vertical spatial hierarchy at the time the behavior occurs;

[0149] The group classification submodule is used to classify bird species into corresponding groups based on their diet.

[0150] The composite index construction module includes:

[0151] The behavioral function assessment submodule is used to generate a behavioral function integrity balance index by using a weighted geometric mean and a range penalty factor.

[0152] The spatial structure assessment submodule is used to generate a vertical structure niche support index based on potential energy gradient and overlapping pressure.

[0153] The stability correction submodule is used to generate community stability correction factors through the hyperbolic tangent mapping of the Simpson dominance index.

[0154] The fusion calculation submodule is used to perform a weighted geometric average of the effectiveness scores of each individual unit according to the square root of the area, and generate a comprehensive index of the ecological behavior benefits of wetland restoration.

[0155] Any combination of one or more computer-readable media may be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0156] The examples described herein are merely preferred embodiments of the invention and are not intended to limit the concept and scope of the invention. Any modifications and improvements made by those skilled in the art to the technical solutions of the invention without departing from the design concept of the invention should fall within the protection scope of the invention.

Claims

1. A method for evaluating the effectiveness of urban wetland restoration based on the frequency distribution of bird behavior, characterized in that, Includes the following steps: Step S1: Obtain remote sensing images and vegetation survey data of the urban wetland to be evaluated. Based on the degree of human intervention and the vertical structure characteristics of vegetation, divide the entire wetland area into natural restoration habitats, semi-natural-semi-artificial restoration habitats, and artificial restoration habitats. Generate a rule-based evaluation grid covering the entire area and assign a corresponding restoration type code to each grid. Step S2: Deploy survey transects or sample points in grids with different repair type codes, and collect individual bird behavior data according to preset behavior state maintenance duration threshold rules and instantaneous scanning rules. The individual bird behavior data includes at least the behavior type, effective behavior frequency, vertical spatial level of the bird when the behavior occurs, and the feeding functional group category to which it belongs. Step S3: Based on the individual bird behavior data, construct a basic frequency table with restoration type encoding, behavior type and vertical spatial hierarchy as dimensions, and extract restoration measure-behavioral response bias feature value, vertical spatial utilization niche potential energy gradient feature value and inter-group spatial niche overlap pressure feature value from the basic frequency table. Step S4: Using the extracted feature values, construct a multidimensional composite assessment model that includes the behavioral function integrity balance index, the vertical structure niche support index, and the community stability correction factor, and calculate the comprehensive index of wetland restoration ecological behavior benefits by using a weighted geometric average method based on area contribution. Step S5: Compare the comprehensive index of ecological behavior benefits of wetland restoration with the preset multi-level diagnostic threshold, and combine the retrospective diagnosis of each sub-feature value to output the wetland restoration effectiveness level judgment result and corresponding restoration management suggestions. Specifically, step S3 includes the following steps: Step S3.1: Construct the remediation measure-behavioral response bias eigenvalue. Based on standardized residuals and introducing the behavioral variance of the semi-natural remediation habitat as a regularization term, this eigenvalue is used to quantify the induction or inhibition strength of different remediation measures on specific behavioral types. The specific formula for the remediation measure-behavioral response bias eigenvalue is as follows: in, This represents the bias eigenvalue representing the remedial action-behavioral response to behavior j for remedial type i. This represents the total effective frequency of behavior j across all meshes of repair type i. This represents the theoretical expected frequency of behavior j in repair type i, under the null hypothesis that repair measures and behavior types are independent of each other. The behavior variance regularization term represents the variance of the frequency of behavior j among different grids within the semi-natural / semi-artificial restoration habitat. Step S3.2: Construct the niche potential energy gradient feature value. By calculating the sum of the square of the difference in the probability of behavior distribution between adjacent vertical layers and the information weight factor of the low utilization layer, assess the continuity of the habitat vertical structure and whether there is an ecological fault that forces the concentration of behavior. Step S3.3: Calculate the vertical spatial overlap of different dietary functional groups in each restoration habitat, and extract the maximum overlap value as the niche overlap pressure characteristic value to reflect the habitat's ability to provide niche separation space. Step S4 specifically includes the following steps: Step S4.1: Calculate the behavioral function integrity balance index, which is obtained by multiplying the weighted geometric mean of the proportions of feeding, locomotion and habitat by an exponential decay penalty factor based on the range of behavior proportions. Step S4.2: Calculate the vertical structure niche support index. The vertical structure niche support index is obtained by multiplying the niche potential energy gradient characteristic value plus one and taking the reciprocal as the vertical continuity factor, and multiplying it with the niche separation factor calculated based on the maximum overlap. Step S4.3: Calculate the community stability correction factor, which is obtained by nonlinear mapping with the reciprocal of the Simpson dominance index as the independent variable. Step S4.4: Multiply the behavioral function integrity balance index, vertical structure niche support index and community stability correction factor to obtain the individual restoration efficiency score of each restoration type. Using the proportion of the square root of the area occupied by each restoration type to the square root of the total area of ​​the whole area as the weight, perform a weighted geometric average on the individual restoration efficiency scores to obtain the comprehensive index of wetland restoration ecological behavior benefits. In step S4.1, the calculation of the behavioral function integrity balance index introduces a behavioral ecological function weight coefficient, wherein the weight of feeding behavior is set higher than that of movement behavior and habitat behavior.

2. The method according to claim 1, characterized in that, Step S1 specifically includes the following steps: Step S1.1: Obtain remote sensing images and vegetation survey data of the urban wetland to be evaluated; Step S1.2: Divide the acquired high-resolution remote sensing image into several habitat patches with land feature boundaries as the boundaries; Step S1.3: For each habitat patch, if there is no direct human modification and the degree of human intervention is lower than the preset threshold, it is coded as a natural restoration habitat; if it is located at the water-land interface and has undergone native plant replanting or micro-topography modification, it is coded as a semi-natural-semi-artificial restoration habitat; if the human intervention exceeds the threshold or includes regular lawns, hard revetments and single vegetation zones, it is coded as an artificial restoration habitat. Step S1.4: Use the Geographic Information System fishing net tool to generate a rule evaluation grid covering the entire area, and assign the restoration type code of the habitat patch where the center point of each grid is located to the grid as its habitat restoration attribute label.

3. The method according to claim 2, characterized in that, Step S2 specifically includes the following steps: Step S2.1: Deploy survey transects or fixed radius sampling points within grids with different habitat restoration attribute labels to ensure that the total length of transects and the number of sampling points within each type of restoration habitat are not less than the preset value. Step S2.2: Define behavior coding rules: record pecking and hunting as feeding behavior, flying, swimming and jumping as movement behavior, and resting, preening and singing as resting behavior. Step S2.3: For trackable individuals, a behavior switch or a single behavior continuously exceeding the threshold is counted as one valid frequency; for flocking birds, the number of individuals of each behavior type is counted as a snapshot value using a fixed-interval instantaneous scanning method. Step S2.4: Synchronously record the vertical spatial level of the bird when the behavior occurs. The vertical spatial level is divided into the ground layer, herbaceous layer, shrub layer, lower canopy layer of trees, upper canopy layer of trees, and aerial layer according to the height above the ground. Step S2.5: Classify bird species into insectivorous, omnivorous, herbivorous, or carnivorous groups based on their diet; Step S2.6: Cross-summarize the data according to the three dimensions of repair type coding, behavior type coding, and vertical hierarchy coding to form a basic frequency table.

4. The method according to claim 3, characterized in that, Step S5 specifically includes the following steps: Step S5.1: Compare the comprehensive index of wetland restoration ecological behavior benefits with the preset excellent threshold, good threshold and general threshold level by level to determine the wetland restoration effectiveness level; Step S5.2: Perform retrospective diagnosis on the sub-feature values ​​corresponding to each restoration type: When the behavioral function integrity balance index of the artificial restoration habitat is higher than the warning threshold of the artificial area behavioral function and the community stability correction factor is lower than the warning threshold of the artificial area community stability, output a suggestion to reduce artificial lawns and replant native shrubs; when the niche potential energy gradient feature value of the semi-natural-semi-artificial restoration habitat is higher than the warning threshold of the vertical fault in the semi-natural area, output a suggestion to check the shrub layer and replant the middle layer vegetation; when the niche overlap pressure feature value of the natural restoration habitat is higher than the warning threshold of the natural area overlap, output a suggestion to increase microhabitat diversity.

5. A system for evaluating the effectiveness of urban wetland restoration based on the frequency distribution of bird behavior, used to implement the method for evaluating the effectiveness of urban wetland restoration based on the frequency distribution of bird behavior as described in any one of claims 1-4, characterized in that, include: The data acquisition and segmentation module is used to acquire remote sensing images and vegetation survey data, segment natural restoration, semi-natural-semi-artificial restoration and artificial restoration habitats, generate rule evaluation grids and assign restoration type codes; The behavioral data acquisition and preprocessing module is used to lay out transects and sample points within the grid, collect individual bird behavioral data according to preset rules, and then summarize the data after cleaning to form a basic frequency table. The feature extraction module is used to extract remediation action-behavioral response bias features, niche potential gradient features, and niche overlap pressure features from the basic frequency table; The composite index construction module is used to calculate the behavioral functional integrity balance index, vertical structure niche support index and community stability correction factor using the above feature values, and to calculate the comprehensive index of wetland restoration ecological behavior benefits by weighted geometric average. The diagnostic output module compares the comprehensive index with preset thresholds and outputs the repair effectiveness level and management suggestions.

6. The system according to claim 5, characterized in that, The behavioral data acquisition and preprocessing module includes: The behavior coding submodule is used to encode observed activities as feeding, locomotion, or resting behaviors; The frequency statistics submodule is used to count individual birds using a duration threshold method and to count flocks of birds using an instantaneous scan method. The spatial attribution submodule is used to record the vertical spatial hierarchy at the time the behavior occurs; The group classification submodule is used to classify bird species into corresponding groups based on their diet.

7. The system according to claim 6, characterized in that, The composite index construction module includes: The behavioral function assessment submodule is used to generate a behavioral function integrity balance index by using a weighted geometric mean and a range penalty factor. The spatial structure assessment submodule is used to generate a vertical structure niche support index based on potential energy gradient and overlapping pressure. The stability correction submodule is used to generate community stability correction factors through the hyperbolic tangent mapping of the Simpson dominance index. The fusion calculation submodule is used to perform a weighted geometric average of the effectiveness scores of each individual unit according to the square root of the area, and generate a comprehensive index of the ecological behavior benefits of wetland restoration.

Images