Urban coastal wetland ecological restoration comprehensive effect evaluation method and system

Abstract

The invention relates to an urban coastal wetland ecological restoration comprehensive effect evaluation method and system. The method comprises the following steps: carrying out space-time alignment and data fusion on multi-source heterogeneous data of the urban coastal wetland to generate an ecological feature-process-service fusion data set; obtaining a causal conduction path among the three types of indexes and calculating a path coefficient; constructing an Ecps chain response network and a system dynamics coupling model by combining the conduction efficiency, inputting different preset repair scene parameters into the system dynamics coupling model, generating a repair scene simulation result and calculating a repair effect comprehensive index, and determining a key repair bottleneck by combining path contribution degree analysis; and generating a comprehensive effect evaluation report based on the repair effect comprehensive index and the key repair bottleneck. The method can effectively solve the problems that a traditional evaluation method cannot reveal an internal chain response mechanism, key bottlenecks are difficult to diagnose, and repair effect evaluation and management decision are disjointed.

Description

A method and system for evaluating the comprehensive effectiveness of urban coastal wetland ecological restoration Technical Field

[0001] This invention belongs to the field of marine ecological assessment technology, and in particular relates to a method and system for evaluating the comprehensive effectiveness of ecological restoration of urban coastal wetlands. Background Technology

[0002] With the rapid urbanization along my country's coast, coastal wetland ecosystems are facing increasing pressure from human activities, making ecological restoration a crucial safeguard for regional ecological security and sustainable development. Scientifically and accurately evaluating the comprehensive effectiveness of ecological restoration is key to optimizing restoration strategies and improving management efficiency. Currently, the evaluation of coastal wetland restoration effectiveness typically relies on monitoring and comprehensively analyzing a range of ecological indicators.

[0003] Traditional technologies often focus on observing the apparent characteristics of ecosystems, such as monitoring vegetation cover changes through remote sensing or analyzing physicochemical parameters like water quality and soil through fixed-point sampling. These methods tend to treat key processes like biodiversity, ecosystem structure (characteristics), material cycling and energy flow, as well as the ecosystem services ultimately provided, such as coastal protection and carbon sequestration, as relatively independent evaluation dimensions, and calculate a comprehensive score using weighted summation or exponential overlay.

[0004] However, current evaluation methods or traditional approaches have significant limitations. First, their evaluation systems are "black boxes," only reflecting static changes in various indicators before and after restoration. They fail to reveal the inherent chain-response mechanism by which improved ecological characteristics drive the optimization of internal ecological processes and further transform into ecosystem services that benefit humans. Second, due to the lack of a quantitative understanding of the causal transmission relationship between "characteristics-processes-services," evaluation results struggle to diagnose key bottlenecks hindering the improvement of restoration effectiveness. This leads to a lack of targeted management decisions and an inability to proactively simulate and optimize restoration measures. Consequently, the evaluation of restoration effectiveness is disconnected from subsequent management decisions, making it difficult to support the needs of precise and adaptive ecological restoration management. Summary of the Invention

[0005] Based on this, it is necessary to provide an evaluation method for urban coastal wetland ecological restoration that can systematically analyze and quantify the intrinsic transmission relationship of "characteristics-process-service" in the process of ecological restoration, thereby enabling comprehensive effectiveness diagnosis and prediction from the mechanism level.

[0006] Firstly, this application provides a method for evaluating the comprehensive effectiveness of urban coastal wetland ecological restoration, including:

[0007] Spatiotemporal alignment of multi-source heterogeneous data from urban coastal wetlands is performed to obtain spatiotemporally aligned heterogeneous data. Data fusion is then performed on the spatiotemporally aligned heterogeneous data to generate an ecological feature-process-service fused dataset. The multi-source heterogeneous data includes space-based data, air-based data, and water-based data.

[0008] Based on the ecological feature-process-service fusion dataset, the causal transmission paths among ecological feature indicators, ecological process indicators, and ecological service indicators are obtained, and the path coefficients of the causal transmission paths are calculated.

[0009] Based on the causal transmission path and the corresponding path coefficients, an Ecps chain response network is generated.

[0010] Based on the Ecps chain response network, a system dynamic coupling model is constructed, and different preset repair scenario parameters are input into the system dynamic coupling model to generate repair scenario simulation results;

[0011] Based on the integrated dataset of ecological characteristics, processes and services and the results of restoration scenario simulation, the weights of each indicator are calculated, and based on each weight, a comprehensive index of restoration effectiveness is calculated.

[0012] Based on the comprehensive index of restoration effectiveness and combined with path contribution analysis, the key restoration bottlenecks that restrict the improvement of restoration effectiveness are identified, and a comprehensive effectiveness evaluation report is generated based on the comprehensive index of restoration effectiveness and the key restoration bottlenecks.

[0013] Furthermore, the method also includes:

[0014] Based on the comprehensive index of repair effectiveness, key repair bottlenecks, and a pre-set retrieval strategy knowledge base, a sequence of repair optimization strategies is generated through multi-criteria decision analysis.

[0015] When the change in the comprehensive index of repair effectiveness is detected to be greater than the preset fluctuation threshold, an early warning signal is generated, and the content of the repair optimization strategy sequence is adjusted based on the early warning signal to obtain the adjusted repair optimization strategy sequence.

[0016] Furthermore, based on the ecological feature-process-service fusion dataset, the causal transmission paths among ecological feature indicators, ecological process indicators, and ecological service indicators are obtained, and the path coefficients of the causal transmission paths are calculated, including:

[0017] Redundancy analysis was performed on the ecological feature indicators, ecological process indicators, and ecological service indicators in the ecological feature-process-service fusion dataset to obtain the core indicators;

[0018] Principal component analysis is performed on each core indicator to reduce its dimensionality, resulting in low-dimensional indicators. These low-dimensional indicators are then arranged and combined in chronological order to obtain a set of low-dimensional indicators.

[0019] Based on a low-dimensional set of indicators, conditional independence is tested to obtain causal transmission paths, and an initial directed acyclic graph is generated based on each causal transmission path; wherein the order of the causal transmission paths is a temporal sequence.

[0020] For each causal transmission path in the initial directed acyclic graph, a structural equation model is constructed, and the path coefficients of each causal transmission path are calculated using the maximum likelihood estimation method based on the structural equation model.

[0021] Furthermore, based on the causal transmission path and the corresponding path coefficients, an Ecps chain response network is generated, including:

[0022] Based on the path coefficients, a transmission strength matrix is ​​constructed. Then, based on the transmission strength matrix and the standard deviations of each indicator in the ecological feature-process-service fusion dataset, the transmission efficiency of each causal transmission path is calculated.

[0023]

[0024] in, The transmission efficiency from indicator i to indicator j, For path coefficients, Let i be the standard deviation of index i. Let j be the standard deviation of index j;

[0025] By combining the causal transmission path, path coefficient, and transmission efficiency, the Ecps chain response network is obtained.

[0026] Furthermore, based on the Ecps chain response network, a system dynamics coupled model is constructed, and different preset repair scenario parameters are input into the system dynamics coupled model to generate repair scenario simulation results, including:

[0027] The index nodes in the Ecps chain response network are mapped to state variables, and the causal transmission path is transformed into rate equations and auxiliary variables connecting the state variables.

[0028] Based on state variables, rate equations, and auxiliary variables, a basic framework for a coupled dynamics model of the system is constructed.

[0029] Based on historical monitoring data, model parameters are estimated to obtain the system dynamic coupling model parameters;

[0030] Based on the system dynamics coupling model parameters and the basic framework of the system dynamics coupling model, the system dynamics coupling model is obtained;

[0031] By combining different preset restoration scenario parameters, multiple scenarios are generated. The different restoration scenario parameters include the annual growth rate of mangrove area, the adjustment range of hydrological exchange rate, and the intensity of ecological facility construction.

[0032] The restoration scenario parameters for each scenario are input into the system dynamics coupling model to generate a dynamic evolution trajectory within a preset time period in the future; the dynamic evolution trajectory includes the dynamic trajectory of ecological characteristic indicators, the dynamic trajectory of ecological process indicators, and the dynamic trajectory of ecological service indicators.

[0033] The Monte Carlo method was used to perform uncertainty analysis on each dynamic evolution trajectory, generate confidence intervals, and combine each confidence interval with the corresponding dynamic evolution trajectory to obtain the simulation results of the repair scenario.

[0034] Furthermore, based on the ecological characteristic-process-service fusion dataset and the results of restoration scenario simulation, the weights of each indicator are calculated, and based on each weight, a comprehensive restoration effectiveness index is calculated, including:

[0035] The ecological characteristic indicators, ecological process indicators, and ecological service indicators in the ecological characteristic-process-service fusion dataset and the restoration scenario simulation results are standardized to obtain the standardized values ​​of each indicator.

[0036] Based on the standardized values, the information entropy value of each indicator is calculated, and based on the information entropy value, the dynamic weight of each indicator is calculated using the entropy weight method.

[0037] Based on the dynamic weights and standardized values ​​of each indicator, a comprehensive index of restoration effectiveness is calculated.

[0038] Secondly, this application also provides a comprehensive evaluation system for the ecological restoration of urban coastal wetlands, including:

[0039] The data fusion module is used to perform spatiotemporal alignment of multi-source heterogeneous data of urban coastal wetlands to obtain spatiotemporally aligned heterogeneous data, and to perform data fusion on the spatiotemporally aligned heterogeneous data to generate an ecological feature-process-service fused dataset; among which, the multi-source heterogeneous data includes space-based data, air-based data and water-based data;

[0040] The path generation module is used to obtain the causal transmission paths between ecological feature indicators, ecological process indicators, and ecological service indicators based on the ecological feature-process-service fusion dataset, and to calculate the path coefficients of the causal transmission paths.

[0041] The network generation module is used to generate Ecps chain response networks based on causal transmission paths and corresponding path coefficients.

[0042] The scenario simulation module is used to construct a system dynamic coupling model based on the Ecps chain response network, and input different preset repair scenario parameters into the system dynamic coupling model to generate repair scenario simulation results;

[0043] The index calculation module is used to calculate the weight of each indicator based on the ecological characteristics-process-service fusion dataset and the results of restoration scenario simulation, and to calculate the comprehensive index of restoration effectiveness based on each weight.

[0044] The report generation module is used to identify key repair bottlenecks that restrict the improvement of repair effectiveness based on the comprehensive repair effectiveness index and path contribution analysis, and to generate a comprehensive effectiveness evaluation report based on the comprehensive repair effectiveness index and key repair bottlenecks.

[0045] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores at least one instruction, at least one program, a code set, or an instruction set, and the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement any of the methods for evaluating the comprehensive effectiveness of ecological restoration of urban coastal wetlands as described in the embodiments of this application.

[0046] Fourthly, this application also provides a computer-readable storage medium storing at least one piece of program code, which is loaded and executed by a processor to implement a comprehensive evaluation method for ecological restoration of urban coastal wetlands as described in any of the embodiments of this application.

[0047] The aforementioned method and system for evaluating the comprehensive effectiveness of urban coastal wetland ecological restoration utilizes integrated multi-source heterogeneous data from the sky, air, and water. Through spatiotemporal alignment and fusion processing, a high-quality ecological characteristic-process-service fusion dataset is generated. Based on this fusion dataset, causal transmission paths among the three types of indicators are identified and path coefficients are quantified. An Ecps chain response network is constructed using transmission efficiency. Based on this network, a system dynamics coupling model is built to simulate the dynamic evolution of ecological indicators under different restoration scenarios and quantify uncertainties. By integrating current data and simulation results, indicator weights are calculated to obtain a comprehensive restoration effectiveness index. Combined with path contribution analysis, key restoration bottlenecks are located, and an evaluation report is generated. This method can dynamically and accurately quantify restoration effectiveness and reveal the transmission relationships and constraints of various indicators during the restoration process from a mechanistic perspective. It provides a scientific and comprehensive decision-making basis for optimizing urban coastal wetland ecological restoration strategies, effectively solving the problems of traditional evaluation methods failing to reveal the inherent chain response mechanism, struggling to diagnose key bottlenecks, and the disconnect between restoration effectiveness evaluation and management decisions. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying 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.

[0049] Figure 1 is a flowchart illustrating a method for evaluating the comprehensive effectiveness of ecological restoration of urban coastal wetlands in one embodiment;

[0050] Figure 2 is a flowchart illustrating the steps of obtaining the causal transmission path between ecological feature indicators, ecological process indicators, and ecological service indicators based on an ecological feature-process-service fusion dataset in one embodiment, and calculating the path coefficient of the causal transmission path.

[0051] Figure 3 is a schematic diagram of the structure of an evaluation system for the comprehensive effectiveness of ecological restoration of urban coastal wetlands in one embodiment. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0053] In one embodiment, a method for evaluating the comprehensive effectiveness of ecological restoration of urban coastal wetlands is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. As shown in Figure 1, in this embodiment, the method includes the following steps:

[0054] Step S101: Spatiotemporally align the multi-source heterogeneous data of urban coastal wetlands to obtain spatiotemporally aligned heterogeneous data, and fuse the spatiotemporally aligned heterogeneous data to generate an ecological feature-process-service fusion dataset; wherein, the multi-source heterogeneous data includes space-based data, air-based data and water-based data.

[0055] Among them, space-based data may include satellite remote sensing imagery, meteorological satellite data, etc.; air-based data may include UAV aerial imagery, lidar point cloud data, etc.; water-based data may include underwater topographic survey data, aquatic biological community monitoring data, etc.; urban coastal wetlands are a special type of wetland located in and around cities, at the junction of land and sea, possessing both wetland ecological attributes and urban spatial connections, and are an important part of the urban ecosystem; spatiotemporal alignment processing is a key step in integrating heterogeneous data from different sources and in different formats into a unified spatiotemporal framework to eliminate data misalignment in time and space dimensions.

[0056] For example, multi-source heterogeneous data such as space-based data, air-based data, and water-based data of urban coastal wetlands can be collected. These multi-source heterogeneous data can be spatiotemporally aligned using methods such as georegistration and temporal interpolation. Data from different sources and with different resolutions can be matched to a unified geographic coordinate system and time scale (such as monthly) to obtain spatiotemporally aligned heterogeneous data. Based on the logical relationship of ecosystem "feature-process-service", the spatiotemporally aligned heterogeneous data can be fused. After eliminating data noise and bias through quality control processes such as outlier removal, missing value imputation, and data standardization, the mapping relationship between ecological feature indicators, ecological process indicators, and ecological service indicators can be established to generate an ecological feature-process-service fused dataset. Georegistration is a technical process that establishes a correspondence between spatial data lacking geographic coordinate information (such as aerial imagery) and spatial data in a known geographic coordinate system, assigning the former precise geographic coordinates. Temporal interpolation is a mathematical method that calculates the corresponding values ​​of unknown time points between two or more known time points in time-series data, based on observations at known time points. A geographic coordinate system is a spatial reference frame used to locate points on the Earth's surface. Through mathematical models, it transforms irregular Earth shapes into calculable coordinates, essentially describing the location of points on the Earth's surface using latitude and longitude. An ecosystem is a unified whole formed by the interaction and interdependence of biological communities and abiotic environments such as air, water, and soil on the Earth's surface through material cycling, energy flow, and information transmission. From a small pond or a forest to the entire Earth's biosphere, all can be considered ecosystems at different scales. The logical relationship of an ecosystem—characteristics-processes-services—is that characteristics support processes, processes produce services, and the three form a progressive, interconnected relationship that collectively reflects the functional value of the ecosystem. Outlier removal is a key step in data preprocessing, aiming to identify and remove anomalous data that deviates from the overall distribution and may be caused by measurement errors or extreme, accidental events. Missing value imputation refers to filling in missing values ​​in some fields (characteristics) of a dataset using appropriate methods. Data standardization is a data preprocessing technique used to eliminate interference caused by differences in units (scales) between different indicators, making data from different dimensions comparable.

[0057] Step S102: Based on the ecological feature-process-service fusion dataset, obtain the causal transmission path between ecological feature indicators, ecological process indicators and ecological service indicators, and calculate the path coefficient of the causal transmission path.

[0058] For example, based on an ecological feature-process-service fusion dataset, redundancy analysis and principal component analysis (PCA) are first performed on the ecological feature indicators, ecological process indicators, and ecological service indicators in the dataset to reduce dimensionality. High-dimensional indicators are mapped to a low-dimensional space, principal components with eigenvalues ​​greater than 1 are extracted, and the low-dimensional indicators are arranged and combined in chronological order to obtain a low-dimensional indicator set. Based on this low-dimensional indicator set, conditional independence tests can be performed using algorithms such as PC causal discovery to identify causal transmission paths from ecological feature indicators to ecological process indicators, and then from ecological process indicators to ecological service indicators. An initial directed acyclic graph (DAG) is generated, and the path coefficients of the causal transmission paths in the initial DAG are calculated. Redundancy analysis is a key step in data preprocessing to eliminate information overlap between indicators and simplify data dimensionality. Its goal is to remove duplicate or highly similar indicators while retaining information with core research value. Principal component analysis (PCA) dimensionality reduction is a commonly used data dimensionality reduction and feature extraction method. While preserving as much of the original data's key information (variance) as possible, it maps high-dimensional data to a low-dimensional space, reducing data dimensionality and redundancy. The PC causal discovery algorithm (Peter-Clark)... Algorithm is a constraint-based causal discovery method that uses statistical tests to rule out impossible causal relationships and infers the causal structure between variables from observed data (such as a causal graph) without relying on prior causal assumptions. Conditional independence test is a statistical method used to verify whether two target variables are independent given one (or more) conditional variables. The core is to eliminate the interference of third-party variables and determine whether the association between variables is mediated or confounded by the conditional variable. For example, for any two indicators X and Y, a third-party indicator set Z (a possible confounding variable) is introduced to test whether "X and Y are still related under the condition of controlling Z". If X and Y are no longer related after controlling Z, it means that the association between the two is mediated by Z and is a spurious association; if they are still related, there is a high probability that there is a direct causal relationship.

[0059] Step S103: Generate an Ecps chain response network based on the causal transmission path and the corresponding path coefficients.

[0060] For example, based on the path coefficients of each causal transmission path, a transmission strength matrix is ​​constructed. Based on this transmission strength matrix and the standard deviations of each indicator in the ecological feature-process-service fusion dataset, the transmission efficiency of each causal transmission path is calculated. The obtained causal transmission paths, the calculated path coefficients, and the calculated transmission efficiency are combined. Indicators are used as network nodes, causal transmission paths as network connecting edges, and path coefficients and transmission efficiency as dual weight attributes of the edges, forming a network structure containing nodes, edges, and dual weights—the Ecps chain response network. Here, Ecps is an abbreviation representing three key dimensions in ecological restoration: Ec represents ecological features, referring to the relatively static and observable structural attributes of an ecosystem; P represents ecological processes, referring to the dynamic functional flows within an ecosystem, driven by the interaction of various ecological features; and S represents ecological services, referring to the benefits provided by the ecosystem to humans, the final result of the well-functioning ecological processes. The overall meaning of Ecps is "ecological features-ecological processes-ecological services."

[0061] Step S104: Based on the Ecps chain response network, construct a system dynamic coupling model, and input the preset parameters of different repair scenarios into the system dynamic coupling model to generate repair scenario simulation results.

[0062] Among them, the system dynamics coupling model is an analytical tool that integrates subsystems of multiple fields, scales or types through coupling mechanisms based on the theoretical framework of system dynamics, in order to simulate the overall behavior and interaction laws of complex systems.

[0063] For example, based on the Ecps chain response network, a basic framework for a system dynamics coupling model is constructed. This framework includes state variables, rate equations connecting each state variable, and auxiliary variables. Simultaneously, positive and negative feedback loops in the model are identified, and the model's boundaries and equation relationships are clarified. Based on historical monitoring data, such as measured data of wetland ecological indicators and environmental factor data from the past 10 years, model parameters are estimated to obtain system dynamics coupling model parameters. Combining these parameters with the basic framework of the system dynamics coupling model, a complete system dynamics coupling model is formed. Preset parameters for different restoration scenarios are input into the system dynamics coupling model, and the simulation period and step size are set. The equations in the model are numerically solved to generate dynamic evolution trajectories within future simulation periods. These trajectories include dynamic trajectories of ecological characteristic indicators, ecological process indicators, and ecological service indicators. Uncertainty analysis is then performed on each dynamic evolution trajectory to obtain confidence intervals for each indicator. These intervals reflect the impact of model parameter uncertainty on the simulation results. Combining each confidence interval with the corresponding dynamic evolution trajectory yields the restoration scenario simulation results. Among them, model parameter estimation is a core step in statistics and machine learning. The goal is to find the parameter values ​​that best fit the data from the observed data. Here, the parameters are predefined but unknown variables in the model, and estimation is the process of inferring the specific values ​​of these variables from the data. The preset parameters for different repair scenarios are parameters that are set in advance and can be directly called for different "repair goals, damage types, and application scenarios". Numerical solution is a method of approximating mathematical problems through numerical calculation (rather than rigorous mathematical derivation). It transforms problems that are difficult to express analytical solutions directly with formulas into numerical calculation processes that can be iteratively approximated by computers or manual calculations. Uncertainty analysis is the process of identifying, quantifying, and evaluating possible unknown, random, or fuzzy factors (i.e., uncertainty).

[0064] Step S105: Based on the ecological characteristics-process-service fusion dataset and the restoration scenario simulation results, calculate the weight of each indicator, and calculate the comprehensive index of restoration effectiveness based on each weight.

[0065] For example, the ecological characteristic indicators, ecological process indicators, and ecological service indicators in the ecological characteristic-process-service fusion dataset and the restoration scenario simulation results are standardized respectively. For positive and negative indicators, different methods are used for standardization to obtain the standardized values ​​of each indicator. Based on the standardized values, the information entropy value of each indicator is calculated. The smaller the information entropy, the greater the degree of variation of the indicator and the more effective information it provides. Based on the information entropy values, the dynamic weight of each indicator can be calculated using the entropy weight method. According to the dynamic weights and the standardized values ​​of each indicator, the comprehensive index of restoration effectiveness can be calculated by weighted summation. Among them, positive indicators refer to those with larger values, which represent better ecological restoration results, such as vegetation coverage, carbon sequestration capacity, and microbial degradation rate; negative indicators refer to those with smaller values, which represent better ecological restoration results, such as pollutant concentration and soil erosion degree; entropy weighting is an objective weighting method, the core idea of ​​which comes from entropy in information theory. The smaller the entropy value, the lower the information uncertainty and the more effective information the data carries, and the greater the corresponding indicator weight should be; conversely, the larger the entropy value, the higher the information disorder and the less effective information, and the smaller the indicator weight should be; weighted summation is a method of calculating multiple data. By assigning different weights (i.e., importance coefficients) to different data, and then calculating the weighted value of each data by "data × corresponding weight", the final result is obtained by adding all weighted values.

[0066] Step S106: Based on the comprehensive index of restoration effectiveness and combined with path contribution analysis, identify the key restoration bottlenecks that restrict the improvement of restoration effectiveness, and generate a comprehensive effectiveness evaluation report based on the comprehensive index of restoration effectiveness and the key restoration bottlenecks.

[0067] Among them, path contribution analysis is a core method in user behavior analysis or conversion path optimization. It quantifies the actual contribution value of different user behavior paths to the final goal (such as conversion or key operations), rather than focusing only on the most popular path.

[0068] For example, based on the comprehensive index of repair effectiveness and combined with path contribution analysis based on the Ecps chain response network, key parameters of each transmission path are extracted from the Ecps network, namely the transmission efficiency from the starting point to the midpoint of the path. The propagation efficiency from the midpoint to the endpoint of the path And calculate the change of the path starting point index ΔSource=z under the repair scenario. j (Simulation)-z j (Current situation), among which, z j (Simulation) To repair the standardized values ​​of indicators in the scenario simulation results, z j(Current Status) represents the standardized value of a centralized indicator in the integrated data set of ecological characteristics, processes, and services. For example, this change is substituted into the contribution formula to calculate the contribution of each path to the change in the comprehensive index. The contribution formula is: , As for contribution level, The higher the value, the greater the contribution of that pathway to improving the repair effectiveness. For example, the average contribution is calculated by statistically analyzing the contributions of all transmission pathways. Where k is the total number of transmission paths, Let this be the contribution of the i-th propagation path. For example, the contribution of each path is... Compared with average contribution In comparison, the marker satisfies The path is a potential bottleneck path. Combining the causal logic of the Ecps network, we trace the root cause of the problem in the potential bottleneck path and analyze whether it is due to insufficient improvement of the starting indicator (small ΔSource) or low transmission efficiency. or Based on the impact of bottleneck paths on the comprehensive index, potential bottlenecks are prioritized to identify key bottlenecks that restrict the improvement of restoration effectiveness. Finally, the relevant information of the comprehensive restoration effectiveness index and key restoration bottlenecks is integrated to generate a comprehensive effectiveness evaluation report. This report is used to clearly present the differences in effectiveness under different restoration scenarios and the core issues restricting the improvement of restoration effectiveness.

[0069] In this embodiment, a high-quality ecological feature-process-service fusion dataset is generated based on integrated multi-source heterogeneous data from the sky, air, and water. This dataset is processed through spatiotemporal alignment and fusion. A causal discovery algorithm is used to identify the causal transmission paths between the three types of indicators and quantify the path coefficients. An Ecps chain response network is constructed based on transmission efficiency. A system dynamics coupling model is built based on this network to simulate the dynamic evolution of ecological indicators under different restoration scenarios and quantify uncertainties. By fusing current data and simulation results, the entropy weight method is used to dynamically calculate indicator weights and obtain a comprehensive restoration effectiveness index. Path contribution analysis is combined to locate key restoration bottlenecks and generate an evaluation report. This approach breaks through the limitations of traditional "black box" evaluation, dynamically and accurately quantifying restoration effectiveness. It provides a scientific and comprehensive decision-making basis for optimizing urban coastal wetland ecological restoration strategies, effectively solving the problems of traditional evaluation methods failing to reveal the inherent chain response mechanism, struggling to diagnose key bottlenecks, and the disconnect between restoration effectiveness evaluation and management decisions.

[0070] In one exemplary embodiment, the method further includes:

[0071] Step S201: Based on the comprehensive index of repair effectiveness, key repair bottlenecks, and a pre-set retrieval strategy knowledge base, a repair optimization strategy sequence is generated through multi-criteria decision analysis.

[0072] The pre-defined retrieval strategy knowledge base is a structured knowledge set that is pre-constructed to improve the efficiency and accuracy of information retrieval. It contains standardized retrieval methods, logical rules, and optimization schemes. This knowledge base stores various restoration technology schemes, management measures, and related implementation parameters corresponding to different wetland types and different bottleneck problems. Multiple Criteria Decision Analysis (MCDA) is a methodology for solving complex decision problems with multiple conflicting or incomparable criteria. It is used to solve problems that require screening, ranking, or selecting the optimal solution from multiple alternatives under multiple conflicting / incomparable criteria.

[0073] For example, the determined comprehensive index of restoration effectiveness and key restoration bottlenecks are retrieved, and a pre-set retrieval strategy knowledge base is loaded. With the core decision-making objectives of maximizing the comprehensive index of restoration effectiveness, minimizing implementation costs, and minimizing ecological risks, the current restoration effect level reflected by the comprehensive index of restoration effectiveness, the targeted solution needs corresponding to key restoration bottlenecks, and the strategies in the pre-set retrieval strategy knowledge base are precisely matched to screen out candidate restoration optimization strategies that can specifically solve bottleneck problems and meet the effectiveness improvement objectives. Multi-criteria decision analysis methods such as the analytic hierarchy process (AHP) can be used to comprehensively evaluate and rank the candidate strategies. By stratifying the decision objectives, evaluation criteria, and candidate strategies, the weights of elements at each level are calculated and consistency checks are performed. Based on the weights, the various indicators of the candidate strategies are weighted and scored. An ordered sequence of restoration optimization strategies is generated based on the scoring results, clarifying the implementation priority of different strategies. Among them, the Analytic Hierarchy Process (AHP) is a system analysis and decision-making method. Its core is to break down complex, multi-objective decision problems into a clear hierarchical structure. By combining subjective judgment with objective calculation, it obtains the priority ranking of each solution. The weight calculation logic of AHP is to compare elements at the same level (such as "effectiveness / cost / risk" in the evaluation criterion layer) pairwise through expert scoring or data support, construct a judgment matrix (e.g., effectiveness is 3 times more important than cost), and then calculate the weight of each element through eigenvalue methods, such as effectiveness weight 0.5, cost 0.3, and risk 0.2. Consistency testing is an analytical method to determine whether multiple sets of data, multiple evaluation results, or multiple decision criteria are consistent and not contradictory.

[0074] Step S202: When the change value of the comprehensive index of repair effectiveness is detected to be greater than the preset fluctuation threshold, an early warning signal is generated, and the content of the repair optimization strategy sequence is adjusted based on the early warning signal to obtain the adjusted repair optimization strategy sequence.

[0075] The preset fluctuation threshold is pre-set based on historical restoration data, ecosystem stability threshold, and restoration target requirements, and is used to determine whether the comprehensive index is within the normal evolution range.

[0076] For example, the dynamic changes of the comprehensive index of restoration effectiveness are continuously monitored, and the changes in the comprehensive index within adjacent monitoring periods are compared with the preset fluctuation threshold in real time. When the changes in the comprehensive index of restoration effectiveness are found to be greater than the preset fluctuation threshold, it indicates that an abnormal situation has occurred in the restoration process, which may cause the restoration effectiveness to deviate from the expected target. An early warning signal is generated. The early warning signal contains key information such as the magnitude of the abnormal fluctuation of the comprehensive index, the time node, and related ecological indicators. Based on the early warning signal, the root cause of the abnormality is traced. Combined with the generated restoration optimization strategy sequence, it is analyzed which contents of the strategy sequence do not match the current abnormal situation or cannot cope with the root cause of the abnormality. Then, the contents of the strategy sequence are adjusted, including replacing inefficient strategies, supplementing targeted new strategies, adjusting the implementation order or parameters of strategies, etc., and finally obtaining the adjusted restoration optimization strategy sequence.

[0077] In this embodiment, based on the comprehensive index of restoration effectiveness and key restoration bottlenecks, and combined with a pre-set retrieval strategy knowledge base, a highly targeted and prioritized sequence of restoration optimization strategies is generated through multi-criteria decision analysis. Changes in the comprehensive index are monitored in real time and compared with a pre-set fluctuation threshold, generating timely early warning signals and dynamically adjusting the strategy sequence. This effectively avoids deviations in restoration effectiveness due to strategy rigidity, rapidly responds to dynamic changes in the ecosystem, ensures that restoration measures always align with actual restoration needs, and significantly improves the scientific rigor, relevance, and flexibility of ecological restoration strategies.

[0078] In one embodiment, as shown in Figure 2, based on the ecological feature-process-service fusion dataset, the causal transmission paths among ecological feature indicators, ecological process indicators, and ecological service indicators are obtained, and the path coefficients of the causal transmission paths are calculated, including:

[0079] Step S301: Redundancy analysis is performed on the ecological feature indicators, ecological process indicators, and ecological service indicators in the ecological feature-process-service fusion dataset to obtain the core indicators.

[0080] For example, the Pearson correlation coefficient method can be used to calculate the correlation between any two indicators in the integrated dataset of ecological characteristics, processes, and services, accurately quantifying the linear association strength between indicators. A reasonable correlation coefficient threshold (e.g., r ≥ 0.85) is preset, and indicators with correlation coefficients higher than this threshold are considered highly redundant. These indicators have extremely strong similarities in characterizing the state or trend of ecosystem changes. Retaining multiple such indicators increases the computational complexity of subsequent data processing and does not significantly improve analytical accuracy. Therefore, these highly redundant indicators are eliminated, retaining only those that reflect the core characteristics, processes, and services of the ecosystem and have low inter-indices. This results in a set of core indicators after redundancy screening. The Pearson correlation coefficient method is a statistical method for measuring the degree of linear correlation between two continuous variables. By quantifying the co-changing trends between variables, it determines whether they have a linear association and the strength of that association. Proposed by statistician Carl Pearson, it is denoted by the symbol r. Its underlying logic is to determine whether two variables, such as "height" and "weight," satisfy the condition that "when one variable increases, the other variable also tends to increase / decrease by a fixed proportion."

[0081] Step S302: Perform principal component analysis to reduce the dimensionality of each core indicator to obtain low-dimensional indicators, and then arrange and combine each low-dimensional indicator in chronological order to obtain a set of low-dimensional indicators.

[0082] For example, all core indicators in the core indicator set are standardized to eliminate the impact of differences in dimensions and orders of magnitude, ensuring that all indicators are at the same comparable level. The covariance matrix of the standardized core indicators is calculated, and the degree of correlation between indicators is analyzed through the covariance matrix. Then, the eigenvalues ​​and eigenvectors of the covariance matrix are solved, and a screening criterion of eigenvalues ​​greater than 1 is set. Principal components that meet this criterion are extracted. These principal components are used to retain the key information of the original core indicators to the greatest extent possible, while effectively compressing the data dimensionality. Typically, the dimensionality-reduced principal component set should retain more than 85% of the information of the original data to ensure the effectiveness of the dimensionality-reduced data. Each extracted principal component is given a physical meaning interpretation, combining the core connotations of ecological characteristics, ecological processes, and ecosystem services to assign a clear ecological meaning to each principal component, transforming it into representative low-dimensional indicators. All low-dimensional indicators are arranged and combined in chronological order to ensure the temporal consistency of the indicator data, forming a low-dimensional indicator set. The covariance matrix is ​​a symmetric square matrix, with rows and columns corresponding to the standardized core indicators. The (i,j)th element in the matrix represents the covariance between the standardized i-th indicator and the j-th indicator, which reflects the direction and strength of their linear correlation.

[0083] Step S303: Based on the low-dimensional index set, perform conditional independence test to obtain causal transmission paths, and generate an initial directed acyclic graph based on each causal transmission path; wherein, the order of the causal transmission paths is a temporal sequence relationship.

[0084] For example, after confirming that all data in the low-dimensional indicator set has been sorted according to time series and meets the requirements of the causal discovery algorithm for the temporal correlation of data, the PC causal discovery algorithm can be used to identify causal transmission paths. By setting a reasonable significance level (such as α=0.05), the correlation between indicators is judged by statistical test to determine whether the correlation is a real causal relationship rather than a false correlation. The specific process is to initialize a completely undirected graph containing all low-dimensional indicator nodes, start with the 0th-order conditional independence test, and gradually increase the dimension of the condition set. For each pair of indicator nodes, under different condition sets, it is tested whether the two satisfy conditional independence. If they satisfy it, the undirected edge between them is deleted. After multiple rounds of testing and screening, the edges with real correlation are retained. Combining the internal logic of the ecosystem and the temporal relationship of the indicator data, the direction of the edge is clarified to ensure that all causal transmission paths follow the temporal order of "ecological characteristic indicators → ecological process indicators → ecological service indicators", eliminating the possibility of reverse causal paths. All the directed edges that have been tested and confirmed are integrated to generate an initial directed acyclic graph that can clearly show the real causal relationship between the three types of indicators. Statistical testing is a scientific method that uses sample data to infer population characteristics and determine whether a research hypothesis is valid. Its core principle is based on the principle of low-probability events, meaning that low-probability events almost never occur in a single trial. Setting a reasonable significance level (e.g., α=0.05) sets a confidence threshold for statistical testing. The specific rules are: if the p-value of the test result is ≤0.05, the probability that "X and Y remain independent after controlling for Z" is ≤5%, which is a low-probability event. Therefore, the hypothesis that "the two are independent" is rejected, and a true causal relationship between X and Y is determined. If the p-value is >0.05, the possibility of "the two are independent" cannot be ruled out, and the association is determined to be a spurious correlation, requiring removal from the transmission path. The inherent logic of an ecosystem revolves around three pillars: "material cycling, energy flow, and information transmission," constructing a dynamic equilibrium system of interdependence and mutual constraint between organisms and between organisms and their abiotic environment.

[0085] Step S304: For each causal transmission path in the initial directed acyclic graph, construct a structural equation model, and calculate the path coefficients of each causal transmission path based on the structural equation model using the maximum likelihood estimation method.

[0086] Among them, the maximum likelihood estimation (MLE) method is one of the core methods of parameter estimation. The core idea is to find a set of model parameters that maximizes the probability of the observed data given the observed data.

[0087] For example, each indicator node in the initial directed acyclic graph is assigned as a latent variable or observed variable in the structural equation model. Observed variables refer to indicators that can be directly monitored and whose data can be obtained; latent variables refer to intrinsic ecosystem attributes that cannot be directly observed but can be indirectly characterized by multiple observed variables. For example, each directed edge in the graph is transformed into a path relationship between variables. Based on these variables and path relationships, the model's hypothesis equation system is defined, including measurement equations and structural equations. The measurement equations describe the representational relationship between observed and latent variables; the structural equations characterize the causal transmission relationship between latent variables. For example, the quantitative data in the low-dimensional indicator set is used as the model fitting sample. The maximum likelihood estimation method is used to solve the unknown parameters in the structural equation model to estimate the path coefficients corresponding to each causal transmission path. The absolute value of the path coefficient directly reflects the strength of the influence of the source indicator on the target indicator. The larger the absolute value, the stronger the influence. The sign of the path coefficient represents the direction of influence. A positive sign indicates a positive correlation, that is, an increase in the value of the source indicator will lead to an increase in the value of the target indicator. A negative sign indicates a negative correlation, that is, an increase in the value of the source indicator will lead to a decrease in the value of the target indicator. During the solution process, the model can be verified for goodness of fit using indicators such as root mean square error. If the goodness of fit is not up to standard, the process returns to step S303 to adjust the path structure of the initial directed acyclic graph, reconstruct the model and solve it until the model meets the statistical requirements. Finally, the quantitative path coefficients of each causal transmission path are obtained. Among them, goodness of fit verification is the core step in evaluating the degree of matching between the model prediction results and the actual observed data. The purpose is to determine whether the model can reasonably reflect the true pattern of the data and avoid overfitting (the model is too close to the training data and has poor generalization ability) or underfitting (the model fails to capture the core trend of the data).

[0088] In this embodiment, redundancy analysis is performed on the three types of indicators in the ecological characteristic-process-service fusion dataset. Highly correlated redundant indicators are eliminated, and core indicators are selected. Principal component analysis is used to reduce dimensionality to obtain a low-dimensional indicator set, which is then arranged chronologically. Based on the low-dimensional indicator set, conditional independence is tested to identify the true causal transmission paths that conform to temporal logic and generate an initial directed acyclic graph. By constructing a structural equation model and using maximum likelihood estimation, the path coefficients of each causal transmission path are calculated. This effectively solves the problems of indicator redundancy, ambiguous causal relationships, and inability to quantify the influence strength between indicators in traditional evaluation methods.

[0089] In one embodiment, an Ecps chain response network is generated based on the causal transmission path and the corresponding path coefficients, including:

[0090] Step S401: Based on the path coefficients, construct the transmission strength matrix, and based on the transmission strength matrix and the standard deviation of each indicator in the ecological feature-process-service fusion dataset, calculate the transmission efficiency of each causal transmission path:

[0091]

[0092] in, The transmission efficiency from indicator i to indicator j, For path coefficients, Let i be the standard deviation of index i. Let be the standard deviation of index j.

[0093] Among them, the standard deviation σᵢ of indicator i and the standard deviation σⱼ of indicator j are used to reflect the dispersion of the original data of indicator i and indicator j, respectively. The larger the standard deviation, the wider the fluctuation range of the indicator data and the more significant its dynamic changes in the ecosystem.

[0094] For example, a transmission strength matrix is ​​constructed based on the path coefficients of each causal transmission path. The rows and columns of the matrix correspond to the ecological characteristic indicators, ecological process indicators, and ecological service indicators in the Ecps chain response network, respectively. Each element value in the matrix is ​​directly assigned the path coefficient of the corresponding causal transmission path, thereby intuitively presenting the strength relationship of causal transmission between different indicators. The original data of each indicator in the ecological characteristic-process-service fusion dataset are extracted, the standard deviation of each indicator is calculated, and a preset transmission efficiency calculation formula is used. This formula is used to comprehensively quantify the actual effectiveness of causal transmission between indicators. The corresponding path coefficient represents the basic influence strength of indicator i on indicator j. σᵢ and σⱼ are the standard deviations of indicators i and j in the fused dataset, used to correct the basic influence strength, making the calculation results more closely reflect the actual dynamic characteristics of the indicators. By substituting the path coefficients of each causal transmission path and the standard deviations of the corresponding indicators into the formula, the transmission efficiency of each causal transmission path is calculated one by one. , The larger the value, the higher the transmission efficiency of the path from indicator i to indicator j, and the more efficient the transmission of ecological impact.

[0095] Step S402: Combine the causal transmission path, path coefficient, and transmission efficiency to obtain the Ecps chain response network.

[0096] For example, three core data categories—causal transmission paths, path coefficients, and transmission efficiencies—are linked and bound together among ecological characteristic indicators, ecological process indicators, and ecological service indicators. These indicators serve as nodes in the network, with identified causal transmission paths acting as directed edges connecting them. Path coefficients and transmission efficiencies are used as dual weight attributes for each directed edge, where the path coefficient weight represents the intensity of influence, and the transmission efficiency weight represents the effectiveness of transmission. Network visualization technology can organically combine these elements to construct a clearly structured and attribute-complete network structure, generating an Ecps chain response network that comprehensively reflects the four-dimensional relationship of "indicator node - causal path - intensity of influence - effectiveness of transmission." Network visualization technology refers to a technical system that transforms abstract network structures and data into intuitive graphics such as nodes, edges, and topology graphs. This helps people quickly understand the relationships, structural characteristics, and dynamic changes of elements in the network, lowering the cognitive threshold for complex network data.

[0097] In this embodiment, a transmission strength matrix is ​​constructed using path coefficients. Combined with the standard deviation of the central indicators in the fused dataset, the transmission efficiency of each causal transmission path is calculated. The causal transmission paths, path coefficients, and transmission efficiency are deeply integrated to construct an Ecps chain response network that possesses both logical correlation and quantitative attributes. This effectively overcomes the limitations of traditional evaluation methods that only focus on the surface correlation of indicators and lack the quantification of the underlying transmission mechanism.

[0098] In one embodiment, a system dynamic coupling model is constructed based on the Ecps chain response network, and different preset repair scenario parameters are input into the system dynamic coupling model to generate repair scenario simulation results, including:

[0099] Step S501: Map each index node in the Ecps chain response network to a state variable, and convert the causal transmission path into a rate equation and auxiliary variables connecting each state variable.

[0100] For example, all indicator nodes in the Ecps chain response network, including ecological characteristic indicators, ecological process indicators, and ecological service indicators, are mapped one by one to state variables in the system dynamics model. The physical meaning and unit of measurement of each state variable are clearly defined; for example, vegetation cover corresponds to an ecological characteristic state variable, ensuring a precise correspondence between indicator nodes and state variables. The transmission paths representing causal relationships between indicators in the network are transformed into rate equations and auxiliary variables connecting each state variable. The rate equations describe the rate of change of state variables over time, and their form and parameters are determined based on quantitative data such as path coefficients and transmission efficiency in the Ecps chain response network. Auxiliary variables include environmental factors (such as temperature and precipitation) and anthropogenic disturbance factors (such as pollution emission intensity and remediation implementation intensity) that affect changes in state variables. For example, through this mapping and transformation, the graphical causal relationships of the Ecps chain response network are transformed into mathematical variable relationships that can be calculated by the system dynamics model. Among them, mapping is a fundamental concept across fields. Its core is the correspondence between two sets or systems. That is, according to certain rules, each element or feature in one set can find a definite corresponding item in another set. Transformation refers to transforming the qualitative-quantitative relationship of "who affects whom and how strong the influence is" into the mathematical expression of the system dynamics model based on the ecological significance and quantitative relationship of the path, such as the path coefficient and transmission efficiency in the Ecps network.

[0101] Step S502: Based on state variables, rate equations, and auxiliary variables, construct the basic framework of the system dynamics coupling model.

[0102] For example, the core structure of the system dynamics coupling model is clearly defined, with state variables as the core computational objects of the model, rate equations as the core driving mechanism for changes in state variables, and auxiliary variables as adjustment factors integrated into the model system. The logical relationships between the variables are clarified to ensure that changes in state variables can accurately reflect the influence of causal transmission paths through rate equations. At the same time, the auxiliary variables are associated with the corresponding state variables or rate equations to clarify the adjustment methods and strengths of the auxiliary variables on changes in state variables. Then, the boundary conditions of the model are set, including spatial boundaries (such as the specific geographical range of the coastal wetlands of the target city) and temporal boundaries (such as the simulation start time and time step of the model). The initial state of the model is defined, that is, the initial values ​​of each state variable at the start of the simulation. These initial values ​​can be determined based on historical monitoring data or current survey data. Finally, the basic framework of the system dynamics coupling model, which includes a variable system, equation relationships, boundary conditions, and initial state, is formed.

[0103] Step S503: Based on historical monitoring data, perform model parameter estimation to obtain the system dynamic coupling model parameters.

[0104] Historical monitoring data includes measured data on ecological characteristic indicators, ecological process indicators, and ecological service indicators over a relatively long period of time (such as the past 10 years), as well as corresponding environmental factors, human activities, and other related data, which serve as the core data support for model parameter estimation.

[0105] For example, based on the basic framework of the constructed system dynamics coupling model, historical monitoring data of urban coastal wetlands can be retrieved. Advanced parameter estimation algorithms such as Markov chain Monte Carlo sampling can be used to substitute the historical monitoring data into the basic framework of the model. Through multiple iterative sampling, such as setting the number of samplings to 10,000 and discarding the first 2,000 as the burning period to eliminate the influence of initial values, the parameters to be estimated in the model, such as the coefficients in the rate equation and the influence weights of auxiliary variables, are continuously adjusted so that the error between the output results simulated by the model based on the current parameters and the historical monitoring data is gradually minimized. At the same time, the stability of the parameter estimation is verified by calculating the convergence index of the parameters (such as the potential scale reduction factor PSRF < 1.1) to ensure that the estimated parameters can truly reflect the actual evolution law of the ecosystem. Finally, a set of stable and reliable system dynamics coupling model parameters is obtained. Markov Chain Monte Carlo (MCMC) is a numerical method for sampling complex probability distributions, addressing the challenge of directly sampling high-dimensional / non-analytical distributions. Essentially, it constructs a reasonable Markov chain so that its stationary distribution equals the target sampling distribution, then uses the chain's samples to approximate the target distribution. Iterative sampling is a common method in data processing, statistical inference, and machine learning. Its core logic is to obtain more accurate samples from the target distribution through multiple iterations and gradual optimization, rather than completing the sampling all at once. Convergence metrics are crucial tools for determining the stability and reliability of Markov Chain Monte Carlo algorithms. The MCMC algorithm gradually approximates the target distribution by constructing Markov chains, and convergence means the chain has traversed the target distribution, allowing subsequent samples to be used for statistical inference (such as calculating the mean and variance). The Potential Scale Reduction Factor (PSRF) is one of the most commonly used convergence metrics. By comparing the overall volatility of multiple parallel Markov chains with the volatility within a single chain, it determines whether the chain has stabilized to the target distribution.

[0106] Step S504: Based on the system dynamics coupling model parameters and the basic framework of the system dynamics coupling model, the system dynamics coupling model is obtained.

[0107] For example, the system dynamics coupling model parameters and the constructed basic framework of the system dynamics coupling model are fused. This is achieved by assigning the estimated parameters one by one to the corresponding estimated terms in the basic framework, such as accurately substituting coefficients in the rate equation and the influence weights of auxiliary variables into the corresponding equations. Simultaneously, the fit between the parameters and the model framework is verified to ensure that the parameter values ​​conform to the logical constraints of the model and the actual situation of the ecosystem. The fused model undergoes preliminary testing to check whether it can calculate normally and whether the output results are reasonable. If abnormal model operation or results that do not conform to ecological logic are found, step S503 is re-estimated, or the structure of the basic framework is adjusted until the model can run stably and the preliminary output results are reasonable, thus forming a complete system dynamics coupling model.

[0108] Step S505: Combine the preset parameters of different restoration scenarios to generate multiple scenarios. The parameters of different restoration scenarios include the annual growth rate of mangrove area, the adjustment range of hydrological exchange rate, and the intensity of ecological facility construction.

[0109] Among them, the preset parameters for different restoration scenarios are set in advance based on the actual needs of urban coastal wetland ecological restoration and common restoration measures; the annual growth rate of mangrove area can be set to different gradients such as 0%, 5%, 10%, and 15%; the adjustment range of hydrological exchange rate can be set to different adjustment ratios such as -20%, -10%, 0, +10%, and +20%, with negative signs indicating reduction and positive signs indicating increase; the intensity of ecological facility construction can be set to three levels: low intensity, medium intensity, and high intensity, corresponding to different specific implementation parameters such as the density of ecological floating beds, the area of ​​artificial wetland construction, and the length of ecological revetments.

[0110] For example, orthogonal experimental design can be used to comprehensively combine different preset restoration scenario parameters to ensure that the generated multiple scenarios can cover the combined effects of different restoration strategies and avoid missing key scenario combinations, such as "high annual growth rate of mangrove area + high hydrological exchange rate + high-intensity ecological facility construction". Each scenario has clearly defined values ​​for each restoration scenario parameter, forming multiple structured restoration scenario schemes. Orthogonal experimental design is an efficient multi-factor experimental optimization method. By using orthogonal arrays, representative experimental points are selected from all possible combinations of comprehensive experiments. With fewer experiments, the influence of multiple factors (variables) on experimental results (indicators) can be analyzed, thereby finding the optimal experimental conditions.

[0111] Step S506: Input the restoration scenario parameters of each scenario into the system dynamic coupling model to generate the dynamic evolution trajectory within a preset time period in the future; wherein, the dynamic evolution trajectory includes the dynamic trajectory of ecological characteristic indicators, the dynamic trajectory of ecological process indicators, and the dynamic trajectory of ecological service indicators.

[0112] For example, a uniform simulation period (e.g., the next 5 years) and simulation step size (e.g., 1 month) are set for each scenario to ensure the comparability of simulation results for different scenarios. The fourth-order Runge-Kutta method can be used to solve the differential equations in the coupled dynamic model of the system. During the model operation, the numerical changes of ecological characteristic indicators, ecological process indicators, and ecological service indicators at each time step are recorded in real time. Through continuous numerical recording, the dynamic evolution trajectories of the three types of indicators over time are generated. These trajectories are presented in the form of a "time-indicator value" sequence, which is used to reflect the trend, rate, and key inflection points of the ecosystem's evolution from the current state to the future state under the influence of specific restoration scenario parameters. The fourth-order Runge-Kutta method is a classic numerical method for solving initial value problems of ordinary differential equations, which approximates the analytical solution of the differential equation through multi-step weighted averaging.

[0113] Step S507: The Monte Carlo method is used to perform uncertainty analysis on each dynamic evolution trajectory, generate confidence intervals, and combine each confidence interval with the corresponding dynamic evolution trajectory to obtain the simulation results of the repair scenario.

[0114] Among them, the Monte Carlo method is a numerical calculation method based on random sampling to solve complex problems. It simulates the probability process through a large number of random experiments and uses statistical results to approximate the exact solution of the problem.

[0115] For example, the Monte Carlo method can be used to perform uncertainty analysis on each dynamic evolution trajectory. First, the key input parameters (such as the mangrove growth rate coefficient) in the system dynamic coupling model are randomly perturbed. Based on the posterior distribution of the parameters, 1000 sets of perturbed parameter sets are generated. Each set of perturbed parameters is substituted into the system dynamic coupling model, the simulation is rerun, and the corresponding ecological index simulation results are output, resulting in a dataset of 1000 sets of dynamic evolution trajectories under different parameter perturbations. Statistical analysis is performed on the 1000 simulated values ​​of each index at each time step, and a 95% confidence interval can be calculated. This interval can intuitively reflect the uncertainty range of the simulation results. The core dynamic evolution trajectory of each index (usually the mean trajectory of 1000 simulations) and the corresponding 95% confidence interval are combined to form a restoration scenario simulation result that includes the dynamic evolution trend and uncertainty range. Random disturbances refer to small disturbances or fluctuations in the system that cannot be accurately predicted in advance and have randomness. These disturbances will break the original ideal stable state of the system, causing its actual behavior to deviate from the theoretical expectation. Statistical analysis refers to sorting 1,000 simulated values ​​of the same indicator and the same time step by size and determining the 95% confidence interval by using the quantile method: taking the 25th value after sorting (2.5% quantile) as the lower limit of the interval and the 975th value (97.5% quantile) as the upper limit of the interval, ensuring that the interval can cover 95% of the simulation results and only excluding 5% of extreme values.

[0116] In this embodiment, by transforming the index nodes and causal transmission paths of the Ecps chain response network into state variables, rate equations, and auxiliary variables of a system dynamics coupled model, a basic model framework is constructed, and parameter estimation is completed based on historical monitoring data to form a reliable and complete model. Multiple scenario schemes covering different combinations of remediation strategies are generated through orthogonal experimental design, and the dynamic evolution trajectories of three types of indicators are obtained by inputting them into the model. The Monte Carlo method is used to quantify uncertainty and generate simulation results with confidence intervals. This approach preserves the inherent causal logic of "feature-process-service" and improves the comprehensiveness and reliability of the results, enabling proactive prediction of the implementation effects of different remediation strategies.

[0117] In one embodiment, based on the ecological feature-process-service fusion dataset and the results of restoration scenario simulation, the weights of each indicator are calculated, and based on each weight, a comprehensive restoration effectiveness index is calculated, including:

[0118] Step S601: Standardize the ecological feature indicators, ecological process indicators, and ecological service indicators in the ecological feature-process-service fusion dataset and the restoration scenario simulation results to obtain standardized values ​​for each indicator.

[0119] For example, by retrieving the fusion dataset of ecological characteristics, processes, and services and the results of restoration scenario simulations, the specific values ​​of the ecological characteristic indicators, ecological process indicators, and ecological service indicators contained in the two types of data are identified. These indicators are then classified and identified, distinguishing between positive and negative indicators. For positive indicators, a formula can be used. For negative indicators, the calculation can be performed using a formula. Calculation. Among them, Let j be the standardized value of the i-th indicator in the i-th sample. Let j be the original value of the j-th indicator in the i-th sample. Let j be the minimum value of the j-th indicator among all samples. Let j be the maximum value of the j-th indicator in all samples. Through this standardization process, the indicator values ​​are uniformly transformed to the interval [0,1], and finally a set of standardized values ​​including all samples and all indicators is obtained.

[0120] Step S602: Based on each standardized value, calculate the information entropy value of each indicator, and based on each information entropy value, calculate the dynamic weight of each indicator using the entropy weight method.

[0121] For example, based on the standardized values ​​of each indicator, the proportion of each sample under the corresponding indicator is calculated using the following formula: ,in, Here, denoted by , and , where n is the total number of samples, the proportion reflects the percentage of the j-th indicator in the i-th sample among all samples for that indicator. =0 indicates that the standardized value of this indicator is 0 in all samples. Therefore, let... To ensure the rationality of the calculation, the weight of each indicator is set as an example. Substitute into the information entropy formula Calculate the information entropy value for each indicator. This information entropy value is an important indicator for measuring the uncertainty of indicator information. The smaller the value, the greater the variability of the indicator, the more effective discriminative information it contains, and the higher its contribution to the evaluation of restoration effectiveness; conversely, the larger the value, the greater the variability. The larger the entropy value, the smaller the variability of the indicator, the less effective information, and the lower the contribution. After obtaining the information entropy value of each indicator, the dynamic weight of each indicator can be calculated using the entropy weight method. This is done first using the formula... Calculate the coefficient of variation for the j-th indicator. This difference coefficient is used to reflect the distinguishing ability of the indicator. The larger the value, the stronger the indicator's ability to distinguish sample differences. This is then further demonstrated by the formula... Calculate the weight of the j-th indicator Where m is the total number of indicators. The sum of the difference coefficients of all indicators is used to normalize the weights, ensuring that the sum of the weights of all indicators is 1.

[0122] Step S603: Calculate the comprehensive index of repair effectiveness based on the dynamic weights and standardized values ​​of each indicator.

[0123] For example, the standardized value of each indicator is multiplied by its corresponding dynamic weight, and all the product results are summed to obtain the final comprehensive index of restoration effectiveness. In the calculation process, the comprehensive index is calculated separately for different evaluation objects. For the current status sample in the ecological feature-process-service fusion dataset, the standardized values ​​of each indicator and the dynamic weight corresponding to the sample are multiplied and summed to obtain the comprehensive index of restoration effectiveness for the current status. For each scenario sample in the restoration scenario simulation results, the standardized values ​​of each indicator under the scenario are multiplied and summed to obtain the comprehensive index of restoration effectiveness for the corresponding scenario, which is used to evaluate the expected effects of different restoration scenarios. The value range of the comprehensive index of restoration effectiveness is [0,1]. The closer the value is to 1, the better the restoration effectiveness, and the closer it is to 0, the worse the restoration effectiveness.

[0124] In this embodiment, the three types of indicators in the ecological characteristic-process-service fusion dataset and the restoration scenario simulation results are standardized. The information entropy value of each indicator is calculated based on the standardized values. Dynamic weights are objectively assigned to each indicator using the entropy weighting method, and a weighted summation method is used to calculate the comprehensive restoration effectiveness index. This effectively solves the problems of inconsistent indicator dimensions, subjective weight setting, and ambiguous evaluation results in traditional evaluation methods, comprehensively and objectively reflecting the differences in effectiveness and the current restoration level under different restoration scenarios.

[0125] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0126] Based on the same inventive concept, this application also provides an urban coastal wetland ecological restoration comprehensive effectiveness evaluation system for implementing the aforementioned method for evaluating the comprehensive effectiveness of urban coastal wetland ecological restoration. The solution provided by this system is similar to the implementation scheme described in the above method. Therefore, the specific limitations of one or more embodiments of the urban coastal wetland ecological restoration comprehensive effectiveness evaluation system provided below can be found in the limitations of the urban coastal wetland ecological restoration comprehensive effectiveness evaluation method described above, and will not be repeated here.

[0127] In an exemplary embodiment, as shown in Figure 3, a comprehensive evaluation system 300 for the ecological restoration of urban coastal wetlands is provided, comprising:

[0128] The data fusion module 301 is used to perform spatiotemporal alignment on multi-source heterogeneous data of urban coastal wetlands to obtain spatiotemporally aligned heterogeneous data, and to perform data fusion on the spatiotemporally aligned heterogeneous data to generate an ecological feature-process-service fusion dataset; wherein, the multi-source heterogeneous data includes space-based data, air-based data and water-based data;

[0129] The path generation module 302 is used to obtain the causal transmission paths between ecological feature indicators, ecological process indicators and ecological service indicators based on the ecological feature-process-service fusion dataset, and to calculate the path coefficients of the causal transmission paths.

[0130] The network generation module 303 is used to generate an Ecps chain response network based on the causal transmission path and the corresponding path coefficients.

[0131] The scenario simulation module 304 is used to construct a system dynamic coupling model based on the Ecps chain response network, and input different preset repair scenario parameters into the system dynamic coupling model to generate repair scenario simulation results.

[0132] The index calculation module 305 is used to calculate the weight of each indicator based on the ecological characteristics-process-service fusion dataset and the restoration scenario simulation results, and to calculate the comprehensive restoration effectiveness index based on each weight.

[0133] The report generation module 306 is used to identify key repair bottlenecks that restrict the improvement of repair effectiveness based on the comprehensive repair effectiveness index and path contribution analysis, and to generate a comprehensive effectiveness evaluation report based on the comprehensive repair effectiveness index and key repair bottlenecks.

[0134] In one exemplary embodiment, the system further includes:

[0135] The strategy sequence generation module is used to generate a sequence of repair optimization strategies based on the comprehensive index of repair effectiveness, key repair bottlenecks, and a preset retrieval strategy knowledge base, through multi-criteria decision analysis.

[0136] The strategy sequence adjustment module is used to generate an early warning signal when the change value of the comprehensive index of repair effectiveness is greater than the preset fluctuation threshold, and to adjust the content of the repair optimization strategy sequence based on the early warning signal to obtain the adjusted repair optimization strategy sequence.

[0137] In one embodiment, the path generation module 302 is further configured to:

[0138] Redundancy analysis was performed on the ecological feature indicators, ecological process indicators, and ecological service indicators in the ecological feature-process-service fusion dataset to obtain the core indicators;

[0139] Principal component analysis is performed on each core indicator to reduce its dimensionality, resulting in low-dimensional indicators. These low-dimensional indicators are then arranged and combined in chronological order to obtain a set of low-dimensional indicators.

[0140] Based on a low-dimensional set of indicators, conditional independence is tested to obtain causal transmission paths, and an initial directed acyclic graph is generated based on each causal transmission path; wherein the order of the causal transmission paths is a temporal sequence.

[0141] For each causal transmission path in the initial directed acyclic graph, a structural equation model is constructed, and the path coefficients of each causal transmission path are calculated using the maximum likelihood estimation method based on the structural equation model.

[0142] In one embodiment, the network generation module 303 is further configured to:

[0143] Based on the path coefficients, a transmission strength matrix is ​​constructed. Then, based on the transmission strength matrix and the standard deviations of each indicator in the ecological feature-process-service fusion dataset, the transmission efficiency of each causal transmission path is calculated.

[0144]

[0145] in, The transmission efficiency from indicator i to indicator j, For path coefficients, Let i be the standard deviation of index i. Let j be the standard deviation of index j;

[0146] By combining the causal transmission path, path coefficient, and transmission efficiency, the Ecps chain response network is obtained.

[0147] In one embodiment, the scenario simulation module 304 is further configured to:

[0148] The index nodes in the Ecps chain response network are mapped to state variables, and the causal transmission path is transformed into rate equations and auxiliary variables connecting the state variables.

[0149] Based on state variables, rate equations, and auxiliary variables, a basic framework for a coupled dynamics model of the system is constructed.

[0150] Based on historical monitoring data, model parameters are estimated to obtain the system dynamic coupling model parameters;

[0151] Based on the system dynamics coupling model parameters and the basic framework of the system dynamics coupling model, the system dynamics coupling model is obtained;

[0152] By combining different preset restoration scenario parameters, multiple scenarios are generated. The different restoration scenario parameters include the annual growth rate of mangrove area, the adjustment range of hydrological exchange rate, and the intensity of ecological facility construction.

[0153] The restoration scenario parameters for each scenario are input into the system dynamics coupling model to generate a dynamic evolution trajectory within a preset time period in the future; the dynamic evolution trajectory includes the dynamic trajectory of ecological characteristic indicators, the dynamic trajectory of ecological process indicators, and the dynamic trajectory of ecological service indicators.

[0154] The Monte Carlo method was used to perform uncertainty analysis on each dynamic evolution trajectory, generate confidence intervals, and combine each confidence interval with the corresponding dynamic evolution trajectory to obtain the simulation results of the repair scenario.

[0155] In one embodiment, the exponent calculation module 305 is further configured to:

[0156] The ecological characteristic indicators, ecological process indicators, and ecological service indicators in the ecological characteristic-process-service fusion dataset and the restoration scenario simulation results are standardized to obtain the standardized values ​​of each indicator.

[0157] Based on the standardized values, the information entropy value of each indicator is calculated, and based on the information entropy value, the dynamic weight of each indicator is calculated using the entropy weight method.

[0158] Based on the dynamic weights and standardized values ​​of each indicator, a comprehensive index of restoration effectiveness is calculated.

[0159] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the comprehensive evaluation method for ecological restoration of urban coastal wetlands as described above.

[0160] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0161] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0162] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for evaluating the comprehensive effectiveness of ecological restoration of urban coastal wetlands, characterized in that, The method includes: spatiotemporally aligning multi-source heterogeneous data of urban coastal wetlands to obtain spatiotemporally aligned heterogeneous data; fusing the spatiotemporally aligned heterogeneous data to generate an ecological feature-process-service fusion dataset; wherein the multi-source heterogeneous data includes space-based data, air-based data, and water-based data; based on the ecological feature-process-service fusion dataset, obtaining causal transmission paths between ecological feature indicators, ecological process indicators, and ecological service indicators, and calculating the path coefficients of the causal transmission paths; generating an Ecps chain response network based on the causal transmission paths and the corresponding path coefficients; constructing a system dynamics coupling model based on the Ecps chain response network, and inputting preset different restoration scenario parameters into the system dynamics coupling model to generate restoration scenario simulation results; calculating the weights of each indicator based on the ecological feature-process-service fusion dataset and the restoration scenario simulation results, and calculating a comprehensive restoration effectiveness index based on each weight; based on the comprehensive restoration effectiveness index, combined with path contribution analysis, identifying key restoration bottlenecks restricting the improvement of restoration effectiveness, and generating a comprehensive effectiveness evaluation report based on the comprehensive restoration effectiveness index and the key restoration bottlenecks.

2. The method according to claim 1, characterized in that, The method further includes: generating a repair optimization strategy sequence based on the comprehensive repair effectiveness index, the key repair bottlenecks, and a preset retrieval strategy knowledge base through multi-criteria decision analysis; generating an early warning signal when the change value of the comprehensive repair effectiveness index is detected to be greater than a preset fluctuation threshold, and adjusting the content of the repair optimization strategy sequence based on the early warning signal to obtain an adjusted repair optimization strategy sequence.

3. The method according to claim 1, characterized in that, The process of obtaining causal transmission paths among ecological feature indicators, ecological process indicators, and ecological service indicators based on the ecological feature-process-service fusion dataset, and calculating the path coefficients of the causal transmission paths, includes: performing redundancy analysis on the ecological feature indicators, ecological process indicators, and ecological service indicators in the ecological feature-process-service fusion dataset to obtain core indicators; performing principal component analysis to reduce the dimensionality of each core indicator to obtain low-dimensional indicators, and arranging and combining each low-dimensional indicator in chronological order to obtain a low-dimensional indicator set; performing conditional independence tests based on the low-dimensional indicator set to obtain causal transmission paths, and generating an initial directed acyclic graph based on each causal transmission path; wherein the order of the causal transmission paths is a chronological sequence; constructing a structural equation model for each causal transmission path in the initial directed acyclic graph, and calculating the path coefficients of each causal transmission path using the maximum likelihood estimation method based on the structural equation model.

4. The method according to claim 1, characterized in that, The step of generating an Ecps chain response network based on the causal transmission path and the corresponding path coefficients includes: constructing a transmission strength matrix based on the path coefficients, and calculating the transmission efficiency of each causal transmission path based on the transmission strength matrix and the standard deviation of each indicator in the ecological feature-process-service fusion dataset. in, The transmission efficiency from indicator i to indicator j, For path coefficients, Let i be the standard deviation of index i. Let j be the standard deviation of index j; combine the causal transmission path, the path coefficient, and the transmission efficiency to obtain the Ecps chain response network.

5. The method according to claim 1, characterized in that, The process involves constructing a system dynamics coupling model based on the ECPS chain response network, inputting preset repair scenario parameters into the system dynamics coupling model, and generating repair scenario simulation results. This includes: mapping each indicator node in the ECPS chain response network to state variables, and converting the causal transmission path into rate equations and auxiliary variables connecting each state variable; constructing a basic framework for the system dynamics coupling model based on the state variables, rate equations, and auxiliary variables; estimating model parameters based on historical monitoring data to obtain system dynamics coupling model parameters; and obtaining system dynamics coupling model parameters and the basic framework of the system dynamics coupling model based on the system dynamics coupling model parameters. A dynamic coupling model is used to generate multiple scenarios by combining preset parameters for different restoration scenarios. These parameters include the annual growth rate of mangrove area, the adjustment range of hydrological exchange rate, and the intensity of ecological facility construction. The restoration scenario parameters of each scenario are input into the system dynamic coupling model to generate dynamic evolution trajectories over a preset future time period. These dynamic evolution trajectories include dynamic trajectories of ecological characteristic indicators, ecological process indicators, and ecological service indicators. The Monte Carlo method is used to perform uncertainty analysis on each dynamic evolution trajectory to generate confidence intervals. The confidence intervals and the corresponding dynamic evolution trajectories are then combined to obtain the simulation results of the restoration scenarios.

6. The method according to claim 1, characterized in that, The process of calculating the weights of each indicator based on the fusion dataset of ecological features, processes, and services and the results of the restoration scenario simulation, and then calculating a comprehensive restoration effectiveness index based on these weights, includes: standardizing the ecological feature indicators, ecological process indicators, and ecological service indicators in the fusion dataset of ecological features, processes, and services and the results of the restoration scenario simulation to obtain standardized values ​​for each indicator; calculating the information entropy value of each indicator based on the standardized values; calculating the dynamic weight of each indicator using the entropy weight method based on the information entropy values; and calculating the comprehensive restoration effectiveness index based on the dynamic weights and the standardized values ​​of each indicator.

7. A comprehensive evaluation system for the ecological restoration of urban coastal wetlands, characterized in that, The system includes: a data fusion module, used to perform spatiotemporal alignment on multi-source heterogeneous data of urban coastal wetlands to obtain spatiotemporally aligned heterogeneous data, and to perform data fusion on the spatiotemporally aligned heterogeneous data to generate an ecological feature-process-service fused dataset; wherein the multi-source heterogeneous data includes space-based data, air-based data, and water-based data; a path generation module, used to obtain causal transmission paths between ecological feature indicators, ecological process indicators, and ecological service indicators based on the ecological feature-process-service fused dataset, and to calculate the path coefficients of the causal transmission paths; and a network generation module, used to generate an Ecps chain response based on the causal transmission paths and the corresponding path coefficients. The system includes: a network; a scenario simulation module for constructing a system dynamics coupling model based on the Ecps chain response network, inputting preset parameters of different restoration scenarios into the system dynamics coupling model, and generating restoration scenario simulation results; an index calculation module for calculating the weights of each indicator based on the ecological feature-process-service fusion dataset and the restoration scenario simulation results, and calculating a comprehensive restoration effectiveness index based on each weight; and a report generation module for identifying key restoration bottlenecks that restrict the improvement of restoration effectiveness based on the comprehensive restoration effectiveness index and path contribution analysis, and generating a comprehensive effectiveness evaluation report based on the comprehensive restoration effectiveness index and the key restoration bottlenecks.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images