Optimization layout method of low impact development measures combined with dynamic landscape hydrology theory

Abstract

The application belongs to the field of environmental engineering and water pollution control technology, and specifically relates to a method for optimizing the layout of low-impact development measures by integrating dynamic landscape hydrology theory. The method includes constructing a hydrological connectivity-land use coupling index to dynamically quantify the pollution transport potential of the landscape; objectively identifying the ecological critical threshold of the index based on machine learning and gradient analysis; establishing a multi-objective optimization model targeting hydrological benefits and costs, with the threshold as a safety constraint; in the optimization solution, dynamically simulating the underlying surface and hydrological connectivity after the implementation of each candidate scheme, recalculating the index value to verify whether the constraint is met; and finally outputting the Pareto optimal solution set that meets the constraint as the optimized layout scheme. The application achieves positive effects such as improving the ecological rationality of the scheme, ensuring long-term safety, and realizing automated and efficient decision-making.

Description

Technical Field

[0001] This application belongs to the field of environmental engineering and water pollution control technology, specifically relating to an optimized layout method for low-impact development measures that integrates dynamic landscape hydrology theory. Background Technology

[0002] In the field of urban stormwater management and non-point source pollution control, the scientific and rational spatial layout of Low Impact Development (LID) measures is a core technical issue in achieving the goals of sponge city construction. Early planning and layout relied heavily on engineers' experience, lacking quantitative basis. Currently, the "simulation-optimization" paradigm, centered on hydrological and hydraulic modeling (such as SWMM) simulation combined with multi-objective optimization algorithms that prioritize both cost and engineering benefits, has become the mainstream technology in this field. This technology improves planning efficiency by simulating the hydrological responses of different layout schemes and automatically searching for the optimal solution using optimization algorithms.

[0003] However, this technology has significant drawbacks that restrict the ecological rationality and long-term safety of the layout scheme. First, its optimization model is disconnected from the dynamic coupling mechanism of landscape hydrology, treating landscape patterns as static background parameters and failing to consider the spatial configuration of the landscape and the resulting hydrological connectivity network as optimizable intrinsic variables. Therefore, it cannot quantify and proactively regulate the dynamic impact of LIDs layout on the entire process of pollutant migration "from source to sink." Second, its optimization process lacks objectively quantified safety boundaries based on landscape ecological mechanisms; constraints largely rely on subjective experience, making it difficult to ensure that the implementation of the scheme does not disrupt the inherent hydrological and ecological safety pattern of the region. Finally, during the iterative search process of the optimization algorithm, existing technology struggles to perform real-time, automated dynamic verification and constraint judgment on the complex landscape hydrological indicators after the implementation of each candidate scheme, potentially causing the optimization search to deviate from the ecologically safe zone.

[0004] In summary, existing technologies have failed to effectively integrate the dynamic coupling mechanism of landscape hydrology. They lack objective and quantifiable landscape ecological security boundaries as optimization constraints and are difficult to verify complex ecological constraints in real time and automatically, which greatly limits their promotion and application in practical engineering. Summary of the Invention

[0005] This application proposes an optimized layout method for low-impact development measures that integrates dynamic landscape hydrology theory. It aims to solve the problems in existing "simulation-optimization" technical solutions, such as the disconnect between the optimization model and the dynamic coupling mechanism of landscape hydrology, the lack of objective and quantifiable landscape ecological security boundaries, and the difficulty in achieving real-time automated verification of complex ecological constraints.

[0006] The first aspect of this application provides an optimized layout method for low-impact development measures that integrates dynamic landscape hydrology theory, including: A hydrological connectivity-land use coupling index is constructed and calculated to dynamically quantify the pollution transport potential of regional landscapes. The index is coupled with a weighted hydrological connectivity parameter reflecting the criticality of hydrological pathways, a weighted parameter reflecting the pollution source-sink intensity of land use types, and a landscape pattern index parameter reflecting the clustering morphology of dominant patches. Based on machine learning and gradient analysis, the ecological critical threshold of the hydrological connectivity-land use coupling index is objectively identified, including: training a pollution load prediction model using a dataset containing the index and a traditional landscape pattern index, calculating the partial dependency curve of the index in the model, and determining the index value corresponding to the maximum gradient value as the ecological critical threshold by calculating the first-order discrete gradient of the partial dependency curve. For the candidate layout schemes of the aforementioned low-impact development measures, a multi-objective optimization model is established with hydrological benefits and costs as optimization objectives and the aforementioned ecological critical threshold as a landscape ecological security constraint. The landscape ecological security constraint means that after the implementation of any candidate layout scheme, its corresponding hydrological connectivity-land use coupling index value must not be greater than the aforementioned ecological critical threshold. The hydrological benefits refer to the comprehensive reduction rate of runoff and pollutants. In the process of solving the multi-objective optimization model, for each candidate layout scheme, the hydrological connectivity-land use coupling index value after the implementation of the scheme is recalculated by dynamically simulating the underlying surface and hydrological connectivity after the scheme is implemented, so as to verify whether it meets the constraints. Output the Pareto optimal solution set that satisfies the landscape ecological security constraints, as an optimized layout scheme for low-impact development measures.

[0007] In some embodiments of this application, prior to constructing and calculating the hydrological connectivity-land use coupling index, a dedicated spatiotemporal database for dynamic coupling analysis of landscape patterns and hydrological processes is constructed. The construction of this dedicated spatiotemporal database includes: Multi-source spatiotemporal data are organized using hydrological response units as basic spatial objects, and the land use attributes, topographic features and associated hydrological parameters of each hydrological response unit are pre-stored. Explicitly construct and store the topological relationships of the hydrological connectivity network between each hydrological response unit and the corresponding path attributes; Configure a scenario simulation interface to dynamically modify the attributes of the affected hydrological response units based on the input candidate layout schemes, and cascade update the associated hydrological connectivity parameters to generate post-planning scenario data.

[0008] In some embodiments of this application, the hydrological connectivity-land use coupling index is calculated as follows: , in, Indicates the sub-watershed area The aforementioned hydrological connectivity-land use coupling index value, This indicates a reflection of the sub-basin watershed. The weighted hydrological connectivity parameter is crucial for hydrological pathways. Indicates the type of land use The weighting parameters for pollution source-sink intensity, Indicates land use type The landscape pattern index parameters mentioned above. Indicated for eliminating sub-catchment areas area The function of the resulting scaling effect.

[0009] In some embodiments of this application, the landscape pattern index parameter is the maximum patch index.

[0010] In some embodiments of this application, the weighted hydrological connectivity parameter The following formula is used to calculate: , in, Represents nodes calculated based on directed hydrological connectivity networks. (Zihuishui District) Improved betweenness centrality of the export (of the export), This represents the slave node calculated based on the estimated flow rate and path length. To the node The path efficiency weight.

[0011] In some embodiments of this application, the method for determining the weighting parameters reflecting the intensity of pollution sources and sinks of land use types includes: Based on synchronous monitoring data from multiple natural sub-catchments with uniform land use within the study area, a predictive model of pollution load per unit area based on landscape characteristics was established using multiple linear regression. The landscape feature values ​​of pure pixels within each land use type are input into the prediction model to calculate the average characteristic pollution load of each type of land use. The mean of the characteristic pollution load is centered and standardized to obtain the weight parameter, wherein the weight parameter is a dimensionless value, the sign of which is used to indicate the "source" or "sink" attribute, and the absolute value reflects its relative intensity.

[0012] In some embodiments of this application, the pollution load prediction model is trained using a random forest regression algorithm.

[0013] In some embodiments of this application, the multi-objective optimization model is solved using a multi-objective genetic algorithm.

[0014] In some embodiments of this application, the cost is the total lifecycle cost of the low-impact development measure.

[0015] In some embodiments of this application, the steps of constructing and calculating the hydrological connectivity-land use coupling index, objectively identifying ecological critical thresholds, establishing and solving multi-objective optimization models, and dynamically simulating and verifying landscape ecological security constraints are integrated into an end-to-end automated decision-making process through programmed scripts.

[0016] In summary, the low-impact development (LID) optimization layout method provided in the embodiments of this application, which integrates dynamic landscape hydrology theory, constructs and calculates a hydrological connectivity-land use coupling index that couples hydrological path criticality, land use "source / sink" intensity, and dominant patch morphology. This index transforms the spatial configuration of the landscape and the resulting hydrological connectivity network from static background parameters into quantifiable dynamic decision variables, thereby characterizing the dynamic impact of LID layout on the entire process of pollutant migration "from source to sink," thus addressing the deficiency of the optimization model being disconnected from the dynamic coupling mechanism of landscape hydrology. Furthermore, through partial dependency analysis and first-order discrete gradient calculation based on machine learning models (such as random forest regression), the ecological critical threshold of the HCI-LU index is objectively identified. This method automatically and objectively derives key critical values ​​indicating abrupt changes in hydrological and ecological states from monitoring data, providing a quantifiable safety boundary derived from landscape mechanisms for the optimization process, thus solving the problem of the lack of objectively quantifiable landscape ecological factors in the optimization process. The shortcomings of safety boundaries are addressed by incorporating the ecological critical threshold as a necessary landscape ecological safety constraint when establishing a multi-objective optimization model. During the optimization iteration, for each candidate scheme, the scenario simulation interface of a dedicated spatiotemporal database is invoked to dynamically simulate the underlying surface and hydrological connectivity after the scheme's implementation, and the HCI-LU index value is recalculated for real-time verification. This mechanism transforms time-consuming offline spatial analysis and manual verification into efficient online automated calculation, enabling complex ecological constraints to be enforced in real-time and in a closed loop during each iteration of the optimization algorithm. This solves the problem of difficulty in achieving real-time, automated verification of complex ecological constraints during optimization iterations. Furthermore, by synergistically integrating the core technical features of index construction, threshold identification, optimization modeling, and dynamic verification, and combining them into an end-to-end automated decision-making process through programmed scripts, a significant improvement in the generation efficiency, scientific rigor, and repeatability of LIDs planning schemes is achieved. Attached Figure Description

[0017] The features and advantages of this application will become clearer with reference to the accompanying drawings, which are illustrative and should not be construed as limiting the application in any way. In the drawings: Figure 1 This is a flowchart illustrating an optimized layout method for low-impact development measures that integrates dynamic landscape hydrology theory, according to some embodiments of this application. Detailed Implementation

[0018] In the following detailed description, numerous specific details of this application are illustrated by example to provide a thorough understanding of the relevant disclosure. However, it will be apparent to those skilled in the art that this application can be practiced without these details. It should be understood that the terms “system,” “apparatus,” “unit,” and / or “module” used in this application are one way of distinguishing different parts, elements, sections, or components at different levels in a sequential arrangement. However, these terms may be replaced with other expressions if other expressions can achieve the same purpose.

[0019] It should be understood that when a device, unit, or module is referred to as being "on," "connected to," or "coupled to" another device, unit, or module, it may be directly connected to or coupled to, or communicate with, other devices, units, or modules, or there may be intermediate devices, units, or modules present, unless the context explicitly indicates otherwise. For example, the term "and / or" as used herein includes any one and all combinations of one or more of the relevant listed items.

[0020] The terminology used in this application is for the purpose of describing specific embodiments only and is not intended to limit the scope of this application. As shown in the specification and claims of this application, unless the context clearly indicates otherwise, words such as "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate that explicitly identified features, integrals, steps, operations, elements, and / or components are included, and such expressions do not constitute an exclusive list, and other features, integrals, steps, operations, elements, and / or components may also be included.

[0021] Referring to the following description and accompanying drawings, these and other features and characteristics, operating methods, functions of related structural elements, combinations of parts, and economics of manufacture of this application can be better understood, wherein the description and drawings form part of the specification. However, it is clearly understood that the drawings are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. It is understood that the drawings are not drawn to scale.

[0022] Various structural diagrams are used in this application to illustrate various variations of the embodiments according to this application. It should be understood that the preceding or following structures are not intended to limit this application. The scope of protection of this application is determined by the claims.

[0023] In the field of urban stormwater management and non-point source pollution control, the core technical issue in achieving the goals of sponge city construction is how to scientifically and rationally spatially deploy Low Impact Development (LID) measures. Current mainstream technologies have formed a "simulation-optimization" paradigm centered on hydrological and hydraulic models (such as SWMM) and supplemented by empirical rules. This paradigm first divides sub-catchments based on static geographic and hydrological data and inputs design rainfall conditions to simulate surface runoff generation, runoff generation, and pipeline transmission processes. Then, LID measures are deployed in the model, and their reduction effects on runoff are simulated by repeatedly adjusting their locations and related parameters. Finally, optimization algorithms are introduced to find feasible layout schemes with a single objective such as "lowest cost" or "highest benefit."

[0024] However, this scheme has significant shortcomings, primarily in the following aspects: the optimization model is disconnected from the dynamic coupling mechanism of landscape hydrology, treating LID measures as static "points" or "surfaces" independent of the background landscape pattern, failing to quantify the nonlinear impact of the dynamic interaction between landscape spatial configuration and hydrological pathway networks on pollution transport; it lacks an objectively quantifiable landscape ecological safety boundary, focusing optimization objectives on engineering benefits and economic costs, lacking quantitative constraints on the long-term health status of the regional landscape hydrological system; and it is difficult to achieve real-time, automated verification of complex ecological constraints during optimization iterations, involving manual operation and trial and error across multiple software programs, resulting in poor repeatability of the decision-making process. These shortcomings collectively constrain the ecological rationality and long-term safety of the LID measure layout scheme.

[0025] To systematically address the aforementioned shortcomings, this application proposes a method for optimizing the layout of low-impact development (LID) measures by integrating dynamic landscape hydrology theory. Its core lies in constructing a dynamic quantitative index (HCI-LU) that couples landscape pattern, hydrological connectivity, and land use pollution potential. Based on machine learning and gradient analysis, it objectively identifies the ecological critical threshold from the data. Finally, this threshold is embedded as a core constraint into a multi-objective optimization model, thereby generating an LID measure layout scheme that is both economical and efficient within a clearly defined landscape ecological security boundary.

[0026] The overall process of the method mainly includes the following core steps that are executed sequentially and tightly coupled: First, we quantified the landscape pattern and hydrological connectivity characteristics and innovatively constructed the hydrological connectivity-land use coupling index (HCI-LU) to dynamically assess the region's pollution transport potential.

[0027] Secondly, key thresholds are identified based on machine learning and gradient analysis. By using a random forest regression model and partial dependency analysis, the ecological critical threshold of the HCI-LU index is objectively mined from historical data.

[0028] Finally, a multi-objective optimization decision-making model integrating landscape ecological constraints is established. During the solution process, the implementation effect of each candidate LID measure layout scheme is dynamically simulated by calling the database interface, and the HCI-LU index is verified in real time to see if it meets the safety threshold constraint, thereby outputting a Pareto optimal solution set that pre-satisfies the landscape ecological safety requirements.

[0029] The synergistic combination of the above-mentioned technical features constitutes a complete, automated technical solution that spans the entire chain, from mechanism quantification and safety threshold identification to intelligent optimization decision-making.

[0030] Figure 1 This is a flowchart illustrating an optimized layout method for low-impact development measures that integrates dynamic landscape hydrology theory, according to some embodiments of this application. Figure 1 As shown, the method specifically includes: S110, Construct and calculate the hydrological connectivity-land use coupling index to dynamically quantify the pollution transport potential of the regional landscape. The index is coupled with a weighted hydrological connectivity parameter reflecting the criticality of hydrological pathways, a weighted parameter reflecting the pollution source-sink intensity of land use types, and a landscape pattern index parameter reflecting the clustering morphology of dominant patches.

[0031] In some embodiments of this application, the construction and calculation of the hydrological connectivity-land use coupling index (HCI-LU) to dynamically quantify the pollution transport potential of a regional landscape is specifically achieved through the following process: First, the basic landscape pattern index is calculated: Based on high-resolution land use data, various landscape pattern indices are calculated. For example, landscape analysis tools such as Fragstats can be used to calculate patch density (PD), edge density (ED), landscape shape index (LPI), and aggregation index (AI).

[0032] Secondly, hydrological connectivity network modeling and quantification are performed: Based on the digital elevation model, the D8 algorithm is used to extract hydrological flow direction and construct a directed weighted graph: , Among them, the node set Indicates the exit point of the sub-catchment area, and the assembly point. Represents hydrological connection paths, weight set The flow resistance is calculated based on factors such as flow direction accumulation, slope, and surface roughness.

[0033] Based on this, graph theory algorithms are used to calculate each node. Improved betweenness centrality : , in, For the source node To the target node The total number of shortest paths, For the nodes The number of shortest paths.

[0034] Further introduce path efficiency weights : , in, This represents the total path length (m). This represents the maximum feature path length (m). To estimate the radial flow rate ( ), The maximum characteristic runoff ( ).

[0035] The final weighted hydrological connectivity is obtained. : , in, Represents nodes calculated based on directed hydrological connectivity networks. (Zihuishui District) Improved betweenness centrality of the export (of the export), This represents the slave node calculated based on the estimated flow rate and path length. To the node The path efficiency weight.

[0036] Next, the land use pollution load weights are determined: Selected study area A natural sub-catchment with synchronous long-term hydrological and water quality monitoring data and uniform internal land use was selected as the modeling sample. For each sample... Calculate the measured annual average pollution load per unit area. (e.g., total phosphorus load,) ), and simultaneously extract its landscape characteristic parameters: average impermeable surface ratio (%), mean normalized vegetation index (Dimensionless) and average soil saturated hydraulic conductivity (cm / / hr), forming a sample set .

[0037] Based on the above sample set, a prediction model of landscape features on pollution load is established using multiple linear regression: , in, This represents the predicted pollution load per unit area at any location ( ); , where are the independent variables of the model, representing the proportion of impermeable surface, normalized vegetation index, and soil saturated hydraulic conductivity at any location, respectively; These are the coefficients obtained from the regression. This model characterizes the quantitative relationship between landscape features and pollution load.

[0038] To assess the inherent pollution potential of various land types, for each land use type Randomly select within its range Each pure pixel (i.e., 100% of that type). Each pure pixel... Landscape feature values Substitute into the forecast model to calculate its forecast load. This type of "characteristic pollution load" Defined as the mean of the predicted load of all clean pixels ( ): , To obtain dimensionless weights suitable for multi-source data coupling comparison and to provide directional identification for "source / sink", the following steps are taken: Standardization is performed. First, the median of all types of characteristic loads is used. Centralize based on a benchmark to distinguish directions: This represents the "source," and vice versa for the "sink." Then, the maximum absolute value after centralization is used. Scaling is performed to obtain the final weights: , in, For dimensionless relative weights, the range is... The plus or minus sign indicates the "source / sink" attribute, and the absolute value reflects its relative strength.

[0039] Finally, the HCI-LU index is integrated: Ultimately, the Zihui Water District The HCI-LU index is defined as follows, and the meaning of each index is the same as in the previous formula.

[0040] , in, Used to eliminate sub-catchment areas area The resulting scale effect.

[0041] The HCI-LU index achieves a ternary coupling of "hydrological pathway criticality + land use pollution potential + dominant patch aggregation morphology", and has a clear dynamic response mechanism and spatial explicit expression capability.

[0042] This application innovatively constructs the hydrological connectivity-land use coupling index (HCI-LU), which for the first time dynamically couples the criticality of hydrological pathway networks, the intensity of land use "source / sink" and the clustering pattern of dominant patches through a mathematical model. This enables a spatial explicit and mechanistic quantitative assessment of the potential for landscape pollution transport, providing a direct theoretical basis for the precise layout of LIDs.

[0043] In some embodiments of this application, prior to performing the step of constructing and calculating the hydrological connectivity-land use coupling index, a dedicated spatiotemporal database for dynamic coupling analysis of landscape patterns and hydrological processes is further constructed. In a preferred embodiment, the construction process of the dedicated spatiotemporal database specifically includes: First, the core spatial object – the “hydrological response unit” – is generated and encoded: First, based on the digital elevation model (DEM), the hydrological analysis tools in the ArcGIS hydrology toolbox were used to fill depressions, calculate flow direction and runoff accumulation, and generate the boundaries of the natural water system and sub-catchment areas of the study area.

[0044] Based on this, sub-catchments are overlaid and intersected with high-precision land use / cover classification maps (obtained through Landsat 8 or Sentinel-2 image interpretation). Continuous patches of the same land use type falling within the same sub-catchment are merged to form the smallest, most uniform hydrological and land use unit (HRU). Each HRU serves as the basic record unit in the database, assigned a unique code, and pre-stored its core attributes: area, perimeter, sub-catchment ID, center point coordinates, dominant land use type and its area proportion (e.g., impervious surface proportion), average slope, and average saturated hydraulic conductivity mapped from soil data. This HRU-based organizational method provides a direct and efficient data structure for subsequent calculations of landscape pattern indices and pollution load weights.

[0045] Next, the explicit construction and storage of the hydrological connectivity network topology will be performed: To enable dynamic hydrological connectivity analysis, the database not only stores spatial entities (HRUs), but also explicitly constructs and stores the hydrological connectivity relationships between them, specifically including: Upstream-downstream relationship table: Based on DEM and flow direction data, the set of directly upstream HRUs and directly downstream HRUs is calculated and recorded for each HRU (or sub-catchment outlet point). This relationship is stored in a relational database in the form of an adjacency list or a topological edge table.

[0046] Path attribute table: For each pair of HRUs (or nodes) with hydrological connectivity, the characteristic hydrological path length (flow path length based on DEM) and default transport resistance coefficient (initial estimate based on land use type roughness and infiltration capacity) are pre-calculated and stored. This table forms the initial template and rapid retrieval basis for subsequent dynamic updates of the hydrological connectivity matrix.

[0047] Finally, the multi-source time series data are correlated and fused: To enable dynamic evaluation, the database design supports efficient association of time-series data, including: Monitoring data correlation: High-frequency monitoring data such as flow rate and water quality (TSS, TN, TP) acquired by field-deployed sensors are linked to the hydrological response unit (HRU) or sub-catchment outlet through location information. The data is stored in time series format and supports slice extraction by rainfall event, providing paired samples of "landscape state-hydrological response" for machine learning models.

[0048] Integration of LIDs facility parameter library: The established LIDs facility parameter library (including type, structural dimensions, design permeability, construction cost, operation and maintenance cost, etc.) not only serves as a static parameter table, but its records are also dynamically associated with HRU type codes. When an HRU is planned as an LIDs facility during optimization, the system can automatically obtain its corresponding hydrological and cost parameters through this association, which are used to dynamically update the HRU's attributes and recalculate the relevant hydrological connectivity resistance.

[0049] And design the dynamic update interface for the database: To support scenario simulations during the optimization process (such as "if a bioretention pond is placed in a certain HRU"), the database provides a programmatic dynamic update interface. When it is necessary to evaluate a candidate LIDs layout scheme, the optimization algorithm calls the database API, passing in a scheme description. The kernel dynamically copies and modifies the land use type code and associated hydrological parameters of the affected HRUs in memory according to the scheme. Then, it cascades and triggers the update of the relevant hydrological connectivity path resistance coefficients, generating a temporary scenario data view for subsequent landscape index calculation and connectivity analysis modules to call. This design avoids the huge overhead of repeatedly processing raw geographic data and reconstructing topology during optimization iterations, and is a key technical guarantee for the entire method to achieve real-time and automated optimization.

[0050] The embodiment described above constructs a structured spatiotemporal database specifically designed for the dynamic coupling analysis of "landscape pattern-hydrological process". Its core is the creation of "Hydrological Response Units (HRUs)" as unified data primitives, achieving an intrinsic fusion of landscape attributes and hydrological characteristics. By explicitly storing the pre-computed hydrological connectivity network topology, complex hydrological relationships are transformed into data relationships that can be quickly queried and updated. A key breakthrough lies in the design of a dynamic update interface supporting "hypothesis analysis," enabling optimization algorithms to generate data snapshots of any planning scenario within seconds via API calls, thereby transforming time-consuming manual GIS operations into efficient automated services. Through proactive pre-association of multi-source data, this database forms a "computationally ready" data hub that drives the efficient operation of subsequent models and optimization algorithms.

[0051] It should be noted that the dedicated spatiotemporal database and its construction and invocation methods in the above embodiments are a preferred technical solution for the efficient and automated implementation of the complete method flow of "constructing the index - identifying the threshold - embedding constraints and optimizing the solution" proposed in this application. It provides structured data primitives for the construction of the hydrological connectivity-land use coupling index (HCI-LU), prepares a standardized dataset for the objective identification of ecological critical thresholds, and ultimately ensures real-time and automated verification of landscape ecological security constraints during optimization iterations through its scenario simulation interface. Those skilled in the art should understand that the core innovation of the method claimed in this application lies in the aforementioned complete method flow and decision-making logic. The specific implementation of each step in the process, especially the functional step of "dynamically simulating the underlying surface and hydrological connectivity status after the implementation of the scheme," can be achieved through various data organization, management, and computation architectures. The dedicated spatiotemporal database is a specific and preferred system implementation method that supports the efficient operation of the entire process.

[0052] S120, Based on machine learning and gradient analysis, objectively identify the ecological critical threshold of the hydrological connectivity-land use coupling index, including: training a pollution load prediction model using a dataset containing the index and a traditional landscape pattern index, calculating the partial dependency curve of the index in the model, and determining the index value corresponding to the maximum gradient value as the ecological critical threshold by calculating the first-order discrete gradient of the partial dependency curve.

[0053] This step aims to objectively uncover inherent patterns from historical landscape patterns and hydrological response data using machine learning methods, and quantitatively identify landscape ecological safety thresholds for controlling the optimization process. The process first constructs a machine learning model capable of accurately predicting runoff pollution load, then assesses the importance of various landscape features through model interpretability techniques, and finally extracts critical thresholds with clear ecological significance from the model response curve. In some embodiments, the steps specifically include: Construction of standardized datasets: Based on the S110 results, a structured dataset was constructed. Each row of the dataset corresponds to a sub-catchment, and its feature columns (input...) It must simultaneously include traditional landscape pattern indices (PD, ED, LPI, AI) and the innovative HCI-LU index proposed in this application, with its target column (output) This refers to the measured pollution load per unit area of ​​the sub-catchment (e.g., TSS load). ).

[0054] Predictive model training: Using the random forest regression algorithm trained on the aforementioned dataset, a model capable of accurately predicting pollution load was obtained. The goal of model training is to obtain a surrogate function that can reliably reflect the complex nonlinear relationship between the HCI-LU index and the pollution load.

[0055] Partial dependence effect of HCI-LU index: To isolate the interference of other factors and evaluate the impact of HCI-LU independently, its partial dependency function is calculated. Specifically, a series of values ​​are uniformly set within the global range of HCI-LU values. For each value Replace the HCI-LU feature values ​​of all samples in the dataset with Meanwhile, keeping other feature values ​​unchanged, and then using the trained model Make a prediction and calculate the average of all predictions. The formula is: , in, This represents the total number of samples. From this, a series of points are obtained. The connection forms a partial dependency curve.

[0056] Identifying mutation thresholds through gradient analysis: The first-order discrete gradient of the aforementioned partial dependence curve is calculated to quantify the sensitivity of the loading to HCI-LU. For adjacent points in the sequence... and Its gradient Calculate using the following formula: , Traverse the entire sequence to obtain the gradient sequence. Ecological critical threshold This is the HCI-LU value corresponding to the maximum value in the gradient sequence: The physical significance of this point is that a small increase in HCI-LU here will lead to the most dramatic rise in predicted pollution load, marking the critical point at which the system transitions from a "low-risk steady state" to a "high-risk sensitive state".

[0057] Ultimately, this threshold will be directly embedded into the subsequent multi-objective optimization model as a key constraint on landscape pattern health, ensuring that the optimized LIDs layout scheme not only meets economic and engineering objectives, but also fundamentally guarantees that the regional landscape is in a state of hydrological safety.

[0058] This application proposes an objective identification method for abrupt change thresholds based on gradient extrema. By analyzing the single-factor response curves generated by machine learning models, the first-order discrete gradient is calculated, and the HCI-LU value corresponding to the maximum gradient point is defined as the ecological critical threshold. This method transforms the abrupt change point of the system state into a computable mathematical criterion, realizing a paradigm shift from setting the threshold based on subjective experience to objectively deriving it from data.

[0059] S130, for the candidate layout schemes of the low-impact development measures, establish a multi-objective optimization model with hydrological benefits and costs as optimization objectives and the ecological critical threshold as landscape ecological security constraints. The landscape ecological security constraints refer to the fact that after any candidate layout scheme is implemented, its corresponding hydrological connectivity-land use coupling index value should not be greater than the ecological critical threshold. The hydrological benefits refer to the comprehensive reduction rate of runoff and pollutants.

[0060] This step is used to establish a multi-objective optimization model to solve for the optimal combination of type, location, and scale of low-impact development measures under the dual constraints of investment budget and landscape ecological security. Specifically, Basic framework of the model: The model's decision variable is the layout scheme of low-impact development measures. This includes information on the location, type, and scale of each facility. Two objective functions are set: maximizing overall hydrological benefits. (Combined reduction rate of runoff and pollutants) and minimizing life-cycle costs Meanwhile, the model includes two core constraints: Budget Constraints: Total Cost Not exceeding the upper limit ; Landscape ecological constraints (core of this application): After the implementation of the scheme, the overall HCI-LU index of the region must not exceed the ecological critical threshold. ,Right now .

[0061] The model is solved using a mature multi-objective genetic algorithm.

[0062] S140, during the solution process of the multi-objective optimization model, for each candidate layout scheme, the hydrological connectivity-land use coupling index value after the scheme is implemented is recalculated by dynamically simulating the underlying surface and hydrological connectivity after the scheme is implemented, so as to verify whether it meets the constraints.

[0063] The key to this application lies in the fact that, during the optimization algorithm iteration process, for each candidate solution... A standardized and automated calculation process is used to verify in real time whether it meets landscape constraints. This process, based on a constructed database, is triggered by the fitness evaluation function of the optimization algorithm during each evaluation, with the solution as its input. The encoding outputs a Boolean value (whether the constraint is satisfied) and a superscalar value used to calculate the penalty value. In the preferred embodiment, the process includes the following four sequentially executed calculation steps: Step 1: Rapid generation and acquisition of scenario data based on a dedicated database The decoder of the optimization algorithm is based on candidate solutions. The system encodes a structured request and initiates a "scenario simulation" call to the constructed landscape hydrology database. This request explicitly lists the Hydrological Response Unit (HRU) identifiers that need adjustment and the target LID facility types to be changed (e.g., [HRU_101: "Bioretention pond", HRU_205: "Permeable pavement"]). Upon receiving the request, the database performs the following operations in real-time in memory via its dynamic update interface: Based on the request list, locate and modify the land use type code and all associated hydrological attribute parameters (such as roughness and infiltration rate) of the relevant HRU.

[0064] Based on the updated HRU properties, the transmission resistance coefficients on the affected hydrological connectivity paths are updated in a cascade manner.

[0065] Generate and return a complete, provisional "post-planning scenario" data package. The core outputs of this package include: "land_use_updated," a raster array of land use types reflecting the underlying surface conditions after the implementation of scenario x; "updated_topology," updated hydrological connectivity network topology data with resistance coefficients; and the associated HRU attribute table. This step, through a single efficient database service call, replaces the tedious process of repeated GIS spatial analysis and manual attribute table operations required in traditional optimization, and is fundamental to ensuring the real-time operation of the entire optimization process.

[0066] Step 2: Dynamic Calculation of Landscape Pattern Index and Hydrological Connectivity Receive the "post-planning scenario" data packet output from step 1 as input, and execute two parallel computing tasks: Landscape pattern index calculation: The function "compute_landscape_indices(land_use_updated)" analyzes a new land use array "land_use_updated". For example, it can be used to calculate the maximum patch index for a specific land use type. This function executes the connected component labeling algorithm to identify relevant patches, calculates their areas, and then... Formula calculation. This process is performed iteratively for all land use types required for HCI-LU calculation, outputting new results. .

[0067] Hydrological connectivity matrix update calculation: The function "update_hydrological_connectivity(updated_topology)" uses the latest network topology and resistance parameters provided by "updated_topology" to recalculate the hydrological transport efficiency between each node (HRU or sub-catchment outlet) using graph theory algorithms (such as the improved Dijkstra algorithm) and generates the updated hydrological connectivity matrix "C_matrix_new".

[0068] Step 3: HCI-LU exponent calculation and constraint determination: The function "compute_HCI_LU(LPI_dict,C_matrix_new,params)" is called. This function reads the pre-stored parameters (plot area). Pollution load weight (parameters, etc.), and integrate the results calculated in step 2. and The aggregation calculation is performed strictly according to the HCI-LU exponent formula defined in this patent, and a scalar value is finally output. To quantify the implementation plan The landscape-hydrological coupling state of the rear area.

[0069] Step 4: Determination and Quantification of Ecological Security Constraints Using the calculation results from step 3, "and the preset ecological critical threshold" The comparison and evaluation logic is as follows: if Then the degree of constraint violation ;otherwise, .Should This will be incorporated as a penalty into the fitness function of the optimization algorithm, thereby automatically avoiding areas that violate landscape ecological security during the search process.

[0070] S150, output the Pareto optimal solution set that satisfies the landscape ecological security constraints, as an optimized layout scheme for low-impact development measures.

[0071] The above four steps are encapsulated in a unified function. In the fitness evaluation of each generation of the NSGA-II optimization algorithm, for each candidate solution in the population... Each step involves calling this function once. Through the efficient call to the dedicated database interface in step 1 and the execution of the standardized algorithm components in steps 2 and 3, this application achieves the transformation of complex landscape ecological constraints into an automated calculation module that can be completed in seconds during the optimization loop.

[0072] The optimization algorithm ultimately outputs a Pareto optimal solution set that satisfies the hard constraints of landscape ecological security. Decision-makers can make their final choice based on their cost-benefit preferences within this safe solution set.

[0073] In summary, the low-impact development (LID) optimization layout method provided in the embodiments of this application, which integrates dynamic landscape hydrology theory, constructs and calculates a hydrological connectivity-land use coupling index that couples hydrological path criticality, land use "source / sink" intensity, and dominant patch morphology. This index transforms the spatial configuration of the landscape and the resulting hydrological connectivity network from static background parameters into quantifiable dynamic decision variables, thereby characterizing the dynamic impact of LID layout on the entire process of pollutant migration "from source to sink," thus addressing the deficiency of the optimization model being disconnected from the dynamic coupling mechanism of landscape hydrology. Furthermore, through partial dependency analysis and first-order discrete gradient calculation based on machine learning models (such as random forest regression), the ecological critical threshold of the HCI-LU index is objectively identified. This method automatically and objectively derives key critical values ​​indicating abrupt changes in hydrological and ecological states from monitoring data, providing a quantifiable safety boundary derived from landscape mechanisms for the optimization process, thus solving the problem of the lack of objectively quantifiable landscape ecological factors in the optimization process. The shortcomings of safety boundaries are addressed by incorporating the ecological critical threshold as a necessary landscape ecological safety constraint when establishing a multi-objective optimization model. During the optimization iteration, for each candidate scheme, the scenario simulation interface of a dedicated spatiotemporal database is invoked to dynamically simulate the underlying surface and hydrological connectivity after the scheme's implementation, and the HCI-LU index value is recalculated for real-time verification. This mechanism transforms time-consuming offline spatial analysis and manual verification into efficient online automated calculation, enabling complex ecological constraints to be enforced in real-time and in a closed loop during each iteration of the optimization algorithm. This solves the problem of difficulty in achieving real-time, automated verification of complex ecological constraints during optimization iterations. Furthermore, by synergistically integrating the core technical features of index construction, threshold identification, optimization modeling, and dynamic verification, and combining them into an end-to-end automated decision-making process through programmed scripts, a significant improvement in the generation efficiency, scientific rigor, and repeatability of LIDs planning schemes is achieved.

[0074] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the devices and modules described above can be referred to the corresponding descriptions in the foregoing device embodiments, and will not be repeated here.

[0075] Although the subject matter described herein is provided in the general context of execution on a computer system in conjunction with an operating system and applications, those skilled in the art will recognize that other implementations can also be executed in conjunction with other types of program modules. Generally, program modules include routines, programs, components, data structures, and other types of structures that perform specific tasks or implement specific abstract data types. Those skilled in the art will understand that the subject matter described herein can be practiced using other computer system configurations, including handheld devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, minicomputers, mainframes, etc., and can also be used in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules may reside on both local and remote memory storage devices.

[0076] Those skilled in the art will recognize that the units and method steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0077] It should be understood that the specific embodiments described above are merely illustrative or explanatory of the principles of this application and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of this application should be included within the protection scope of this application. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

Claims

1. An optimized layout method for low-impact development measures integrating dynamic landscape hydrology theory, characterized in that, include: A hydrological connectivity-land use coupling index is constructed and calculated to dynamically quantify the pollution transport potential of regional landscapes. The index is coupled with a weighted hydrological connectivity parameter reflecting the criticality of hydrological pathways, a weighted parameter reflecting the pollution source-sink intensity of land use types, and a landscape pattern index parameter reflecting the clustering morphology of dominant patches. Based on machine learning and gradient analysis, the ecological critical threshold of the hydrological connectivity-land use coupling index is objectively identified, including: training a pollution load prediction model using a dataset containing the index and a traditional landscape pattern index, calculating the partial dependency curve of the index in the model, and determining the index value corresponding to the maximum gradient value as the ecological critical threshold by calculating the first-order discrete gradient of the partial dependency curve. For the candidate layout schemes of the aforementioned low-impact development measures, a multi-objective optimization model is established with hydrological benefits and costs as optimization objectives and the aforementioned ecological critical threshold as a landscape ecological security constraint. The landscape ecological security constraint means that after the implementation of any candidate layout scheme, its corresponding hydrological connectivity-land use coupling index value must not be greater than the aforementioned ecological critical threshold. The hydrological benefits refer to the comprehensive reduction rate of runoff and pollutants. In the process of solving the multi-objective optimization model, for each candidate layout scheme, the hydrological connectivity-land use coupling index value after the implementation of the scheme is recalculated by dynamically simulating the underlying surface and hydrological connectivity after the scheme is implemented, so as to verify whether it meets the constraints. Output the Pareto optimal solution set that satisfies the landscape ecological security constraints, as an optimized layout scheme for low-impact development measures; The hydrological connectivity-land use coupling index is calculated as follows: ， in, Indicates the sub-watershed area The aforementioned hydrological connectivity-land use coupling index value, This indicates a reflection of the sub-basin watershed. The weighted hydrological connectivity parameter, which is critical to hydrological pathways, is calculated using the following formula: ,in, Represents nodes calculated based on directed hydrological connectivity networks. Improved betweenness centrality, the node For Zihui District exports, This represents the slave node calculated based on the estimated flow rate and path length. To the node Path efficiency weights Indicates the type of land use The weighting parameters for pollution source-sink intensity, Indicates land use type The landscape pattern index parameters mentioned above. Indicated for eliminating sub-catchment areas area The function of the resulting scaling effect.

2. The method according to claim 1, characterized in that, Before constructing and calculating the hydrological connectivity-land use coupling index, a dedicated spatiotemporal database for dynamic coupling analysis of landscape patterns and hydrological processes is also constructed. The construction of the dedicated spatiotemporal database includes: Multi-source spatiotemporal data are organized using hydrological response units as basic spatial objects, and the land use attributes, topographic features and associated hydrological parameters of each hydrological response unit are pre-stored. Explicitly construct and store the topological relationships of the hydrological connectivity network between each hydrological response unit and the corresponding path attributes; Configure a scenario simulation interface to dynamically modify the attributes of the affected hydrological response units based on the input candidate layout schemes, and cascade update the associated hydrological connectivity parameters to generate post-planning scenario data.

3. The method according to claim 1, characterized in that, The landscape pattern index parameter is the maximum patch index.

4. The method according to claim 1, characterized in that, The method for determining the weighting parameters reflecting the intensity of pollution sources and sinks based on land use type includes: Based on synchronous monitoring data from multiple natural sub-catchments with uniform land use within the study area, a predictive model of landscape characteristics on pollution load per unit area was established using multiple linear regression. The landscape feature values ​​of pure pixels within each land use type are input into the prediction model to calculate the average characteristic pollution load of each type of land use. The mean of the characteristic pollution load is centered and standardized to obtain the weight parameter, wherein the weight parameter is a dimensionless value, the sign of which is used to indicate the "source" or "sink" attribute, and the absolute value reflects its relative intensity.

5. The method according to claim 1, characterized in that: The pollution load prediction model was trained using the random forest regression algorithm.

6. The method according to claim 1, characterized in that: The multi-objective optimization model is solved using a multi-objective genetic algorithm.

7. The method according to claim 1, characterized in that: The cost is the total lifecycle cost of the low-impact development measure.

8. The method according to claim 1 or 6, characterized in that: The steps of constructing and calculating the hydrological connectivity-land use coupling index, objectively identifying ecological critical thresholds, establishing and solving multi-objective optimization models, and dynamically simulating and verifying landscape ecological security constraints are integrated into an end-to-end automated decision-making process through programmed scripts.

Images