Urban inland inundation risk comparison method and device based on spatial sensitivity

Abstract

The invention relates to an urban inland inundation risk comparison method and device based on spatial sensitivity. The method comprises the following steps: comprehensively collecting weather, geographic information, population, traffic and historical disaster data; the method comprises the following steps: integrating multi-source heterogeneous data, constructing a risk-sensitivity-vulnerability three-dimensional coupling waterlogging risk assessment model, dynamically adjusting a weight coefficient in combination with a machine learning algorithm, introducing physical constraints to optimize surface runoff simulation precision, and generating a gridding risk unit; fusing the multi-dimensional risk index with geographic space data to generate a gridding risk distribution map; and constructing a waterlogging risk deduction model under rainfall scenes in different recurrence periods, realizing multi-layer comparative analysis through a GIS spatial superposition technology, and performing visual output. Through a multi-dimensional comprehensive evaluation system, dynamic weight adjustment and high-resolution spatial mapping, the accuracy and dynamic adaptability of waterlogging risk evaluation are remarkably improved, and scientific and efficient decision support is provided for urban waterlogging treatment.

Description

Technical Field

[0001] The present application belongs to the technical field of urban area risk assessment, and in particular relates to a method and device for comparing urban waterlogging risks based on spatial sensitivity. Background Art

[0002] Currently, most existing urban flooding risk assessment systems simulate only a single rainfall scenario and lack the ability to simultaneously compare different rainfall conditions (such as short bursts of heavy rain versus sustained rainfall). Users must manually switch between different rainfall scenarios during use, preventing them from visually observing the differences in risk distribution under different rainfall conditions from a single perspective, significantly reducing decision-making efficiency.

[0003] Urban flooding risk is a complex, multi-factor system. Assessing it from a single dimension alone cannot accurately and comprehensively reflect the actual risk situation. Existing technologies for urban flooding risk comparison have the following shortcomings:

[0004] (1) Insufficient integration of spatial sensitivity factors: Existing waterlogging risk comparison methods have defects in the integration of spatial sensitivity factors. Most models (such as U-RNN and traditional hydrodynamic models) fail to systematically quantify the dynamic impact of spatial sensitivity factors such as land use change and microtopography.

[0005] (2) Lack of scenario comparison function: Most systems use a fixed rainfall threshold (e.g., once in 50 years), which makes it impossible to simultaneously compare the different impacts of short-term heavy rain and continuous rainfall.

[0006] (3) Weak real-time dynamic update capability: Existing methods (such as KMeans-ANN) mainly rely on historical static data and are difficult to integrate real-time IoT sensor data (such as water level, rainfall intensity, etc.) for waterlogging risk assessment.

[0007] In summary, although existing technologies have made certain progress in waterlogging prediction speed and multi-source data fusion, they still have shortcomings in the dynamic integration of spatial sensitivity factors and real-time map comparison. Summary of the Invention

[0008] In order to address the above problems, this application proposes a new urban waterlogging risk comparison method and device based on spatial sensitivity by coupling an innovative sensitivity coefficient model with a high-resolution geographic information system (GIS), aiming to provide more accurate and explainable decision support for urban waterlogging management.

[0009] The present invention mainly adopts the following technical strategies:

[0010] 1. Construction of a multi-dimensional comprehensive risk assessment system

[0011] The present invention is based on a multi-factor fusion assessment, integrating multi-dimensional sensitive factors such as the elasticity coefficient of land use change, the terrain relief index, and hydro-meteorological parameters (such as rainfall intensity and runoff coefficient) to construct a three-dimensional dynamic coupling evaluation system of "hazard-sensitivity-vulnerability". By adopting machine learning algorithms (such as LightGBM) to dynamically adjust the weights of each factor, and combining physical constraints (such as the Manning equation) to ensure the scientificity and adaptability of the model, the dynamic weight distribution of factors affecting urban waterlogging is achieved. At the same time, based on GIS technology, the city is divided into 100-meter grid units, combined with the building unit and road network scale, to achieve accurate positioning of the spatial distribution of disaster-prone bodies and risk superposition analysis.

[0012] Compared with traditional models, this invention breaks through the limitations of traditional static weight mechanisms and single-dimensional evaluation (such as based only on rainfall intensity), and innovatively introduces spatially sensitive factors such as land use change elasticity coefficient and terrain undulation index, significantly improving the accuracy and adaptability of risk assessment.

[0013] 2. Unique index calculation model and dynamic correction of sensitivity coefficient

[0014] The dynamic sensitivity coefficient correction model employed in this study is centered around quantitative formulas for the land use change elasticity coefficient and the terrain relief index. Using a nonlinear regression model, it modifies traditional waterlogging grade calculations and dynamically generates a waterlogging risk index. Combining historical rainfall event probability models with real-time data, a composite waterlogging risk probability index based on rainfall thresholds and waterlogging depth is established, supporting comparisons of scenarios with different return periods. Furthermore, a multi-algorithm collaborative optimization parameter tuning approach is employed to ensure the model's accuracy and reliability, using risk index generation logic based on both historical data and real-time scenarios.

[0015] Compared with the existing technology, the "sensitivity coefficient correction-multi-algorithm collaboration" dual-engine model proposed in this invention solves the problem that traditional methods relying on empirical formulas or a single algorithm (such as the AHP hierarchical analysis method) are unable to quantify the dynamic impact of sensitive factors, and significantly improves the scientific nature and dynamic adaptability of urban flooding risk assessment.

[0016] 3. Spatial overlay analysis and physical constraint enhancement

[0017] This method uses Kriging interpolation to fill data gaps, generating a 30-meter gridded risk distribution map. It also supports GIS overlay comparison of historical disaster layers and real-time heat maps. Furthermore, it employs multi-scale coupled analysis to support multi-scale geographic information data simulation and conduct refined risk assessments for urban lifeline projects, such as underpasses and subway stations.

[0018] Compared with the existing technology, the present invention significantly improves the level of refinement and physical connectivity of risk assessment through high-resolution spatial overlay analysis and physical constraint enhancement, providing more scientific and accurate decision-making support for urban waterlogging control.

[0019] In summary, the main steps of the method of the present invention include:

[0020] 1. Data Collection

[0021] Comprehensively collect various data, including meteorological data, water system data, transportation data, disaster data, production risk data, and municipal-level geographic information basic data. Through multiple channels, we obtain historical flood disaster data and information on flood risk points, providing rich material for subsequent analysis, building a detailed flood risk data foundation, and ensuring the comprehensiveness and objectivity of the assessment.

[0022] 2. Multi-dimensional data coupling waterlogging risk modeling

[0023] Integrate multi-source heterogeneous data such as meteorology, topography, land use, and population density to construct a three-dimensional coupled risk assessment model of "hazard-sensitivity-vulnerability". Combine machine learning algorithms to dynamically adjust weight coefficients, and introduce physical constraints such as the Manning equation to optimize the accuracy of surface runoff simulation. Based on the spatial overlay technology of the geographic information system (GIS), generate risk cells with a gridding of 100 meters, and use Kriging interpolation to fill data gaps. Combine real-time sensor data and historical disaster information to dynamically modify the risk index, and support horizontal comparability analysis of rainfall scenarios with different recurrence periods. By quantifying highly sensitive factors such as land use elasticity coefficient and terrain relief, the risk prediction error rate is reduced to below 12%, providing a scientific decision-making basis for accurately identifying vulnerable areas and optimizing emergency resource scheduling.

[0024] 3. Risk Index Space Mapping

[0025] Based on GIS technology, the multidimensional risk index is integrated with geospatial data, and data gaps are filled using the Kriging interpolation algorithm to generate a 30-meter gridded risk distribution map. A dynamic weight allocation mechanism is used, combining real-time IoT sensor data (such as water level and rainfall) with historical disaster data to dynamically correct risk units at the minute level. Sensitive factors such as terrain undulation and drainage network density are superimposed to construct spatial heat maps and vector layers. This system supports the comparison of scenarios with different rainfall return periods (such as 10-year and 50-year return periods), achieving accurate visualization of the spatial distribution of risk and providing a scientific spatial decision-making basis for urban waterlogging management.

[0026] 4. Scenario comparison and visualization output

[0027] Construct a waterlogging risk simulation model under different rainfall scenarios (e.g., 10-year and 50-year rainfall events), and implement multi-layer comparative analysis on the same screen through GIS spatial overlay technology. Use the WebGL engine to render dynamic risk heat maps, overlaying vector data such as real-time traffic flow and emergency resource locations. Optimize rendering efficiency through a multi-window sliding window warm-up training paradigm, supporting second-level refresh and interactive zooming of 100-meter grids. Combined with an interpretable algorithm, generate factor contribution heat maps, intuitively displaying the dominant risk factors in sensitive areas and providing high-precision visualization support for urban waterlogging emergency decision-making.

[0028] Specifically, this application provides the following technical solutions:

[0029] A first aspect of the present application provides a method for comparing urban flooding risks based on spatial sensitivity, the method comprising:

[0030] Data collection: Comprehensively collect meteorological data, geographic information data, population information and evacuation data, traffic flow data, and historical disaster data;

[0031] Multi-dimensional data-coupled waterlogging risk modeling: This integrates collected multi-source heterogeneous data (such as meteorological, topographic, land use, and population density data) to construct a three-dimensional coupled risk assessment model of "hazard-sensitivity-vulnerability." This model uses machine learning algorithms to dynamically adjust weight coefficients, introduces physical constraints to optimize surface runoff simulation accuracy, and generates gridded risk cells.

[0032] Risk index spatial mapping: Based on GIS (Geographic Information System) technology, multi-dimensional risk index is integrated with geographic spatial data to generate a gridded risk distribution map;

[0033] Scenario comparison and visualization output: Construct a waterlogging risk simulation model under rainfall scenarios with different return periods, conduct a comparative analysis of multiple layers on the same screen using GIS spatial overlay technology, and visualize the analysis results.

[0034] Furthermore, in the method of this application, the data collection step includes:

[0035] Meteorological data collection: Real-time temperature, humidity, rainfall, wind speed, and wind direction data are collected through ground-based weather stations. Information on precipitation intensity and wind fields is collected from Doppler radar, as well as satellite remote sensing data.

[0036] Geographic information data collection: obtaining the city's overall DEM elevation data, administrative division data, road network data, water system and drainage network data, land classification data, and location information of urban waterlogging points;

[0037] Population information and shelter data collection: Obtain real-time population movement trajectories through mobile operators, and integrate shelter point of interest data and shelter material data provided by emergency management departments;

[0038] Traffic flow data collection: Connecting to cameras, ground sensors, and electronic police equipment on the city road network to obtain real-time traffic flow and speed data;

[0039] Collection of historical disaster data: Obtain historical flooding records from water and emergency management departments, disaster investigation reports, road flooding monitoring records from traffic control and municipal archives, and drainage facility operation logs.

[0040] Furthermore, the present application method also includes:

[0041] Meteorological data collection: The collected meteorological data is cleaned and preprocessed, missing values are handled using Kriging interpolation, and the data units are standardized to form a unified unit format. The standardized data is then subjected to spatiotemporal alignment and feature extraction, and a time series database of meteorological data is constructed using a machine learning model (LSTM).

[0042] Geographic information data collection: spatial calibration, topological analysis, spatial overlay, and map service publishing of collected geographic information data;

[0043] Population information and evacuation data collection: Population density analysis and credibility marking based on real-time population flow trajectories are used to analyze population gathering and evacuation routes during heavy rains. Shelter point of interest data and evacuation material data are used to assess population accessibility during flooding.

[0044] Traffic flow data collection: The collected traffic flow data is cleaned and combined with road network information to reconstruct the network topology, and a heat map of traffic congestion is output for waterlogging risk diffusion deduction;

[0045] Collection of historical disaster data: Clean the collected historical disaster data, remove invalid data, and normalize them into historical disaster events (event level, occurrence time, and location).

[0046] Furthermore, in the method of the present application, the multi-dimensional data coupling waterlogging risk modeling step includes:

[0047] (1) The calculation of water ripple hazard H is based on the simulation of water depth using a hydrodynamic model, solving the shallow water equation using numerical methods, and dynamically calculating the waterlogging process in combination with actual terrain and rainfall data, including:

[0048] Data preparation and model input: preprocess terrain data, divide unstructured grids, and load rainfall data;

[0049] Hydrodynamic model construction and equation discretization: The shallow water equation is solved by numerical methods. The shallow water equation is:

[0050]

[0051] in, is the rate of change of water depth over time, is the water flow transport in the x direction, is the water flow transport in the y direction, and q is the external source and sink term; is the inertia term, and is the convection term, is the pressure gradient term, g(S 0x -S fx ) represents the balance between driving force and resistance; h is the water depth, u is the velocity component of the water flow along the x direction, v is the velocity component of the water flow along the y direction, g is the acceleration of gravity, q is the flow rate per unit width, S 0x is the bottom slope in the x direction, S 0y is the bed slope in the y direction, S fx is the friction slope in the x direction, S fy is the friction slope in the y direction;

[0052] Numerical solution and dynamic simulation: The system of equations is solved using the direction-varying implicit method, and the time step is dynamically adjusted according to the Courant-Friedrichs-Lewy condition;

[0053] Waterlogging depth indicator output: Generate a gridded waterlogging depth field, overlay the pipe network overflow data to extract the maximum waterlogging depth; output indicators include flooded area, peak water depth, and waterlogging duration;

[0054] Quantification of hydrological hazard H: Risk levels are divided according to water depth thresholds, and the risk weight is adjusted by adding the traffic island index;

[0055] (2) Calculation of spatial sensitivity K, including:

[0056] (a) Calculation of land use change sensitivity:

[0057]

[0058] Among them, S j is the land use change sensitivity index of grid j. The higher the value, the higher the sensitivity of the grid to land use change. ijis the area of the i-th land class in grid j (such as the actual area of cultivated land, forest land or water body in grid j), Aj is the benchmark area of the i-th land class in the study area (such as the total area of the land class in the whole region or the planned target area), Wi is the weight coefficient of the i-th land class, reflecting the contribution of the land class to the overall sensitivity, i is the land use index (i = 1: cultivated land (Cropland), i = 2: forest (Forest), i = 3: water body (Waterbody), traversing the three land classes of cultivated land, forest land and water body);

[0059] (b) Terrain slope sensitivity calculation:

[0060] Using soil loss equation

[0061]

[0062] Where Q is the terrain slope sensitivity before adjustment, and θ is the slope percentage;

[0063] And perform terrain roughness correction:

[0064]

[0065] Among them, Q adjusted is the adjusted terrain slope sensitivity, t is the river network density;

[0066] (c) Drainage capacity sensitivity calculation:

[0067] D=log(A*tanθ+1)

[0068] Where D is the drainage capacity sensitivity before adjustment, A is the catchment area, and θ is the slope angle;

[0069] And carry out drainage facility regulation:

[0070]

[0071] Among them, D adjusted is the adjusted drainage capacity sensitivity, Wi is the facility type weight, n is the number of facility types, and fi is the facility coverage rate;

[0072] (d) Comprehensive sensitivity weighting:

[0073] K=w1*S j +w2*Q adjusted +w3*D adjusted

[0074] Among them, W1, W2, and W3 are respectively the land use change sensitivity Sj, terrain slope sensitivity Qj determined by the hierarchical analysis method adjusted , drainage capacity sensitivity D adjusted The weight of

[0075] (3) Calculation of the waterlogging risk index R, including:

[0076] The water ripple hazard H and spatial sensitivity K are dynamically coupled to form the waterlogging risk index R:

[0077] R=α*H+β*K

[0078] Among them, α is the hydrological hazard coupling coefficient, which represents the contribution weight of H to the comprehensive risk R; β is the spatial sensitivity coupling coefficient, which represents the contribution weight of K to the comprehensive risk R, 0≤α≤1 and α+β≤1;

[0079] Adjustment coefficient: Use the LightGBM model to optimize the coupling coefficients α and β, and prevent overfitting through regularization terms. Input historical flooding event records (labels) and the H and K indices (features) of the corresponding areas, set α = 0.6 and β = 0.3 as the initial values of LightGBM training, and sparse coefficients to automatically remove insignificant parameters to prevent α and β from overfitting and ensure model generalization. Finally, take the average after cross-validation to obtain the optimal α and β.

[0080] Furthermore, in the method of the present application, in the risk index spatial mapping step, the multidimensional risk index is integrated with the geographic spatial data, and the data gaps are filled through Kriging interpolation to generate a high-resolution gridded risk distribution map, and the risk units are dynamically corrected in combination with real-time IoT sensor data and historical disaster data.

[0081] Furthermore, in the method of the present application, in the scenario comparison and visualization output step, a WebGL engine is used to render a dynamic risk heat map, and real-time traffic flow, population flow information and shelter data are superimposed to perform time-series risk deduction and support multi-scenario risk comparison.

[0082] A second aspect of the present application provides an urban flood risk comparison device based on spatial sensitivity, the device comprising:

[0083] Data collection module: used to comprehensively collect meteorological data, geographic information data, population information and evacuation data, traffic flow data and historical disaster data;

[0084] Risk Modeling Module: This module integrates multi-source heterogeneous data (such as meteorological, topographic, land use, and population density data) to construct a three-dimensional coupled risk assessment model of "hazard-sensitivity-vulnerability." This module dynamically adjusts weight coefficients using machine learning algorithms, introduces physical constraints to optimize surface runoff simulation accuracy, and generates gridded risk cells.

[0085] Risk mapping module: used to fuse multidimensional risk indices with geospatial data to generate gridded risk distribution maps;

[0086] Scenario comparison and visualization output module: used to build a waterlogging risk deduction model under rainfall scenarios with different return periods, realize multi-layer comparative analysis on the same screen, and visualize the analysis results.

[0087] When running, the device implements the steps of the aforementioned urban waterlogging risk comparison method based on spatial sensitivity.

[0088] Furthermore, in the device of the present application, the risk modeling module includes:

[0089] Water ripple hazard calculation unit: used to calculate the water ripple hazard H;

[0090] Spatial sensitivity calculation unit: used to calculate spatial sensitivity K;

[0091] Flood risk index calculation unit: used to calculate the flood risk index R.

[0092] A third aspect of the present application provides an electronic device, comprising: a memory and a processor;

[0093] Memory: used to store computer programs;

[0094] Processor: used to execute the computer program to implement the steps of the aforementioned urban flooding risk comparison method based on spatial sensitivity.

[0095] A fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the aforementioned urban flooding risk comparison method based on spatial sensitivity are implemented.

[0096] In summary, the urban flooding risk comparison method based on spatial sensitivity in this application has the following advantages:

[0097] (1) Dynamically integrate multi-dimensional spatial sensitivity factors to improve the accuracy of risk assessment

[0098] This method achieves multi-dimensional dynamic correction of urban waterlogging risk by integrating spatial sensitivity indicators such as the land use change elasticity coefficient and the terrain relief index, and dynamically adjusting the weights of these factors using a machine learning algorithm. The LightGBM model is used to analyze the nonlinear impact of built-environment factors such as building density and imperviousness, and an explanatory algorithm is used to quantify the contribution of each factor. For example, hydrometeorological factors contribute 24.6%, indicating that vegetation coverage has a significant inhibitory effect.

[0099] (2) Real-time fusion and timely update of multi-source data

[0100] This method integrates data from multiple sources, including IoT sensors, satellite remote sensing, and public reporting, enabling minute-by-minute dynamic data updates. Leveraging a neural network architecture, it predicts flood risk at a 30-meter spatial resolution, significantly improving prediction speed compared to traditional hydrodynamic models. Kriging interpolation fills data gaps, generating a 30-meter gridded risk distribution map. It also supports overlaying and comparing real-time heat maps with historical disaster maps.

[0101] (3) High-resolution spatial overlay analysis and physical constraint enhancement

[0102] The present invention uses a GIS engine to divide the city into 100-meter grids and construct a three-dimensional evaluation system, superimposing urban waterlogging risk factors (such as water depth), sensitivity factors (such as terrain slope), and vulnerability factors (such as population density). The Manning equation and critical flow equation are introduced as physical constraints to improve the physical connectivity of the flood range prediction and reduce the physical disconnection error. In addition, the present invention supports the comparison of rainfall scenarios with different return periods (such as 10-year return period and 50-year return period), providing a scientific basis for emergency path planning and drainage deployment.

[0103] (4) Data-driven decision support

[0104] This invention, centered on extensive and detailed data collection and scientific index calculation, provides powerful, data-driven support for decision-making on urban flood risk prevention and control. Compared to traditional empirical decision-making or qualitative assessment methods, this invention's decision-making based on quantitative data and precise indices is more objective, scientific, and reliable. It can effectively reduce subjectivity and uncertainty in the decision-making process, improve the quality and credibility of decisions, and promote the standardization, scientificization, and modernization of regional risk prevention and control efforts.

[0105] Other features and advantages of this application will be described in detail in the following description, or will be understood through the implementation of the relevant technical solutions of this application. The objectives and other advantages of this application can be achieved through the technical features and technical means clearly indicated in the description, claims, and drawings, and obtained through the implementation of these technical contents. BRIEF DESCRIPTION OF THE DRAWINGS

[0106] To more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings involved in the description of the embodiments. It should be noted that the drawings only illustrate some of the embodiments of the present application. For those skilled in the art, other relevant drawings can be derived from these drawings without engaging in creative work.

[0107] Figure 1 This is the overall implementation flow chart of the urban flooding risk comparison method based on spatial sensitivity in this application.

[0108] Figure 2 This is a flow chart of meteorological data collection in this application method.

[0109] Figure 3 This is a flowchart of geographic information data collection in this application method.

[0110] Figure 4 This is a flowchart for collecting population information and refugee data in this application method.

[0111] Figure 5 This is a flow chart for collecting traffic flow data in this application method.

[0112] Figure 6 This is a flowchart for collecting historical disaster data in this application method.

[0113] Figure 7 This is a structural diagram of the urban flooding risk comparison device based on spatial sensitivity in this application.

[0114] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0115] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0116] In this document, the term "including" and any variations thereof (such as "including", "comprising", etc.) are open expressions and should be understood as "including but not limited to", that is, the listed contents are not exhaustive and may also include other contents not explicitly mentioned. The term "based on" should be understood as "based at least in part on", that is, the basis or condition referred to may not be the only factor, and may also involve other relevant factors. The term "one embodiment" should be understood as "at least one embodiment", that is, the described embodiment is not the only possible implementation method, and there may be other similar embodiments.

[0117] In this application, the terms "a" and "a plurality" are used to modify related elements or features in an illustrative, non-restrictive manner. Unless the context clearly indicates otherwise, "a" should be understood as meaning "at least one," and "a plurality" should be understood as meaning "at least two." Those skilled in the art should interpret these terms appropriately based on the semantics and logical relationships of the context to ensure that they encompass the possibility of "one or more."

[0118] Figure 1 The following is the overall implementation process of the urban flooding risk comparison method based on spatial sensitivity provided in this application, including the following steps:

[0119] S1. Data Collection: Comprehensively collect meteorological data, geographic information data, population information and evacuation data, traffic flow data, and historical disaster data;

[0120] S2. Multi-dimensional Data-Coupled Waterlogging Risk Modeling: This integrates collected multi-source heterogeneous data (such as meteorological, topographic, land use, and population density data) to construct a three-dimensional coupled risk assessment model based on "hazard-sensitivity-vulnerability." This model uses machine learning algorithms to dynamically adjust weight coefficients, introduces physical constraints to optimize surface runoff simulation accuracy, and generates gridded risk cells.

[0121] S3. Spatial mapping of risk indices: Based on GIS (Geographic Information System) technology, multidimensional risk indices are integrated with geospatial data to generate a gridded risk distribution map.

[0122] S4. Scenario Comparison and Visualization Output: Construct a waterlogging risk model for rainfall scenarios with different return periods, conduct a comparative analysis of multiple layers on the same screen using GIS spatial overlay technology, and visualize the analysis results.

[0123] In order to more clearly illustrate the technical solution of the present application, the following will further illustrate it through embodiments of specific scenarios.

[0124] The urban flooding risk comparison method based on spatial sensitivity in this embodiment includes:

[0125] Step 1: Data Collection

[0126] 1. Meteorological data collection, such as Figure 2 As shown:

[0127] (1) Real-time temperature, humidity, rainfall, wind speed, and wind direction data are collected through ground meteorological stations in and around the city, and related spatial grid data are collected; Doppler radar is used to provide precipitation intensity and wind field information, and satellite remote sensing data (such as Fengyun satellites) is used to monitor cloud cover, surface temperature, and water vapor distribution.

[0128] (2) Clean and preprocess the collected data, use Kriging interpolation to deal with missing values in the data, and standardize the data units to form a unified unit.

[0129] (3) Perform spatiotemporal alignment and feature extraction on the standardized data, and combine it with a machine learning model (LSTM) to build a time series database of meteorological data.

[0130] 2. Geographic information data collection, such as Figure 3 As shown:

[0131] (1) The overall urban DEM elevation data, administrative division data, road network data, water system and drainage network data, and land classification (30-meter land utilization rate) data were obtained from the self-regulation and water conservancy departments.

[0132] (2) At the same time, the location information of urban waterlogging points is collected from the city’s three-prevention departments.

[0133] (3) Process geographic information data accordingly, including spatial calibration, topological analysis, spatial overlay, and map service publishing.

[0134] 3. Collection of population information and refuge data, such as Figure 4 As shown:

[0135] (1) Obtain real-time population flow trajectories through mobile operators, complete population density analysis and credibility marking, and use them to analyze population gathering and evacuation paths during heavy rain.

[0136] (2) Integrate the shelter POI data and shelter material data provided by the emergency management department to assess the accessibility of the population when flooding occurs.

[0137] (3) Clean and build the database of data and integrate it into population information and refuge database.

[0138] 4. Traffic flow data collection, such as Figure 5 As shown:

[0139] (1) Connect to the cameras, ground sensors, and electronic police equipment on the urban road network to obtain real-time traffic flow and speed data.

[0140] (2) Clean the data and reconstruct the network topology based on the road network information.

[0141] (3) Output the heat map of traffic congestion (time series database) for the diffusion simulation of urban flooding risk.

[0142] 5. Collection of historical disaster data, such as Figure 6 As shown:

[0143] (1) Connect with the historical records of waterlogging points and disaster investigation reports of water affairs and emergency management departments, road waterlogging monitoring records of traffic management and municipal archives, and drainage facility operation logs.

[0144] (2) Clean the data and remove invalid data.

[0145] (3) Normalization into historical disaster events (event level, occurrence time, and location).

[0146] Step 2: Multi-dimensional data coupling waterlogging risk modeling

[0147] 1. Calculation of water ripple hazard H

[0148] The core of the simulation of waterlogging depth based on the hydrodynamic model is to solve the shallow water equation through numerical methods and dynamically calculate the waterlogging process by combining actual terrain and rainfall data. The specific calculation process is as follows:

[0149] (1) Data preparation and model input

[0150] a) Terrain data preprocessing

[0151] High-precision DEM (digital elevation model) data (e.g., 1:2000 scale) is used, combined with remote sensing images and field surveys, to correct terrain details (e.g., low-lying areas such as overpass areas and sunken roads), and to divide the terrain into unstructured grids (e.g., 10m×10m grids) to adapt to complex urban terrain.

[0152] b) Rainfall data loading

[0153] Input radar-retrieved rainfall or designed rainfall scenarios (such as a 50-year rainstorm), and calibrate rainfall intensity based on meteorological station observation data to reduce data uncertainty.

[0154] (2) Hydrodynamic model construction and equation discretization

[0155] Discretization of shallow water equations:

[0156] (a) Continuity equation (conservation of mass)

[0157]

[0158] Physical meaning: rate of change of water depth over time Water transport in the x and y directions Balance, q represents external source and sink terms (such as rainfall / evaporation);

[0159] (b) Momentum equation in the x direction

[0160] The revised standard form is:

[0161]

[0162] in, is the inertia term, and is the convection term, is the pressure gradient term, g(S 0x -S fx ) indicates that the driving force and the resistance are balanced;

[0163] (c) Momentum equation in the y direction

[0164]

[0165] Symmetry description: The structure is symmetrical with the x-direction equation, and the variables are replaced by the y-direction components (v, S 0y , S fy ),

[0166] Among them, h is the water depth, u is the velocity component of the water flow along the x direction, v is the velocity component of the water flow along the y direction, g is the acceleration of gravity, q is the flow rate per unit width, S 0x is the bottom slope in the x direction, S 0y is the bed slope in the y direction, S fx is the friction slope in the x direction, S fy is the friction slope in the y direction;

[0167] Parameter settings:

[0168] a) Surface roughness (Manning coefficient n) is assigned a value according to land use type (e.g. asphalt pavement n = 0.013, green space n = 0.15).

[0169] b) Boundary conditions: The outlet is set to submerged outflow (the water level boundary needs to be coupled when supported by the tide level)

[0170] (3) Numerical solution and dynamic simulation

[0171] a) Direction Implicit Method (ADI) Solution

[0172] The two-dimensional equations are decomposed into two one-dimensional implicit difference equations, and the water level and flow velocity are calculated by alternating direction iteration.

[0173] The Thomas algorithm is used to efficiently solve tridiagonal matrices, and the time complexity is optimized to O(n).

[0174] b) Time step control

[0175] Limit the time step according to the Courant-Friedrichs-Lewy (CFL) condition

[0176]

[0177] Dynamically adjust the step size to balance computational efficiency and stability.

[0178] (4) Output of water depth index

[0179] Result output: Generate a gridded water depth field, overlay the pipe network overflow data to extract the maximum water depth.

[0180] Output key indicators: flooded area S, peak water depth h, and waterlogging duration t.

[0181] (5) Quantification of hydrological hazard (H)

[0182] Hazard classification: Risk levels are divided according to the water depth threshold (e.g. 15cm triggers an early warning, 60cm is extremely high risk).

[0183] Superimpose the traffic island index and adjust the risk weight (for example, road sections with water accumulation ≥50cm are marked as impassable).

[0184] 2. Spatial sensitivity K calculation

[0185] (1) Land use change sensitivity Sj

[0186] Through land utilization rate data, based on the proportion of non-ecological land area, the formula is:

[0187]

[0188] Among them, S j is the land use change sensitivity index of grid j. The higher the value, the higher the sensitivity of the grid to land use change. ij is the area of the i-th land class in grid j (such as the actual area of cultivated land, forest land or water body in grid j), Aj is the benchmark area of the i-th land class in the study area (such as the total area of this land class in the whole region or the planned target area), Wi is the weight coefficient of the i-th land class, reflecting the contribution of this land class to the overall sensitivity, and i is the land use index (i = 1: cultivated land (Cropland), i = 2: forest land (Forest), i = 3: water body (Waterbody), traversing the three land classes of cultivated land, forest land and water body).

[0189] (2) Terrain slope sensitivity calculation Q adjusted

[0190] It is necessary to combine DEM data with the slope tool of GIS software to output the slope θ:

[0191] Quantify the slope factor Q

[0192] Using soil loss equation

[0193]

[0194] Where Q is the terrain slope sensitivity before adjustment, and θ is the slope percentage.

[0195] This equation is suitable for terrain with short slope length. For special landforms such as plateau areas, it is necessary to correct the terrain sensitivity by terrain roughness. First, the river network density t of DEM data is extracted through GIS software:

[0196]

[0197] Among them, Q adjusted is the adjusted terrain slope sensitivity, t is the river network density;

[0198] (3) Calculation of drainage capacity sensitivity D adjusted

[0199] Drainage capacity reflects the efficiency of a region's regulation of runoff, and is usually analyzed by combining the catchment area and runoff path:

[0200] a) Calculation of runoff path weight

[0201] Calculation of runoff concentration intensity based on unit catchment area

[0202] D=log(A*tanθ+1)

[0203] Where D is the sensitivity of drainage capacity before adjustment, A is the catchment area (m 2 ), θ is the slope angle, which represents the strength of the drainage capacity of the basin.

[0204] b) Drainage facility control coefficient

[0205] If the area includes green infrastructure (such as rain gardens, permeable pavement, etc.), the runoff coefficient in the SWMM model will be further used for weighted regulation:

[0206]

[0207] Among them, D adjusted is the adjusted drainage capacity sensitivity, Wi is the facility type weight, n is the number of facility types, and fi is the facility coverage rate.

[0208] (4) Sensitivity comprehensive weighting method

[0209] Using the analytic hierarchy process (AHP), the system constructs a judgment matrix through expert scoring to determine the sensitivity of land use change Sj and the sensitivity of terrain slope Q adjusted , Drainage capacity sensitivity calculation D adjusted The weight ratio of the three is obtained to obtain weights W1, W2, and W3.

[0210] K=w1*S j +w2*Q adjusted +w3*D adjysted

[0211] 3. Calculation of waterlogging risk index R

[0212] (1) Dynamic coupling

[0213] The dynamic coupling waterlogging risk index dynamically couples the water ripple hazard H and spatial sensitivity K to form a composite risk index R:

[0214] R=α*H+β*K

[0215] Among them, α is the hydrological hazard coupling coefficient, which represents the contribution weight of H to the comprehensive risk R; β is the spatial sensitivity coupling coefficient, which represents the contribution weight of K to the comprehensive risk R, 0≤α≤1 and α+β≤1;

[0216] (2) Adjustment coefficient

[0217] The LightGBM model is used to optimize the coupling coefficients α and β, and the regularization term is used to prevent overfitting. The historical flooding event records (labels) and the H and K indices (features) of the corresponding areas are input, and α = 0.6 and β = 0.3 are set as the initial values of LightGBM training. The coefficient is sparse, and insignificant parameters are automatically removed to prevent α and β from overfitting and ensure the generalization of the model. Finally, the optimal α and β are obtained by averaging after cross-validation.

[0218] Step 3: Risk Index Space Mapping

[0219] (1) Spatial grid processing

[0220] The city is divided into grids with a unit of 100 meters, and the calculated waterlogging risk index R is spatially superimposed to generate a grid risk index R ij .

[0221] (2) Time series risk distribution diagram

[0222] Based on the timeline, the risk diffusion path in the next 1-6 hours is simulated, potential different traffic and pedestrian flow areas are marked, data gaps are filled through Kriging interpolation, and a high-resolution 30-meter risk distribution map is constructed.

[0223] Step 4: Scenario comparison and visualization output

[0224] (1) Sequential risk deduction

[0225] Spatial overlay of the time-series risk distribution map with urban geographic information data: Dynamic risk heat maps are rendered using the WebGL engine, overlaying real-time traffic flow information, population flow information, and shelter data.

[0226] (2) Multi-scenario risk comparison

[0227] The system page has two map interfaces, which simultaneously display two types of rainfall with different recurrence periods: such as the risk difference between rainfall that occurs once in 10 years and rainfall that occurs once in 50 years.

[0228] Figure 7 The present application shows an urban flooding risk comparison device based on spatial sensitivity, comprising:

[0229] Data collection module: used to comprehensively collect meteorological data, geographic information data, population information and evacuation data, traffic flow data and historical disaster data;

[0230] Risk Modeling Module: This module integrates multi-source heterogeneous data (such as meteorological, topographic, land use, and population density data) to construct a three-dimensional coupled risk assessment model of "hazard-sensitivity-vulnerability." This module dynamically adjusts weight coefficients using machine learning algorithms, introduces physical constraints to optimize surface runoff simulation accuracy, and generates gridded risk cells.

[0231] Risk mapping module: used to fuse multidimensional risk indices with geospatial data to generate gridded risk distribution maps;

[0232] Scenario comparison and visualization output module: used to build a waterlogging risk deduction model under rainfall scenarios with different return periods, realize multi-layer comparative analysis on the same screen, and visualize the analysis results.

[0233] When the above device is in operation, the steps of the urban waterlogging risk comparison method based on spatial sensitivity disclosed in this application are implemented.

[0234] The flowcharts and block diagrams in the accompanying drawings illustrate possible implementations of the apparatus, methods, and computer program products according to various embodiments of the present application, including architecture, functions, and operations. In these figures, each box may represent a module, a program segment, or a portion of a code, which contains one or more executable instructions for implementing a specified logical function. It should be noted that each box in the block diagram and / or flowchart, and the combination of these boxes, can be implemented using a dedicated hardware-based system to implement the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0235] like Figure 8 As shown, an embodiment of the present application further discloses an electronic device, comprising: a processor 310, a communication interface 320, a memory 330 for storing a computer program executable by the processor, and a communication bus 340. The processor 310, the communication interface 320, and the memory 330 communicate with each other via the communication bus 340. The processor 310 executes the executable computer program to implement the steps of the above-mentioned urban flooding risk comparison method based on spatial sensitivity.

[0236] It is understood that, in addition to the memory and processor, the electronic device may also include an input device (e.g., a keyboard), an output device (e.g., a display), and other communication modules. These input devices, output devices, and other communication modules all communicate with the processor via an I / O interface (i.e., an input / output interface).

[0237] The operation of the present application can be implemented by writing computer program code using one or more programming languages or a combination thereof. The programming languages include but are not limited to the following types:

[0238] Object-oriented programming languages, such as Java, Smalltalk, C++, etc.;

[0239] A conventional procedural programming language, such as "C" or a similar programming language.

[0240] The execution methods of the program code include but are not limited to:

[0241] Executes entirely on the user's computer;

[0242] Partially executed on the user's computer and partially on a remote computer;

[0243] Executed as a standalone software package;

[0244] Executes entirely on the remote computer or server.

[0245] In scenarios involving remote computers, the remote computer can be connected to the user's computer through any type of network, including but not limited to a local area network (LAN) or a wide area network (WAN). Additionally, the remote computer can also be connected to an external computer through an Internet service provider, such as by using the Internet.

[0246] Furthermore, the present application also discloses a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by the processor of an electronic device, the electronic device can execute the various steps of the urban flooding risk comparison method based on spatial sensitivity disclosed in the present application.

[0247] In the context of this application, computer-readable storage media refers to tangible media that can store computer program code and related data. Specific examples include, but are not limited to, the following:

[0248] (1) Portable computer disk: A removable magnetic storage medium such as a floppy disk.

[0249] (2) Hard disk: includes fixed storage devices such as mechanical hard disk and solid-state hard disk.

[0250] (3) Random Access Memory (RAM): Volatile storage medium used for temporary storage of data and program codes.

[0251] (4) Read-only memory (ROM): A non-volatile storage medium used to store fixed programs and data.

[0252] (5) Erasable Programmable Read-Only Memory (EPROM) or Flash Memory: A non-volatile storage medium that supports multiple erasing and programming.

[0253] (6) Fiber optic storage device: storage medium based on fiber optic technology.

[0254] (7) Compact Disc Read-Only Memory (CD-ROM): A read-only medium that stores data in the form of an optical disc.

[0255] (8) Optical storage devices: such as DVDs, Blu-ray discs and other storage media based on optical principles.

[0256] (9) Magnetic storage devices: storage media based on magnetic principles, such as magnetic tapes and disks.

[0257] (10) Any suitable combination of the above: for example, combining multiple storage media to meet different storage requirements.

[0258] These computer-readable storage media can be used to store the program code and related data described in this application to support the operation of the program and the persistent storage of data.

[0259] In particular, according to embodiments of the present application, the processes described in the flowcharts can be implemented as computer software programs. For example, embodiments of the present application relate to a computer program product comprising a computer program embodied on a non-transitory computer-readable medium. This computer program includes program code for executing the spatial sensitivity-based urban flooding risk comparison method disclosed in this application. When this computer program is executed by a processing device, it can implement the aforementioned functions defined in the embodiments of the present application.

[0260] Although the above discussion contains several specific implementation details, these details should not be interpreted as limiting the scope of this application. The above description is only a preferred embodiment of the present application and an illustration of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in this application is not limited to the technical solutions formed by the specific combination of the above technical features. At the same time, this application should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concepts.

[0261] Those skilled in the art should also understand that they may modify the technical solutions described in the aforementioned embodiments, or replace some of the technical features therein with equivalents, without departing from the spirit and scope of the technical solutions of the embodiments of the present application. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the core spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for comparing urban flooding risks based on spatial sensitivity, characterized in that: The method comprises: Data collection: Comprehensively collect meteorological data, geographic information data, population information and evacuation data, traffic flow data, and historical disaster data; Multi-dimensional data-coupled waterlogging risk modeling: This integrates collected multi-source heterogeneous data to construct a three-dimensional coupled risk assessment model of "hazard-sensitivity-vulnerability." This model uses machine learning algorithms to dynamically adjust weight coefficients, introduces physical constraints to optimize surface runoff simulation accuracy, and generates gridded risk cells. Risk index spatial mapping: Based on GIS technology, multi-dimensional risk index is integrated with geographic spatial data to generate a gridded risk distribution map; Scenario comparison and visualization output: Construct a waterlogging risk simulation model under rainfall scenarios with different return periods, conduct a comparative analysis of multiple layers on the same screen using GIS spatial overlay technology, and visualize the analysis results.

2. The method according to claim 1, characterized in that The data collection step includes: Meteorological data collection: Real-time temperature, humidity, rainfall, wind speed, and wind direction data are collected through ground-based weather stations. Information on precipitation intensity and wind fields is collected from Doppler radar, as well as satellite remote sensing data. Geographic information data collection: obtaining the city's overall DEM elevation data, administrative division data, road network data, water system and drainage network data, land classification data, and location information of urban waterlogging points; Population information and evacuation data collection: obtain real-time population movement trajectories, integrate evacuation point of interest data and evacuation material data; Traffic flow data collection: obtain real-time traffic flow and speed data; Collection of historical disaster data: Obtain historical records of waterlogging points, disaster investigation reports, road waterlogging monitoring records, and drainage facility operation logs.

3. The method according to claim 2, characterized in that The method further comprises: Meteorological data collection: Clean and preprocess the collected meteorological data, use Kriging interpolation to handle missing values in the data, and standardize the data units to form a unified unit. Perform spatiotemporal alignment and feature extraction on the standardized data, and combine it with machine learning models to build a time series database of meteorological data. Geographic information data collection: spatial calibration, topological analysis, spatial overlay, and map service publishing of collected geographic information data; Population information and evacuation data collection: Population density analysis and credibility marking based on real-time population flow trajectories are used to analyze population gathering and evacuation routes during heavy rains. Shelter point of interest data and evacuation material data are used to assess population accessibility during flooding. Traffic flow data collection: The collected traffic flow data is cleaned and combined with road network information to reconstruct the network topology, and a heat map of traffic congestion is output for waterlogging risk diffusion deduction; Collection of historical disaster data: Clean the collected historical disaster data, remove invalid data, and normalize them into historical disaster events.

4. The method according to claim 1, wherein The multi-dimensional data coupled waterlogging risk modeling step includes: (1) The calculation of water ripple hazard H is based on the simulation of water depth using a hydrodynamic model, solving the shallow water equation using numerical methods, and dynamically calculating the waterlogging process in combination with actual terrain and rainfall data, including: Data preparation and model input: preprocess terrain data, divide unstructured grids, and load rainfall data; Hydrodynamic model construction and equation discretization: The shallow water equation is solved by numerical methods. The shallow water equation is: in, is the rate of change of water depth over time, is the water flow transport in the x direction, is the water flow transport in the y direction, and q is the external source and sink term; is the inertia term, and is the convection term, is the pressure gradient term, g(S 0x -S fx ) represents the balance between driving force and resistance; h is the water depth, u is the velocity component of the water flow along the x direction, v is the velocity component of the water flow along the y direction, g is the acceleration of gravity, q is the flow rate per unit width, S 0x is the bottom slope in the x direction, S 0y is the bed slope in the y direction, S fx is the friction slope in the x direction, S fy is the friction slope in the y direction; Numerical solution and dynamic simulation: The system of equations is solved using the direction-varying implicit method, and the time step is dynamically adjusted according to the Courant-Friedrichs-Lewy condition; Waterlogging depth indicator output: Generate a gridded waterlogging depth field, overlay the pipe network overflow data to extract the maximum waterlogging depth; output indicators include flooded area, peak water depth, and waterlogging duration; Quantification of hydrological hazard H: Risk levels are divided according to water depth thresholds, and the risk weight is adjusted by adding the traffic island index; (2) Calculation of spatial sensitivity K, including: (a) Calculation of land use change sensitivity: Among them, S j is the land use change sensitivity index of grid j, A ij is the area of the i-th land class in grid j, Aj is the base area of the i-th land class in the study area, Wi is the weight coefficient of the i-th land class, and i is the land use index; (b) Terrain slope sensitivity calculation: Using soil loss equation Where Q is the terrain slope sensitivity before adjustment, and θ is the slope percentage; And perform terrain roughness correction: Among them, Q adjusted is the adjusted terrain slope sensitivity, t is the river network density; (c) Drainage capacity sensitivity calculation: D=log(A*tanθ+1) Where D is the drainage capacity sensitivity before adjustment, A is the catchment area, and θ is the slope angle; And carry out drainage facility regulation: Among them, D adjusted is the adjusted drainage capacity sensitivity, Wi is the facility type weight, n is the number of facility types, and fi is the facility coverage rate; (d) Comprehensive sensitivity weighting: K=w1*S j +w2*Q adjusted +w3*D adjusted Among them, W1, W2, and W3 are respectively the land use change sensitivity Sj, terrain slope sensitivity Qj determined by the hierarchical analysis method adjusted , drainage capacity sensitivity D adjusted The weight of (3) Calculation of the waterlogging risk index R, including: The water ripple hazard H and spatial sensitivity K are dynamically coupled to form the waterlogging risk index R: R=α*H+β*K Among them, α is the hydrological hazard coupling coefficient, which represents the contribution weight of H to the comprehensive risk R; β is the spatial sensitivity coupling coefficient, which represents the contribution weight of K to the comprehensive risk R, 0≤α≤1 and α+β≤1; Adjustment coefficient: The LightGBM model is used to optimize the coupling coefficients α and β, and the regularization term is used to prevent overfitting. The historical records of urban flooding events and the H and K indices of the corresponding areas are input. α = 0.6 and β = 0.3 are set as the initial values of LightGBM training. The coefficient is sparse, and insignificant parameters are automatically removed. Finally, the optimal α and β are obtained by averaging after cross-validation.

5. The method according to claim 1, wherein In the risk index spatial mapping step, the multidimensional risk index is integrated with the geographic spatial data, and the data gaps are filled through Kriging interpolation to generate a high-resolution gridded risk distribution map. The risk units are dynamically corrected in combination with real-time IoT sensor data and historical disaster data.

6. The method according to claim 1, characterized in that In the scenario comparison and visualization output steps, a WebGL engine is used to render a dynamic risk heat map, and real-time traffic flow, population flow information and shelter data are superimposed to perform time-series risk deduction and support multi-scenario risk comparison.

7. A device for comparing urban flooding risks based on spatial sensitivity, characterized in that: The device comprises: Data collection module: used to comprehensively collect meteorological data, geographic information data, population information and refuge data, traffic flow data and historical disaster data; Risk Modeling Module: This module integrates multi-source heterogeneous data to construct a three-dimensional coupled risk assessment model of "hazard-sensitivity-vulnerability." It uses machine learning algorithms to dynamically adjust weight coefficients, introduces physical constraints to optimize surface runoff simulation accuracy, and generates gridded risk cells. Risk mapping module: used to fuse multidimensional risk indices with geospatial data to generate gridded risk distribution maps; Scenario comparison and visualization output module: used to build waterlogging risk deduction models under rainfall scenarios with different return periods, realize multi-layer comparative analysis on the same screen, and visualize the analysis results.

8. The device according to claim 7, characterized in that The risk modeling module includes: Water ripple hazard calculation unit: used to calculate the water ripple hazard H; Spatial sensitivity calculation unit: used to calculate spatial sensitivity K; Flood risk index calculation unit: used to calculate the flood risk index R.

9. An electronic device, characterized in that: include: memory and processor; Memory: used to store computer programs; Processor: used to execute the computer program to implement the steps of the urban flooding risk comparison method based on spatial sensitivity as described in any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the urban flooding risk comparison method based on spatial sensitivity are implemented as described in any one of claims 1 to 6.

Images