River overflow flood routing rapid simulation method based on river adaptive segmentation and strong water balance constraint

Abstract

The invention discloses a river spill flood routing rapid simulation method based on river adaptive segmentation and strong water balance constraint, and the method comprises the steps: collecting and processing basic data of a drainage basin, building a rapid calculation model of river spill flood inundation, and carrying out the rapid prediction and risk analysis of the river spill flood. Rapid and fine simulation of watershed river overflow flood is realized, four-pre requirements of watershed flood control fine simulation, real-time early warning and forecasting, rapid scene deduction and the like are met, and the method can also serve watershed flood control planning and design.

Description

Technical Field

[0001] The present invention relates to the technical field of river basin flood warning and forecasting, and in particular to a method for rapidly simulating river overflow flood evolution based on river adaptive segmentation and strong water balance constraints. Background Art

[0002] River overflow floods are sudden, highly disastrous, and have a wide impact range. For flood disaster emergency response, the rapid generation of dynamic flood inundation risk maps and the early issuance of risk avoidance warnings are urgent needs to reduce casualties and disaster losses. Traditional river overflow flood simulations are based on hydrodynamic models. When solving the calculation, grid cells are divided to discretize and generalize the terrain. The number of grid cells is generally proportional to the simulation accuracy. The more grid cells there are, the smaller the generalization error of the terrain and the higher the calculation accuracy. However, as the number of grids increases, the computational complexity also increases, consuming a lot of time and computing resources, resulting in insufficient computational timeliness and making it difficult to support real-time simulation and prediction of flood inundation. In addition, hydrodynamic modeling relies on a large amount of underlying surface data such as large river sections. Basins with insufficient or no data cannot be modeled, which also restricts the realization of flood inundation simulation and prediction.

[0003] Currently, scholars at home and abroad are committed to exploring lightweight and fast methods. One method is based on a data-driven deep learning model, which generates training data through a hydrodynamic model, and then uses the deep learning model to construct a mapping relationship between different water inflows and flooding results in the upper reaches of the river, thereby conducting a rapid analysis of flood inundation. This method has high computational efficiency, but relies on hydrodynamic model simulation to generate a large amount of sample data for training. It still faces problems such as difficulty in modeling, high training costs, and insufficient interpretability of physical mechanisms, especially when applied to large-scale watersheds. Another method is to improve computational efficiency by reducing model complexity, such as the Height Above Nearest Drainage (HAND) model and the GeoFlood model. The computational efficiency is greatly improved compared to the hydrodynamic model, and the physical meaning of the model is clear. The key technology lies in: (1) reasonably dividing the length and number of river sections and conducting river segment modeling; (2) converting the flow of river segments into segment water depths, and dynamically calculating the flooding situation by comparing the relationship between segment water depths and flooding thresholds. However, current technology still has shortcomings. First, river segment modeling uses specified segment lengths, which is highly subjective. Second, the conversion of segment flow into segment depth based on the segment water level-flow relationship curve lacks water balance constraints, resulting in inundation simulation results that are too large in flooding sections and too small in receding sections. These issues make it impossible to accurately simulate the dynamic evolution of river overflow floods, making it insufficient to support the creation of detailed, real-time dynamic risk maps. Summary of the Invention

[0004] The purpose of the present invention is to provide a rapid simulation method for river overflow flood evolution based on river adaptive segmentation and strong water balance constraints, so as to solve the above-mentioned problems existing in the prior art.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A fast simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints, including:

[0007] Construction of a rapid calculation model for river overflow flood inundation: Based on the processed basin elevation data, a HAND matrix composed of HAND values ​​that characterize the correlation between the basin slope units and the river is calculated. A clustering algorithm is used to perform cluster analysis on the ROI values ​​that characterize the inundation capacity of the river catchment area to achieve adaptive river segmentation; the total river water storage in each time period is calculated based on the water balance principle, and the flood inundation water volume of each river section catchment area is calculated based on the water allocation coefficient AC; the river channel water depth-water volume relationship curve of each river section catchment area is constructed, and the river section flood inundation water volume in each time period is converted into the section river channel water depth in each time period. By comparing the size relationship between the section river channel depth and the slope HAND value, the flood inundation water depth of each section river section catchment area is dynamically calculated, thereby drawing a dynamic flood inundation map, and finally realizing the construction of a rapid calculation model for river overflow flood inundation.

[0008] Preferably, the construction of a rapid calculation model for river overflow flooding specifically includes the following contents:

[0009] S11. Generation of river network data in the watershed: Use the hydrological analysis toolbox of GIS software to perform depression filling calculation, flow direction calculation, pixel cumulative flow calculation, river linking, river network classification, and raster river network vectorization operations on the processed watershed elevation data to obtain the watershed river network distribution data;

[0010] S12, HAND matrix calculation: Calculate the elevation difference between the slope grid point in the watershed and its nearest confluence grid point, i.e., the HAND value, to obtain the HAND matrix of the watershed;

[0011] S13. ROI-based adaptive river segmentation: Calculate the ROI value of the river's overtopping capacity for each grid. Use the Bisecting K-means clustering algorithm to perform cluster analysis on the ROI values ​​in the order from upstream to downstream of the river. Continuous river grids with similar inundation capacity are divided into one segment, thus forming a continuous river section.

[0012] S14. Dynamic calculation of flood inundation volume in river section catchments based on strong water balance constraints: Calculate the flood volume per unit time within the river section catchment based on the water balance principle. Construct a water allocation coefficient AC based on the topographic characteristics of the river section catchment area. Use the AC to allocate the total river water volume to each river section, and obtain the flood inundation volume of each river section outward from the river channel.

[0013] S15. Dynamic mapping of river overbank floods: Based on the constructed segmented river channel water depth-water volume relationship curve, the flood inundation water volume of each river section is converted into segmented river channel water depth. When the segmented water depth is greater than the slope HAND value, the flood inundation depth of the river section is the difference between the segmented water depth and the slope HAND value. Otherwise, the flood inundation depth of the river section is 0. In this way, the flood inundation depth of each river section into the river catchment area is obtained, and a dynamic flood inundation map is drawn to realize the construction of a rapid calculation model for river channel overbank flood inundation.

[0014] Preferably, the calculation formula of the ROI value is:

[0015]

[0016] Where D0 is the constructed river equidistant water depth sequence; is the maximum possible water depth of the river; n is the number of water depth values ​​in the water depth sequence D0; is the i-th water depth value in the water depth sequence D0; V i For the river depth The amount of water overflowing the river channel at that time; F is the area of ​​a single grid; N i For the river depth The number of grids that overflow and flood outside the river channel.

[0017] Preferably, in step S14,

[0018] The calculation formula for flood volume per unit time period in a river catchment area is:

[0019] W T =W T-1 +W T,I -W T,O (2)

[0020]

[0021] Among them, W T is the flood volume at time T; W T-1 is the flood volume at time T-1; W T,I is the river inflow at time T; W T,Ois the river outflow at time T; Q(t) is the measured or predicted flow sequence data of the upstream hydrological station; n0 is the roughness value; A is the cross-sectional area of ​​the most downstream river channel; R is the hydraulic radius of the most downstream river channel section; S is the gradient of the most downstream river section; Δt is the time difference;

[0022] The calculation formula of the water distribution coefficient of each river section is:

[0023]

[0024] Among them, IEF k is the inundation range index of the k-th river section; D is the constructed equidistant water depth sequence of the k-th river section, D n is the maximum possible water depth of the k-th river section; m represents the number of water depth values ​​in the water depth sequence D; N i Indicates that the water depth in the river is D i When , the number of slope grids that are submerged outside the river channel in the kth river section; m0 is the number of river sections;

[0025] The calculation formula for the flood water volume of each river section outward is:

[0026] V k =W T AC k (7)

[0027] Among them, V k is the flood inundation volume of the k-th river catchment area.

[0028] Preferably, in step S15, the process of constructing the segmented water depth-water volume relationship curve is as follows:

[0029] According to the equidistant water depth sequence D = [0,…D i-1 ,D i ,D i+1 …,D m ], calculate the flood water volume corresponding to each water depth, the formula is:

[0030]

[0031] Among them, V i The river depth is D i When , the flood water volume of the catchment area of ​​the river section is inundated; n′ is the number of grid cells in the catchment area of ​​the river section; V i,j When the river depth is D i When , the flood water volume of the jth grid unit in the catchment area of ​​the river section is; H j is the HAND value of the jth grid cell; F j is the area of ​​the jth grid cell;

[0032] Based on the equidistant water depth sequence of the river section D=[0,…D i-1 ,D i ,D i+1 …,D m ] and the corresponding water volume sequence V=[0,…V i-1 ,V i ,V i+1 …,V m ], we can get the river section water depth-water volume relationship curve.

[0033] Preferably, between step S11 and step S12, the following is further included:

[0034] The calculated river network distribution data of the basin was corrected using Tiandi Map remote sensing images to make the river distribution consistent with the actual location.

[0035] Preferably, after the rapid calculation model of river overflow flooding is constructed, the method further includes:

[0036] Rapid prediction and risk analysis of river overflow floods: Based on the constructed rapid analysis model for river overflow floods and access to future hydrological forecast data, the development and evolution of river floods in the basin are predicted in real time, and the dynamic inundation range and water depth of river overflow floods in the basin are obtained; flood inundation risk indicators are established based on the dynamic inundation range and water depth of river overflow floods in the basin, and a dynamic flood risk map of the basin is drawn based on the flood inundation risk indicators.

[0037] Preferably, the rapid prediction and risk analysis of river overflow floods specifically includes the following:

[0038] S21. River Basin Flood Inundation Prediction: Based on the flow forecast results of upstream river sections, a rapid analysis model for river overflow flood inundation is driven to predict the flood inundation range and water depth data of the river basin in real time;

[0039] S22. Construction of flood inundation risk indicators: Taking into account the threat posed by flood inundation depths to the safety of life and property, as well as social and economic security, as well as the standards for compiling flood risk maps, for extreme flood events, flood inundation depths of 0.5m, 1m, 2m, and 3m are set as indicators of low, medium, high, and very high flood inundation risks in the basin, respectively.

[0040] S23. Real-time flood risk prediction and risk analysis of key protected objects: Based on real-time inundation prediction data and combined with flood inundation risk indicators, a dynamic flood risk map is drawn; geographic spatial information layers such as flood control projects, lifeline projects, and key protected objects are superimposed on the dynamic flood risk map to draw a thematic risk map.

[0041] Preferably, before constructing the rapid calculation model for river overflow flooding, the method further includes:

[0042] Collection and processing of basic watershed data: Collection and processing of high-resolution digital elevation data, river embankment data and hydrological data of the watershed.

[0043] Preferably, the collection and processing of basin basic data specifically includes the following contents:

[0044] S01. Download meter-level digital elevation data (DEM) of the basin through a map downloader, with a spatial resolution of 30m, 12.5m, or 5m. Collect river embankment elevation data in the basin and modify it based on the digital elevation data to ensure that the river channel and embankment elevation data are more consistent with the actual topography of the basin.

[0045] S02. Collect measured flow data from river hydrological stations in the basin and unify the flow data time interval to 1 hour.

[0046] The beneficial effects of the present invention are as follows: 1. The present invention constructs a river overtopping and flooding capacity index to characterize the flooding potential of a river catchment area, and uses a cluster analysis method to achieve adaptive segmentation modeling of the river. Compared to traditional manual segmentation methods, the adaptive segmentation method based on the river overtopping and flooding capacity index provides a more scientific and convenient segmentation modeling method, which can eliminate subjective errors caused by human factors, ensure the objectivity and rationality of the segmentation results, and significantly improve the efficiency of river segmentation. In addition, the traditional method converts the segment flow rate into the river depth based on the river section water level-flow curve. This curve is a single relationship and cannot depict the actual rope-shaped non-single relationship affected by flood fluctuations. The water balance constraint is insufficient, resulting in large errors in the flood simulation results and an inability to accurately simulate the dynamic evolution of floods. The water volume-water depth curve proposed in the present invention and the dynamic calculation method of flood inundation water volume in river section catchments based on strong water balance constraints improve the accuracy of flood dynamic evolution simulation and can accurately simulate the full spatiotemporal dynamic process of floods from rising to receding. 2. Unlike traditional hydrodynamic models, which are primarily used for small-scale, refined simulations, flood control planning, and design, the proposed rapid simulation method for river overtopping flood evolution, based on adaptive river segmentation and strong water balance constraints, not only meets the needs of large-scale, refined simulations, flood control planning, and design, but can also be applied to real-time early warning and rapid simulation of river basin floods. While maintaining similar accuracy to hydrodynamic models, the proposed method improves computational efficiency to seconds, significantly improving computational timeliness. Furthermore, the proposed method requires only upstream cross-sectional flow data and DEM data for rapid and lightweight modeling. Compared to hydrodynamic models, its reliance on modeling data is significantly reduced, providing a novel solution for simulating flood inundation and evolution in data-scarce or data-free areas. Furthermore, unlike deep learning-based inundation calculation methods, the proposed method does not require the costly training of large numbers of samples, and the model's physical meaning is clear, making modeling more convenient and lightweight. Combined with flood inundation depth risk indicators, it can support relevant departments in publishing real-time dynamic flood risk maps, as well as thematic risk maps for river basin flood control projects, lifeline projects, and key protected areas, providing technical support for emergency response to sudden flood disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flow chart of a simulation method according to an embodiment of the present invention;

[0048] Figure 2 Schematic diagram of adaptive river segmentation in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0050] In view of the current technical difficulties of insufficient timeliness of classical dynamic models, complex modeling of deep learning models, and inability of simplified models to simulate accurately, we focus on the pain points and difficult problems faced by lightweight and fast models, and fully explore and integrate the advantages of topography, hydrology, big data technology and other methods. In this embodiment, a rapid simulation method for river overflow flood evolution based on river adaptive segmentation and strong water balance constraints is provided to achieve rapid and precise simulation of river overflow floods in the basin, meet the "four prediction" requirements of basin flood control precise simulation, real-time warning and forecast, rapid scenario deduction, etc., and can also serve the basin flood control planning and design. Figure 1 As shown, the method of the present invention mainly includes the following three parts:

[0051] 1. Collection and processing of basic watershed data

[0052] Collect and process high-resolution digital elevation data (DEM), river embankment data and hydrological data of the basin (study area) to provide a data basis for constructing a rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints.

[0053] 1.1. Collection and processing of watershed elevation data

[0054] Download meter-level digital elevation data (DEM) using a map downloader like BIGEMAP, with spatial resolutions of 30m, 12.5m, and 5m. Collect river levee crest elevation data and modify it based on the DEM data to ensure that the river channel and levee elevation data better match the actual basin topography.

[0055] 1.2 Hydrological data collection and processing

[0056] Collect the measured flow data from river hydrological stations and unify the flow data time interval to 1 hour.

[0057] 2. Construction of a rapid calculation model for river overflow flooding

[0058] Based on the corrected basin elevation data from step 1.1, generate basin river network distribution data and calculate the HeightAboveNearestDrainage matrix (hereinafter referred to as the "HAND matrix"). Use the HAND matrix to construct a physical characteristic index that characterizes the flooding capacity of the river grid unit, and use a clustering algorithm to perform cluster analysis on the index to perform adaptive river segmentation. Based on the water balance principle and Manning's formula, calculate the total water storage of the river in each time period, and construct the water distribution coefficient according to the topographic characteristics of the river section catchment area to calculate the flood inundation volume of each river section catchment area. Construct a river channel water depth-water volume relationship curve for each river section catchment area, convert the flood inundation volume of the river section (river section) in each time period into the segmented river channel water depth in each time period, and finally dynamically calculate the flood inundation depth of the grid outside each segmented river channel by comparing the relationship between the segmented river channel water depth and the inundation threshold, thereby drawing a dynamic flood inundation map.

[0059] 2.1. Generation of river network data

[0060] Use the hydrological analysis toolbox of GIS software to perform operations such as depression filling calculation, flow direction calculation, pixel cumulative flow calculation, river linking, river network classification, and raster river network vectorization on the DEM data corrected in step 1.1 to obtain the basin river network distribution data.

[0061] At the same time, in order to improve the accuracy of river network data, Tiandi Map remote sensing images are used to correct the calculated river network data so that the river distribution is consistent with the actual location.

[0062] 2.2 HAND matrix calculation

[0063] The HAND value is the elevation difference between a slope grid point and its nearest confluence grid point within a watershed. It is a physical indicator that characterizes the relationship between a slope unit and a river in the watershed. By traversing and calculating the HAND value of each grid point in the DEM grid of the study area, a HAND matrix with the same dimension as the DEM data can be obtained, where the HAND value of the river grid is 0. Based on the assumption of vertical water pressure balance in the slope confluence path and the assumption that there is no lateral flow in the confluence path, the HAND matrix and river water depth data can be used to identify the water supply inundation range and flood inundation depth. That is, the area where the HAND value of the slope grid is lower than the corresponding river grid water depth is the inundation area, and the corresponding flood inundation depth is the difference between the river grid water depth and the slope grid HAND value.

[0064] 2.3 ROI-based adaptive river segmentation

[0065] Determining the grid depth values ​​for the river channel in step 2.2 is crucial for calculating flood inundation range and depth. Due to the lack of grid-by-grid depth data for river channels, the average depth of each segment is typically used when applying the HAND method for inundation calculations. Artificial river segment division is subjective and uncertain, and the length of river segments can affect model simulation accuracy. Excessively long segments can over-homogenize the river channel topography, while too short segments can over-amplify local errors.

[0066] The present invention adopts ROI-based river adaptive segmentation technology, that is, automatically segmenting the river (river channel) according to the terrain characteristics, merging the areas with similar terrain in the river catchment area into one section. This method ensures the efficiency and rationality of segmentation. The principle diagram of the method is shown in the figure below. Figure 2 The River Overflow Index (ROI) is used to characterize the flooding potential of a river catchment area. The calculation method is as follows:

[0067]

[0068] Where D0 is the constructed equidistant water depth sequence of the river channel, represents the maximum possible water depth of the river (m), which can be obtained through flood surveys or flood frequency calculations; n represents the number of water depth values ​​in the water depth sequence D0; represents the i-th water depth value in the water depth sequence D0 (m); V i Indicates that the water depth in the river is When the water level is 0.05, the amount of water that overflows from the river channel (m3); F represents the area of ​​a single grid (m2); N i Indicates that the water depth in the river is When , the number of grids that overflow and submerge outside the river channel (in units).

[0069] ROI reflects the comprehensive flooding capacity of a river to overflow outside the river channel under different water depth scenarios, with a maximum value of 1. The closer the ROI value is to 1, the stronger the river's flooding capacity is.

[0070] The ROI value of each grid of the river was calculated, and the Bisecting K-means clustering algorithm was used to perform cluster analysis on the ROI values ​​in the order from upstream to downstream of the river. The continuous river grids with similar flooding capacity were divided into one section, thus forming a continuous river section.

[0071] 2.4 Dynamic calculation of flood inundation volume in river catchments based on strong water balance constraints

[0072] Based on the principle of water balance, the flood volume per unit time period in a river catchment area can be calculated by the following formula:

[0073] W T =WT-1 +W T,I -W T,O (2)

[0074] Where W T Indicates the flood volume at time T (m 3 );W T-1 Indicates the flood volume at time T-1 (m 3 );W T,I represents the river inflow at time T (m 3 );W T,O represents the river outflow at time T (m 3 ).

[0075] W T,I It can be calculated based on the flow data of the upstream hydrological station using the following formula:

[0076]

[0077] Where Q(t) represents the measured or predicted flow sequence data of the upstream hydrological station (m 3 / s).

[0078] W T,O It can be calculated according to Manning's formula:

[0079]

[0080] Where n is the roughness value; A is the cross-sectional area of ​​the most downstream river channel; R is the hydraulic radius of the most downstream river channel section; S is the gradient of the most downstream river section; Δt is the time difference.

[0081] In formula 4, A and R are calculated based on the river depth at the most downstream section at time T-1. The water depth is the flood volume (W) of the river section at time T-1 calculated using formula (2). T-1 ), obtained by interpolating the depth-water volume relationship curve of the segmented river channel.

[0082] The total flood volume of the river (W T ) After that, use the water allocation coefficient AC (Allocation Coefficient) to allocate the total water volume of the river to each river section divided in step 2.3. The water allocation coefficient AC of the kth river section is k The calculation formula is as follows:

[0083]

[0084] In the formula, IEF k is the inundation extent factor of the k-th river section; D is the equidistant water depth sequence constructed for the k-th river section (common to all river sections), Dn represents the maximum possible water depth of the river channel at the kth river section (m), which can be obtained through flood surveys or flood frequency calculations; m represents the number of water depth values ​​in the water depth sequence D; N i Indicates that the water depth in the river is D i When , the number of slope grids that are submerged outside the river channel in the kth river section is m0; m0 represents the number of river sections.

[0085] After calculating the water distribution coefficient of each river section, the flood water volume V that the kth river section floods out of the river channel is k The calculation formula is as follows:

[0086] V k =W T AC k (7)

[0087] 2.5. Dynamic mapping of river bank overflow floods

[0088] To draw a flood inundation map for each river section, the flood inundation volume for each river section calculated in step 2.4 needs to be converted into segmented river channel depths. Then, by comparing the segmented river channel depths with the slope HAND value (i.e., the inundation threshold of each slope grid), the flood inundation depth of each river section toward the slope grid (catchment area) outside the river channel is dynamically calculated, thereby drawing a dynamic flood inundation map.

[0089] The section-wise river depth-water volume relationship curve is used to convert the flood inundation volume of the river catchment area into the section-wise river depth. The construction method of the section-wise river depth-water volume relationship curve is as follows:

[0090] For each river section, construct an equidistant water depth sequence D = [0,…D i-1 ,D i ,D i+1 …,D m ], and then calculate the flood water volume corresponding to each water depth, the formula is as follows:

[0091]

[0092] Where V i Indicates that the water depth of the river section is D i When the flood water volume in the catchment area of ​​this river section is (m 3 ); n′ represents the number of grid cells in the catchment area; V i,j Indicates that when the river depth is D i When the flood water volume of the jth grid unit in the catchment area of ​​the river section is (m 3 );H j represents the HAND value of the jth grid cell (m); F j Represents the area of ​​the jth grid cell (m2 ).

[0093] Based on the equidistant water depth sequence of the river section D=[0,…D i-1 ,D i ,D i+1 …,D m ] and the corresponding water volume sequence V=[0,…V i-1 ,V i ,V i+1 …,V m ], we can get the river section water depth-water volume relationship curve.

[0094] Using the above method, a water depth-water volume relationship curve is constructed for each river section. Formulas (2) and (7) are used to calculate the amount of water overflowing and inundating the river channel at time T for each river section. This water depth-water volume relationship curve is then converted into the average water depth of each river section at time T. Formulas (2) and (7) are then iterated to obtain the average water depth of each river section at each moment. Then, by comparing the average water depth of each river section with the HAND value (i.e., the inundation threshold of the slope grid) of the corresponding catchment area at each moment, the flood inundation depth of each river section into the slope grid (catchment area) outside the river channel is dynamically calculated, thereby drawing a dynamic flood inundation map.

[0095] The calculation formula for flood inundation depth is as follows:

[0096]

[0097] Where Z i represents the flood depth of grid cell i (m); D represents the river depth (m); H i Represents the HAND value (m) of grid cell i.

[0098] 3. Rapid prediction and risk analysis of river overflow floods

[0099] Based on the constructed rapid analysis model for river overflow flooding, and incorporating future hydrological forecast data (upstream cross-section flow), the development and evolution of river floods in the basin can be predicted in real time, and the dynamic inundation range and depth of river overflow floods in the basin can be determined. Risk indicators can be established, and a dynamic flood risk map for the basin can be drawn.

[0100] 3.1. Basin flood inundation prediction

[0101] Based on the flow forecast results of the upstream section of the river, the rapid inundation analysis model constructed in step 2 is driven to predict the flood inundation range and water depth data of the basin in real time.

[0102] 3.2 Construction of flood risk indicators

[0103] With full reference to the threat degree of flood inundation depth to the safety of life and property of the people and social and economic security, and the standards for compiling flood risk maps, for extreme flood events, flood inundation depths of 0.5m, 1m, 2m and 3m are set as indicators of low risk, medium risk, high risk and extremely high risk of flood in the basin respectively.

[0104] 3.3 Real-time flood risk prediction and risk analysis of key protected areas

[0105] Based on real-time inundation forecast data and combined with the aforementioned flood risk indicators, a dynamic flood risk map is created. Geospatial information layers, such as those for flood control projects, lifeline projects, and key protected areas, are overlaid on this risk map to create thematic risk maps, providing technical support for emergency response to sudden flood disasters.

[0106] By adopting the above technical solution disclosed in the present invention, the following beneficial effects are obtained:

[0107] This invention provides a rapid simulation method for river overtopping flood inundation based on adaptive river segmentation and strong water balance constraints. This method constructs a river overtopping flood capacity index to characterize the flooding potential of a river catchment area and uses cluster analysis to achieve adaptive segmented river modeling. Compared to traditional manual segmentation methods, the adaptive segmentation method based on the river overtopping flood capacity index provides a more scientific and convenient segmented modeling approach, eliminating subjective errors caused by human factors, ensuring the objectivity and rationality of segmentation results, and significantly improving river segmentation efficiency. Furthermore, traditional methods convert segmented discharges into segment depths based on the river level-discharge curve. This curve is a single relationship and cannot capture the actual non-singular, loop-like relationships affected by flood fluctuations. This insufficient water balance constraint leads to large errors in inundation simulation results and an inability to accurately simulate the dynamic evolution of floods. The proposed water volume-depth curve and the dynamic calculation method for flood inundation in river catchments based on strong water balance constraints improve the accuracy of flood dynamic simulations and can accurately simulate the full spatiotemporal dynamics of floods from rising to receding. Unlike traditional hydrodynamic models, which are primarily used for small-scale, refined simulations, flood control planning, and design, the proposed rapid simulation method for river overtopping flood evolution, based on adaptive river segmentation and strong water balance constraints, not only meets the needs of large-scale, refined simulations, flood control planning, and design, but can also be applied to real-time early warning and rapid simulation of river basin floods. While maintaining similar accuracy to hydrodynamic models, the proposed method improves computational efficiency to seconds, significantly improving computational timeliness. Furthermore, the proposed method requires only upstream cross-sectional flow data and DEM data for rapid and lightweight modeling. Compared to hydrodynamic models, its reliance on modeling data is significantly reduced, providing a novel solution for simulating flood inundation and evolution in data-scarce or data-free areas. Furthermore, unlike deep learning-based inundation calculation methods, the proposed method does not require costly training with large numbers of samples, and the model's physical meaning is clear, making modeling more convenient and lightweight. Combined with flood inundation depth risk indicators, it can support relevant departments in publishing real-time dynamic flood risk maps, as well as thematic risk maps for river basin flood control projects, lifeline projects, and key protected areas, providing technical support for emergency response to sudden flood disasters.

[0108] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints, characterized by: include, Construction of a rapid calculation model for river overflow flooding: Based on processed basin elevation data, a HAND matrix consisting of HAND values ​​representing the relationship between basin slope units and rivers is calculated. A clustering algorithm is used to perform cluster analysis on the ROI values ​​representing the flooding capacity of the river catchment area to achieve adaptive river segmentation. The total water storage of the river in each time period is calculated based on the water balance principle, and the flood inundation water volume of each river section catchment area is calculated based on the water allocation coefficient AC; the river channel water depth-water volume relationship curve of each river section catchment area is constructed, and the river section flood inundation water volume in each time period is converted into the section river channel water depth in each time period. By comparing the size relationship between the section river channel water depth and the slope HAND value, the flood inundation water depth of each section river section catchment area is dynamically calculated, thereby drawing a dynamic flood inundation map, and finally realizing the construction of a rapid calculation model for river overflow flood inundation.

2. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to claim 1 is characterized by: The construction of a rapid calculation model for river overflow flood inundation specifically includes the following: S11. Generation of river network data in the watershed: Use the hydrological analysis toolbox of GIS software to perform depression filling calculation, flow direction calculation, pixel cumulative flow calculation, river linking, river network classification, and raster river network vectorization operations on the processed watershed elevation data to obtain the watershed river network distribution data; S12, HAND matrix calculation: Calculate the elevation difference between the slope grid point in the watershed and its nearest confluence grid point, i.e., the HAND value, to obtain the HAND matrix of the watershed; S13. ROI-based adaptive river segmentation: Calculate the ROI value of the river's overtopping capacity for each grid. Use the Bisecting K-means clustering algorithm to perform cluster analysis on the ROI values ​​in the order from upstream to downstream of the river. Continuous river grids with similar inundation capacity are divided into one segment, thus forming a continuous river section. S14. Dynamic calculation of flood inundation volume in river section catchments based on strong water balance constraints: Calculate the flood volume per unit time within the river section catchment based on the water balance principle. Construct a water allocation coefficient AC based on the topographic characteristics of the river section catchment area. Use the AC to allocate the total river water volume to each river section, and obtain the flood inundation volume of each river section outward from the river channel. S15. Dynamic mapping of river overbank floods: Based on the constructed segmented river channel water depth-water volume relationship curve, the flood inundation water volume of each river section is converted into segmented river channel water depth. When the segmented water depth is greater than the slope HAND value, the flood inundation depth of the river section is the difference between the segmented water depth and the slope HAND value. Otherwise, the flood inundation depth of the river section is 0. In this way, the flood inundation depth of each river section into the river catchment area is obtained, and a dynamic flood inundation map is drawn to realize the construction of a rapid calculation model for river channel overbank flood inundation.

3. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to claim 2 is characterized by: The calculation formula of the ROI value is: Where D0 is the constructed river equidistant water depth sequence; is the maximum possible water depth of the river; n is the number of water depth values ​​in the water depth sequence D0; is the i-th water depth value in the water depth sequence D0; V i For the river depth The amount of water overflowing the river channel at that time; F is the area of ​​a single grid; N i For the river depth The number of grids that overflow and flood outside the river channel.

4. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to claim 2 is characterized by: In step S14, The calculation formula for flood volume per unit time period in a river catchment area is: IN T =In T-1 +W T,I -IN T,O (2) Among them, W T is the flood volume at time T; W T-1 is the flood volume at time T-1; W T,I is the river inflow at time T; W T,o is the river outflow at time T; Q(t) is the measured or predicted flow sequence data of the upstream hydrological station; n0 is the roughness value; A is the cross-sectional area of ​​the most downstream river channel; R is the hydraulic radius of the most downstream river channel section; S is the gradient of the most downstream river section; Δt is the time difference; The calculation formula for the water distribution coefficient of each river section is: Among them, IEF k is the inundation range index of the k-th river section; D is the constructed equidistant water depth sequence of the k-th river section, D n is the maximum possible water depth of the k-th river section; m represents the number of water depth values ​​in the water depth sequence D; N i Indicates that the water depth in the river is D i When , the number of slope grids that are submerged outside the river channel in the kth river section; m0 is the number of river sections; The calculation formula for the flood water volume of each river section outward is: V k =W T ·AM k (7) Among them, V k is the flood inundation volume of the k-th river catchment area.

5. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to claim 2 is characterized by: In step S15, the process of constructing the segmented water depth-water volume relationship curve is as follows: According to the equidistant water depth sequence D = [0,…D i-1 ,D i ,D i+1 …,D m ], calculate the flood water volume corresponding to each water depth, the formula is: Among them, V i The river depth is D i When , the flood water volume of the catchment area of ​​the river section is inundated; n′ is the number of grid cells in the catchment area of ​​the river section; V i,j When the river depth is D i When , the flood water volume of the jth grid unit in the catchment area of ​​the river section is; H j is the HAND value of the jth grid cell; F j is the area of ​​the jth grid cell; Based on the equidistant water depth sequence of the river section D=[0,…D i-1 ,D i ,D i+1 …,D m ] and the corresponding water volume sequence V=[0,…V i-1 ,V i ,V i+1 …,V m ], we can get the river section water depth-water volume relationship curve.

6. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to claim 2 is characterized by: Also included between step S11 and step S12, The calculated river network distribution data of the basin was corrected using Tiandi Map remote sensing images to make the river distribution consistent with the actual location.

7. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to any one of claims 1 to 6, characterized in that: After the rapid calculation model of river overflow flood inundation is built, it also includes: Rapid prediction and risk analysis of river overflow floods: Based on the constructed rapid analysis model for river overflow floods and access to future hydrological forecast data, the development and evolution of river floods in the basin are predicted in real time, and the dynamic inundation range and water depth of river overflow floods in the basin are obtained; flood inundation risk indicators are established based on the dynamic inundation range and water depth of river overflow floods in the basin, and a dynamic flood risk map of the basin is drawn based on the flood inundation risk indicators.

8. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to claim 7 is characterized by: The rapid prediction and risk analysis of river overflow floods specifically include the following: S21. River Basin Flood Inundation Prediction: Based on the flow forecast results of upstream river sections, a rapid analysis model for river overflow flood inundation is driven to predict the flood inundation range and water depth data of the river basin in real time; S22. Construction of flood inundation risk indicators: Taking into account the threat posed by flood inundation depths to the safety of life and property, as well as social and economic security, as well as the standards for compiling flood risk maps, for extreme flood events, flood inundation depths of 0.5m, 1m, 2m, and 3m are set as indicators of low, medium, high, and very high flood inundation risks in the basin, respectively. S23. Real-time flood risk prediction and risk analysis of key protected objects: Based on real-time inundation prediction data and combined with flood inundation risk indicators, a dynamic flood risk map is drawn; geographic spatial information layers such as flood control projects, lifeline projects, and key protected objects are superimposed on the dynamic flood risk map to draw a thematic risk map.

9. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to any one of claims 1 to 6, characterized in that: Before the construction of the rapid calculation model of river overflow flood inundation, it also includes: Collection and processing of basic watershed data: Collection and processing of high-resolution digital elevation data, river embankment data and hydrological data of the watershed.

10. The rapid simulation method for river overflow flood evolution based on adaptive river segmentation and strong water balance constraints according to claim 9, characterized in that: The collection and processing of basin basic data specifically includes the following: S01. Download meter-level digital elevation data (DEM) of the basin through a map downloader, with a spatial resolution of 30m, 12.5m, or 5m. Collect river embankment elevation data in the basin and modify it based on the digital elevation data to ensure that the river channel and embankment elevation data are more consistent with the actual topography of the basin. S02. Collect measured flow data from river hydrological stations in the basin and unify the flow data time interval to 1 hour.

Images