A method for identifying storm surge overbank flooding and inundation based on cellular automata

Abstract

The present invention proposes a method for identifying storm surge overtopping and inundation based on cellular automata, which includes the following steps: gridifying the study area, preparing geographical, hydrological, and meteorological data, and constructing an initial grid structure; based on cellular automata and the dry-wet grid algorithm, designing dry-wet grid transformation rules based on the two-dimensional shallow water momentum equation, and performing iterative calculations on the grid structure; using GIS technology to visually display the influence range and degree of storm surge overtopping and inundation. The present invention mainly aims at the problem that the existing methods for identifying storm surge overtopping and inundation cannot simultaneously meet the requirements of high efficiency, stability, and accuracy. While achieving rapid identification, it also takes into account the identification accuracy, providing a new option for scientific research and practical applications of storm surge overtopping, so as to better meet different needs.

Description

Technical Field

[0001] The present invention relates to the fields of marine science and marine engineering technology, and particularly relates to a storm surge overtopping and inundation model based on cellular automata. Background Art

[0002] As a region with dense population and economy, the coastal areas have long been threatened by various marine disasters, among which the storm surge disaster is the most serious, causing a large number of casualties and economic losses. A storm surge refers to the phenomenon of abnormal sea level rise caused by strong atmospheric disturbances, such as strong winds and sudden changes in air pressure. Along with astronomical tides and ocean waves, it can cause a sharp rise in the tide level, leading to land inundation and disasters. According to the statistics of the China Marine Disaster Bulletin, a total of 16 storm surges occurred in the coastal waters of China in 2021, causing direct economic losses of 2.47 billion yuan, accounting for 80% of all marine disaster losses. With the rapid development of the marine economy and the increasingly severe climate change situation, the risks of marine disasters, especially storm surge disasters, in coastal areas are becoming increasingly prominent, posing higher requirements for disaster risk assessment and disaster prevention and mitigation.

[0003] A disaster is the consequence of a physical process imposing a load on the environment, manifested as property losses and casualties. For storm surge disasters, overtopping and inundation caused by rising tide levels are the main physical processes affecting human society. In order to formulate appropriate disaster prevention and mitigation strategies, it is very necessary to conduct risk assessment. Estimating the scope and intensity of a disaster is the first and most important step in risk assessment. Therefore, establishing an accurate and efficient storm surge overtopping and inundation model to estimate disaster-causing factors including inundation area, inundation depth, and flow velocity is an important tool to support storm surge disaster risk assessment work.

[0004] At present, storm surge overland inundation identification models can generally be divided into two categories. One is the numerical model based on solving the shallow water equations, which has relatively accurate simulation results because it comprehensively reflects the hydrodynamic principles. Commercial or non-commercial storm surge numerical models such as FVCOM, MIKE21, and ADCIRC are widely used in the field of storm surge research. However, numerical models have high requirements for computing resources, and the time cost of simulating a storm surge process for a relatively large area is high. Generally, it is difficult to meet the timeliness requirements of storm surge disaster forecasting, and the stability is insufficient. Although various means, including optimizing the program operation method and simplifying the model basis, have been applied to improve the model operation efficiency, the overall effect is still quite limited. Different from numerical models, another type of overland inundation identification method, which can be collectively referred to as the conceptual model, is mainly known for its high efficiency. Conceptual models can be further divided into two categories in terms of principle. One is the model based on the digital elevation model (DEM), which only considers the source water level and terrain connectivity to identify the inundation range and depth. The other is the model based on cellular automata (CA), which designs the evolution rules of wet and dry grids based on the continuity equation to simulate the inundation process. Conceptual models are simpler in design principle, so the identification efficiency is higher. However, relatively speaking, the accuracy of the identification results of conceptual models is lower. Especially for the overland inundation process caused by storm surges, external forces such as wind stress and bottom friction have a certain influence on the movement of water flow. Therefore, conceptual models that do not consider external forces will produce relatively large errors.

[0005] Therefore, in order to better meet the needs of storm surge disaster risk assessment, it is necessary to design a storm surge overland inundation and flooding model that takes into account both accuracy and high efficiency and stability, providing new options for storm surge overland inundation scientific research and practical applications to better meet different needs. Summary of the Invention

[0006] The problem to be solved by the present invention is to provide a storm surge overland inundation and flooding identification method based on cellular automata to achieve rapid and accurate identification of the influence range and degree of storm surge overland inundation in view of the defects of existing storm surge overland inundation and flooding identification methods.

[0007] To achieve the above object, the present invention adopts the following technical solutions to implement:

[0008] A storm surge overland inundation and flooding identification method based on cellular automata, comprising the following steps:

[0009] S1. Collect geographical, hydrological, and meteorological data, and construct an initial grid structure of the area to be identified;

[0010] S2. Based on the cellular automaton and the dry-wet grid algorithm, establish the dry-wet grid transition rules based on the two-dimensional shallow water momentum equation; iteratively calculate the grid structure in the area, predict and update the dry-wet state of the grid, and obtain the prediction and recognition results: the scope and degree of the impact of storm surge overtopping and inundation.

[0011] The construction of the initial grid structure includes the following steps:

[0012] a. Collect data: The geographical data includes terrain elevation raster data, land cover type raster data, river and levee line vector data, the hydrological data includes the highest water level and flow velocity data of the coastal storm surge, and the meteorological data includes the wind field data during the storm surge.

[0013] b. Determine the grid size and rasterize the area to be measured; resample the terrain elevation data to the raster model, and determine the initial dry-wet state raster according to the terrain elevation with 0m as the threshold.

[0014] c. According to the land cover data standard, convert the land cover type into the corresponding Manning coefficient and resample it to the model raster to obtain the surface Manning coefficient raster; interpolate the maximum wind speed and average wind direction during the storm surge process into the model raster to obtain the wind field raster, which affects the overtopping and inundation processes as external forcing factors.

[0015] d. Interpolate the highest water level and flow velocity during the storm surge process into the raster with the initial state of wet in the model raster, and assign 0 to the raster with the initial state of dry to obtain the water level raster and the flow velocity raster, which drive the overtopping process as boundary conditions.

[0016] The specific derivation process of the constructed dry-wet grid transition rules is as follows:

[0017] Two-dimensional shallow water momentum equation (x-direction):

[0018]

[0019] Among them, v is the flow velocity in the x-direction, η is the water level, τ a is the wind stress in the x-direction, τ b is the bottom friction force in the x-direction, ρ is the seawater density, g is the acceleration due to gravity, and D is the water depth;

[0020] Neglect the local acceleration term in the two-dimensional shallow water momentum equation, and rewrite the two-dimensional shallow water momentum equation (x-direction) into the energy form, whose physical meaning is the change in energy height caused by external forcing:

[0021]

[0022] Use the finite difference method to discretize formula (2) between the target grid and the adjacent grid:

[0023]

[0024] Among them, v tar is the flow velocity in the x direction of the target grid, η tar is the target grid water level, v ner is the flow velocity in the x direction of the adjacent grid, η ner is the water level of the adjacent grid, Δx is the distance between the adjacent grid and the center of the target grid;

[0025] Assume that the Friedman numbers for the flow in the target grid and the neighboring grids are approximately equal:

[0026]

[0027] Where Fr is the Fred number, h tar is the target grid elevation, h ner is the neighboring grid elevation;

[0028] Combine equations (3)-(4) to obtain the target grid water level and flow velocity:

[0029]

[0030] According to the wetting condition that the grid water level is higher than the grid elevation, the condition for the dry grid to become a wet grid is obtained:

[0031]

[0032] Among them, the effects of wind stress and bottom friction on storm surge floodplain and inundation are considered: wind stress τ a and bottom friction τ b Respectively expressed as:

[0033] τ a =ρ a C d v 2

[0034] C d =(0.75+0.067|v wind |)×10 -3

[0035] τ b =ρC f |v wind |v wind

[0036]

[0037] Among them, τ a is the wind stress, τ b is the bottom friction, ρ is the seawater density, ρ a is the air density, v windis the projection of the relative wind speed in the flow direction, C d is the drag coefficient, C f is the bottom friction coefficient, and n is the Manning roughness coefficient of the ground surface, which is related to the ground surface cover type.

[0038] The iterative calculation of the grid structure in the area, predicting and updating the wet / dry state of the grid, and obtaining the prediction and recognition result include the following steps:

[0039] Traverse all grids, and loop and iterate the following steps a-d to continuously iterate and evolve the grid structure until all grids no longer change. The final wet grid area obtained is the maximum possible inundation range, and the inundation water depth and flow velocity are obtained.

[0040] Step a: If the target grid is a wet grid or there is no adjacent wet grid, the state of the target grid remains unchanged.

[0041] Step b: If the target grid is a dry grid and there is at least one adjacent wet grid, calculate the remaining energy height and Froude number of the adjacent wet grid.

[0042] Step c: If the wet / dry grid transformation rule is satisfied, update the state of the current target grid to a wet grid; otherwise, the state of the current target grid remains unchanged.

[0043] Step d: Update and calculate the water level and flow velocity of the wet grid.

[0044] The calculation of the remaining energy height and Froude number of the adjacent wet grid includes:

[0045]

[0046] Among them, H ner is the remaining energy height of the adjacent grid, Fr ner is the Froude number of the adjacent grid, v ner is the flow velocity of the adjacent grid in the x direction, η ner is the water level of the adjacent grid, τ a is the wind stress in the x direction, τ b is the bottom friction force in the X direction, ρ is the seawater density, g is the acceleration due to gravity, D is the water depth, and Δx is the distance between the center of the adjacent grid and the target grid.

[0047] If there are multiple adjacent wet grids, the remaining energy height and Froude number are averaged and regarded as a single grid for calculation.

[0048] This method further includes: S3. Using GIS technology to visually display the range and degree of storm surge overbank flooding and inundation impacts.

[0049] Compared with the existing storm surge overbank flooding and inundation identification methods, the present invention has the following beneficial effects:

[0050] 1. Based on the two-dimensional shallow water momentum equations, fully considering the effects of wind stress and bottom friction on storm surge overbank flooding and inundation, the identification results of the model are more accurate than those of existing rapid identification models.

[0051] 2. Design the model from the perspective of energy, not only considering the impact of storm surge overbank flooding and inundation driving factors on water level, but also considering their impact on flow velocity. The identification results include not only the inundation area and inundation depth, but also the flow velocity, an important disaster-causing factor of storm surge.

[0052] 3. Design the model based on the principle of cellular automata, which has higher efficiency and stability compared to a complete numerical model. Description of the Drawings

[0053] Figure 1 is the implementation flow chart of the present invention for storm surge overbank flooding and inundation identification;

[0054] Figure 2 is the program flow chart of the storm surge overbank flooding and inundation identification model of the present invention;

[0055] Figure 3 is the selected area and typhoon path of the verification experiment of the present invention;

[0056] Figure 4 is the comparison chart between the present invention and the actual disaster investigation results (Cangzhou, Hebei);

[0057] Figure 5 is the comparison chart between the present invention and the actual disaster investigation results (Shenzhen, Guangdong);

[0058] Figure 6 is the comparison chart between the present invention and the results of the overbank flooding model (ADCIRC-SWAN). Detailed Embodiments

[0059] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following describes in detail the specific implementation methods of the present invention with reference to the accompanying drawings. Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the invention. Therefore, the present invention is not limited by the specific implementations disclosed below.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0061] A storm surge floodplain and inundation model based on cellular automata, including the construction of model principles and the implementation of model programs.

[0062] The core of the model principle is to construct the dry-wet grid transformation rules, which includes the following derivation process:

[0063] Two-dimensional shallow water momentum equation (x direction):

[0064]

[0065] Where v is the flow velocity in the x direction, η is the water level, and τ a is the wind stress in the x direction, τ b is the bottom friction in the x direction, ρ is the seawater density, g is the acceleration due to gravity, and D is the water depth;

[0066] Ignoring the local acceleration term in the two-dimensional shallow water momentum equation, the two-dimensional shallow water momentum equation (x direction) is rewritten in energy form. Its physical meaning is that the energy height changes due to external force:

[0067]

[0068] The finite difference method is used to discretize formula (2) between the target grid and the adjacent grids:

[0069]

[0070] Among them, v tar is the flow velocity in the x direction of the target grid, η tar is the target grid water level, v ner is the flow velocity in the x direction of the adjacent grid, η ner is the water level of the adjacent grid, Δx is the distance between the adjacent grid and the center of the target grid;

[0071] Assume that the Friedman numbers for the flow in the target grid and the neighboring grids are approximately equal:

[0072]

[0073] Where Fr is the Fred number, h tar is the target grid elevation, h ner is the neighboring grid elevation;

[0074] Combine equations (3)-(4) to obtain the target grid water level and flow velocity:

[0075]

[0076] According to the wetting condition that the grid water level is higher than the grid elevation, the condition for the dry grid to become a wet grid is obtained:

[0077]

[0078] Among them, the wind stress τ a and the bottom friction τ b are respectively expressed as:

[0079] τ a = ρ a C d v 2

[0080] C d =(0.75 + 0.067|v wind |)×10 -3

[0081] τ b = ρC f |v wind |v wind

[0082]

[0083] Among them, τ a is the wind stress, τ b is the bottom friction, ρ is the seawater density, ρ a is the air density, v wind is the projection of the relative wind speed in the flow direction, C d is the drag coefficient, C f is the bottom friction coefficient, and n is the surface Manning roughness coefficient, which is related to the surface cover type.

[0084] As Figure 1 shown, the implementation of the model program includes the following modules:

[0085] 1) Data preparation module; grid the study area, prepare geographical, hydrological, and meteorological data, and construct an initial grid structure as the model input;

[0086] 2) Floodplain and inundation identification module; perform iterative calculations on the grid structure based on the cellular automaton and wet-dry grid algorithms;

[0087] 3) Result visualization module; use GIS technology to visually display the scope and degree of the storm surge floodplain and inundation impact.

[0088] The data preparation module specifically includes the following steps:

[0089] 1) Determine the study area and storm surge process, and obtain geographical, hydrological, and meteorological data during the storm surge process in the study area. Among them, the geographical data includes terrain elevation grid data, land cover type grid data, river and levee line vector data of the study area; the hydrological data includes the highest water level and flow velocity data during the coastal storm surge process, which are simulated based on the storm surge numerical model. The ADCIRC-SWAN coupled model is used in this model; the meteorological data includes wind field data during the storm surge, including wind speed and wind direction;

[0090] 2) Determine the grid size, rasterize the study area to obtain the model grid. Resample the terrain elevation grid data to the model grid. According to the characteristic horizontal scale of the river and levee line vector data, obtain the buffer zone surface vector data, and convert the buffer zone into grid data according to its elevation and resample it to the model grid. Correct the terrain elevation of the grid where the river and levee are located to obtain the corrected terrain elevation grid;

[0091] 3) Take 0m as the threshold, and determine the initial dry-wet state of the grid according to the terrain elevation. In the program, the dry-wet binary state of the grid is represented by assignment. The grid state value with an elevation greater than or equal to 0m is set to 0, indicating a dry grid; the grid state value with an elevation less than 0m is set to 255, indicating a wet grid. Obtain the initial dry-wet state grid;

[0092] 4) According to the land cover data standard, convert the land cover type to the corresponding Manning coefficient and resample it to the model grid to obtain the surface Manning coefficient grid;

[0093] 5) Interpolate the maximum wind speed and average wind direction during the storm surge process into the model grid to obtain the wind field grid, which is used as the external force to affect the floodplain process;

[0094] 6) Interpolate the highest water level and maximum flow velocity during the storm surge process into the grids with an initial state of wet in the model grid, and assign 0 to the grids with an initial state of dry to obtain the water level grid and flow velocity grid, which are used as boundary conditions to drive the floodplain process.

[0095] Thus, the initial grid structure is obtained. The grid attributes include dry-wet state, elevation, surface Manning coefficient, water level, flow velocity, wind speed, and wind direction, which are used as the input of the model. Among them, the dry-wet state, water level, and flow velocity of the grid evolve with iteration.

[0096] As Figure 2 shown, the floodplain and inundation identification module includes the following steps:

[0097] 1) Input the initial grid structure;

[0098] 2) Traverse all grids and execute the following program:

[0099] If the target grid is a wet grid or there is no adjacent wet grid, the target grid remains unchanged. The adjacent wet grid is defined as the Moore neighborhood, that is, the eight grids adjacent to the target grid;

[0100] If condition 1 is satisfied: the target grid is a dry grid and there is at least one adjacent wet grid, then the following steps are executed:

[0101] Calculate the remaining energy height and Froude number of the adjacent wet grid:

[0102]

[0103] where, H ner is the remaining energy height of the adjacent grid, Fr ner is the Froude number of the adjacent grid, v ner is the flow velocity in the x - direction of the adjacent grid, η ner is the water level of the adjacent grid, τ a is the wind stress in the x - direction, τ b is the bottom friction in the x - direction, ρ is the seawater density, g is the acceleration due to gravity, D is the water depth, and Δx is the distance between the center of the adjacent grid and the target grid.

[0104] For multiple adjacent wet grids, the remaining energy height and Froude number are averaged and regarded as a single grid for calculation;

[0105] If condition 2 is satisfied: the target grid satisfies formula (7), then the target grid changes from a dry grid to a wet grid, and the water level and flow velocity after the change are solved according to formulas (5) - (6); if the target grid does not satisfy formula (7), the target grid remains unchanged;

[0106] After traversing all grids, the grid state changes, and the iterated grid structure is obtained;

[0107] 3) Loop the above iterative program to continuously iterate and evolve the grid structure until all grids no longer change;

[0108] 4) Extract the grids that are dry in the initial state and wet in the final state to obtain the inundation range, which is converted into surface vector data. Calculate the difference between the grid water level and the elevation to obtain the inundation water depth. Extract the inundation water depth and flow velocity within the inundation range. Output the results, including the inundation range surface vector data, inundation water depth raster data, and flow velocity raster data;

[0109] The result visualization module includes the following steps:

[0110] Use GIS technology to overlay the inundation range, inundation water depth, and flow velocity on the map to visually reflect the influence range and degree of storm surge overtopping and inundation.

[0111] The verification of the accuracy of the present invention includes the following comparative experiments:

[0112] To verify the accuracy of the recognition results of this model, two types of comparative experiments were designed, which were compared with the actual disaster investigation results and the numerical model results respectively. The comparison with the actual disaster investigation results is to illustrate the reliability of the model for practical applications, and the comparison with the numerical model results is to illustrate the rationality of the model principle.

[0113] The comparison with the actual disaster investigation results includes the following experiments:

[0114] Hebei Cangzhou and Guangdong Shenzhen were selected as the study areas from the Bohai Sea and the South China Sea respectively, and typhoon storm surge processes that had a greater impact on the two areas were selected respectively. Typhoon No. 1909 landed in Shandong Province on October 11, 2019, with a maximum central wind speed of level 9, and caused serious storm surge disasters in the Bohai Bay and Laizhou Bay. After the disaster, the National Marine Environmental Forecasting Center organized an investigation on the affected areas on the south bank of the Bohai Bay. Therefore, the storm surge of Typhoon No. 1909 was selected as the experimental storm surge process in Cangzhou City. Typhoon No. 9216 originated in the western North Pacific on August 20, 2017. Due to its moving path and intensity, the Pearl River Estuary was affected by strong winds and serious storm surges. The Shenzhen Marine Monitoring and Forecasting Center carried out an investigation on the key affected areas of this storm surge process. Therefore, the storm surge of Typhoon No. 1713 was selected as the experimental storm surge process on the west coast of Shenzhen. The experimental areas and the paths of the experimental typhoons are as Figure 3 shown.

[0115] Since the actual disaster investigation data is post-disaster investigation, there are errors in the inundation situation during the disaster, especially the specific values of the inundation depth. Therefore, the comparison with the actual disaster investigation results only involves the comparison of the inundation range, and does not compare the inundation depth.

[0116] The comparison with the numerical model results includes the following experiments:

[0117] The comparison with the numerical model results takes the coastal area of the Laizhou Bay as the study area, and the ADCIRC-SWAN coupled model is selected as the control. Due to the special geographical location, the storm surge water level rise is relatively high, and it is frequently affected by storm surge disasters. The storm surges of Typhoon No. 1909 and No. 9216 are selected as the experimental storm surge processes, representing two types of typhoon paths affecting the Shandong Peninsula respectively. Among them, the center of Typhoon No. 1909 passed directly through the Laizhou Bay, and Typhoon No. 9216 moved roughly along the southern part of the Shandong Peninsula.

[0118] The maximum water level and flow velocity of the model boundary input are taken from the coastal simulation results of the ADCIRC-SWAN model to ensure that the same boundary conditions are used as in the model. In addition, considering that the expression of the bottom friction coefficient in the ADCIRC-SWAN model is different from that in this model, it is related to the water depth but not to the Manning coefficient. Therefore, in the comparative test, the Manning coefficient in this model will be replaced by an equivalent Manning coefficient related to the water depth, so that the model has the same underlying surface conditions as the model. The equivalent Manning coefficient is obtained by combining formulas (11) and (14).

[0119] The bottom friction coefficient expression in ADCIRC-SWAN mode:

[0120]

[0121] Among them, C f is the bottom friction coefficient, is the minimum bottom friction coefficient, H break is the breaking depth, H is the water depth, θ is a dimensionless parameter indicating the speed at which the bottom friction coefficient approaches its limit, and γ is a dimensionless parameter indicating the speed at which the bottom friction coefficient increases with decreasing water depth; and γ are set to 0.0015, 1, 10, and 1 / 3, respectively, in ADCIRC-SWAN mode.

[0122] The equivalent Manning coefficient expression used in this model in the comparative experiment is:

[0123]

[0124] Where n is the Manning roughness coefficient of the surface, g is the acceleration due to gravity, and D is the water depth.

[0125] The comparison with the numerical model results includes both the flooding range and the flooding depth.

[0126] Considering that this model and ADCIRC-SWAN mode use different computational grids, the comparison of simulation results between this model and ADCIRC-SWAN mode is based on the computational nodes of ADCIRC-SWAN mode. The simulation results of this model are interpolated to the nodes, and quantitative comparison is performed in terms of flooding range and water depth.

[0127] The comparison of the flooded range uses the consistency rate δ as a measurement indicator, which is defined as the ratio of the number of intersection nodes to the number of union nodes of the flooded nodes of the model and pattern recognition:

[0128]

[0129] Among them, N a is the number of submerged nodes identified by this model, N mis the number of inundated nodes recognized by the ADCIRC-SWAN model, N o represents the number of inundated nodes where the two methods overlap.

[0130] For the comparison of the inundation water depths at the overlapping inundated nodes, the correlation coefficient R 2 and the root mean square error RMSE are used as the measurement indicators:

[0131]

[0132] where represents the inundation water depth recognized by this model at node i, represents the inundation water depth recognized by the ADCIRC-SWAN model at node i, represents the mean value of the inundation water depths recognized by this model at the overlapping nodes, represents the mean value of the inundation water depths recognized by the ADCIRC-SWAN model at the overlapping nodes.

[0133] The results of the comparative experiment are as follows:

[0134] From Figure 4 it can be seen that the simulation results of this model are highly consistent with the inundation range boundary line caused by the actual No. 1909 typhoon storm surge in Cangzhou City. The coastal sea dikes in Cangzhou City have played a significant role, and most of the inundation boundary lines are consistent with the trend of the sea dikes. Among them, the part where the inundation boundary line bends towards the land side is caused by the seawater backflow in the river. The inundation survey of the No. 1713 typhoon storm surge on the west coast of Shenzhen only includes some local affected locations, and three locations are more severely affected. As Figure 5 it can be seen, the three affected locations are basically consistent with the inundation range simulated by the model.

[0135] Except for the difference between the DEM data and the real terrain, and the errors caused by other factors such as rainfall, the simulation results of this model are relatively consistent with the actual disaster investigation results.

[0136] From Figure 6 it can be seen that the inundation range recognized by this model is basically consistent with the recognition results of the ADCIRC-SWAN model, and the coincidence rates of the inundation ranges during the two typhoon processes are 0.92 and 0.95 respectively. From the comparison of the inundation water depth scatter plots, it can be seen that the recognition water depth error between this model and the ADCIRC-SWAN model is basically within 0.3 m. Among them, for the water depth comparison during the No. 1909 typhoon storm surge process, R 2 = 0.96, NMSE = 0.13 m, and for the water depth comparison during the No. 9216 typhoon storm surge process, R 2 = 0.96, NMSE = 0.12 m, indicating that the inundation water depths recognized by this model are basically consistent with those of the ADCIRC-SWAN model.

[0137] The difference between this model and the numerical model mainly comes from the difference in grid structure and the simplification of hydrodynamic principles in this model. However, for the design purpose of this model, this degree of error is within an acceptable range. In addition, different typhoon paths will also cause differences in comparison consistency. Judging from the two typhoon processes in this experiment, 9216 is more consistent than 1909. The reason is that the center of 1909 passed directly through Laizhou Bay, and the hydrological and meteorological environment was more complex and changeable in time. The model uses extreme value and averaging processing to produce relatively larger errors.

[0138] The above comparative experimental results are sufficient to illustrate the rationality of the design principle of this model and the accuracy of the recognition results.

[0139] In addition, by comparing the expression switches of wind stress and bottom friction in the model, a sensitivity experiment was designed to illustrate the significance of considering external forcing in the model on the accuracy of simulation results. The sensitivity experiment design and results are shown in Table 1.

[0140] Table 1

[0141]

[0142] Experiment 1 represents the model that considers wind stress and bottom friction. The results are the results of the above comparison test, and the consistency is good. Experiment 2 represents the model that does not consider wind stress and bottom friction. The results show that the consistency of flooding range and water depth is lower than that of Experiment 1. Experiments 3 and 4 represent the models that only consider wind stress and bottom friction, respectively, and the consistency is reduced. Among them, ignoring wind stress greatly reduces the accuracy of the model.

[0143] In summary, based on the above sensitivity experimental results, this model considers wind stress and bottom friction, which is more consistent with the hydrodynamic principles of the storm surge inundation process, thus producing more accurate simulation results.

[0144] Based on the conclusions of the above comparative tests and sensitivity experiments, the accuracy of the present invention in identifying storm surge floodplains is verified. The relatively more complete hydrodynamic theoretical basis makes it more accurate than the traditional storm surge inundation conceptual model. At the same time, the computational efficiency of the model program based on cellular automata is higher and more stable, making it faster to identify than the storm surge floodplain model. This confirms the unique advantages of the present invention in the field of storm surge floodplain identification.

[0145] The above is 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 be regarded as within the scope of protection of the present invention.

Claims

1. A method for identifying storm surge overtopping and inundation based on cellular automata, characterized in that The following steps are involved: S1, collect geographical, hydrological and meteorological data and construct the initial grid structure of the area to be identified; S2. Based on cellular automata and dry-wet grid algorithms, the dry-wet grid transition rules are established based on the two-dimensional shallow water momentum equation; the grid structure in the region is iteratively calculated, and the dry-wet state of the grid is updated in advance to obtain the prediction and identification results: the scope and degree of storm surge floodplain and inundation impact; The specific derivation process of the constructed dry-wet grid transformation rule is as follows: Two-dimensional shallow water momentum equation (x direction): Where v is the flow velocity in the x direction, η is the water level, and τ a is the wind stress in the x direction, τ b is the bottom friction in the x direction, ρ is the density of seawater, g is the acceleration due to gravity, and d is the water depth; Ignoring the local acceleration term in the two-dimensional shallow water momentum equation, the two-dimensional shallow water momentum equation (x direction) is rewritten in energy form. Its physical meaning is that the energy height changes due to external force: The finite difference method is used to discretize formula (2) between the target grid and the adjacent grids: Among them, v tar is the flow velocity in the x-direction of the target grid, and η tar is the water level of the target grid. v ner is the flow velocity in the x-direction of the adjacent grid, and η ner is the water level of the adjacent grid. Δx is the distance between the center of the adjacent grid and the target grid; Assume that the Fried numbers of the water flow in the target grid and the neighboring grids are equal: where Fr is the Fred number, h tar is the target grid elevation, h ner is the adjacent grid elevation; Combine equations (3)-(4) to obtain the target grid water level and flow velocity: According to the wetting condition that the grid water level is higher than the grid elevation, the condition for the dry grid to become a wet grid is obtained: Among them, the effects of wind stress and bottom friction on storm surge overtopping and inundation are considered: the wind stress τ a and the bottom friction τ b are respectively expressed as: τ a = ρ a C d v 2 (8) C d = (0.75 + 0.067|v wind |) × 10 -3 (9) τ b = ρC f |v wind |v wind (10) where τ a is the wind stress, τ b is the bottom friction, ρ is the seawater density, ρ a is the air density, v wind is the projection of the relative wind speed in the flow velocity direction, C d is the drag coefficient, C f is the bottom friction coefficient, n is the surface Manning roughness coefficient, which is related to the surface cover type; the iterative calculation of the grid structure in the area, predicting and updating the wet-dry state of the grid, and obtaining the prediction and recognition result include the following steps: Traverse all grids and iterate the following steps ad in a loop to make the grid structure continuously evolve iteratively until all grids no longer change, and obtain the final wet grid area, which is the maximum flooding range, and obtain the flooding water depth and flow velocity; Step a: If the target grid is a wet grid or there is no adjacent wet grid, the state of the target grid remains unchanged; Step b: if the target grid is a dry grid and there is at least one adjacent wet grid, calculate the residual energy height and Friedman number of the adjacent wet grid; Step c: if the dry-wet grid transformation rule is met, the state of the current target grid is updated to a wet grid; otherwise, the state of the current target grid remains unchanged; Step d: Update and calculate the water level and flow rate of the wet grid.

2. The method for identifying storm surge overbank flooding and inundation based on cellular automata according to claim 1, wherein The construction of the initial grid structure comprises the following steps: a. Collected data: Geographic data includes terrain elevation raster data, land cover type raster data, river and levee line vector data, hydrological data includes coastal storm surge maximum water level and flow velocity data, meteorological data includes wind field data during storm surge; b. Determine the grid size and rasterize the area to be tested; resample the terrain elevation data to the grid model, and use 0m as the threshold to determine the initial dry and wet state grid according to the terrain elevation; c. According to the land cover data standard, the surface cover type is converted into the corresponding Manning coefficient and resampled to the model grid to obtain the surface Manning coefficient grid; the maximum wind speed and average wind direction of the storm surge process are interpolated into the model grid to obtain the wind field grid; it acts as an external forcing factor to affect the floodplain and inundation process; d. Interpolate the highest water level and flow velocity of the storm surge process into the model grid. The grids with an initial wet state are assigned a value of 0, and the grids with an initial dry state are assigned a value of 0. The water level grid and flow velocity grid are obtained as boundary conditions to drive the floodplain process.

3. A storm surge overbank flooding and inundation identification method based on cellular automata according to claim 1, characterized in that The calculation of the residual energy height and Fred number of the adjacent wet grid includes: Among them, H ner is the remaining energy height of the adjacent grid, Fr ner is the Froude number of the adjacent grid, v ner is the flow velocity in the x-direction of the adjacent grid, η ner is the water level of the adjacent grid, τ a is the wind stress in the x-direction, τ b is the bottom friction in the x-direction, ρ is the seawater density, g is the acceleration due to gravity, D is the water depth, and Δx is the distance between the adjacent grid and the center of the target grid.

4. A method for identifying storm surge overbank flooding and inundation based on cellular automata according to claim 1, characterized in that If there are multiple adjacent wet grids, the residual energy height and Friedel number are averaged and regarded as a single grid for calculation.

5. A method for identifying storm surge overbank flooding and inundation based on cellular automata according to any one of claims 1-4, characterized in that, The method further comprises: S3. Use GIS technology to visualize the scope and extent of storm surge floodplains and inundation.

Images