Flood disaster risk grade prediction method

Abstract

The invention discloses a flood disaster risk level prediction method, and relates to the field of disaster risk management and control, and the method comprises the steps: evaluating the flood disaster risk levels of a first preset year and a second preset year, and obtaining the flood disaster risk level distribution grid images of the first preset year and the second preset year; calculating to obtain a flood risk level state transition matrix according to the flood disaster risk level distribution grid images of the first preset year and the second preset year; generating a flood disaster risk level change suitability atlas through characteristic analysis of the flood disaster risk; and inputting the flood disaster risk level distribution grid image of the second preset year, the flood risk level state transition matrix and the flood disaster risk level transition suitability atlas into a flood disaster risk level prediction model to obtain a flood disaster risk level distribution map of a third preset year. The method can improve the accuracy of flood disaster risk level prediction.

Description

Technical Field

[0001] The present invention relates to the field of disaster risk control, and particularly to a method for predicting the risk level of flood disasters. Background Art

[0002] Flood is a common natural phenomenon globally. Flood disasters are common disasters that can cause serious social and economic consequences. Flood disasters threaten hundreds of millions of people worldwide every year, resulting in incalculable property losses. With the growth of the population and the rapid advancement of urbanization, the harm and impact of floods will further deepen. Due to the wide range of influence and great destructive power of flood disasters, forward-looking measures are needed for prevention. Predicting the risk level of flood disasters is particularly important and is also the deficiency in the current research on flood disaster risks.

[0003] Researchers have studied the regional flood risk assessment in terms of the hazard, exposure, and vulnerability of flood disasters, and formed a set of index systems to predict the risk level of flood disasters. Existing technologies usually integrate various indexes into flood risk research and combine the indexes with appropriate weight settings. Moreover, since meteorological factors change cyclically, and human activities and social and economic development have changed the exposure and vulnerability patterns, the existing technologies do not consider the influence of the change of meteorological factors and human activities and social and economic development on flood disaster prediction, resulting in low accuracy in predicting the risk level of flood disasters.

[0004] Therefore, there is an urgent need for a method to improve the accuracy of predicting the risk level of flood disasters. Summary of the Invention

[0005] Based on this, it is necessary to provide a method for predicting the risk level of flood disasters for the above technical problems. This method can improve the accuracy of predicting the risk level of flood disasters.

[0006] The present invention adopts the following technical solutions:

[0007] The present invention provides a method for predicting the risk level of flood disasters, including:

[0008] Respectively evaluate the flood disaster risk levels of the first preset year and the second preset year to obtain the flood disaster risk level distribution grid images of the first preset year and the second preset year;

[0009] According to the flood disaster risk level distribution grid images of the first preset year and the second preset year, calculate the flood risk level state transition matrix;

[0010] Through the characteristic analysis of the flood disaster risk, generate a flood disaster risk level transformation suitability atlas;

[0011] Input the raster image of the flood disaster risk level distribution in the second preset year, the flood risk level state transition matrix, and the flood disaster risk level transformation suitability atlas into the flood disaster risk level prediction model to obtain the flood disaster risk level distribution map in the third preset year.

[0012] Preferably, evaluate the flood disaster risk levels in the first preset year and the second preset year respectively to obtain the raster images of the flood disaster risk level distributions in the first preset year and the second preset year, specifically including:

[0013] Taking either the first preset year or the second preset year as an example, obtain the initial data of multiple evaluation factors affecting flood disasters, and convert the initial data to generate multiple raster data;

[0014] Perform standardization processing on the multiple raster data to obtain multiple evaluation factor data;

[0015] Determine the weights of the multiple evaluation factors, and determine the flood disaster risk index values of multiple grids based on the multiple evaluation factor data and the weights of the multiple evaluation factors;

[0016] Divide the flood disaster risk levels of each unit grid according to the flood disaster risk index values of the multiple grids to obtain the raster image of the flood disaster risk level distribution.

[0017] Preferably, the evaluation factors include hazard evaluation factors, exposure evaluation factors, and vulnerability evaluation factors; the hazard evaluation factors include the annual average precipitation, 24-hour heavy precipitation frequency, 72-hour heavy precipitation frequency, the exposure evaluation factors include digital elevation, terrain slope, elevation standard deviation, river network density, and water buffer zone, and the vulnerability evaluation factors include population per kilometer grid, gross domestic product per kilometer grid, and land type;

[0018] The standardization formulas for the annual average precipitation, 24-hour heavy precipitation frequency, 72-hour heavy precipitation frequency, terrain slope, river network density, water buffer zone, population per kilometer grid, and gross domestic product per kilometer grid are:

[0019]

[0020] where, μ A (x) is the evaluation factor data, x is the initial evaluation factor data, and a and b are the lower and upper limits of the evaluation factor thresholds respectively;

[0021] The standardization formulas for digital elevation, elevation standard deviation, and land type are:

[0022]

[0023] where, μ B(x) is the evaluation factor data, x is the initial evaluation factor data, and a and b are the lower and upper limits of the evaluation factor threshold respectively.

[0024] Preferably, determine the weights of multiple evaluation factors, specifically including:

[0025] Construct a judgment matrix based on the importance of each evaluation factor to the evaluation result;

[0026] Conduct a consistency test on the judgment matrix;

[0027] If the consistency check of the judgment matrix passes, normalize the eigenvector corresponding to the largest eigenvalue of the judgment matrix to obtain the weights of multiple evaluation factors.

[0028] Preferably, the calculation formula for the flood disaster risk index value is:

[0029]

[0030] Among them, μ A is the flood disaster risk index value, μ A (x i ) is the evaluation factor data of the evaluation factor x i , ω i is the weight of the evaluation factor x i , and n is the number of evaluation factors.

[0031] Preferably, divide the flood disaster risk levels of each unit grid according to the flood disaster risk index values of multiple grids, specifically including:

[0032] When the flood disaster risk index value is less than or equal to the first preset threshold, determine that the flood disaster risk level is extremely low risk;

[0033] When the flood disaster risk index value is greater than the first preset threshold and less than or equal to the second preset threshold, determine that the flood disaster risk level is low risk;

[0034] When the flood disaster risk index value is greater than the second preset threshold and less than or equal to the third preset threshold, determine that the flood disaster risk level is medium risk;

[0035] When the flood disaster risk index value is greater than the third preset threshold and less than or equal to the fourth preset threshold, determine that the flood disaster risk level is high risk;

[0036] When the flood disaster risk index value is greater than the fourth preset threshold, determine that the flood disaster risk level is extremely high risk.

[0037] Preferably, the flood disaster risk level prediction model includes a meta-cellular automaton model and a Markov model; inputting the flood disaster risk level distribution raster image of the second preset year, the flood risk level state transition matrix, and the flood disaster risk level transformation suitability atlas into the flood disaster risk level prediction model to obtain the flood disaster risk level distribution map of the third preset year, specifically including:

[0038] Taking the unit raster in the flood disaster risk level distribution raster image of the second preset year as the meta-cell in the meta-cellular automaton model;

[0039] For each raster in the flood disaster risk level distribution raster image of the second preset year, determine the flood risk level of the raster in the flood disaster risk level distribution raster image of the third preset year according to the flood risk level of the raster in the flood disaster risk level distribution raster image of the second preset year, the meta-cell conversion rules between the second preset year and the third preset year, and the neighborhood of the meta-cell;

[0040] The Markov model in the flood disaster risk level prediction model determines the change in the quantity of the flood risk level in the third preset year according to the flood risk level quantity state in the second preset year and the flood risk level state transition matrix;

[0041] Determine the flood disaster risk level distribution map of the third preset year according to the flood risk levels of all the rasters in the flood disaster risk level distribution raster image of the third preset year, the change in the quantity of the flood risk level in the third preset year, and the flood disaster risk level transformation suitability atlas;

[0042] Preferably, the meta-cellular automaton model is:

[0043] S T+1 = f(S T , N);

[0044] Among them, S T , S T+1 respectively represent the flood risk levels of the unit raster at the T and T+1 moments of the meta-cell, f is the meta-cell conversion rule between the two moments, and N is the neighborhood of the meta-cell;

[0045] The Markov model is used to represent the change in the quantity of different flood risk levels, and the Markov model is:

[0046] Y T+1 = P T Y T ;

[0047] Among them, Y T , Y T+1 respectively represent the flood risk level quantity states at the T and T+1 moments, and P TRepresents the flood risk level quantity transfer matrix between time T and T+1.

[0048] Preferably, the method further includes:

[0049] Based on the predicted flood disaster risk level distribution map, draw a receiver operating characteristic curve;

[0050] Calculate the area under the curve of the receiver operating characteristic curve;

[0051] Evaluate the prediction accuracy of the flood disaster risk level prediction model through the area under the curve.

[0052] Preferably, evaluating the prediction accuracy of the flood disaster risk level prediction model through the area under the curve specifically includes:

[0053] When the area under the curve is less than or equal to the first prediction threshold, determine that the prediction accuracy of the flood disaster risk level prediction model is poor;

[0054] When the area under the curve is greater than the first prediction threshold and less than or equal to the second prediction threshold, determine that the prediction accuracy of the flood disaster risk level prediction model is average;

[0055] When the area under the curve is greater than the second prediction threshold, determine that the prediction accuracy of the flood disaster risk level prediction model is high.

[0056] The present invention provides a flood disaster risk level prediction device, including:

[0057] An acquisition module, configured to evaluate the flood disaster risk levels of the first preset year and the second preset year respectively, and obtain the flood disaster risk level distribution raster images of the first preset year and the second preset year;

[0058] A calculation module, configured to calculate a flood risk level state transition matrix according to the flood disaster risk level distribution raster images of the first preset year and the second preset year;

[0059] A generation module, configured to generate a flood disaster risk level transformation suitability atlas through the feature analysis of the flood disaster risk;

[0060] A prediction module, configured to input the flood disaster risk level distribution raster image of the second preset year, the flood risk level state transition matrix, and the flood disaster risk level transformation suitability atlas into the flood disaster risk level prediction model to obtain the flood disaster risk level distribution map of the third preset year.

[0061] The present invention provides a computer-readable storage medium, and the storage medium stores a computer program, and when the computer program is executed by a processor, it implements the above-mentioned flood disaster risk level prediction method.

[0062] The present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the above-mentioned method for predicting the flood disaster risk level is implemented.

[0063] The above at least one technical solution adopted by the present invention can achieve the following beneficial effects:

[0064] The present invention respectively evaluates the flood disaster risk levels of the first preset year and the second preset year, obtains the grid images of the flood disaster risk level distributions of the first preset year and the second preset year. The obtained grid images of the flood disaster risk level distributions are more intuitive, easier to understand and have high resolution, and can more accurately predict the flood disaster risk level; according to the grid images of the flood disaster risk level distributions of the first preset year and the second preset year, a flood risk level state transition matrix is calculated. This state transition matrix combines the historical data between the first preset year and the second preset year, and refers to the dynamic evolution process of the flood disaster levels between the first preset year and the second preset year when predicting the flood disaster risk level of the third preset year, with higher prediction accuracy; through the characteristic analysis of the flood disaster risk, a suitability atlas for the transformation of the flood disaster risk level is generated. The suitability atlas for the transformation of the flood disaster risk level more accurately reflects the risk levels of flood disasters in different regions and the suitability of their transformations; the grid image of the flood disaster risk level distribution of the second preset year, the flood risk level state transition matrix, and the suitability atlas for the transformation of the flood disaster risk level are input into the flood disaster risk level prediction model to obtain the flood disaster risk level distribution map of the third preset year. This method can improve the prediction accuracy of the flood disaster risk level. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The drawings described herein are used to provide a further understanding of the present invention, and constitute a part of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention, and do not constitute an improper limitation to the present invention. In the drawings:

[0066] Figure 1 is a schematic flow chart of a method for predicting the flood disaster risk level provided by the present invention;

[0067] Figure 2 is a technical framework diagram of a method for evaluating the flood disaster risk level provided by the present invention;

[0068] Figure 3 is a schematic flow chart of a technical method for verifying the prediction accuracy of the CA-Markov model for the evolution of flood disaster risks provided by the present invention;

[0069] Figure 4 is a schematic flow chart of a technical method for predicting the flood disaster risk level distribution map of a future horizontal year provided by the present invention;

[0070] Figure 5 Grid image map of 11 evaluation factors for evaluating flood disaster risk provided by the embodiments of the present invention;

[0071] Figure 6 Grid image map of 11 evaluation factors for evaluating flood disaster risk provided by the embodiments of the present invention after being standardized;

[0072] Figure 7 Technical flow chart of the AHP method for determining evaluation factors provided by the present invention;

[0073] Figure 8 Frame diagram of the hierarchical structure model for evaluating flood disaster risk level provided by the present invention;

[0074] Figure 9 Schematic diagram of the ROC curve for verifying the prediction model accuracy provided by the embodiments of the present invention;

[0075] Figure 10 Grid image of the flood disaster risk distribution in the study area in 2025 and schematic diagram of the proportion of risk areas provided by the embodiments of the present invention for implementing predictions;

[0076] Figure 11 Schematic diagram of a device for predicting flood disaster risk level provided by the present invention;

[0077] Figure 12 Schematic diagram of a computer device for implementing a method for predicting flood disaster risk level provided by the present invention. Detailed implementation manners

[0078] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0079] Devices such as desktop computers, servers, and laptop computers that execute the solutions of the present invention. For the convenience of description, only the server is used as the execution subject for illustration below.

[0080] In existing flood risk studies, there is already a relatively mature method system for evaluating the current situation of flood risk disasters. However, due to the periodic changes in meteorological factors that cause flood disaster hazards, as well as subsequent human activities and social and economic development that will change the existing exposure and vulnerability patterns of regional flood disasters, the current evaluation results cannot fully guide future flood control and disaster reduction strategies. In addition, due to the wide range of influence and great destructive power of flood disasters, forward-looking measures are needed for prevention, and predicting the development and changes of flood risks is particularly important and is also the deficiency in current flood disaster risk research.

[0081] The following will, in conjunction with the accompanying drawings, elaborate on the technical solutions provided by each embodiment of the present invention.

[0082] Figure 1 It is a schematic flowchart of a method for predicting the flood disaster risk level in the present invention, specifically including the following steps:

[0083] S101: Evaluate the flood disaster risk levels of the first preset year and the second preset year respectively to obtain the flood disaster risk level distribution grid images of the first preset year and the second preset year.

[0084] In an exemplary embodiment, evaluating the flood disaster risk levels of the first preset year and the second preset year to obtain the flood disaster risk level distribution grid images of the first preset year and the second preset year specifically includes: taking either the first preset year or the second preset year as an example, obtaining the initial data of multiple evaluation factors affecting flood disasters, and converting the initial data to generate multiple grid data; performing standardization processing on the multiple grid data to obtain multiple evaluation factor data; determining the weights of the multiple evaluation factors, and determining the flood disaster risk index values of multiple grids according to the multiple evaluation factor data and the weights of the multiple evaluation factors; dividing the flood disaster risk levels of each unit grid according to the flood disaster risk index values of the multiple grids to obtain the flood disaster risk level distribution grid images.

[0085] Specifically, the settings of the first preset year and the second preset year are determined according to specific engineering practices. For example, the first preset year is 2015 and the second preset year is 2020.

[0086] Specifically, as Figure 2 shown, the evaluation factors for flood disasters include hazard evaluation factors, exposure evaluation factors, and vulnerability evaluation factors; the hazard evaluation factors include the average annual precipitation, the frequency of heavy precipitation in 24 hours, and the frequency of heavy precipitation in 72 hours, the exposure evaluation factors include digital elevation, terrain slope, elevation standard deviation, river network density, and water buffer zone, and the vulnerability evaluation factors include the population per kilometer grid, the gross domestic product per kilometer grid, and land type.

[0087] Specifically, the factors determining flood disasters can generally be classified into three types, namely, hazard factors, exposure factors, and vulnerability factors. Each factor further includes: Hazard factors refer to the decisive conditions that induce floods. Generally speaking, precipitation is the decisive factor in inducing floods. The precipitation intensity and frequency determine the scale of floods, and thus affect the severity of flood disasters. In this application, the mean annual precipitation (MAP), the frequency of 24-hour heavy precipitation (FHP–24h), and the frequency of 72-hour heavy precipitation (FHP–72h) are selected as the evaluation factors for the hazard of flood disasters. By referring to Document 2, the 24-hour heavy precipitation defined in this application means that the precipitation in 24 hours exceeds 100 mm, and the 72-hour heavy precipitation means that the precipitation in 72 hours exceeds 200 mm.

[0088] Exposure factors include topographic and geomorphic factors such as those that induce precipitation to develop into floods, including conditions such as elevation, slope, and water system. In this application, the digital elevation model (DEM), topographic slope (TS), standard deviation of elevation (SDE), drainage density (DD), and buffer zone of water (BZW) are selected as the evaluation factors for the exposure of flood disasters.

[0089] The vulnerability of floods is reflected by the objects that bear flood disasters, generally including the population and economy in the disaster-stricken areas. In this application, the population grid (PG), the gross domestic product of the kilometer grid (GDPG), and the land type (LT) are selected as the evaluation factors for the vulnerability of flood disasters.

[0090] The initial data obtained above is converted and then raster data is generated and stored in the created flood disaster risk level assessment database.

[0091] Specifically, the steps for generating the raster image of the flood disaster risk level distribution include:

[0092] 1. Establish a flood disaster risk level assessment database. According to the present invention, flood disasters are determined by the accumulation of various factors causing floods. Relevant data is selected, evaluation factors are calculated, and a flood disaster risk level assessment database is established. All data is generated through projection transformation and spatial resampling to obtain raster data with the same projection information, the same spatial range, and the same resolution.

[0093] 2. Data standardization processing. The geographical information data stored in the database in step 1 is standardized, and through fuzzy processing (Fuzzy), it is transformed into dimensionless data between 0 and 1.

[0094] 3. Determine the weights of each evaluation factor. The Analytic Hierarchy Process (AHP) is used to determine the weights of each evaluation factor. The AHP method synthesizes the results of expert scoring and quantitatively processes the qualitative evaluation results through a judgment matrix. This method is systematic and clear, can effectively eliminate individual human errors, and can more accurately reflect the influence magnitude of each factor.

[0095] 4. Evaluate the flood risk in the current year and divide the risk levels. According to the standardized evaluation factor data in step 2 and the weights of each evaluation factor obtained in step 3, the comprehensive scoring method is used to calculate the flood disaster risk index values of each evaluation unit in the study area, and then the risk levels are divided according to the risk index values.

[0096] In an exemplary embodiment, the fuzzy processing (Fuzzy) method is used to standardize the raster images of the flood disaster risk level distribution in the first preset year and the second preset year. The evaluation factors include hazard evaluation factors, exposure evaluation factors, and vulnerability evaluation factors; the hazard evaluation factors include the annual average precipitation, the frequency of heavy precipitation in 24 hours, and the frequency of heavy precipitation in 72 hours. The exposure evaluation factors include digital elevation, terrain slope, elevation standard deviation, river network density, and water buffer zone. The vulnerability evaluation factors include population per kilometer grid, GDP per kilometer grid, and land type.

[0097] The standardization processing methods for the annual average precipitation, the frequency of heavy precipitation in 24 hours, the frequency of heavy precipitation in 72 hours, terrain slope, river network density, water buffer zone, population per kilometer grid, and GDP per kilometer grid are shown in formula (1):

[0098]

[0099] Among them, μ A (x) is the evaluation factor data, x is the initial evaluation factor data, and a and b are the lower and upper limits of the evaluation factor thresholds respectively.

[0100] The standardization processing methods for digital elevation, elevation standard deviation, and land type are shown in formula (2):

[0101]

[0102] Among them, μ B (x) is the evaluation factor data, x is the initial evaluation factor data, and a and b are the lower and upper limits of the evaluation factor threshold respectively.

[0103] In an exemplary embodiment, determining the weights of multiple evaluation factors specifically includes: constructing a judgment matrix based on the importance degree of the evaluation factors to the evaluation result; performing a consistency check on the judgment matrix; if the consistency check of the judgment matrix passes, normalizing the eigenvector corresponding to the maximum eigenvalue of the judgment matrix to obtain the weights of the evaluation factors.

[0104] Specifically, if the consistency check of the judgment matrix fails, first re-examine and analyze the importance between each factor in the judgment matrix, eliminate logical contradictions, and then perform the consistency check again until the check passes.

[0105] Determine the weights of each evaluation factor, use the AHP method, invite industry experts to score and determine according to the pre-specified quantification results of the importance between pairwise factors, construct a judgment matrix, and calculate the weights of each factor.

[0106] In an exemplary embodiment, according to the obtained standardized data of the evaluation factors and the weights of the evaluation factors, adopt the comprehensive scoring method to calculate the flood disaster risk index value of each unit grid. The calculation method of the flood disaster risk index value is shown in formula (3):

[0107]

[0108] Among them, μ A is the flood disaster risk index value, μ A (x i ) is the evaluation factor data of the evaluation factor x i , ω i is the weight of the evaluation factor x i , and n is the number of evaluation factors.

[0109] In an exemplary embodiment, according to the flood disaster risk index values of multiple grids, the flood disaster risk levels of each unit grid are divided, specifically including: when the flood disaster risk index value is less than or equal to the first preset threshold, the flood disaster risk level is determined to be an extremely low risk; when the flood disaster risk index value is greater than the first preset threshold and less than or equal to the second preset threshold, the flood disaster risk level is determined to be a low risk; when the flood disaster risk index value is greater than the second preset threshold and less than or equal to the third preset threshold, the flood disaster risk level is determined to be a medium risk; when the flood disaster risk index value is greater than the third preset threshold and less than or equal to the fourth preset threshold, the flood disaster risk level is determined to be a high risk; when the flood disaster risk index value is greater than the fourth preset threshold, the flood disaster risk level is determined to be an extremely high risk.

[0110] Specifically, the first preset threshold, the second preset threshold, the third preset threshold, and the fourth preset threshold are set according to specific engineering practices.

[0111] Specifically, the area corresponding to the flood risk index range of 0 - 0.2 is an extremely low risk area, the area corresponding to 0.2 - 0.4 is a low risk area, the area corresponding to 0.4 - 0.6 is a medium risk area, the area corresponding to 0.6 - 0.8 is a high risk area, and the area corresponding to 0.8 - 1.0 is an extremely high risk area.

[0112] S102: According to the flood disaster risk level distribution grid images of the first preset year and the second preset year, calculate the flood risk level state transition matrix.

[0113] Specifically, based on the flood disaster risk level distribution grid graphics of the first preset year and the second preset year, the flood disaster risk level quantity transition matrix is generated by the Markov tool of the IDRISI software.

[0114] S103: Through the characteristic analysis of the flood disaster risk, generate the flood disaster risk level transformation suitability atlas.

[0115] Specifically, the adaptability files of each level of flood disaster risk are generated by the Decision Wizard tool of the IDRISI software, and then the adaptability files of each level of flood disaster risk are integrated by the Collection Editor tool of the IDRISI software to generate the flood disaster risk level transformation suitability atlas.

[0116] S104: Input the flood disaster risk level distribution grid image of the second preset year, the flood risk level state transition matrix, and the flood disaster risk level transformation suitability atlas into the flood disaster risk level prediction model to obtain the flood disaster risk level distribution map of the third preset year.

[0117] The third preset year can be set according to specific engineering practices. For example, the third preset year is 2025.

[0118] The flood disaster risk level prediction model is a cellular automata-Markov chain (CA-Markov) model, and the flood risk level change is predicted according to the flood disaster risk level prediction model.

[0119] In an exemplary embodiment, the flood disaster risk level prediction model includes a cellular automata model and a Markov model; the flood disaster risk level distribution raster image, the flood risk level state transition matrix, and the flood disaster risk level transformation suitability atlas of the second preset year are input into the flood disaster risk level prediction model to obtain the flood disaster risk level distribution map of the third preset year, specifically including: taking the unit raster in the flood disaster risk level distribution raster image of the second preset year as the cell in the cellular automata model; for each raster in the flood disaster risk level distribution raster image of the second preset year, determining the flood risk level of the raster in the flood disaster risk level distribution raster image of the third preset year according to the flood risk level of the raster in the flood disaster risk level distribution raster image of the second preset year, the cell conversion rules between the second preset year and the third preset year, and the neighborhood of the cell; determining the quantity change of the flood risk level in the third preset year according to the quantity state of the flood risk level in the second preset year and the flood risk level state transition matrix; and determining the flood disaster risk level distribution map of the third preset year according to the flood risk levels of all the rasters in the flood disaster risk level distribution raster image of the third preset year, the quantity change of the flood risk level in the third preset year, and the flood disaster risk level transformation suitability atlas.

[0120] Specifically, in the flood disaster risk level prediction model, the cellular automata model is as shown in formula (4):

[0121] S T+1 = f(S T , N) (4)

[0122] where S T , S T+1 respectively represent the flood risk levels of the unit raster at the T and T+1 moments of the cell, f is the cell conversion rule between the two moments, and N is the neighborhood of the cell;

[0123] The Markov model is used to represent the quantity change of different flood risk levels, and the Markov model is as shown in formula (5):

[0124] Y T+1 = P T Y T (5)

[0125] where YT and Y T+1 respectively represent the flood risk level quantity status at times T and T+1, and P T represents the flood risk level quantity transfer matrix between times T and T+1.

[0126] According to the flood risk levels of all grids in the grid image of the flood disaster risk level distribution in the third preset year, the quantity change of the flood risk level in the third preset year, and the atlas of the suitability of flood disaster risk level transformation, determine the flood disaster risk level distribution map of the third preset year. The particularity of the numerical values of some evaluation factors in the atlas of the suitability of flood disaster risk level transformation determines that some grids will not suffer from flood disasters. For example, if the terrain slope in the evaluation factor is at a very high altitude, the corresponding grid is marked as not suffering from flood disasters in the flood disaster risk level distribution map of the third preset year.

[0127] Specifically, the specific principle of the Cellular Automata-Markov Chain (CA-Markov) model is as follows:

[0128] Principle of the CA model: The CA model refers to the Cellular Automata model. The CA system includes discrete state units and finite state units. A CA has four basic components, namely discrete time steps, discrete spatial units, neighborhood, and transition rules. This model utilizes the discrete characteristics presented by cells in space and state, and performs unit transformation in the discrete time dimension according to certain local rules. The cells in the present invention are each grid in the flood risk level grid image.

[0129] Principle of the Markov model: The Markov chain is a continuous random process, where the state at time T+1 is only affected by the state at time T and is independent of the states before time T. The Markov model transmits information through the formation of a transition matrix, and the transition matrix reflects the change probability between time periods. In the present invention, the Markov model is mainly used to represent the quantity change of different flood risk levels.

[0130] The present invention uses the CA-Markov coupling model to dynamically predict the flood disaster risk level in the study area. This coupling model integrates the advantages of the CA model in storing and transporting information for discrete spatial units and the characteristics of the Markov model in transmitting the quantity change of flood disaster risk levels between time periods, thereby realizing the dynamic evolution of the flood disaster risk level distribution in space and time.

[0131] In an exemplary embodiment, the embodiment includes: based on the predicted flood disaster risk level distribution map, draw a Receiver Operating Characteristic (ROC) curve; calculate the area under the ROC curve; evaluate the prediction accuracy of the flood disaster risk level prediction model through the area under the curve.

[0132] Specifically, taking the flood disaster risk level distribution map, the transition matrix, and the suitability atlas as input items, the flood disaster risk level distribution map for the required year is predicted through the CA-Markov model module of the IDRISI software. Based on the flood disaster risk level distribution map generated by the evaluation in the corresponding year, by comparing the predicted image and the evaluated image of the predicted map, the Receiver Operator Characteristic (ROC) curve of the predicted image data is drawn, and the prediction model accuracy is evaluated according to the Area Under the Curve (AUC) value.

[0133] In an exemplary embodiment, the prediction accuracy of the flood disaster risk level prediction model is evaluated by the area under the curve, specifically including: when the area under the curve is less than or equal to the first prediction threshold, it is determined that the prediction accuracy of the flood disaster risk level prediction model is poor; when the area under the curve is greater than the first prediction threshold and less than or equal to the second prediction threshold, it is determined that the prediction accuracy of the flood disaster risk level prediction model is average; when the area under the curve is greater than the second prediction threshold, it is determined that the prediction accuracy of the flood disaster risk level prediction model is high.

[0134] Specifically, to verify the prediction accuracy of the flood disaster risk level prediction model, based on the evaluation results, by comparing the prediction results and the evaluation results, the ROC curve of the flood disaster risk level prediction data is drawn, and the prediction accuracy of the model is evaluated by calculating the AUC value. When AUC > 0.85, it indicates that the prediction result is good; when 0.70 < AUC < 0.85, it indicates that the prediction accuracy is average; when AUC < 0.70, it indicates that the prediction accuracy is poor.

[0135] The present invention uses the CA-Markov coupling model to dynamically predict the flood disaster risk level in the study area. This coupling model integrates the advantages of the CA model in processing information storage and transportation in discrete spatial units and the characteristics of the Markov in transferring the quantity change of the flood disaster risk level between time periods, so as to realize the dynamic evolution of the flood disaster risk level distribution in space and time.

[0136] The key technical part of this coupling model lies in the conversion rules of the CA model. The transfer rules consist of two parts: one is the quantity state transition matrix of flood disaster risk levels calculated by the Markov model, which is achieved by the Markov model tool in IDRISI software through calculating the flood disaster risk level distribution images at two moments; the other is the suitability atlas for the transformation of flood disaster risk levels, which is composed of the suitability files for each level of flood risk. First, analyze the characteristics of flood risks at each level, and use the Decision Wizard tool in IDRISI software to generate the suitability files for each level of flood disaster risk by setting restrictive factors and influential factors. After the suitability files for each level of flood disaster risk are completed, integrate the suitability atlas for the transformation of flood disaster risk levels through the Collection Editor tool in IDRISI software.

[0137] Take the flood disaster risk level distribution raster image in the starting year of prediction, the quantity state transition matrix of flood disaster risk levels generated by the Markov model, and the suitability atlas for the transformation of flood disaster risk levels as input items, and use the CA-Markov model module in IDRISI software to predict the flood disaster risk level distribution map for the required year.

[0138] Verify the prediction accuracy of the flood disaster risk level prediction model. Based on the evaluation results, draw the ROC curve of the flood disaster risk level prediction data by comparing the prediction results and the evaluation results, and evaluate the prediction accuracy of the model by calculating the AUC value. When AUC > 0.85, it indicates that the prediction results are good; when 0.70 < AUC < 0.85, it indicates that the prediction accuracy is average; when AUC < 0.70, it indicates that the prediction accuracy is poor.

[0139] Predict the flood disaster risk level distribution map for the future horizontal year in the study area, update the data in the starting year, and repeat the operation process of the CA-Markov model in step five to predict and form the flood disaster risk level distribution map for the future horizontal year.

[0140] Predict the flood disaster risk level distribution in the future horizontal year in the study area. Update the corresponding input items and repeat the operation process of the CA-Markov model to predict the flood disaster risk level distribution in the future year.

[0141] The present invention provides a method for a regional flood disaster risk prediction system. On the basis of fully considering the natural and social attributes of flood disasters, through studying the changes in meteorology, hydrology, topography, river systems, land use over time and the development trend of social economy in the study area, use the CA-Markov model to simulate and predict the dynamic evolution of the flood disaster risk level in the future horizontal year of the region.

[0142] In an exemplary embodiment, the main content of the technical solution of this invention application mainly includes three parts:

[0143] I. Collect relevant data and conduct a flood disaster risk level assessment for the study area. The process framework is as Figure 2 shown. The factors determining flood disasters can generally be summarized into three types, namely hazard factors, exposure factors, and vulnerability factors. Hazard factors refer to the decisive conditions for inducing floods. Generally speaking, precipitation is the decisive factor for inducing floods. Precipitation intensity and frequency determine the scale of floods, and thus affect the magnitude of flood disasters. In this invention, MAP, FHP–24h, and FHP–72h are selected as the evaluation factors for the hazard of flood disasters. Exposure factors include topographical and geomorphic factors such as those that induce precipitation to develop into floods, including conditions such as elevation, slope, and water system. In this invention, DEM, TS, SDE, DD, and BZW are selected as the exposure evaluation factors for flood disasters. The vulnerability of floods is reflected by the objects that bear flood disasters, generally including conditions such as the population and economy in the disaster-stricken area. In this invention, PG, GDPG, and LT are selected as the vulnerability evaluation factors for flood disasters.

[0144] II. Verify the accuracy of the prediction model. Based on the flood disaster risk level assessment, use the CA-Markov model to predict the flood disaster risk level distribution in the study area in the current year, and use the flood disaster risk level assessment in the study area in the current year as the data basis to verify the accuracy of the model. The technical framework is as Figure 3 shown. According to the flood disaster risk level distribution map of the study area in 2010 and the flood disaster risk level distribution map of the study area in 2015, use the Markov tool in IDRISI software to generate a flood disaster risk level quantity transfer matrix; use the Decision Wizard tool in IDRISI software to set restrictive factors and influencing factors to generate an adaptability file for each level of flood disaster risk, and use the Collection Editor tool in IDRISI software to generate a flood disaster risk level transformation suitability atlas for this adaptability file; according to the flood disaster risk level transformation suitability atlas, the flood disaster risk level quantity transfer matrix, and the flood disaster risk level distribution map of the study area in 2015, use the CA-Markov model in IDRISI software to predict the flood disaster risk level distribution map of the study area in 2020; judge the accuracy of the prediction model based on the flood disaster risk level distribution map of the study area in 2020 and the predicted flood disaster risk level distribution map of the study area in 2020.

[0145] III. Simulate and predict the flood disaster risk level distribution in the future horizontal year of the study area, as Figure 4As shown in the figure, according to the flood disaster risk level distribution map of the study area in 2015 and the flood disaster risk level distribution map of the study area in 2020, determine the flood disaster risk level quantity transfer matrix; according to the flood disaster risk level quantity transfer matrix, the flood disaster risk level transformation suitability atlas, and the flood disaster risk level distribution map of the study area in 2020, predict the flood disaster risk level distribution map of the study area in 2025.

[0146] Next, the technical solution of the present invention will be decomposed and specifically described through the embodiments of this application:

[0147] Step S1, establish a flood disaster risk level assessment database. The original data required by the present invention includes precipitation data of meteorological stations in the study area, DEM raster data, river system vector data, and population, GDP, and land use raster data.

[0148] Sub-step S1.1, the daily precipitation data of the precipitation data of meteorological stations in the study area is obtained by downloading from a certain website. Calculate MAP, FHP-24h, and FHP-72h of each station through the obtained data. Among them, FHP-24h is the number of times the precipitation in 24 hours exceeds 100 mm, and FHP-72h is the number of times the precipitation in 72 hours exceeds 200 mm. Through the longitude and latitude of the meteorological station, directly import the calculated relevant data into the geographic information system software ArcGIS. Use the Define Projection tool in ArcGIS software to define the projection coordinates, and then use the Kriging tool in ArcGIS software to obtain the raster images of MAP, FHP-24h, and FHP-72h in the study area by Kriging interpolation method.

[0149] Sub-step S1.2, the DEM image data of the study area is downloaded from a certain website. Calculate the raster image of TS in the study area through the Slop tool in ArcGIS software, and calculate the raster image of SDE in the study area through the Block Statistics tool in ArcGIS software.

[0150] Sub-step S1.3, the river system vector data of the study area is downloaded from a certain website. Calculate the raster image of DD in the study area through the LineDensity tool in ArcGIS software, and calculate the raster image of BZW in the study area through the Buffer tool in ArcGIS software.

[0151] Sub-step S1.4, the population and GDP raster data of the study area is downloaded from a certain website to obtain the PG and GDPG raster images of the study area.

[0152] Sub-step S1.5: The land use grid data of the study area was downloaded from a certain website. According to the importance of land categories, using the Raster Reclass tool in ArcGIS software, the construction land, cultivated land, forest land, grassland, unused land, and water body in the grid data were respectively assigned the values of 1, 2, 3, 4, 5, and 6 to obtain the LT grid image of the study area.

[0153] The obtained grid image data of MAP, FHP-24h, FHP-72h, DEM, TS, SDE, DD, BZW, PG, GDPG, and LT in the study area were processed by the Define Projection tool and the Resample tool in ArcGIS software, and finally formed grid image data with the same spatial projection and consistent spatial resolution.

[0154] The 11 grid image data obtained in the embodiment of the present invention through step S1 are as Figure 5 shown.

[0155] Step S2: Data standardization processing. For the grid data obtained in step S1, the grid data was standardized by the Fuzzy method, and the Raster Calculator tool in ArcGIS software was used to perform standardization processing on 11 grid data respectively.

[0156] For the grid data standardized by the Fuzzy method, the maximum Value of the grid is 1, the minimum is 0, and the rest are between 0 and 1.

[0157] The 11 grid image data obtained in the embodiment of the present invention through the standardization processing in step S2 are as Figure 6 shown.

[0158] Step S3: Determine the weights of each evaluation factor. The AHP method is used to determine the weights of each evaluation factor for flood disaster risk. The process of this method is as Figure 7 shown. Analyze the causes, select evaluation indicators; construct a judgment matrix by pairwise comparison; solve the eigenvector, solve the maximum eigenvalue, and perform a consistency check on the judgment matrix. If the consistency check passes, the normalized weight vector is adopted. If the consistency check fails, adjust the judgment matrix.

[0159] Sub-step S3.1: Establish a hierarchical structure framework for flood risk level assessment. By analyzing the causes of floods, the influencing factors of flood disasters, and the bearing objects of flood disasters, a hierarchical structure framework for flood risk level assessment is constructed, as Figure 8As shown in the figure, the target layer is flood disaster risk assessment, the criterion layer includes hazard, exposure, and vulnerability. Among the index layer, the hazard includes MAP, FHP-24h, and FHP-72h, the exposure includes DEM, TS, SDE, DD, and BZW, and the vulnerability includes PG, GDPG, and LT. Due to the natural and social attributes of flood disasters, flood disasters are manifested as hazard, exposure, and vulnerability. The hazard of flood disasters uses MAP, FHP-24h, and FHP-72h as evaluation factors, the exposure uses DEM, TS, SDE, DD, and BZW as evaluation factors, and the vulnerability uses PG, GDPG, and LT as evaluation factors.

[0160] Sub-step S3.2, make pairwise comparisons between evaluation factors to construct a judgment matrix. Make pairwise comparisons between evaluation indicators, quantify according to a pre-specified scale by judging the importance of each evaluation indicator for the evaluation result, and then construct a judgment matrix. The results of the judgment matrix are generally obtained from expert scoring.

[0161] Sub-step S3.3, solve the weight vector of evaluation factors. The eigenvector Wi corresponding to the maximum eigenvalue λmax of the judgment matrix, after normalization, is the weight vector ωi of each evaluation indicator. Before this, in order to verify the scientificity of the judgment matrix, it is necessary to conduct a consistency test on it.

[0162] Step S4, evaluate the flood disaster risk in the study area in the current year, calculate the disaster risk index, and divide the risk levels.

[0163] Sub-step S4.1, according to the standardized evaluation factor raster image obtained in step S2 and the factor weights calculated in step S3, use the comprehensive scoring method to calculate the flood risk index of each raster in the study area. This step is implemented by the RasterCalculator tool in ArcGIS software.

[0164] Sub-step S4.2, divide the flood disaster risk levels. From the calculated flood risk index, according to the range where the index is located, use the Raster Reclass tool in ArcGIS software to divide the raster risk levels: when μA ≤ 0.2, it is a very low risk; when 0.2 < μA ≤ 0.4, it is a low risk; when 0.4 < μA ≤ 0.6, it is a medium risk; when 0.6 < μA ≤ 0.8, it is a high risk; when μA > 0.8, it is an extremely high risk.

[0165] Step S5, verify the prediction accuracy of the flood disaster risk level prediction model. Predict the raster data of the flood disaster risk level distribution through the CA-Markov model. Based on the raster data of the flood disaster risk level distribution obtained from the evaluation, draw the ROC curve of the prediction data, and evaluate the prediction model accuracy through the AUC value of the curve. As Figure 4As shown in the figure, the embodiment of the present invention is based on the evaluation data of the flood risk level distribution in the study area in 2020. By comparing the predicted data and the evaluation data of the flood risk level distribution in 2020, the accuracy of the prediction model is evaluated. Based on the raster images of the flood disaster risk level distribution in the study area in 2010 and 2015 obtained through the evaluation steps S1 - S4, the flood disaster risk level distribution in 2020 is predicted to obtain predicted raster data. Through the comparison of the predicted data and the evaluation data, the accuracy of the prediction model is evaluated.

[0166] Sub - step S5.1, generate the quantity transfer matrix of flood disaster risk levels. Based on the raster images of the flood disaster risk level distribution in the study area in 2010 and 2015 obtained through evaluation, use the Markov tool in IDRISI software to generate the quantity transfer matrix of flood risk levels.

[0167] Sub - step S5.2, generate the suitability atlas of flood disaster risk level transitions.

[0168] Sub - step S5.2.1, generate the suitability files for each level of flood disaster risk. Through the characteristic analysis of each level of flood risk, use the Decision Wizard tool in IDRISI software to set the restrictive parameters and influencing parameter factors for each level of flood disaster risk, and generate the suitability files for each level of flood disaster risk.

[0169] Sub - step S5.2.2, generate the suitability atlas of flood disaster risk level transitions. Use the Collection Editor tool in IDRISI software to integrate the suitability files for each level of flood disaster risk and generate the suitability atlas of flood disaster risk level transitions.

[0170] Sub - step S5.3, predict the raster image of the flood risk level distribution in the current year of the study area. Use the CA - Markov module tool in IDRISI software, taking the raster image of the flood disaster risk level distribution in the study area in 2015 as the starting year, and inputting the quantity transfer matrix of flood risk levels and the suitability atlas of flood disaster risk level transitions to simulate and predict the raster image of the flood disaster risk level in the study area in 2020.

[0171] Sub - step S5.4, evaluate the accuracy of the prediction model.

[0172] Sub - step S5.4.1, use the Raster Reclass tool in ArcGIS software to perform binary classification on the raster image of the flood disaster risk level in the study area in 2020. By comparing the simulated and predicted raster image in the study area with the evaluated image data, if the simulated prediction level value of the raster is the same as the evaluated level value, assign a value of 1, otherwise assign a value of 0.

[0173] Sub-step S5.4.2: Using ArcGIS software, extract and export the raster image data after binary classification and the raster image data of the risk index value of the study area calculated in sub-step S4.1 into a txt file. Use the ROC curve tool in SPSS software to generate an ROC curve and calculate the AUC value of the curve. When AUC > 0.85, it indicates that the prediction result is good; when 0.70 < AUC < 0.85, it indicates that the prediction accuracy is average; when AUC < 0.70, it indicates that the prediction accuracy is poor. The ROC curve generated in the embodiment of this application is as Figure 9 shown.

[0174] Step S6: Predict the flood disaster risk level distribution map of the study area in the future horizontal year. Repeat steps S5.1 - S5.3 to simulate and generate the raster image map of the flood disaster risk level distribution in the study area in 2025. Among them, the flood disaster risk level quantity transfer matrix in sub-step S5.2.1 is generated from the evaluation data of 2015 and 2020, and the starting year data in sub-step S5.3 is replaced with the raster image data of the flood disaster risk level distribution in the study area generated by evaluation in 2020. The technical process of simulating the raster image of the flood risk level distribution in 2025 in the embodiment of this application is as Figure 4 shown. The simulation and prediction results in the embodiment of this application are as Figure 10 shown.

[0175] Compared with the existing related technologies, in the flood disaster risk research of this application, the characteristics that the flood disaster risk changes due to the changes of meteorological elements, environmental elements and social and economic elements are fully considered, and the dynamic evolution of the flood disaster risk level in the simulated prediction area is carried out. For flood disasters, the achievements obtained can provide certain scientific support for flood disaster prevention departments to pre-deploy flood prevention measures and can more effectively prevent flood disasters.

[0176] When applying a flood disaster risk level prediction method provided by the present invention, it is not necessary to execute according to the Figure 1 sequence of each step shown. The specific execution sequence of each step can be determined as needed, and the present invention does not limit this.

[0177] The above is a flood disaster risk level prediction method provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding flood disaster risk level prediction device, as Figure 11 shown.

[0178] Figure 11 is a schematic diagram of a flood disaster risk level prediction device provided by the present invention, including:

[0179] An acquisition module 1101, configured to evaluate the flood disaster risk levels of a first preset year and a second preset year, and obtain flood disaster risk level distribution raster images of the first preset year and the second preset year.

[0180] A calculation module 1102, configured to calculate a flood risk level state transition matrix according to the flood disaster risk level distribution raster images of the first preset year and the second preset year.

[0181] A generation module 1103, configured to generate a flood disaster risk level transition suitability atlas by analyzing the characteristics of flood disaster risks.

[0182] A prediction module 1104, configured to input the flood disaster risk level distribution raster image of the second preset year, the flood risk level state transition matrix, and the flood disaster risk level transition suitability atlas into a flood disaster risk level prediction model, and obtain a flood disaster risk level distribution map of a third preset year.

[0183] For the specific limitations on a flood disaster risk level prediction device, reference can be made to the limitations on a flood disaster risk level prediction method in the foregoing text, which will not be elaborated here. Each module in the above flood disaster risk level prediction device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0184] The present invention also provides a computer-readable storage medium, which stores a computer program that can be used to execute the Figure 1 provided flood disaster risk level prediction method.

[0185] The present invention also provides Figure 12 a schematic structural diagram of the computer device shown in Figure 12 As shown, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, other hardware required for other services may also be included. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the Figure 1 provided flood disaster risk level prediction method.

[0186] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0187] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as within the scope recorded by the present invention.

Claims

1. A method for predicting the risk level of flood disasters, characterized in that, Including: Evaluating the flood disaster risk levels of the first preset year and the second preset year respectively to obtain the grid images of the flood disaster risk level distributions of the first preset year and the second preset year; Calculating a flood risk level state transition matrix based on the grid images of the flood disaster risk level distributions of the first preset year and the second preset year; Generating an atlas of the suitability for changes in the flood disaster risk level through the characteristic analysis of the flood disaster risk; Inputting the grid image of the flood disaster risk level distribution of the second preset year, the flood risk level state transition matrix, and the atlas of the suitability for changes in the flood disaster risk level into a flood disaster risk level prediction model to obtain the flood disaster risk level distribution map of the third preset year.

2. The method according to claim 1, wherein The step of evaluating the flood disaster risk levels of the first preset year and the second preset year respectively to obtain the grid images of the flood disaster risk level distributions of the first preset year and the second preset year specifically includes: Taking any one of the first preset year and the second preset year as an example, obtaining the initial data of multiple evaluation factors affecting flood disasters and converting the initial data to generate multiple grid data; Performing standardization processing on the multiple grid data to obtain multiple evaluation factor data; Determining the weights of multiple evaluation factors and determining the flood disaster risk index values of multiple grids according to the multiple evaluation factor data and the weights of multiple evaluation factors; Dividing the flood disaster risk levels of each unit grid according to the flood disaster risk index values of multiple grids to obtain the grid image of the flood disaster risk level distribution.

3. The method according to claim 2, wherein The evaluation factors include hazard evaluation factors, exposure evaluation factors, and vulnerability evaluation factors; the hazard evaluation factors include the annual average precipitation, the frequency of heavy precipitation in 24 hours, and the frequency of heavy precipitation in 72 hours, the exposure evaluation factors include digital elevation, terrain slope, elevation standard deviation, river network density, and water area buffer zone, and the vulnerability evaluation factors include population per kilometer grid, gross domestic product per kilometer grid, and land type; The standardization formulas for the annual average precipitation, the frequency of heavy precipitation in 24 hours, the frequency of heavy precipitation in 72 hours, the terrain slope, the river network density, the water area buffer zone, the population per kilometer grid, and the gross domestic product per kilometer grid are: where μ A (x) is the evaluation factor data, x is the initial evaluation factor data, and a and b are the lower and upper limits of the evaluation factor threshold, respectively; The standardization formulas for the digital elevation, the elevation standard deviation, and the land type are: Among them, μ B (x) is the evaluation factor data, x is the initial evaluation factor data, and a and b are the lower and upper limits of the evaluation factor threshold, respectively.

4. The method according to claim 2, wherein The step of determining the weights of multiple evaluation factors specifically includes: Constructing a judgment matrix based on the importance degree of each evaluation factor to the evaluation result; Performing a consistency test on the judgment matrix; If the consistency check of the judgment matrix passes, normalizing the eigenvector corresponding to the maximum eigenvalue of the judgment matrix to obtain the weights of the multiple evaluation factors.

5. The method according to claim 2, characterized in that, The calculation formula for the flood disaster risk index value is: Among them, μ A is the flood disaster risk index value, μ A (x i ) is the evaluation factor data of the evaluation factor x i , ω i is the weight of the evaluation factor x i , and n is the number of evaluation factors.

6. The method according to claim 2, wherein The step of dividing the flood disaster risk levels of each unit grid according to the flood disaster risk index values of multiple grids specifically includes: When the flood disaster risk index value is less than or equal to the first preset threshold, determining that the flood disaster risk level is an extremely low risk; When the flood disaster risk index value is greater than the first preset threshold and less than or equal to the second preset threshold, determine that the flood disaster risk level is a low risk; When the flood disaster risk index value is greater than the second preset threshold and less than or equal to the third preset threshold, determine that the flood disaster risk level is a medium risk; When the flood disaster risk index value is greater than the third preset threshold and less than or equal to the fourth preset threshold, determine that the flood disaster risk level is a high risk; When the flood disaster risk index value is greater than the fourth preset threshold, determine that the flood disaster risk level is an extremely high risk.

7. The method according to claim 1, characterized in that, The flood disaster risk level prediction model includes a cellular automata model and a Markov model; the process of inputting the flood disaster risk level distribution raster image of the second preset year, the flood risk level state transition matrix, and the flood disaster risk level transformation suitability atlas into the flood disaster risk level prediction model to obtain the flood disaster risk level distribution map of the third preset year specifically includes: Regarding the unit raster in the flood disaster risk level distribution raster image of the second preset year as the cell in the cellular automata model; For each raster in the flood disaster risk level distribution raster image of the second preset year, determine the flood risk level of the raster in the flood disaster risk level distribution raster image of the third preset year according to the flood risk level of the raster in the flood disaster risk level distribution raster image of the second preset year, the cell conversion rules between the second preset year and the third preset year, and the neighborhood of the cell; Determine the change in the quantity of the flood risk level in the third preset year according to the quantity state of the flood risk level in the second preset year and the flood risk level state transition matrix; Determine the flood disaster risk level distribution map of the third preset year according to the flood risk levels of all the rasters in the flood disaster risk level distribution raster image of the third preset year, the change in the quantity of the flood risk level in the third preset year, and the flood disaster risk level transformation suitability atlas.

8. The method according to claim 7, wherein The cellular automata model is: S T+1 = f(S T , N); where S T and S T+1 represent the flood risk levels of the unit grid at the time of the meta-cell T and T+1 respectively, f is the meta-cell conversion rule between the two times, and N is the neighborhood of the meta-cell; The Markov model is used to represent the change in the quantity of different flood risk levels, and the Markov model is: Y T+1 = P T Y T ; Among them, Y T and Y T+1 represent the flood risk level quantity status at times T and T+1 respectively, and P T represents the flood risk level quantity transition matrix between times T and T+1.

9. The method according to claim 1, characterized in that, The method further includes: Based on the predicted flood disaster risk level distribution map, draw a receiver operating characteristic curve; Calculate the area under the curve of the receiver operating characteristic curve; Evaluate the prediction accuracy of the flood disaster risk level prediction model through the area under the curve.

10. The method according to claim 9, characterized in that, The process of evaluating the prediction accuracy of the flood disaster risk level prediction model through the area under the curve specifically includes: When the area under the curve is less than or equal to the first prediction threshold, determine that the prediction accuracy of the flood disaster risk level prediction model is poor; When the area under the curve is greater than the first prediction threshold and less than or equal to the second prediction threshold, determine that the prediction accuracy of the flood disaster risk level prediction model is average; When the area under the curve is greater than the second prediction threshold, determine that the prediction accuracy of the flood disaster risk level prediction model is high.

Images