Coastal area tropical cyclone disaster vulnerability quantitative evaluation method and system

Abstract

The invention provides a coastal area tropical cyclone disaster vulnerability quantitative evaluation method and system, and relates to the technical field of tropical cyclone disaster vulnerability evaluation. According to the method, a time-space association data set is constructed by integrating historical cyclone paths, social economy and disaster loss data, and a time-space distribution rule is extracted to generate a time-space evolution characteristic spectrum. A random forest model is used for screening key factors of exposure, sensitivity and adaptive capacity, a comprehensive model quantization parameter coupling relation is established, and a dynamic weight set is generated by combining subjective and objective fusion weighting. And calculating a vulnerability index based on multiple time nodes and space unit division, and outputting a space-time differentiation feature matrix. According to the invention, a scientific basis is provided for coastal disaster risk management and ocean space planning, and the accuracy and operability of an evaluation result are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tropical cyclone disaster vulnerability assessment, and in particular to a coastal area tropical cyclone disaster vulnerability quantitative assessment method and system. BACKGROUND

[0002] The current coastal area tropical cyclone disaster vulnerability assessment method has significant limitations: first, the data dimension is single: traditional methods rely more on meteorological data or static social and economic data, and do not correlate historical tropical cyclone paths, social and economic properties and disaster loss data in space and time, resulting in evaluation results that cannot reflect the dynamic interaction of disasters and disaster-bearing systems. Second, static analysis defects: existing researches mostly use static spatio-temporal analysis, which is difficult to capture the spatio-temporal evolution of tropical cyclones (such as seasonal frequency characteristics, regional impact differences), and cannot provide support for dynamic risk assessment. Third, index redundancy and subjectivity: vulnerability assessment indicators mostly rely on expert experience for screening, and lack objective methods such as machine learning to sort the importance of exposure, sensitivity and adaptability factors, resulting in missing or redundant key factors. Fourth, one-sided weight allocation: traditional weighting methods mostly use single methods such as subjective weighting methods (such as analytic hierarchy process) or objective weighting methods (such as entropy weighting method), without combining subjective and objective weights, resulting in insufficient scientificity of the evaluation results. Fifth, lack of spatio-temporal dynamic output: the vulnerability index output by existing methods is mostly static value or single spatial unit result, which cannot quantitatively show the differentiation characteristics of vulnerability in time and space, and is difficult to directly guide the targeted optimization of marine spatial planning.

[0003] The above problems make it difficult for existing evaluation methods to meet the needs of dynamic, precision and operability of marine spatial planning, and there is an urgent need for a systematic, multi-dimensional and dynamic quantitative evaluation method. SUMMARY

[0004] In view of the above problems existing in the prior art, the first aspect of the present application proposes a coastal area tropical cyclone disaster vulnerability quantitative assessment method, comprising: Step 1, obtaining historical tropical cyclone path data, social and economic data and disaster loss data of the target coastal area, forming a spatio-temporal correlation data set; Step 2, based on the spatio-temporal correlation data set, extracting the spatio-temporal distribution rule of the tropical cyclone and the disaster influence range, and generating a spatio-temporal evolution characteristic map; Step 3, inputting the spatio-temporal evolution characteristic map into a random forest model, sorting the importance of the exposure, sensitivity and adaptability factors of the disaster-bearing system, and outputting a vulnerability factor set; Step 4, establishing a comprehensive model including exposure parameters, sensitivity parameters and adaptability parameters according to the vulnerability factor set; Step 5: Combining subjective and objective weighting methods, assign weights to the exposure parameters, sensitivity parameters, and adaptability parameters in the integrated model to form a dynamic weight set; Step 6: Based on the dynamic weight set and the integrated model, perform vulnerability calculations on the target coastal area at multiple time nodes and in multiple spatial units, and output the spatiotemporal differentiation feature matrix.

[0005] In conjunction with the first aspect, in some implementations of the first aspect, step 1 includes: Step 11: Extract the center location, movement path, and intensity data of tropical cyclones in the target coastal area from the meteorological database to form a cyclone path dataset; Step 12: Extract population density, infrastructure distribution, and emergency resource data from the statistics department to form a socioeconomic attribute dataset; Step 13: Extract historical disaster economic losses, casualties, and recovery cycle data from the disaster database to form a disaster loss record set; Step 14: Perform spatiotemporal alignment and standardization on the cyclone path dataset, socioeconomic attribute dataset, and disaster loss record set to generate a spatiotemporal correlated dataset.

[0006] In conjunction with the first aspect, in some implementations of the first aspect, step 2 includes: Step 21: Analyze the frequency and intensity changes of tropical cyclones in different years, seasons and months from the spatiotemporal correlation dataset to generate a time distribution series; Step 22: Based on the landfall location and movement path of the tropical cyclone, delineate the disaster-affected area and generate a spatial distribution heat map; Step 23: Overlay the time distribution sequence with the spatial distribution heatmap to generate a spatiotemporal evolution feature map.

[0007] In conjunction with the first aspect, in some implementations of the first aspect, the average rate of change of the time distribution series is calculated based on the following formula: , in, Let x be the average rate of change of the time distribution series, and Median be the median. j Let x be the numerical value of the time distribution sequence at time j. k Let j be the value of the time distribution sequence at time k, where j > k > 1.

[0008] In conjunction with the first aspect, in some implementations of the first aspect, step 4 includes: Step 41: Based on the exposure factors in the vulnerability factor set, calculate the distribution density and exposure probability of the disaster-affected body within the disaster impact range; Step 42: Based on the sensitivity factors in the vulnerability factor set, quantify the physical structure disturbance resistance properties and ecological buffering capacity of the disaster-bearing body; Step 43: Based on the adaptability factor in the vulnerability factor set, determine the emergency response speed and resource recovery efficiency of the disaster-bearing body; Step 44: Input the exposure parameters, sensitivity parameters, and adaptability parameters into the integrated model to establish the coupling relationship between the parameters.

[0009] In conjunction with the first aspect, in some implementations of the first aspect, step 5 includes: Step 51: Subjective weights are assigned to the exposure parameters, sensitivity parameters, and adaptability parameters in the integrated model using the analytic hierarchy process (AHP) to generate a set of subjective weights. Step 52: Objective weights are assigned to the exposure parameters, sensitivity parameters, and adaptability parameters in the integrated model using the entropy weight method to generate an objective weight set; Step 53: Based on the principle of minimum relative information entropy, the subjective weight set and the objective weight set are combined and optimized to generate a dynamic weight set.

[0010] In conjunction with the first aspect, in some implementations of the first aspect, combinatorial optimization is calculated based on the following formula: , Among them, w j Let w be the weight after optimization of the j-th parameter combination, where n is 3 and j is one of the n parameters. 1j w represents the subjective weight of the first parameter. 2j This is the objective weight of the second parameter.

[0011] In conjunction with the first aspect, in some implementations of the first aspect, step 6 includes: Step 61: Divide the target coastal area into multiple spatial units according to administrative divisions or ecological functional zones; Step 62: Select multiple historical years or disaster cycles as time nodes based on the seasonal characteristics of tropical cyclones. Step 63: Based on the dynamic weight set and the integrated model, the exposure parameters, sensitivity parameters and adaptability parameters of each spatial unit at each time node are weighted and calculated to generate a vulnerability index. Step 64: Arrange the vulnerability indices according to spatial units and time nodes to generate a spatiotemporal differentiation feature matrix.

[0012] In conjunction with the first aspect, in some implementations of the first aspect, the vulnerability index is calculated based on the following formula: , Where V is the vulnerability index, E is the exposure parameter, S is the sensitivity parameter, and A is the adaptability parameter.

[0013] Secondly, the present invention provides a quantitative assessment system for the vulnerability of coastal areas to tropical cyclone disasters, comprising: The data acquisition module is used to acquire historical tropical cyclone path data, socio-economic data, and disaster loss data for the target coastal area, forming a spatiotemporal correlated dataset; The spatiotemporal analysis module, connected to the data acquisition module, is used to extract the spatiotemporal distribution patterns and disaster impact range of tropical cyclones based on spatiotemporal correlation datasets, and generate spatiotemporal evolution feature maps. The factor identification module, connected to the spatiotemporal analysis module, is used to input the spatiotemporal evolution feature map into the random forest model, rank the importance of the influencing factors of the exposure, sensitivity and adaptability of the disaster-bearing system, and output a set of vulnerability factors. The model building module, connected to the factor identification module, is used to build a comprehensive model that includes exposure parameters, sensitivity parameters, and adaptability parameters based on the set of vulnerability factors. The weight allocation module, connected to the model building module, is used to combine subjective and objective weighting methods to allocate weights to exposure parameters, sensitivity parameters, and adaptability parameters in the comprehensive model, forming a dynamic weight set. The vulnerability assessment module, connected to the weight allocation module, is used to perform vulnerability calculations on the target coastal area at multiple time nodes and in multiple spatial units based on a dynamic weight set and a comprehensive model, and outputs a spatiotemporal differentiation feature matrix.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: Step 1 integrates historical tropical cyclone path data, socio-economic data, and disaster loss data to form a spatiotemporal correlated dataset, breaking through the limitations of traditional single data dimensions and providing a foundation for the dynamic correlation of multi-source heterogeneous data, ensuring the multidimensionality and completeness of the assessment input. Step 2 extracts the spatiotemporal distribution patterns of tropical cyclones and the scope of disaster impact based on the spatiotemporal correlated dataset, generating a spatiotemporal evolution feature map, solving the problem that static analysis cannot capture the dynamic evolution of disasters, and providing spatiotemporal dynamic input for vulnerability mechanism analysis. For Step 3, the random forest model is used to rank the importance of exposure, sensitivity, and adaptability influencing factors, eliminating redundant index interference, screening the set of key vulnerability factors, and avoiding bias or omissions caused by traditional subjective screening. Step 4 establishes a comprehensive model containing exposure parameters, sensitivity parameters, and adaptability parameters based on the vulnerability factor set, realizing the coupling and quantification of multidimensional parameters, overcoming the one-sidedness of traditional models that only focus on a single dimension. Step 5 integrates subjective and objective weighting methods to generate a dynamic weight set, balancing expert experience with data-driven logic. This addresses the limitations of single weighting methods in weight allocation and enhances the scientific rigor and adaptability of parameter weights. Step 6, based on the dynamic weight set and the integrated model, performs vulnerability calculations for multiple time nodes and spatial units, outputting a spatiotemporal differentiation feature matrix. This overcomes the shortcomings of traditional static assessments, providing quantifiable and visualized spatiotemporal dynamic results that directly support targeted optimization in marine spatial planning.

[0015] Steps 1-2 achieve dynamic integration and spatiotemporal feature extraction of multi-source data, providing structured input for subsequent analysis; Steps 3-4 use machine learning to screen key factors and build a comprehensive model, eliminating redundant index interference and ensuring the scientific nature and multidimensionality of the parameter system; Steps 5-6 use subjective and objective fusion weights and spatiotemporal dynamic calculations to output refined evaluation results, forming a closed loop of the entire process from data integration → dynamic analysis → model construction → weight optimization → quantitative evaluation.

[0016] This invention systematically solves problems such as single data dimension, static analysis defects, redundant indicators, one-sided weighting, and static results, significantly improving the dynamism, accuracy, and operability of the assessment results, and providing direct scientific basis for disaster risk management and marine spatial planning. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 The diagram shown is a flowchart illustrating a method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters, according to an embodiment of the present invention.

[0019] Figure 2 The diagram shown is a structural schematic of a quantitative assessment system for the vulnerability of coastal areas to tropical cyclone disasters, provided by an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0021] The specific embodiments of the present invention will be described below.

[0022] Example 1 like Figure 1 As shown, this invention proposes a method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters, including: Step 1: Obtain historical tropical cyclone path data, socio-economic data, and disaster loss data for the target coastal area to form a spatiotemporal correlation dataset; Step 2: Based on the spatiotemporal correlation dataset, extract the spatiotemporal distribution patterns and disaster impact range of tropical cyclones, and generate a spatiotemporal evolution feature map; Step 3: Input the spatiotemporal evolution feature map into the random forest model, rank the importance of the influencing factors of the exposure, sensitivity and adaptability of the disaster-bearing system, and output the vulnerability factor set. Step 4: Based on the set of vulnerability factors, establish a comprehensive model that includes exposure parameters, sensitivity parameters, and adaptability parameters; Step 5: Combining subjective and objective weighting methods, assign weights to the exposure parameters, sensitivity parameters, and adaptability parameters in the integrated model to form a dynamic weight set; Step 6: Based on the dynamic weight set and the integrated model, perform vulnerability calculations on the target coastal area at multiple time nodes and in multiple spatial units, and output the spatiotemporal differentiation feature matrix.

[0023] Specifically, step 1, by integrating multi-source data and constructing a dynamic model, achieves a spatiotemporal dynamic quantitative analysis of disaster vulnerability in coastal areas. First, historical tropical cyclone path data, socioeconomic data, and disaster loss data undergo spatiotemporal alignment and standardization to form a spatiotemporally correlated dataset. This process ensures the multidimensionality and completeness of the data, overcoming the limitations of traditional single-dimensional data. For example, combining tropical cyclone path data from the meteorological database with population density data from the statistics department can reflect the correlation between disaster impact and socioeconomic distribution.

[0024] Step 2, the generation of spatiotemporal evolution feature maps, relies on extracting the spatiotemporal distribution patterns and disaster impact ranges of tropical cyclones. By statistically analyzing the frequency and intensity changes of tropical cyclones in different years, seasons, and months, a temporal distribution sequence is generated. Combined with the landfall location and movement path of tropical cyclones, the disaster-affected areas are delineated, and a spatial distribution heat map is generated. The overlay analysis of these two data not only reveals the spatiotemporal dynamic characteristics of the disaster but also provides structured input for subsequent model construction. For example, overlay analysis reveals a significant increase in the frequency of tropical cyclones in a certain region during the rainy season; combined with socioeconomic data, the potential risks in that region can be further identified.

[0025] Step 3 involves inputting the spatiotemporal evolution feature map into the model and then filtering out a set of key vulnerability factors by ranking the importance of factors influencing exposure, sensitivity, and adaptability. For example, the model might identify population density, infrastructure wind resistance level, and emergency resource distribution as key factors, eliminating interference from redundant indicators. This step avoids the subjective bias of traditional expert experience-based screening and improves the scientific rigor of factor selection.

[0026] Step 4 further integrates exposure parameters, sensitivity parameters, and adaptability parameters into the comprehensive model. Exposure parameters are quantified by calculating the distribution density and exposure probability of the affected body within the disaster's impact area; sensitivity parameters focus on assessing the physical structure's resistance to disturbance and its ecological buffering capacity; and adaptability parameters are defined by combining emergency response speed and resource recovery efficiency. The coupling relationship among these three parameters is established through mathematical models, such as using linear or nonlinear equations to reflect the interactions between parameters.

[0027] Step 5: Weight allocation employs a combined subjective and objective approach. A subjective weight set is generated using the analytic hierarchy process (AHP), while an objective weight set is generated using the entropy weighting method. Finally, the weights are optimized based on the principle of minimum relative information entropy. This dynamic weight set balances expert experience with data-driven logic. For example, in economically developed coastal regions, objective weights may emphasize the disaster resilience of infrastructure, while subjective weights may focus on the impact of emergency management policies.

[0028] Step 6: Vulnerability calculation across multiple time points and spatial units is achieved through a dynamic weight set and integrated model. For example, the target coastal area is divided into multiple spatial units according to administrative divisions, and historical years are selected as time points based on the seasonal characteristics of tropical cyclones, ultimately generating a spatiotemporal differentiation feature matrix. This matrix can intuitively display the vulnerability changes of different regions at different times. For example, the vulnerability index of a certain region increases significantly during the typhoon season, providing a direct basis for targeted optimization of marine spatial planning.

[0029] In this embodiment of the invention, the comprehensiveness and dynamic correlation of data dimensions are improved by integrating multi-source data and aligning it with spatiotemporal data; the spatiotemporal evolution feature map combined with the random forest model enables the objective screening of key factors; the fusion and optimization of subjective and objective weights enhances the scientificity and adaptability of parameter weights; and the output of the spatiotemporal differentiation feature matrix supports dynamic risk assessment and precise planning and decision-making.

[0030] In conjunction with the first aspect, in some implementations of the first aspect, step 1 includes: Step 11: Extract the center location, movement path, and intensity data of tropical cyclones in the target coastal area from the meteorological database to form a cyclone path dataset; Step 12: Extract population density, infrastructure distribution, and emergency resource data from the statistics department to form a socioeconomic attribute dataset; Step 13: Extract historical disaster economic losses, casualties, and recovery cycle data from the disaster database to form a disaster loss record set; Step 14: Perform spatiotemporal alignment and standardization on the cyclone path dataset, socioeconomic attribute dataset, and disaster loss record set to generate a spatiotemporal correlated dataset.

[0031] In practice, the cyclone track dataset is extracted from meteorological databases, including the center location, movement path, and intensity data of tropical cyclones. For example, the optimal track data for the North Indian Ocean provided by a typhoon warning center records tropical cyclone information since a certain time period, with a temporal resolution of 6 hours and spatial precision in latitude and longitude coordinates.

[0032] The socioeconomic attribute dataset encompasses population density, infrastructure distribution, and emergency resource data. For example, census data provided by the Bangladesh Statistics Department includes population size, education level, and number of shelters in each administrative district. Combined with Geographic Information System (GIS) technology, this data can generate population density heat maps and spatial distribution maps of infrastructure. The disaster loss record set is obtained from a global disaster data platform and includes data on economic losses, casualties, and recovery cycles from historical disasters.

[0033] Spatiotemporal alignment and standardization are core aspects of data integration. The temporal and spatial scales of different datasets need to be unified. For example, timestamps for tropical cyclone path data need to be converted to local Bangladesh time, and administrative divisions of socioeconomic data need to be matched with the spatial extent of disaster-affected areas. Standardization includes data normalization and missing value imputation. For instance, interpolation can be used to complete disaster loss data for certain years, or Z-score standardization can be used to eliminate dimensional differences.

[0034] Alternatively, if some data is missing, remote sensing data can be used as a substitute, such as using satellite imagery to invert population density or infrastructure distribution; or using machine learning models (such as LSTM) to predict historical disaster loss data.

[0035] In this embodiment of the invention, the efficient integration of multi-source heterogeneous data ensures the spatiotemporal consistency of input data; data standardization processing improves the stability and reliability of model calculation; and spatiotemporal alignment technology enhances the correlation analysis capability of multi-dimensional data.

[0036] In conjunction with the first aspect, in some implementations of the first aspect, step 2 includes: Step 21: Analyze the frequency and intensity changes of tropical cyclones in different years, seasons and months from the spatiotemporal correlation dataset to generate a time distribution series; Step 22: Based on the landfall location and movement path of the tropical cyclone, delineate the disaster-affected area and generate a spatial distribution heat map; Step 23: Overlay the time distribution sequence with the spatial distribution heatmap to generate a spatiotemporal evolution feature map.

[0037] In practice, the time distribution series is generated by statistically analyzing the interannual, seasonal, and monthly variations of tropical cyclones. For example, frequency statistics of tropical cyclone data in the Bay of Bengal region over the past 50 years show that September to November is the peak season for cyclones, and their intensity shows an upward trend.

[0038] Spatial distribution heatmaps are generated based on the landfall locations and movement paths of tropical cyclones. Using GIS spatial analysis tools, cyclone path data is converted into kernel density estimation maps, visually displaying hotspots of disaster impact. For example, the Khurna region on the southwest coast of Bangladesh, frequently affected by cyclone landfalls, has a significantly higher kernel density value than other areas.

[0039] The spatiotemporal evolution characteristic map is achieved by overlaying temporal distribution sequences with spatial distribution heatmaps. For example, overlaying seasonal cyclone frequency changes with spatial kernel density maps reveals that cyclones not only increase in frequency at the end of the rainy season but also expand their impact range northward and inland. This map provides spatiotemporal dynamic input for vulnerability mechanism analysis; for instance, if a region transitions from a low-risk area to a high-risk area over time, it should be prioritized for inclusion in the key control areas of marine spatial planning.

[0040] Alternatively, if the data resolution is insufficient, spatial interpolation methods (such as Kriging interpolation) can be used to optimize the heatmap; or wavelet analysis can be introduced to reveal the multi-scale periodicity of the time series.

[0041] In this embodiment of the invention, the combination of temporal and spatial analysis comprehensively captures the dynamic evolution of disasters; kernel density estimation and GIS technology enhance spatial visualization; and overlay analysis reveals spatiotemporal interactions, supporting refined risk assessment.

[0042] In conjunction with the first aspect, in some implementations of the first aspect, the average rate of change of the time distribution series is calculated based on the following formula: , in, Let x be the average rate of change of the time distribution series, and Median be the median. j Let x be the numerical value of the time distribution sequence at time j. k Let j be the value of the time distribution sequence at time k, where j > k > 1.

[0043] Specifically, the median method is used to calculate the average rate of change of the time distribution series to avoid interference from extreme values. In practice, the median of the numerical differences between adjacent time points in the time series is selected as the average rate of change. For example, for the annual frequency data of tropical cyclones over the past 30 years, the frequency change rate between every two years is calculated in chronological order, and then the median is taken to reflect the overall trend.

[0044] This method effectively reduces the impact of outliers (such as an abnormally high frequency of cyclones in a given year) on the calculation results by leveraging the robustness of the median. For example, if the frequency of cyclones surges in a given year due to El Niño, the median method can still accurately reflect the long-term trend, rather than the simple arithmetic mean which may be distorted by outliers.

[0045] Alternatively, if the data distribution is close to normal, the mean can be used instead of the median; or a weighted average method can be introduced, assigning different weights according to the time interval.

[0046] In this embodiment of the invention, the median method improves the anti-interference capability of the rate of change calculation; the long-term trend analysis is more in line with the actual disaster evolution characteristics; and at the same time, it provides reliable time series input parameters for the spatiotemporal dynamic model.

[0047] In conjunction with the first aspect, in some implementations of the first aspect, step 4 includes: Step 41: Based on the exposure factors in the vulnerability factor set, calculate the distribution density and exposure probability of the disaster-affected body within the disaster impact range; Step 42: Based on the sensitivity factors in the vulnerability factor set, quantify the physical structure disturbance resistance properties and ecological buffering capacity of the disaster-bearing body; Step 43: Based on the adaptability factor in the vulnerability factor set, determine the emergency response speed and resource recovery efficiency of the disaster-bearing body; Step 44: Input the exposure parameters, sensitivity parameters, and adaptability parameters into the integrated model to establish the coupling relationship between the parameters.

[0048] Step 4 achieves coupled quantification of multidimensional parameters by constructing a comprehensive model of exposure, sensitivity, and adaptability parameters. Exposure parameters are calculated using the distribution density of disaster-bearing bodies and exposure probability. For example, the exposure probability of the population in each region can be calculated by overlaying population density data with the cyclone's impact range; or the degree of infrastructure exposure can be assessed by the spatial intersection of infrastructure distribution maps and disaster paths.

[0049] Sensitivity parameters focus on physical resilience and ecological buffering capacity. For example, the wind resistance rating of building structures and the coverage of mangrove forests in coastal areas can serve as sensitivity indicators. Adaptability parameters, on the other hand, are quantified by combining emergency response speed (such as the arrival time of shelters) and resource recovery efficiency (such as the proportion of funds invested in post-disaster reconstruction). For example, an area equipped with sufficient emergency supplies and a rapid response mechanism will have a higher adaptability parameter value.

[0050] The coupling relationships in the comprehensive model are realized through mathematical equations. For example, the synergistic effect of exposure and sensitivity is expressed in a product form, and its inhibitory effect on vulnerability is reflected by the inverse relationship of the adaptability parameter. This model structure can intuitively demonstrate the interaction between various parameters; for example, high exposure and low adaptability together lead to a significant increase in the vulnerability index.

[0051] Alternatively, if nonlinear relationships need to be considered, machine learning models (such as neural networks) can be used to replace linear equations; or fuzzy logic can be introduced to handle uncertainties.

[0052] In this embodiment of the invention, the systematic integration of multi-dimensional parameters overcomes the one-sidedness of single-dimensional analysis; the mathematical model clearly expresses the coupling mechanism between parameters; and the quantitative output supports precise vulnerability classification management.

[0053] In conjunction with the first aspect, in some implementations of the first aspect, step 5 includes: Step 51: Subjective weights are assigned to the exposure parameters, sensitivity parameters, and adaptability parameters in the integrated model using the analytic hierarchy process (AHP) to generate a set of subjective weights. Step 52: Objective weights are assigned to the exposure parameters, sensitivity parameters, and adaptability parameters in the integrated model using the entropy weight method to generate an objective weight set; Step 53: Based on the principle of minimum relative information entropy, the subjective weight set and the objective weight set are combined and optimized to generate a dynamic weight set.

[0054] Specifically, step 5 generates a dynamic weight set by integrating subjective and objective weighting methods, overcoming the limitations of traditional single weighting methods. In implementation, the subjective weight allocation employs the Analytic Hierarchy Process (AHP). A hierarchical model is designed for exposure parameters, sensitivity parameters, and adaptability parameters. Experts in fields such as disaster risk management and marine planning are invited to compare and score each parameter pairwise, ultimately calculating the subjective weight set. For example, experts might consider emergency response speed more important than resource recovery efficiency in the adaptability parameter, thus assigning it a higher weight. This process fully incorporates expert experience, but subjective biases may arise due to differences in expert backgrounds.

[0055] The objective weight allocation employs the entropy weighting method, calculating information entropy based on the original data distribution of each parameter in the comprehensive model. For example, if the exposure parameter data for a certain region differs significantly at different time points, its entropy value is low, indicating that the parameter has high discriminative power for vulnerability assessment and should be assigned a higher weight. The entropy weighting method relies entirely on data-driven approaches, avoiding human interference, but may overlook policy or management needs in practical applications.

[0056] Combinatorial optimization, based on the principle of minimum relative information entropy, merges subjective and objective weight sets into a dynamic weight set. Specifically, by calculating the geometric mean of the two weights and normalizing it, the final weights reflect both expert experience and data patterns. For example, in economically developed coastal regions, objective data on infrastructure disaster resilience may show high importance, while experts may focus more on long-term climate adaptation strategies. Combinatorial optimization balances the weights of both. This dynamic adjustment mechanism can adapt to the assessment needs of different regions and periods. For instance, during periods of frequent disasters, objective weights may emphasize real-time data changes, while subjective weights reinforce the priority of emergency response strategies.

[0057] Alternatively, if simplified calculation is required, linear weighting can be used instead of geometric mean; or fuzzy comprehensive evaluation can be introduced to handle weight uncertainty.

[0058] In this embodiment of the invention, the integration of subjective and objective weights enhances the scientific rigor and practicality of the evaluation results; the dynamic weight set adapts to the evaluation needs of different spatiotemporal scenarios; and the minimum relative information entropy optimization ensures the rationality and stability of the weight allocation.

[0059] In conjunction with the first aspect, in some implementations of the first aspect, combinatorial optimization is calculated based on the following formula: , Among them, w j Let w be the weight after optimization of the j-th parameter combination, where n is 3 and j is one of the n parameters. 1j w represents the subjective weight of the first parameter. 2jThis is the objective weight of the second parameter.

[0060] Specifically, the combinatorial optimization formula effectively integrates subjective and objective weights through mathematical methods. In its implementation, the weight calculation in the formula is divided into two stages: First, the geometric mean of the subjective and objective weights for each parameter is calculated separately; second, the geometric mean of all parameters is normalized to ensure that the total weight sums to 1. For example, if the subjective weight of the exposure parameter is 0.4 and the objective weight is 0.6, then its geometric mean is 0.49, and after normalization, the final combined weights are obtained.

[0061] The core advantage of this method lies in its ability to balance extreme differences between subjective and objective weights using geometric mean. For example, if a parameter has an extremely low subjective weight (e.g., 0.1) and an extremely high objective weight (e.g., 0.9), the geometric mean will be 0.3, neither completely biased towards the data nor entirely reliant on expert judgment, thus avoiding the bias of a single weighting method. Furthermore, normalization ensures the comparability of parameter weights; for instance, among multiple parameters, those with higher combined weights will have a greater impact on vulnerability calculations.

[0062] In this embodiment of the invention, the geometric mean balances subjective and objective weights to avoid extreme biases; normalization ensures the overall consistency of the weight system; and formulaic operations enhance the repeatability and transparency of weight allocation.

[0063] In conjunction with the first aspect, in some implementations of the first aspect, step 6 includes: Step 61: Divide the target coastal area into multiple spatial units according to administrative divisions or ecological functional zones; Step 62: Select multiple historical years or disaster cycles as time nodes based on the seasonal characteristics of tropical cyclones. Step 63: Based on the dynamic weight set and the integrated model, the exposure parameters, sensitivity parameters and adaptability parameters of each spatial unit at each time node are weighted and calculated to generate a vulnerability index. Step 64: Arrange the vulnerability indices according to spatial units and time nodes to generate a spatiotemporal differentiation feature matrix.

[0064] Specifically, step 6 outputs a spatiotemporal differentiation feature matrix through vulnerability calculations across multiple time nodes and spatial units. In practice, spatial unit division is based on the administrative divisions or ecological functional zone boundaries of the target coastal area. For example, the coastal area of ​​Bangladesh can be divided into administrative districts such as Khurna and Barisal, or ecological functional units such as mangrove reserves and port economic zones. When dividing, it is necessary to ensure that the socio-economic attributes and disaster exposure characteristics within the units are relatively homogeneous; for example, the population density and infrastructure distribution within the same administrative district should have relatively small differences.

[0065] The selection of time points is based on the seasonal characteristics of tropical cyclones and historical disaster cycles. For example, years with significant El Niño events in the past 20 years are selected as high-disaster-intensity points, or the rainy season and non-rainy season are divided according to the monsoon cycle. Each time point corresponds to a different climate background and socio-economic status; for example, the adaptability parameters of a certain region may be significantly improved during the post-disaster reconstruction phase.

[0066] Vulnerability calculation is performed using a dynamic weight set and a comprehensive model. For each spatial unit and time node, the weighted values ​​of exposure parameters, sensitivity parameters, and adaptability parameters are calculated to generate a vulnerability index. For example, a port economic zone may have a high exposure parameter during the typhoon season, but its adaptability (such as flood control facilities) is strong; therefore, its vulnerability index after comprehensive calculation may be lower than that of an agricultural-dominated area. Finally, all results are arranged into a matrix according to spatial units and time nodes to visually demonstrate the spatiotemporal differences in vulnerability.

[0067] In this embodiment of the invention, multi-dimensional spatiotemporal partitioning supports refined risk assessment; dynamic weights combined with the model enhance computational adaptability; and matrix-based output provides a visual decision-making basis for targeted planning.

[0068] In conjunction with the first aspect, in some implementations of the first aspect, the vulnerability index is calculated based on the following formula: , Where V is the vulnerability index, E is the exposure parameter, S is the sensitivity parameter, and A is the adaptability parameter.

[0069] Specifically, the vulnerability index calculation formula quantifies comprehensive vulnerability through the mathematical relationship between exposure parameters, sensitivity parameters, and adaptability parameters. In practice, the formula uses the product of exposure (E) and sensitivity (S) to reflect the potential loss of the affected body, and then divides it by adaptability (A) to reflect its disaster reduction effect. For example, if a region has high exposure (E=0.8), high sensitivity (S=0.7), and low adaptability (A=0.3), then the vulnerability index V=(0.8×0.7) / 0.3≈1.87, indicating that the region is extremely vulnerable.

[0070] The formula's design logic is based on the idea that the synergistic effect of exposure and sensitivity amplifies disaster risk, while resilience mitigates risk through emergency response and recovery measures. For example, even if a region has high exposure, if its resilience (such as a robust early warning system) is strong enough, the vulnerability index can still be kept at a low level. The inverse relationship of the formula emphasizes the non-linear inhibitory effect of resilience on vulnerability; for example, a 10% increase in resilience may lead to a decrease in the vulnerability index of more than 10%.

[0071] In this embodiment of the invention, the product-reciprocal relationship accurately characterizes the interaction mechanism between parameters; the formula is concise and has a clear physical meaning, making it easy to apply in practice; the nonlinear model is more in line with the actual response law of disaster vulnerability.

[0072] Example 2 like Figure 2 As shown, in a second aspect, the present invention provides a quantitative assessment system for the vulnerability of coastal areas to tropical cyclone disasters, comprising: The data acquisition module is used to acquire historical tropical cyclone path data, socio-economic data, and disaster loss data for the target coastal area, forming a spatiotemporal correlated dataset; The spatiotemporal analysis module, connected to the data acquisition module, is used to extract the spatiotemporal distribution patterns and disaster impact range of tropical cyclones based on spatiotemporal correlation datasets, and generate spatiotemporal evolution feature maps. The factor identification module, connected to the spatiotemporal analysis module, is used to input the spatiotemporal evolution feature map into the random forest model, rank the importance of the influencing factors of the exposure, sensitivity and adaptability of the disaster-bearing system, and output a set of vulnerability factors. The model building module, connected to the factor identification module, is used to build a comprehensive model that includes exposure parameters, sensitivity parameters, and adaptability parameters based on the set of vulnerability factors. The weight allocation module, connected to the model building module, is used to combine subjective and objective weighting methods to allocate weights to exposure parameters, sensitivity parameters, and adaptability parameters in the comprehensive model, forming a dynamic weight set. The vulnerability assessment module, connected to the weight allocation module, is used to perform vulnerability calculations on the target coastal area at multiple time nodes and in multiple spatial units based on a dynamic weight set and a comprehensive model, and outputs a spatiotemporal differentiation feature matrix.

[0073] This system corresponds to the method provided in Example 1, and will not be described in detail here.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters, characterized in that, include: Step 1: Obtain historical tropical cyclone path data, socio-economic data, and disaster loss data for the target coastal area to form a spatiotemporal correlation dataset; Step 2: Based on the spatiotemporal correlation dataset, extract the spatiotemporal distribution patterns and disaster impact range of tropical cyclones, and generate a spatiotemporal evolution feature map; Step 3: Input the spatiotemporal evolution feature map into the random forest model, rank the importance of the influencing factors of the exposure, sensitivity and adaptability of the disaster-bearing system, and output the vulnerability factor set; Step 4: Based on the set of vulnerability factors, establish a comprehensive model that includes exposure parameters, sensitivity parameters, and adaptability parameters; Step 5: Combining subjective and objective weighting methods, assign weights to the exposure parameters, sensitivity parameters, and adaptability parameters in the comprehensive model to form a dynamic weight set; Step 6: Based on the dynamic weight set and the integrated model, perform vulnerability calculations on the target coastal area at multiple time nodes and in multiple spatial units, and output a spatiotemporal differentiation feature matrix.

2. The method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters according to claim 1, characterized in that, Step 1 includes: Step 11: Extract the center location, movement path, and intensity data of tropical cyclones in the target coastal area from the meteorological database to form a cyclone path dataset; Step 12: Extract population density, infrastructure distribution, and emergency resource data from the statistics department to form a socioeconomic attribute dataset; Step 13: Extract historical disaster economic losses, casualties, and recovery cycle data from the disaster database to form a disaster loss record set; Step 14: Perform spatiotemporal alignment and standardization on the cyclone path dataset, the socioeconomic attribute dataset, and the disaster loss record set to generate the spatiotemporal correlation dataset.

3. The method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters according to claim 1, characterized in that, Step 2 includes: Step 21: Statistically analyze the frequency and intensity changes of tropical cyclones in different years, seasons, and months from the spatiotemporal correlation dataset to generate a time distribution sequence; Step 22: Based on the landfall location and movement path of the tropical cyclone, delineate the disaster-affected area and generate a spatial distribution heat map; Step 23: Overlay the time distribution sequence with the spatial distribution heatmap to generate the spatiotemporal evolution feature map.

4. The method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters according to claim 3, characterized in that, The average rate of change of the time distribution sequence is calculated based on the following formula: , in, Let x be the average rate of change of the time distribution series, and Median be the median. j Let x be the numerical value of the time distribution sequence at time j. k Let j be the value of the time distribution sequence at time k, where j > k > 1.

5. The method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters according to claim 1, characterized in that, Step 4 includes: Step 41: Based on the exposure factors in the vulnerability factor set, calculate the distribution density and exposure probability of the disaster-affected body within the disaster impact range; Step 42: Based on the sensitivity factors in the vulnerability factor set, quantify the physical structural disturbance resistance and ecological buffering capacity of the disaster-bearing body; Step 43: Based on the adaptability factors in the vulnerability factor set, determine the emergency response speed and resource recovery efficiency of the disaster-bearing body; Step 44: Input the exposure parameter, the sensitivity parameter, and the adaptability parameter into the integrated model to establish the coupling relationship between the parameters.

6. The method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters according to claim 1, characterized in that, Step 5 includes: Step 51: Subjective weights are assigned to the exposure parameters, sensitivity parameters, and adaptability parameters in the comprehensive model using the analytic hierarchy process (AHP) to generate a set of subjective weights. Step 52: Objectively assign weights to the exposure parameters, sensitivity parameters, and adaptability parameters in the integrated model using the entropy weight method to generate an objective weight set; Step 53: Based on the principle of minimum relative information entropy, the subjective weight set and the objective weight set are combined and optimized to generate the dynamic weight set.

7. A method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters according to claim 6, characterized in that, Combinatorial optimization is calculated based on the following formula: ， Among them, w j The weights are the optimized weights for the j-th parameter combination, where j=1, 2, 3 correspond to the exposure parameter, sensitivity parameter, and adaptability parameter, respectively. 1j w represents the subjective weight of the j-th parameter (obtained through the analytic hierarchy process). 2j The objective weight of the j-th parameter (obtained via entropy weighting); n is 3, indicating there are three parameters; for the exposure parameter, j=1, w 11 w represents the subjective weight of the exposure parameter. 21 This represents the objective weight of the exposure parameter; similarly, for the sensitivity and adaptability parameters, j=2 and 3, respectively.

8. The method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters according to claim 1, characterized in that, Step 6 includes: Step 61: Divide the target coastal area into multiple spatial units according to administrative divisions or ecological functional zones; Step 62: Select multiple historical years or disaster cycles as time nodes based on the seasonal characteristics of tropical cyclones. Step 63: Based on the dynamic weight set and the integrated model, the exposure parameters, sensitivity parameters and adaptability parameters of each spatial unit at the time node are weighted and calculated to generate a vulnerability index; Step 64: Arrange the vulnerability indices according to the spatial units and the time nodes to generate the spatiotemporal differentiation feature matrix.

9. A method for quantitatively assessing the vulnerability of coastal areas to tropical cyclone disasters according to claim 8, characterized in that, The vulnerability index is calculated based on the following formula: ， Where V is the vulnerability index, E is the exposure parameter, S is the sensitivity parameter, A is the adaptability parameter, and w E Let w be the combined weight of E. S Let w be the combined weight of S. A Let A be the combined weight.

10. A quantitative assessment system for the vulnerability of coastal areas to tropical cyclone disasters, characterized in that, The system employs the method described in any one of claims 1 to 9, comprising: The data acquisition module is used to acquire historical tropical cyclone path data, socio-economic data, and disaster loss data for the target coastal area, forming a spatiotemporal correlated dataset; The spatiotemporal analysis module, connected to the data acquisition module, is used to extract the spatiotemporal distribution patterns and disaster impact range of tropical cyclones based on the spatiotemporal correlation dataset, and generate a spatiotemporal evolution feature map. The factor identification module, connected to the spatiotemporal analysis module, is used to input the spatiotemporal evolution feature map into the random forest model, rank the importance of the influencing factors of the exposure, sensitivity and adaptability of the disaster-bearing system, and output a set of vulnerability factors. The model building module, connected to the factor identification module, is used to build a comprehensive model including exposure parameters, sensitivity parameters, and adaptability parameters based on the vulnerability factor set. The weight allocation module, connected to the model construction module, is used to combine subjective weighting and objective weighting methods to allocate weights to the exposure parameters, sensitivity parameters, and adaptability parameters in the comprehensive model, forming a dynamic weight set. The vulnerability assessment module, connected to the weight allocation module, is used to perform vulnerability calculations on the target coastal area at multiple time nodes and in multiple spatial units based on the dynamic weight set and the integrated model, and output a spatiotemporal differentiation feature matrix.

Images