Regional differentiation-based short-time heavy rain disaster risk zoning evaluation system

Abstract

The present application relates to the technical field of short-time heavy rain disaster risk assessment, in particular to a short-time heavy rain disaster risk zoning assessment system based on regional differentiation, which is precisely adapted around regional characteristics, and breaks through the limitation of unified standards through hierarchical differentiated design, whole-process parameter calibration, multi-dimensional verification optimization, realizes the fine evaluation of one region one strategy from the three core links of regional division, index system and evaluation model; thereby solve the problem that the existing evaluation method does not fully consider the differences of different regions in topography, surface permeability, drainage capacity, disaster-bearing body characteristics and other differences, and the unified evaluation standard and threshold lead to insufficient accuracy of risk assessment.

Description

Technical Field

[0001] This invention relates to the field of disaster risk assessment technology for short-duration heavy rainfall, specifically to a regionally differentiated disaster risk assessment system for short-duration heavy rainfall. Background Technology

[0002] Against the backdrop of global warming, extreme weather events are becoming more frequent, more intense, and more severe. Extreme precipitation is increasingly concentrated within short timeframes, and the growth rate of intra-diurnal extreme precipitation has surpassed that of diurnal extreme precipitation. Short-duration heavy rainfall is characterized by its short duration, high intensity, suddenness, and unpredictability. It easily triggers disasters such as torrential rains and floods, urban waterlogging, landslides, and mudslides, posing a significant threat to economic development and the safety of people's lives and property.

[0003] In terms of short-duration heavy rainfall monitoring, traditional climate monitoring relies heavily on daily precipitation data, which tends to overestimate the intensity of prolonged, continuous, weak precipitation and underestimate the intensity of short-duration heavy precipitation. Furthermore, hourly precipitation data struggles to capture the short-duration rainfall patterns required for urban flooding. Secondly, existing assessment methods do not adequately consider differences in topography, surface permeability, drainage capacity, and the characteristics of disaster-bearing structures across different regions. Uniform assessment standards and thresholds lead to insufficient accuracy in risk assessment. Additionally, some assessments focus only on a single link in the disaster hazard or the vulnerability of disaster-bearing structures, failing to cover the entire disaster chain and lacking comprehensive consideration of factors such as risk ambiguity and the cumulative effect over time. Finally, existing research largely relies on short-term disaster data, which cannot comprehensively reflect historical disaster situations, affecting the scientific rigor and comprehensiveness of disaster risk classification.

[0004] Therefore, the present invention provides a regionally differentiated risk assessment system for disasters caused by short-term heavy rainfall to solve the above problems. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, this invention provides a regionally differentiated risk assessment system for short-term heavy rainfall disasters. This system addresses the problem that existing assessment methods do not fully consider the differences in topography, surface permeability, drainage capacity, and disaster-bearing characteristics of different regions, and that the lack of accuracy in risk assessment is caused by the use of uniform assessment standards and thresholds.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A regionally differentiated zoning assessment system for the disaster risk caused by short-term heavy rainfall includes:

[0008] The data acquisition module retrieves multi-source related data;

[0009] The data processing module preprocesses the multivariate correlated data to obtain preprocessed data;

[0010] The region partitioning module sequentially performs feature extraction, cluster analysis, boundary correction, and region delimitation on the preprocessed data to obtain the region partitioning results.

[0011] The indicator analysis module calculates the data of each key secondary indicator based on the regional division results; and performs weighted summation of each secondary indicator to obtain the primary indicator data.

[0012] The risk assessment module performs a detailed evaluation of the primary indicator data and preprocessed data to obtain risk zoning assessment results.

[0013] Preferably, the preprocessing of multivariate correlated data to obtain preprocessed data includes: supplementing missing data using linear interpolation; identifying and calibrating outlier data using the 3σ principle; and performing hierarchical and standardization processing on the calibrated data to obtain preprocessed data.

[0014] Preferably, the step of sequentially performing feature extraction, cluster analysis, boundary correction, and region demarcation on the preprocessed data to obtain the region division result includes: selecting corresponding quantitative data from the preprocessed data based on preset core influencing factors to obtain multidimensional feature data; performing cluster analysis on the multidimensional feature data using the K-means method to obtain initial cluster division results; performing boundary correction and region demarcation on the initial cluster division results to obtain second cluster division results; performing statistical analysis on the core influencing factors of each sub-region in the second cluster division results to obtain key disaster-causing features of the region; and integrating the second cluster division results and the key disaster-causing features of the region to obtain the region division result.

[0015] Preferably, the preset core influencing factors include precipitation intensity factor, topographic feature factor, permeability factor, disaster-bearing body factor, drainage capacity factor, and disaster frequency factor.

[0016] Preferably, the step of using the K-means method to perform cluster analysis on multidimensional feature data to obtain the initial clustering result includes: determining the optimal number of clusters using the elbow rule; selecting the number of clusters from all grids using random sampling as the initial cluster centers; iteratively calculating the Euclidean distance between the feature vector of each grid and the initial cluster center, assigning the grid to the nearest cluster, recalculating the mean of each cluster as the new cluster center, and when the contour coefficient meets the preset contour condition, exiting the loop and outputting the preliminary clustering result.

[0017] Preferably, the step of correcting the boundaries and defining the regions of the initial clustering results to obtain the second clustering results includes: adjusting the boundaries of the initial clustering results and adapting them to administrative boundaries based on geographical barriers to obtain the second clustering results.

[0018] Preferably, the calculation of key secondary indicator data based on the regional division results includes: calculating the surface permeability index, drainage efficiency index, and return period threshold based on the quantitative data in the regional division results; the surface permeability index... The calculation method is as follows:

[0019] ,

[0020] in, For slope, For roughness, Soil water storage coefficient, This is the land cover coefficient. The correction constants for different regions are used; the drainage coefficient corresponding to the drainage efficiency index is processed by the weighted summation method to calculate the drainage efficiency index; the generalized extreme value distribution method is used to process the precipitation sequence in the quantitative data for data screening, parameter estimation and preset threshold calculation formula to obtain the return period threshold.

[0021] Preferably, the detailed evaluation of the primary indicator data and preprocessed data to obtain the risk zoning assessment results includes: obtaining precipitation thresholds for different return periods in each region; fusing the primary indicator data and precipitation thresholds using fuzzy rules to generate preliminary risk quantification values; analyzing the return period benchmark and fuzzy inference clear values ​​based on preset regional differentiated grading rules to obtain preliminary risk levels; and optimizing the preliminary risk levels based on the time cumulative effect to obtain the risk zoning assessment results.

[0022] Preferably, the step of using fuzzy rules to fuse the primary indicator data and precipitation thresholds to generate a preliminary risk quantification value includes: fuzzifying the primary indicator data and precipitation thresholds to obtain multiple fuzzy subsets; using triangular membership functions to describe each fuzzy subset and determining an initial threshold interval; performing regional adaptation on the initial threshold interval based on the regional division results to determine the final threshold interval of the triangular membership functions; analyzing each triangular membership function using the Mamdani inference method based on a preset fuzzy rule base to obtain fuzzy output; and using the centroid method to convert the fuzzy output into a clear value to obtain a preliminary risk quantification value.

[0023] Preferably, the optimization of the preliminary risk level by time accumulation effect to obtain the risk zoning assessment result includes: calculating the time accumulation index according to the preset accumulation rule; predicting the probability of precipitation on the next day based on Bayes' theorem, using the average precipitation of the previous seven days and the average precipitation of the previous three days as prior information; correcting the risk level of each region according to the relationship between the time accumulation index, the probability of precipitation and the preset level correction conditions to obtain the final risk level; and integrating the information of each region that has determined the final risk level to obtain the risk zoning assessment result.

[0024] The beneficial effects of this invention are as follows:

[0025] 1. This invention focuses on precise adaptation based on regional characteristics. Through hierarchical differentiated design, full-process parameter calibration, and multi-dimensional verification and optimization, it breaks through the limitations of unified standards in three core aspects: regional division, indicator system, and evaluation model, and achieves refined evaluation with a one-policy-per-region approach. This solves the problem that existing evaluation methods do not fully consider the differences in topography, surface permeability, drainage capacity, and disaster-bearing body characteristics of different regions, and that unified evaluation standards and thresholds lead to insufficient accuracy in risk assessment.

[0026] 2. This invention takes the consistency of precipitation characteristics and the significant differences in disaster-causing factors as its two core elements, while taking into account the regional differentiation of natural geographical features, socio-economic level and disaster prevention and mitigation capabilities. This ensures that the regional division results not only conform to the spatiotemporal distribution pattern of short-term heavy precipitation, but also accurately match the disaster-causing risk characteristics of different regions.

[0027] 3. This invention focuses on resolving indicator ambiguity, organically integrating multiple factors, and adapting to regional differences. Through a four-step process—input variable fuzzification, fuzzy inference based on a pre-set fuzzy rule base, and defuzzified output—it transforms the high, medium, and low qualitative descriptions of four primary indicators—precipitation characteristics, underlying surface characteristics, vulnerability of disaster-bearing bodies, and adaptability of disaster-bearing bodies—into quantitative risk values. Simultaneously, it adjusts the fuzzy subset threshold and rule base according to the disaster-causing characteristics of different regions, ensuring that the fusion results accurately reflect the actual disaster-causing patterns in each region. This solves the problems of difficulty in quantifying qualitative descriptions, lack of logical factor fusion, and insufficient regional adaptability in traditional multi-factor assessments. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of a regionally differentiated short-term heavy rainfall disaster risk zoning assessment system according to the present invention. Detailed Implementation

[0029] The following will refer to the attached reference. Figure 1 The various embodiments of the present invention will be described in detail below. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0030] A regionally differentiated risk assessment system for disasters caused by short-term heavy rainfall, as shown in the attached document. Figure 1As shown, it includes: a data acquisition module, which acquires multivariate correlation data; multivariate correlation data includes precipitation data, topographic data, socioeconomic data, infrastructure data, and historical disaster data; among which, precipitation data includes hourly precipitation data from the time the station was established to the current time; topographic data includes altitude, topographic relief, slope, land use type, and soil type; socioeconomic data includes total regional population, population density, ratio of elderly to young population, percentage of female population, percentage of illiterate population, human development index, and per capita GDP; infrastructure data includes drainage pipeline coverage, number and efficiency of drainage equipment, distribution of transportation network, and distance to water system / fault zone, etc.; historical disaster data includes disaster data such as floods and geological disasters caused by rainstorms over the years.

[0031] The data processing module preprocesses the multivariate correlated data to obtain preprocessed data; it uses linear interpolation to supplement missing hourly precipitation data for a single day, and uses inverse distance weighted (IDW) interpolation to fill in missing station data for more than 3 consecutive days, combined with data from surrounding stations.

[0032] Abnormal precipitation data (such as extreme values ​​exceeding three standard deviations of the average for the same period) are identified using the 3σ principle. This is then combined with meteorological observation logs to determine if it is due to observational error. If so, the average for the same period from surrounding stations is used instead. Correlation analysis is performed on the same indicator data from different sources (such as precipitation data from meteorological stations and radar). Data with a correlation coefficient below 0.8 require verification of accuracy until data consistency is achieved.

[0033] The data was then processed into five levels, with land use types categorized as forest, grassland, farmland, water bodies, and others. Numerical indicators such as precipitation frequency and altitude were classified using the quantile method, while topographic relief and soil erosion were classified using the equal-interval method. The secondary data for indicators such as extreme precipitation triggering (T), disaster-bearing body vulnerability (V), and disaster-bearing body adaptability (A) were processed using the min-max normalization formula to obtain standardized preprocessed data, ensuring the uniformity and comparability of subsequent analyses.

[0034] The region segmentation module sequentially performs feature extraction, cluster analysis, boundary correction, and region demarcation on the preprocessed data to obtain the region segmentation results. The main process of the region segmentation module is as follows: Key features of each grid point are extracted from the preprocessed dataset, such as the 95th percentile threshold of 1-hour precipitation, 3-hour precipitation return period, topographic relief, surface permeability index, population density, and drainage efficiency index, forming a 6-dimensional feature vector. The K-means clustering algorithm is used to cluster the feature vector, with the number of clusters determined by the elbow rule. During clustering, the normalized values ​​of each feature are used as input to iteratively optimize the cluster centers until the clustering error converges. Boundary correction is performed on the clustering results by incorporating administrative boundaries and geographical barriers (such as mountains and rivers) to ensure the integrity and management convenience of the sub-regions, ultimately forming differentiated sub-regions and obtaining the region segmentation results.

[0035] The indicator analysis module calculates the data of each key secondary indicator based on the regional division results; it then performs a weighted summation of each secondary indicator to obtain the primary indicator data.

[0036] The risk assessment module performs a detailed evaluation of the primary indicator data and preprocessed data to obtain risk zoning assessment results.

[0037] This invention focuses on precise adaptation based on regional characteristics. Through hierarchical differentiated design, full-process parameter calibration, and multi-dimensional verification and optimization, it breaks through the limitations of unified standards in three core aspects: regional division, indicator system, and assessment model, and achieves refined assessment with a tailored approach for each region. This solves the problem that existing assessment methods do not fully consider the differences in topography, surface permeability, drainage capacity, and disaster-bearing body characteristics of different regions, and that unified assessment standards and thresholds lead to insufficient accuracy in risk assessment.

[0038] This invention takes the consistency of precipitation characteristics and the significant differences in disaster-causing factors as its two core elements, while taking into account the regional differentiation of natural geographical features, socio-economic level and disaster prevention and mitigation capabilities. This ensures that the regional division results not only conform to the spatiotemporal distribution pattern of short-term heavy precipitation, but also accurately match the disaster-causing risk characteristics of different regions.

[0039] In one embodiment of the present invention, preprocessing of multivariate correlation data to obtain preprocessed data includes: supplementing missing data using linear interpolation; identifying and calibrating outlier data using the 3σ principle; and classifying and standardizing the calibrated data to obtain preprocessed data.

[0040] Linear interpolation was used to supplement missing hourly precipitation data for a single day. For station data missing for more than three consecutive days, inverse distance weighted (IDW) interpolation was used in combination with data from surrounding stations to fill the gaps.

[0041] Abnormal precipitation data (such as extreme values ​​exceeding three standard deviations of the average for the same period) are identified using the 3σ principle. This is then combined with meteorological observation logs to determine if it is due to observational error. If so, the average for the same period from surrounding stations is used instead. Correlation analysis is performed on the same indicator data from different sources (such as precipitation data from meteorological stations and radar). Data with a correlation coefficient below 0.8 require verification of accuracy until data consistency is achieved.

[0042] The data was then processed into five levels, with land use types categorized as forest, grassland, farmland, water bodies, and others. Numerical indicators such as precipitation frequency and altitude were classified using the quantile method, while topographic relief and soil erosion were classified using the equal-interval method. The secondary data for indicators such as extreme precipitation triggering (T), disaster-bearing body vulnerability (V), and disaster-bearing body adaptability (A) were processed using the min-max normalization formula to obtain standardized preprocessed data, ensuring the uniformity and comparability of subsequent analyses.

[0043] In one embodiment of the present invention, the step of sequentially performing feature extraction, cluster analysis, boundary correction, and region delimitation on the preprocessed data to obtain region division results includes: selecting corresponding quantitative data from the preprocessed data based on preset core influencing factors to obtain multidimensional feature data; performing cluster analysis on the multidimensional feature data using the K-means method to obtain initial cluster division results; performing boundary correction and region delimitation on the initial cluster division results to obtain second cluster division results; performing statistical analysis on the core influencing factors of each sub-region in the second cluster division results to obtain key disaster-causing features of the region; and integrating the second cluster division results and the key disaster-causing features of the region to obtain the region division result.

[0044] Specifically, the preset core influencing factors include precipitation intensity factor, topographic feature factor, permeability factor, disaster-bearing body factor, drainage capacity factor, and disaster frequency factor. Among them, the precipitation intensity factor is quantified by the 95th percentile threshold of 1-hour precipitation, the topographic feature factor is quantified by the standardized value of topographic relief, the permeability factor is quantified by the surface permeability index, the disaster-bearing body factor is quantified by the standardized value of population density, the drainage capacity factor is quantified by the drainage efficiency index, and the disaster frequency factor is quantified by the annual average number of short-duration heavy rainfall events.

[0045] Quantitative data of the above six core influencing factors are extracted from the preprocessed data to construct six-dimensional multidimensional feature data. The preprocessed data is in the form of 1km×1km grid data. The K-means method is used to perform cluster analysis on the multidimensional feature data to obtain the initial clustering results, including: determining the optimal number of clusters using the elbow rule; selecting the number of clusters from all grids as the initial cluster centers using random sampling; cyclically calculating the Euclidean distance between the feature vector of each grid and the initial cluster center, assigning the grid to the nearest cluster, recalculating the mean of each cluster as the new cluster center, and when the contour coefficient meets the preset contour condition, the loop is exited and the preliminary clustering results are output.

[0046] In K-means clustering analysis, the elbow rule is used to determine the optimal number of clusters K. The sum of squared clustering errors is calculated from K=2 to K=6. As K increases, the sum of squared clustering errors gradually decreases. When K=3, the rate of decrease in the sum of squared clustering errors slows down significantly, forming the "elbow" inflection point. Furthermore, the clustering results can distinguish different disaster-causing regions to the greatest extent. Therefore, K=3 is determined.

[0047] Three representative grids were selected from all grids using random sampling as initial cluster centers, ensuring that the feature vectors of the initial centers differed significantly. The Euclidean distance between the feature vector of each grid and the three cluster centers was calculated, and the grid was assigned to the nearest cluster. The mean of each cluster was recalculated as the new cluster center, and this process was repeated iteratively until the cluster centers stabilized. The clustering effect was verified by the silhouette coefficient, which ranged from 0.65 to 0.78, indicating high similarity within clusters and significant differences between clusters.

[0048] After iterative analysis, preliminary cluster analysis results are output. For example, cluster 1 represents a dense plain area with high F4 (population density) and F5 (drainage efficiency) factor values ​​(mean ≥ 0.6) and low F2 (topographic relief) factor values ​​(mean ≤ 0.3), corresponding to 12 cities including Anyang and Hebi in Henan Province.

[0049] Cluster 2 represents a mountainous and hilly area, with significant differences in the F2 (topographic relief) and F3 (permeability) factor values ​​(mean F2 ≥ 0.7, mean F3 ≤ 0.4), corresponding to Sanmenxia, ​​Luoyang, and Jiyuan.

[0050] Cluster 3 represents a humid transition zone with high F6 (disaster frequency) and F3 (permeability) factor values ​​(mean ≥ 0.5), corresponding to Xinyang, Zhumadian, and Nanyang.

[0051] The process of correcting the boundaries and defining regional boundaries of the initial clustering results to obtain the second clustering results includes: adjusting the boundaries of the initial clustering results based on geographical barriers and adapting them to administrative boundaries. Specifically, it involves adjusting the cluster boundaries using geographical barriers such as mountains and rivers, using rivers like the Yellow River and Huai River as natural boundaries to correct unreasonable grid allocations in cross-river clusters. Mountains like the Funiu Mountains and Dabie Mountains are used as boundaries to ensure clear boundaries between mountainous and plain areas, preventing the same mountain range from being split into different regions. While considering administrative convenience, minor adjustments are made to the overlapping parts of cluster boundaries and county-level administrative boundaries to ensure that individual county-level administrative regions belong to the same sub-region as much as possible, avoiding management inconvenience caused by a single region spanning multiple domains. For clusters spanning multiple administrative regions, their affiliation is determined according to the "core grid percentage" principle; administrative regions with a core grid (internal similarity ≥ 0.8) percentage ≥ 60% are assigned to the corresponding sub-region.

[0052] Then, the boundary vector data of the three major sub-regions are output to clarify the administrative scope and number of grids covered by each region, determine the final boundary, and output the second clustering result.

[0053] Furthermore, statistical analysis was conducted on six core influencing factors in each sub-region to extract key disaster-causing characteristics of the region, providing a basis for subsequent adjustment of the indicator system and setting of model parameters.

[0054] In this embodiment, the present invention takes the consistency of precipitation characteristics and the significant differences in disaster-causing factors as its dual cores, while also considering regional differentiation based on natural geographical features, socio-economic levels, and disaster prevention and mitigation capabilities. This ensures that the regional division results not only conform to the spatiotemporal distribution patterns of short-duration heavy precipitation but also accurately match the disaster-causing risk characteristics of different regions. Through the setup of this embodiment, the present invention breaks through traditional administrative boundaries, defining sub-regions with similar precipitation characteristics, underlying surface conditions, and disaster-causing patterns. Furthermore, it customizes differentiated assessment indicator weights, disaster-causing thresholds, and risk classification standards for each sub-region, supporting subsequent refined risk assessments and the formulation of targeted disaster prevention and mitigation measures, thereby improving the practicality and operability of the assessment results.

[0055] In one embodiment of the present invention, the calculation of key secondary indicator data based on the regional division results includes: calculating the surface permeability index, drainage efficiency index, and return period threshold based on the quantitative data in the regional division results; the surface permeability index... The calculation method is as follows:

[0056] ,

[0057] in, For slope, For roughness, Soil water storage coefficient, This is the land cover coefficient. The correction constants are for different regions: 0.9 for densely populated plains, 1.1 for mountainous and hilly areas, and 1.0 for humid transitional areas.

[0058] The drainage efficiency index is calculated by processing the drainage coefficient corresponding to the drainage efficiency index using a weighted summation method; the return period threshold is calculated by using a generalized extreme value distribution method to screen the precipitation sequence in the quantitative data, estimate parameters, and calculate the return period threshold using a preset threshold calculation formula.

[0059] Specifically, the primary indicators include four: precipitation characteristic indicators, underlying surface characteristic indicators, disaster-bearing body vulnerability indicators, and disaster-bearing body adaptability indicators. Among them, precipitation characteristic indicators are used to quantify the extreme degree and frequency of short-term heavy precipitation, which are the triggering sources of disaster risk; underlying surface characteristic indicators reflect the surface's response to precipitation infiltration and runoff, affecting the speed and scope of disaster formation; disaster-bearing body vulnerability indicators reflect the tolerance of the regional population and social structure to disasters; and disaster-bearing body adaptability indicators characterize the basic conditions and response capabilities of regional disaster prevention, mitigation, and relief.

[0060] Each primary indicator has several secondary indicators, with weights adjusted based on regional disaster-causing characteristics to ensure indicator suitability. These include precipitation characteristic indicators such as 1-hour / 3-hour / 6-hour extreme precipitation values, precipitation percentile thresholds, frequency of short-duration heavy rainfall, and return period thresholds; underlying surface characteristic indicators such as topographic relief, surface permeability index, land use type, distance to fault zones, and soil water storage coefficient; disaster-bearing body vulnerability indicators such as population density, ratio of elderly to children, and ratio of illiterate population; and disaster-bearing body adaptability indicators such as drainage efficiency index, transportation advantage, and human development index.

[0061] A weighted summation method was used to process the drainage coefficients corresponding to the drainage efficiency index, and the drainage efficiency index was calculated. The drainage coefficients corresponding to the drainage efficiency index include drainage pipeline coverage coefficient, equipment efficiency coefficient, equipment quantity coefficient, equipment renewal coefficient, and equipment utilization rate coefficient. According to the above drainage coefficients, the weights corresponding to the dense plain area are 0.3, 0.25, 0.2, and 0.15 respectively; the weights corresponding to the mountainous and hilly area are 0.2, 0.25, 0.2, 0.2, and 0.15 respectively; and the weights corresponding to the humid transition area are 0.25, 0.25, 0.2, 0.15, and 0.15 respectively. The above weights are determined based on expert experience.

[0062] The generalized extreme value distribution method is used to process precipitation sequences in the quantified data through data screening, parameter estimation, and calculation of a preset threshold formula to obtain the return period threshold. The 1-hour, 3-hour, and 6-hour precipitation sequences are extracted from the quantified data. The scale, location, and shape parameters of the generalized extreme value distribution are solved using the maximum likelihood estimation method. The return period threshold, i.e., the precipitation threshold, is then calculated using the preset threshold formula. The calculation method for the precipitation threshold is as follows:

[0063]

[0064] in, For precipitation threshold, The location parameter of the generalized extreme value distribution. The size parameter of the generalized extremum distribution. The shape parameter of the generalized extremum distribution. The set number of years is 50, which can be adjusted according to actual needs.

[0065] Then, a weighted summation method is used to synthesize the primary indicators, thus obtaining the primary indicator data.

[0066] In this embodiment, the present invention constructs an indicator system that combines a four-level architecture with regional weights. The indicators cover multiple dimensions such as nature, society, economy, and infrastructure, comprehensively reflecting the formation mechanism of disaster risks. Furthermore, it breaks through the limitations of traditional unified indicators, achieving a single system that is adaptable to multiple regions, thereby improving the relevance of the indicators. The calculation of key indicators integrates multi-source data and scientific models, taking into account both data availability and computational scientificity.

[0067] In one embodiment of the present invention, the detailed evaluation of primary indicator data and preprocessed data to obtain risk zoning assessment results includes: obtaining precipitation thresholds for different return periods in each region; fusing primary indicator data and precipitation thresholds using fuzzy rules to generate preliminary risk quantification values; analyzing the return period benchmark and fuzzy inference clear values ​​based on preset regional differentiated grading rules to obtain preliminary risk levels; and optimizing the preliminary risk levels based on time cumulative effects to obtain risk zoning assessment results.

[0068] Specifically, obtain the precipitation thresholds of different return periods in each region and each primary index data in the primary index data and preprocessed data; use fuzzy rules to fuse each primary index data and precipitation threshold to generate a preliminary risk quantification value, including: fuzzify each primary index data and precipitation threshold to obtain multiple fuzzy subsets; use the triangular membership function to describe each fuzzy subset and determine the initial threshold interval; according to the regional division results, perform regional adaptation on the initial threshold interval to determine the final threshold interval of the triangular membership function; based on a preset fuzzy rule base, use the Mamdani reasoning method to analyze each triangular membership function to obtain a fuzzy output; use the centroid method to convert the fuzzy output into a crisp value to obtain a preliminary risk quantification value.

[0069] That is, take the precipitation thresholds of different return periods in each region and each primary index data as input variables for fuzzification; each input variable is divided into three fuzzy subsets: "Low (L), Medium (M), High (H)"; then use the triangular membership function to describe each subset and determine the initial threshold interval; according to the regional division results, perform regional adaptation on the initial threshold interval to determine the final threshold interval of the triangular membership function; based on the objective law that "the higher the precipitation extremeness, the more sensitive the underlying surface response, the more vulnerable the disaster-bearing body, and the weaker the adaptation ability, the higher the risk", a number of fuzzy rules covering all variable combinations are formulated to form a preset fuzzy rule base; for example, if P = H and S = H and V = H and A = L, then the risk = extremely high; if P = M and S = M and V = L and A = M, then the risk = medium; if P = L and S = L and V = any and A = any, then the risk = extremely low, without actual risk, and subsequent classification and merging; the character meanings include Low (L), Medium (M), High (H), precipitation characteristic index (P), underlying surface characteristic index (S), disaster-bearing body vulnerability index (V), and disaster-bearing body adaptation ability index (A).

[0070] The triangular membership functions are as follows:

[0071] The low membership function is: ,

[0072] The medium membership function is: ,

[0073] The high membership function is: ,

[0074] Where a, b, and c are the precipitation threshold parameters of each index and different return periods in each region (a < b < c), b is the peak point of the "Medium" subset, and a and c are the demarcation points of "Low-Medium" and "Medium-High" respectively, and X represents a certain index in the primary index.

[0075] The precipitation threshold is the "skeleton" of the membership function. The precipitation threshold clarifies the division boundary of the three fuzzy subsets of "low, medium and high" and determines the segmented range of the domain of the membership function.

[0076] The Mamdani inference method is used, calculating the trigger strength of each rule through AND operations and aggregating the outputs of all rules through OR operations. The centroid method is employed to transform the fuzzy outputs into clear values, i.e., the preliminary risk quantification value Y. The formula for the centroid method is as follows:

[0077] ,

[0078] in, Let i be the trigger strength of the i-th rule. Let i be the risk quantification value corresponding to the i-th rule. This represents the total number of fuzzy rules.

[0079] Based on pre-defined regional differentiated classification rules, the return period benchmark and fuzzy inference clear values ​​are analyzed to obtain preliminary risk levels. The main process involves combining the return period benchmark and fuzzy inference clear values, and referring to historical disaster data, to classify four risk levels: "medium," "relatively high," "high," and "extremely high," ensuring that the classification has both quantitative basis and practical significance. Specific classification benchmarks include a core benchmark based on the return period threshold: 5a ≤ T < 20a corresponds to the "medium risk" benchmark, 20a ≤ T < 50a corresponds to the "relatively high risk" benchmark, 50a ≤ T < 100a corresponds to the "high risk" benchmark, and T ≥ 100a corresponds to the "extremely high risk" benchmark. Fuzzy inference clear values ​​are used as auxiliary benchmarks: 2.0 ≤ Y < 3.0 corresponds to "medium," 3.0 ≤ Y < 4.0 corresponds to "relatively high," 4.0 ≤ Y < 4.5 corresponds to "high," and Y ≥ 4.5 corresponds to "extremely high."

[0080] Based on the regional division results, the corresponding weighted fusion rules are used for fusion to obtain the fusion result; when the fusion result meets the corresponding preset conditions, the risk level is adjusted. For example: Dense plain area: The fusion result is determined by the weighted fusion rule of "recurrence period benchmark × 0.6 + clarity value benchmark × 0.4". When the fusion result is ≥3.5 and the recurrence period threshold (T) is ≥30 years, it is upgraded to high risk.

[0081] Mountainous and hilly areas: The weighted fusion rule of "recurrence period benchmark × 0.5 + clarity value benchmark × 0.5" is adopted. When the fusion result is ≥ 3.0 and T ≥ 25 years, it is upgraded to a higher risk.

[0082] Wet transition zone: The weighted fusion rule of “recurrence period benchmark × 0.55 + clarity value benchmark × 0.45” is adopted. When the fusion result is ≥3.3 and T≥28a, it is upgraded to high risk.

[0083] For grids with complex terrain but sparse population, if the clarity value is <3.0, even if T≥50a, the risk level is downgraded to "higher risk" to avoid over-evaluation.

[0084] Furthermore, the optimization of the preliminary risk level by time accumulation effect to obtain the risk zoning assessment result includes: calculating the time accumulation index according to the preset accumulation rule; predicting the probability of precipitation on the next day based on Bayes' theorem, using the average precipitation of the previous seven days and the average precipitation of the previous three days as prior information; correcting the risk level of each region according to the relationship between the time accumulation index, the precipitation probability and the preset level correction conditions to obtain the final risk level; and integrating the information of each region that determined the final risk level to obtain the risk zoning assessment result.

[0085] For scenarios involving continuous extreme precipitation, the cumulative effect of precipitation over time is considered (continuous precipitation easily leads to soil saturation, drainage system overload, and a continuous increase in risk), and the initial risk level is dynamically revised. The specific revision process includes: calculating the cumulative time index according to preset cumulative rules; the cumulative time index includes a daily precipitation disaster prediction index and a continuous precipitation cumulative index; the formula for calculating the daily precipitation disaster prediction index is as follows: ,

[0086] in, This is the standardized value for daily precipitation. This represents the response value of the underlying surface. Vulnerability value, The adaptability value is represented by i, which indicates the i-th day. The cumulative continuous precipitation index is the sum of the products of the number of consecutive precipitation days and the corresponding weighted coefficients.

[0087] Based on Bayes' theorem, the average precipitation over the previous seven days and the average precipitation over the previous three days are used as prior information to predict the probability of precipitation on the next day. The risk level is corrected according to the relationship between the cumulative time index, the probability of precipitation and the preset level correction conditions. After the analysis and integration of all processes in this embodiment, the risk zoning assessment results are output, including the final risk level (medium / relatively high / high / extremely high), the basis for level classification (recurrence period benchmark, fuzzy inference clear value, cumulative effect correction).

[0088] In this embodiment, the present invention focuses on "quantifying the degree of extreme precipitation, integrating multi-dimensional indicators, and adapting to regional differences," constructing a four-level assessment model that includes return period estimation, fuzzy inference fusion, risk classification, and time cumulative effect optimization. First, the return period is used to quantify the extreme precipitation baseline. Then, fuzzy inference is used to integrate multi-dimensional indicators, and risk levels are classified based on regional disaster-causing characteristics. Finally, time cumulative effect correction is applied to achieve a refined and dynamic assessment of the disaster risk caused by short-term heavy precipitation.

[0089] This invention overcomes the limitations of single-indicator assessment, achieving the organic integration of multiple dimensions such as precipitation, underlying surface, disaster-bearing body, and adaptability. Secondly, it addresses the ambiguity of risk assessment by quantifying qualitative descriptions such as "high," "medium," and "low," thus improving the objectivity of the assessment results. Simultaneously, this invention can adapt to the disaster sensitivities of different regions, ensuring that the risk level classification matches the actual disaster impact in the region. Furthermore, it considers the cumulative effect of extreme precipitation over time, improving the accuracy of risk assessment under long-term continuous precipitation scenarios.

[0090] This invention focuses on resolving indicator ambiguity, organically integrating multiple factors, and adapting to regional differences. Through a four-step process—fuzzification of input variables, fuzzy inference based on a pre-set fuzzy rule base, and defuzzification output—it transforms the high, medium, and low qualitative descriptions of four primary indicators—precipitation characteristics, underlying surface characteristics, vulnerability of disaster-bearing bodies, and adaptability of disaster-bearing bodies—into quantitative risk values. Simultaneously, it adjusts the fuzzy subset threshold and rule base according to the disaster-causing characteristics of different regions, ensuring that the fusion results closely match the actual disaster-causing patterns of the region. This solves the problems of difficulty in quantifying qualitative descriptions, lack of logical factor fusion, and insufficient regional adaptability in traditional multi-factor assessments.

[0091] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0092] It should be noted that in the description of this invention, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0093] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0094] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0095] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0096] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0097] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0098] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A regional differentiation-based short-time heavy rain disaster risk zoning evaluation system, characterized in that, include: The data acquisition module retrieves multi-source related data; The data processing module preprocesses the multivariate correlated data to obtain preprocessed data; The region partitioning module sequentially performs feature extraction, cluster analysis, boundary correction, and region delimitation on the preprocessed data to obtain the region partitioning results. The indicator analysis module calculates the data of each key secondary indicator based on the regional division results; The primary indicator data is obtained by weighted summation of the secondary indicators; The risk assessment module performs a detailed evaluation of the primary indicator data and preprocessed data to obtain risk zoning assessment results. 2.The short-time heavy rain disaster risk zoning assessment system according to claim 1, characterized in that, The preprocessing of multivariate correlated data to obtain preprocessed data includes: using linear interpolation to supplement missing data; using the 3σ principle to identify and calibrate outlier data; and performing hierarchical and standardization processing on the calibrated data to obtain preprocessed data. 3.The short-time heavy rain disaster risk zoning assessment system according to claim 1, characterized in that, The process of sequentially performing feature extraction, cluster analysis, boundary correction, and region delimitation on preprocessed data to obtain region division results includes: selecting corresponding quantitative data from preprocessed data based on preset core influencing factors to obtain multidimensional feature data; performing cluster analysis on the multidimensional feature data using the K-means method to obtain initial clustering results; performing boundary correction and region delimitation on the initial clustering results to obtain second clustering results; performing statistical analysis on the core influencing factors of each sub-region in the second clustering results to obtain key disaster-causing features of the region; and integrating the second clustering results and the key disaster-causing features of the region to obtain the final region division result.

4. The short-time heavy rain disaster risk zoning assessment system according to claim 3, characterized in that, The preset core influencing factors include precipitation intensity factor, topographic feature factor, permeability factor, disaster-bearing body factor, drainage capacity factor, and disaster frequency factor.

5. The short-time heavy rain disaster risk zoning assessment system according to claim 3, characterized in that, The method of using K-means to perform cluster analysis on multidimensional feature data to obtain initial clustering results includes: determining the optimal number of clusters using the elbow rule; selecting the number of clusters from all grids using random sampling as the initial cluster centers; iteratively calculating the Euclidean distance between the feature vector of each grid and the initial cluster center, assigning the grid to the nearest cluster, recalculating the mean of each cluster as the new cluster center, and exiting the loop when the contour coefficient meets the preset contour conditions, outputting the preliminary clustering results. 6.The short-time heavy rain disaster risk zoning assessment system according to claim 3, characterized in that, The process of correcting the boundaries and defining the regions of the initial clustering results to obtain the second clustering results includes: adjusting the boundaries of the initial clustering results and adapting them to administrative boundaries based on geographical barriers to obtain the second clustering results. 7.The short-time heavy rain disaster risk zoning assessment system according to claim 1, characterized in that, The calculation of each key secondary index data based on the region division result includes: calculating the surface permeability index, the drainage efficiency index, and the return period threshold according to the quantitative data in the region division result The calculation method of the surface permeability index is: ， wherein, is the slope, is the roughness, is the soil water storage coefficient, is the surface cover coefficient, is the correction factor for different areas; The drainage efficiency index is calculated by processing the drainage coefficient corresponding to the drainage efficiency index using a weighted summation method. The return period threshold is obtained by using the generalized extreme value distribution method to process the precipitation sequence in the quantitative data through data screening, parameter estimation, and calculation of the preset threshold formula. 8.The short-time heavy rain disaster risk zoning assessment system according to claim 1, characterized in that, The detailed evaluation of primary indicator data and preprocessed data to obtain risk zoning assessment results includes: obtaining precipitation thresholds for different return periods in each region; fusing primary indicator data and precipitation thresholds using fuzzy rules to generate preliminary risk quantification values; analyzing the return period benchmark and fuzzy inference clear values ​​based on preset regional differentiated grading rules to obtain preliminary risk levels; and optimizing the preliminary risk levels based on the time cumulative effect to obtain risk zoning assessment results.

9. The short-time heavy rain disaster risk zoning assessment system according to claim 8, characterized in that, The method of using fuzzy rules to fuse primary indicator data and precipitation thresholds to generate preliminary risk quantification values ​​includes: fuzzifying the primary indicator data and precipitation thresholds to obtain multiple fuzzy subsets; using triangular membership functions to describe each fuzzy subset and determining an initial threshold interval; performing regional adaptation on the initial threshold interval based on the regional division results to determine the final threshold interval of the triangular membership functions; analyzing each triangular membership function using the Mamdani inference method based on a preset fuzzy rule base to obtain fuzzy outputs; and using the centroid method to convert the fuzzy outputs into clear values ​​to obtain preliminary risk quantification values.

10. The short-time heavy rain disaster risk zoning assessment system according to claim 8, characterized in that, The optimization of the preliminary risk level by time accumulation effect to obtain the risk zoning assessment result includes: calculating the time accumulation index according to the preset accumulation rule; predicting the probability of precipitation on the next day based on Bayes' theorem, using the average precipitation of the previous seven days and the average precipitation of the previous three days as prior information; correcting the risk level of each region according to the relationship between the time accumulation index, the probability of precipitation and the preset level correction conditions to obtain the final risk level; and integrating the information of each region that has determined the final risk level to obtain the risk zoning assessment result.

Images