Early warning method for rainstorm and flood composite disaster based on meteorological big data

Abstract

The invention discloses a weather big data-based rainstorm and flood composite disaster early warning method, which comprises the following steps of: extracting periodic characteristics and abnormal fluctuation information of rainstorm frequency by acquiring long-term weather data and performing hierarchical processing, and analyzing and determining the change trend of the rainstorm frequency in combination with a time sequence. And constructing an incidence matrix based on the rainstorm frequency change trend and the topographic hydrological characteristic data, determining an initial parameter range of flood probability assessment, introducing dynamic correlation analysis, fusing meteorological data and environmental conditions, and obtaining a dynamic adjustment coefficient of probability assessment. Through mapping the adjustment coefficient and the rainstorm frequency change trend, a quantitative expression of a rainstorm and flood coupling structure is constructed, and an early warning model is obtained based on the quantitative expression. And finally generating multi-scene early warning signal distribution by performing adaptive adjustment on the model response and verifying the accuracy of the model response. According to the invention, accurate early warning of flood disasters in a complex environment is realized, and the timeliness and reliability of early warning are significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of disaster early warning technology, and in particular relates to a method for early warning of combined rainstorm and flood disasters based on meteorological big data. Background Technology

[0002] Early warning of combined rainstorm and flood disasters is a crucial aspect of disaster prevention and mitigation, and its research is directly related to the safety of people's lives and property and social stability. With the intensification of climate change and the increasing frequency of extreme weather events, flood disasters triggered by rainstorms are exhibiting greater complexity and destructive power, becoming a critical issue that urgently needs to be addressed. Against this backdrop, research on early warning methods based on meteorological big data is particularly urgent and necessary. Currently, although various methods have attempted to predict and warn of rainstorm and flood disasters, most schemes have significant shortcomings in integrating multi-source data and dynamic correlation analysis. Many traditional methods often overlook the deep interaction between meteorological conditions and topographic and hydrological features, resulting in a lack of accuracy and adaptability in predictions when facing complex disaster scenarios, especially in combined disasters where rainstorms and floods overlap, where the timeliness and reliability of early warnings are often limited. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention proposes a method for early warning of combined rainstorm and flood disasters based on meteorological big data, thereby resolving the issues present in the existing technologies.

[0004] To achieve the above objectives, this invention provides a method for early warning of combined rainstorm and flood disasters based on meteorological big data, comprising:

[0005] Long-term meteorological data is acquired and processed in layers. Based on the processed data, periodic characteristics and abnormal fluctuations of rainstorm frequency are obtained. The trend distribution of rainstorm frequency changes is obtained by combining time series methods.

[0006] Based on the trend distribution of rainfall frequency changes and topographic and hydrological feature data, a correlation matrix between rainfall frequency and flood probability is obtained, and an initial parameter range for flood probability assessment is obtained based on the correlation matrix.

[0007] Adaptive constraints for meteorological data analysis and environmental conditions are constructed. Based on the initial parameter range and adaptive constraints for flood probability assessment, and combined with dynamic correlation analysis, dynamic adjustment coefficients for probability assessment are obtained.

[0008] Based on the dynamic adjustment coefficient of probability assessment and the trend distribution of rainstorm frequency changes, the coupling structure of rainstorm and flood is obtained and quantified, and the quantitative expression of the coupling relationship is obtained.

[0009] The basic information set under the complex disaster scenario is obtained and corrected. Combined with the quantitative expression of the coupling relationship, the prediction module of the early warning model is constructed. The response characteristics of the prediction module are obtained, the early warning model is adjusted based on the response characteristics, and the prediction output of the adjusted early warning model is verified.

[0010] Based on the verified prediction output, early warning signal distributions under multiple scenarios are generated, and an early warning information mapping table is obtained. Based on the early warning information mapping table, early warning of combined rainstorm and flood disasters is realized.

[0011] Optionally, the process of obtaining the trend distribution of changes in the frequency of rainstorms includes:

[0012] The long-term meteorological data is cleaned to obtain basic data. A hierarchical processing method is used to classify the basic data, dividing it according to time and geographical dimensions to obtain hierarchical data groups. Based on these hierarchical data groups, the periodic characteristics of rainstorm frequency are extracted, and the time series is decomposed using Fourier transform to obtain the periodic characteristics of rainstorm frequency. If the fluctuation amplitude of rainstorm frequency exceeds a preset threshold within a certain time period, it is marked as an abnormal fluctuation point, and the abnormal fluctuation distribution is obtained. After integrating the abnormal fluctuation distribution and periodic characteristics, the time series of frequency changes is smoothed using a moving average method. The trend distribution of rainstorm frequency changes is obtained based on the smoothed frequency change curve.

[0013] Optionally, the process of obtaining the initial parameter range for flood probability assessment includes:

[0014] Correlation analysis was used to calculate the correlation strength between rainfall frequency and topographic and hydrological features, and a set of significantly correlated indicators was selected. Based on the set of significantly correlated indicators, a correlation matrix between rainfall frequency and flood probability was constructed, and the weight distribution of each indicator in the correlation matrix was obtained to obtain the weighted correlation relationship. Based on the weighted correlation relationship, key influencing factors were obtained. Based on the key influencing factors and information processing, the corresponding topographic and hydrological features were selected to obtain the distribution range of the key influencing factors. Combined with the assessment requirements of flood probability, the initial parameter range for flood probability assessment was obtained.

[0015] Optionally, the process of constructing adaptive constraints for meteorological data analysis and environmental conditions, based on the initial parameter range and adaptive constraints of flood probability assessment, and combined with dynamic correlation analysis, to obtain the dynamic adjustment coefficients for probability assessment includes:

[0016] To meet the requirements of flood probability assessment, meteorological data and environmental conditions are integrated to obtain a comprehensive environmental dataset. Based on this dataset and dynamic correlation analysis, the interaction between meteorological data and environmental conditions is extracted to obtain the distribution characteristics of key triggering factors. If the correlation of a certain data point is lower than a preset threshold, that data point is removed, resulting in an optimized set of triggering factors. Based on the optimized set of triggering factors and the initial parameter range, a mapping relationship between flood probability and key triggering factors is constructed, and weights are assigned to this mapping relationship to obtain a weighted impact distribution. Based on the weighted impact distribution and adaptive constraints, the limiting relationship between environmental conditions and flood probability is obtained. If the weight of a certain adaptive constraint is lower than a preset threshold, its impact ratio is adjusted to obtain an optimized constraint distribution. For the optimized constraint distribution, a logistic regression model is used to quantify the dynamic adjustment of flood probability assessment, obtaining a candidate set of adjustment coefficients, which is then filtered to obtain the final coefficient range. Based on the final coefficient range and the actual needs of probability assessment, the dynamic adjustment coefficients are divided into intervals, and the applicability of each interval is judged. Based on the applicability results and integrated analysis, the dynamic adjustment coefficients for probability assessment are obtained.

[0017] Optionally, based on the dynamic adjustment coefficients of probability assessment and the trend distribution of rainstorm frequency changes, the coupling structure of rainstorms and floods is obtained and quantified. The process of obtaining the quantitative expression of the coupling relationship includes:

[0018] The dynamic adjustment coefficient is correlated and matched with the rainfall and trend distribution using a preset mapping rule to determine the initial coefficient adjustment range. Based on the coefficient adjustment range and the construction requirements of the coupling framework, a coupling structure between rainfall and flood is built by integrating flood correlation data. According to the coupling structure, multivariate regression analysis is applied to quantify the relationship between rainfall frequency, flood correlation and dynamic adjustment coefficient, and obtain a quantitative expression of the coupling relationship.

[0019] Optionally, the process of acquiring and correcting the basic information set under the complex disaster scenario, and then constructing the prediction module of the early warning model by combining the quantitative expression of the coupling relationship includes:

[0020] The system acquires and initially classifies basic information sets under complex disaster scenarios to obtain classified data groups. Based on these data groups, a pre-established rule base is used for matching to determine key data subsets related to the disaster scenario. For these key data subsets, the system constructs the input layer structure of the early warning model, obtains the mapping relationship between the input layer and environmental conditions, and assesses the stability of the mapping relationship. If the stability of the mapping relationship is lower than a preset threshold, the key data subset is corrected through information processing to obtain a corrected data subset. Based on the corrected data subset and a quantification expression, the system constructs the prediction module of the early warning model.

[0021] Optionally, the process of obtaining the response features of the prediction module and adjusting the early warning model based on the response features includes:

[0022] The process involves acquiring the response characteristics of the prediction module in the early warning model, extracting response values, determining whether the response values ​​exceed the specified range, and dynamically correcting the parameters of the core module based on response values ​​exceeding the threshold range. The magnitude of parameter adjustments is calibrated through information processing to determine the adjusted parameter configuration. The prediction module of the early warning model is updated with the adjusted parameter configuration until the stability of the response characteristic change trend meets the requirements, resulting in an optimized parameter set. Based on the optimized parameter set, the prediction function's operating environment is reloaded, and the performance of the prediction module on the updated response characteristics is obtained to determine whether the response values ​​have returned to the threshold range. If the response values ​​still exceed the threshold range, the triggering conditions are analyzed through information processing to obtain a condition correction scheme. For the condition correction scheme, the triggering condition settings of the early warning model are adjusted, and a support vector machine algorithm is used to model and analyze the fluctuation pattern of the response values, obtaining the distribution characteristics of the fluctuation pattern and determining whether the distribution characteristics meet the expected standards. The prediction output of the prediction module is updated based on the judgment results.

[0023] Optionally, the process of validating the prediction output of the adjusted early warning model includes:

[0024] The prediction output of the early warning model is analyzed using uncertainty quantification methods to obtain fluctuation data and determine its distribution characteristics. Based on these characteristics, uncertainty factors are decomposed using information processing to obtain a set of influencing factors. It is then determined whether this set of factors meets a preset range standard. If the set exceeds the preset range standard, the adjustment results are corrected through information processing to obtain the corrected parameter configuration. Based on the corrected parameter configuration, the prediction module of the early warning model is updated, and the updated model performance data is obtained. Finally, it is determined whether the model performance data meets the timeliness and reliability requirements.

[0025] Optionally, the process of generating early warning signal distributions for multiple scenarios and obtaining an early warning information mapping table based on the validated prediction output results is as follows:

[0026] Based on the verified prediction output results and the characteristics of the warning signals, a data integration tool is used to initially classify the signal distribution under multiple scenarios, resulting in classified signal distribution groups. According to the classified signal distribution groups, combined with environmental and adaptive conditions, the degree of matching between each group and the conditions is compared. If the matching degree does not reach a preset threshold, the signal distribution groups are reclassified to determine the adjusted distribution groups. Based on the adjusted distribution groups, feature extraction is performed on the signal distribution characteristics under different scenarios to obtain the extracted feature set. Based on the extracted feature set, combined with the warning signals and signal distribution, information processing is performed on the correlation between the feature set and the distribution map. If the feature set cannot be completely mapped to the distribution map, the feature set is supplemented to obtain a supplemented feature mapping. Based on the supplemented feature mapping, combined with the information mapping and the warning set, the mapping content is integrated using a data fusion tool to determine whether the integrated mapping content meets the condition matching requirements, obtaining an integrated warning information set. Based on the integrated warning information set, combined with condition matching and scenario analysis, an information verification tool is used to perform consistency checks on the set content. After passing the consistency check, combined with the signal distribution and distribution map, a warning information mapping table is obtained.

[0027] Compared with the prior art, the present invention has the following advantages and technical effects:

[0028] This invention discloses a method for early warning of combined rainstorm and flood disasters based on meteorological big data. By analyzing long-term meteorological data, it extracts the frequency variation characteristics of rainstorms and constructs a correlation matrix based on topographic and hydrological features to determine flood probability assessment parameters. Dynamic correlation analysis is introduced to integrate meteorological data with environmental conditions, determine flood triggering conditions, and obtain dynamic adjustment coefficients for probability assessment. These adjustment coefficients are mapped to the trend of rainstorm frequency variation to construct a rainstorm-flood coupling framework and determine a quantitative expression. Based on this, a core prediction module for the early warning model is constructed, and the model response is adaptively adjusted. The accuracy is verified using uncertainty quantification methods, and finally, a multi-scenario early warning signal distribution is generated. This invention achieves accurate early warning of flood disasters in complex environments, improves the timeliness and reliability of early warnings, and provides strong support for disaster prevention and mitigation. Attached Figure Description

[0029] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0030] Figure 1 This is a flowchart of a method according to an embodiment of the present invention. Detailed Implementation

[0031] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0032] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0033] Example 1

[0034] A thorough analysis of the challenges in this field reveals that the core difficulty lies in the frequency characteristics of rainstorm events and their correlation with flood occurrence. The long-term patterns of rainstorm frequency are often difficult to quantify accurately, and this uncertainty directly impacts the probability assessment of flood disasters. Without effectively establishing a coupling relationship between the two, reliable quantitative data for disaster early warning cannot be provided. This lack of conversion from frequency characteristics to probability makes early warning models ill-suited for handling changing environmental conditions in practical applications, thus limiting the accuracy and practicality of disaster early warnings. Therefore, how to construct a coupling framework between rainstorm frequency and flood probability based on long-term meteorological data, and how to transform uncertainty into quantifiable early warning data through appropriate methods, has become a key issue in improving the ability to issue early warnings for complex disasters.

[0035] like Figure 1 As shown, this embodiment provides a method for early warning of combined rainstorm and flood disasters based on meteorological big data, including:

[0036] Long-term meteorological data is acquired and processed in layers. Based on the processed data, periodic characteristics and abnormal fluctuations of rainstorm frequency are obtained. The trend distribution of rainstorm frequency changes is obtained by combining time series methods.

[0037] As one specific implementation method, the process of obtaining the trend distribution of changes in the frequency of rainstorms includes:

[0038] Long-term meteorological data is cleaned to obtain basic data. A hierarchical processing method is used to classify the basic data, dividing it according to time and geographical dimensions to obtain stratified data groups. Based on these stratified data groups, the periodic characteristics of rainstorm frequency are extracted, and the time series is decomposed using Fourier transform to obtain the periodic characteristics of rainstorm frequency. If the fluctuation amplitude of rainstorm frequency exceeds a preset threshold within a certain time period, it is marked as an abnormal fluctuation point, obtaining the abnormal fluctuation distribution. After integrating the abnormal fluctuation distribution and periodic characteristics, the time series of frequency changes is smoothed using the moving average method. The trend distribution of rainstorm frequency changes is obtained based on the smoothed frequency change curve.

[0039] For example, when constructing the initial dataset, rainfall records from the past 50 years can be obtained from weather stations in a certain region, including daily rainfall, time, and location information. During the initial data cleaning, if missing rainfall data or negative values ​​are found for some dates, these can be removed to ensure data integrity and reasonableness. This effectively improves the accuracy of subsequent analysis and avoids biases caused by erroneous data.

[0040] In one possible implementation, the stratified processing of the basic data can be grouped by year and province. For example, the data can be divided into two time periods: 1990-2000 and 2001-2010, and further subdivided by region, such as East China and South China, resulting in multiple data groups. Next, the frequency of heavy rainfall within each group—that is, the number of days with rainfall exceeding 50 mm per year—can be statistically analyzed to determine its distribution. This stratified approach helps reveal the characteristics of heavy rainfall at different times and in different regions, providing a fine-grained perspective for subsequent analysis.

[0041] When extracting the periodic characteristics of rainstorm frequency, Fourier transform can be applied to 30 years of data from a specific region to decompose the time series and identify periodic fluctuations occurring every 5 or 10 years. This method can help discover regular changes in rainstorm frequency, laying the foundation for predicting long-term trends. When analyzing anomalous fluctuation patterns, if the rainstorm frequency suddenly increases from an average of 10 times to 20 times in a certain year, exceeding a preset threshold of 15 times, it is marked as an anomaly, and its distribution is recorded for further investigation into the causes of the anomaly.

[0042] In one possible implementation, after integrating the abnormal fluctuation distribution and periodic characteristics, the moving average method can be used to smooth the time series.

[0043] For example, by using a 5-year window, the average frequency of rainstorms at a certain point in time can be calculated to obtain a smoothed curve. This method can effectively reduce the impact of short-term fluctuations, highlight long-term trends, and improve data readability. Based on the smoothed curve, if the frequency of rainstorms has gradually increased from 8 times per year to 12 times over the past 20 years, it can be judged as an upward trend, reflecting that rainstorm events may be increasing due to the influence of climate change.

[0044] When analyzing long-term trend distribution, if data from South China shows a continuous upward trend while data from East China remains relatively stable, it can be inferred that the former is more affected by tropical cyclones. This analysis helps local governments develop targeted flood control measures and improve disaster response capabilities. Through the above multi-dimensional analysis, from data cleaning to trend judgment, each link supports each other, forming a complete logical chain to ensure that the research on changes in rainstorm frequency is scientific and practical.

[0045] Based on the trend distribution of rainfall frequency variation and topographic and hydrological feature data, a correlation matrix between rainfall frequency and flood probability is obtained, and the initial parameter range for flood probability assessment is obtained based on the correlation matrix.

[0046] As one specific implementation method, the process of obtaining the initial parameter range for flood probability assessment includes:

[0047] Correlation analysis was used to calculate the correlation strength between rainfall frequency and topographic and hydrological features, and a set of significantly correlated indicators was selected. Based on the set of significantly correlated indicators, a correlation matrix between rainfall frequency and flood probability was constructed, and the weight distribution of each indicator in the correlation matrix was obtained to obtain the weighted correlation relationship. Based on the weighted correlation relationship, key influencing factors were obtained. Based on the key influencing factors and information processing, the corresponding topographic and hydrological features were selected to obtain the distribution range of the key influencing factors. Combined with the assessment requirements of flood probability, the initial parameter range for flood probability assessment was obtained.

[0048] In the application of correlation analysis methods, statistical tools can be used to calculate the correlation strength between rainfall frequency and topographic and hydrological characteristics. For example, if the correlation coefficient between slope and rainfall frequency in a certain area is 0.8, it indicates a high correlation between the two, while the correlation coefficient for soil moisture content is only 0.2, which can be eliminated to filter out the set of significant correlation indicators. For the construction of the correlation matrix, the relationship between rainfall frequency and flood probability can be quantified as a weighted distribution. Assuming the weight of slope is 0.6 and river density is 0.3, and if a preset threshold of 0.5 is set, then slope is marked as a key influencing factor. This method helps to focus on the main driving factors. When screening the distribution range of key influencing factors, it can be found that areas with slopes greater than 15 degrees have significantly higher rainfall frequencies than other areas, and the probability of flooding also increases accordingly. Determining this distribution range provides a spatial basis for subsequent modeling.

[0049] In the classification process of logistic regression models, data can be divided into high-risk and low-risk categories, with the initial candidate parameter range potentially set to a slope of 10 to 20 degrees. Through further interval division, if the sum of weights in the 10 to 12 degree interval is less than a threshold of 0.4, that interval is discarded, ultimately locking the parameter range to 12 to 20 degrees. This approach ensures the applicability of the parameters.

[0050] The combined analysis of interval division and weight distribution can help identify which terrain conditions are more likely to trigger floods. For example, a mountainous area with steep slopes and high river density may increase the flood probability from 0.3 to 0.7. This refined analysis provides data support for disaster prevention planning, improves the accuracy of forecasts, and helps optimize resource allocation and early warning mechanisms.

[0051] Adaptive constraints for meteorological data analysis and environmental conditions are constructed. Based on the initial parameter range and adaptive constraints for flood probability assessment, and combined with dynamic correlation analysis, dynamic adjustment coefficients for probability assessment are obtained.

[0052] As a specific implementation method, the process of constructing adaptive constraints for meteorological data analysis and environmental conditions, and obtaining the dynamic adjustment coefficients for probability assessment based on the initial parameter range and adaptive constraints of flood probability assessment, combined with dynamic correlation analysis, includes:

[0053] To meet the requirements of flood probability assessment, meteorological data and environmental conditions are integrated to obtain a comprehensive environmental dataset. Based on this dataset and dynamic correlation analysis, the interaction between meteorological data and environmental conditions is extracted to obtain the distribution characteristics of key triggering factors. If the correlation of a certain data point is lower than a preset threshold, that data point is removed, resulting in an optimized set of triggering factors. Based on the optimized set of triggering factors and the initial parameter range, a mapping relationship between flood probability and key triggering factors is constructed, and weights are assigned to this mapping relationship to obtain a weighted impact distribution. Based on the weighted impact distribution and adaptive constraints, the limiting relationship between environmental conditions and flood probability is obtained. If the weight of a certain adaptive constraint is lower than a preset threshold, its impact ratio is adjusted to obtain an optimized constraint distribution. For the optimized constraint distribution, a logistic regression model is used to quantify the dynamic adjustment of flood probability assessment, obtaining a candidate set of adjustment coefficients, which is then filtered to obtain the final coefficient range. Based on the final coefficient range and the actual needs of probability assessment, the dynamic adjustment coefficients are divided into intervals, and the applicability of each interval is judged. Based on the applicability results and integrated analysis, the dynamic adjustment coefficients for probability assessment are obtained.

[0054] For example, in addressing the need for flood probability assessment, data fusion technology can be used to integrate meteorological data with environmental conditions. Meteorological data, such as rainfall and wind speed, can be fused with environmental conditions, such as soil moisture and topographic slope, to form a comprehensive environmental dataset. Assuming a region experiences 80 mm of rainfall in 24 hours and soil moisture saturation reaches 90%, the fused dataset can visually reflect the impact of rainfall on soil infiltration, laying the foundation for subsequent analysis.

[0055] When applying dynamic correlation analysis, the interaction between meteorological data and environmental conditions can be extracted. Assuming that the correlation between rainfall and soil moisture is as high as 0.8, exceeding the preset threshold of 0.5, it is listed as a key triggering factor. Wind speed data with a correlation of only 0.2 is removed, and the dataset is refined to focus on the core impact.

[0056] When constructing the mapping relationship between flood probability and key triggering factors, the potential impact range can be determined based on rainfall and soil moisture. If rainfall exceeds 50 mm and soil moisture exceeds 85%, the flood probability may rise to 70%. Through weight allocation, with rainfall weighted at 0.6 and soil moisture at 0.4, a weighted impact distribution is formed.

[0057] When incorporating adaptive constraints, environmental conditions such as terrain slope may limit the probability of flooding. If the slope is less than 5 degrees, the risk of water accumulation is high, and the weight is set to 0.7. If the condition is below the threshold of 0.3, the ratio is adjusted to optimize the constraint distribution and ensure that the analysis fits the actual terrain characteristics.

[0058] When using a logistic regression model to quantify dynamic adjustments, a candidate set of adjustment coefficients can be generated based on historical data. Assuming the candidate values ​​range from 0.5 to 1.5, by filtering out values ​​that do not conform to terrain constraints, a coefficient range of 0.8 to 1.2 is finally determined, improving the adaptability of the assessment.

[0059] In the interval division and applicability judgment, if the applicability of a certain interval is lower than the threshold of 0.6, the interval is removed, and the set of adjustment coefficients with higher applicability is retained. For example, if the applicability of the interval from 0.8 to 1.0 is 0.9, it will be retained first for subsequent dynamic update mechanism construction.

[0060] Specifically, during the division process, if the applicability of a certain interval is lower than a preset threshold, that interval is removed, and a set of adjustment coefficients with higher applicability is obtained. For the set of adjustment coefficients with higher applicability, the results of the integrated analysis methods are used to construct a dynamic update mechanism for flood probability assessment, and the final assessment adjustment parameters are determined.

[0061] When constructing a dynamic update mechanism for flood probability assessment, the above analysis results can be integrated, and parameters can be updated periodically based on the set of adjustment coefficients. It is assumed that adjustments are made monthly based on the latest meteorological data to ensure that the assessment results are synchronized with actual environmental changes, thereby enhancing the timeliness and reliability of forecasts.

[0062] Based on the dynamic adjustment coefficient of probability assessment and the trend distribution of rainstorm frequency changes, the coupling structure of rainstorm and flood is obtained and quantified, and the quantitative expression of the coupling relationship is obtained.

[0063] As a specific implementation method, the process of obtaining and quantifying the coupling structure between rainstorms and floods based on the dynamic adjustment coefficient of probability assessment and the trend distribution of rainstorm frequency changes, and obtaining the quantitative expression of the coupling relationship, includes:

[0064] The dynamic adjustment coefficient is correlated and matched with the rainfall and trend distribution using a preset mapping rule to determine the initial coefficient adjustment range. Based on the coefficient adjustment range and the construction requirements of the coupling framework, a coupling structure between rainfall and flood is built by integrating flood correlation data. According to the coupling structure, multivariate regression analysis is applied to quantify the relationship between rainfall frequency, flood correlation and dynamic adjustment coefficient, and obtain a quantitative expression of the coupling relationship.

[0065] For example, when analyzing the distribution trend of rainstorm frequency changes, one can start with long-term observation of the data to understand its periodicity and fluctuation characteristics. Suppose that historical data for a certain region shows a year-on-year increasing trend in rainstorm frequency over the past 10 years, especially during the rainy season from June to August, when the average number of rainy days per month increases from 3 to 5. This trend provides a basis for subsequent dynamic adjustment coefficient correlation. Through pre-established mapping rules, this trend is linked to the adjustment coefficient; for example, for every one-day increase in frequency, the coefficient may be adjusted upwards by 0.2 to reflect a higher flood risk.

[0066] To correlate the dynamic adjustment coefficients with the trend distribution, we can start from regional characteristics and consider the differences in rainfall patterns across different areas. A coastal region, due to its topography, faces a higher risk of flooding after heavy rains; therefore, its coefficient adjustment range might be set between 0.5 and 1.5, while for inland plains, it might be between 0.3 and 1.0. This matching method ensures the targeted nature of the coefficient adjustment and also lays the data foundation for the subsequent construction of the coupling framework.

[0067] When establishing the initial coupling structure between rainstorms and floods, a comprehensive correlation distribution map can be formed by integrating multi-source data, such as rainfall, soil moisture, and river levels. Suppose an analysis reveals that the probability of flooding increases significantly when rainfall exceeds 100 mm and soil moisture reaches 80%. If certain characteristic values ​​are below preset thresholds, such as missing or inaccurate soil moisture data, outlier data is removed through information processing, and the structure is corrected using historical averages to optimize the coupling structure. This approach improves the reliability of the structure and provides more accurate input for subsequent analyses.

[0068] For multivariate regression analysis, we can start by examining the interactions between variables to quantify the relationship between rainfall frequency, flood correlation, and dynamic adjustment coefficients. Assuming that in a certain analysis, rainfall frequency is the primary variable with a weight of 60%, while river water level is the secondary variable with a weight of 30%, the dynamic adjustment coefficient is adjusted based on the combined effect of both. This quantification helps clarify the magnitude of each factor's influence, ensuring the final expression closely reflects the actual scenario and providing a scientific basis for the dynamic assessment of flood probability.

[0069] When constructing quantization expressions, changes over time can be further considered.

[0070] Suppose that the flood risk in a region differs at the beginning and end of the rainy season. Initially, due to the soil's high water absorption capacity, the risk is lower, and the coefficient might be adjusted to 0.4. However, at the end, when the soil is saturated, the risk increases, and the coefficient is adjusted upwards to 0.8. This time-varying quantification method allows the assessment system to be more flexible and adaptable to environmental conditions at different stages.

[0071] Specifically, a preliminary coupling structure between rainstorms and floods is established to obtain a structured correlation distribution. If some characteristic values ​​of the correlation distribution are lower than a preset threshold, the data is filtered and corrected through information processing to obtain an optimized coupling structure.

[0072] The optimized coupling structure can be further analyzed from the perspective of data integrity. For example, if some rainfall records are missing due to equipment malfunction after data filtering, interpolation using rainfall data from neighboring areas can correct this, ensuring the integrity of the coupling structure. This approach not only improves data usability but also provides more comprehensive support for subsequent flood warnings. Through the above multi-faceted analysis and examples, it can be seen that from the trend of rainstorm frequency to the optimization of the coupling structure and the construction of quantitative relationships, each step closely revolves around the core needs of flood probability assessment, ensuring the scientific rigor and practicality of the analysis.

[0073] The basic information set under the complex disaster scenario is obtained and corrected. Combined with the quantitative expression of the coupling relationship, the prediction module of the early warning model is constructed. The response characteristics of the prediction module are obtained, the early warning model is adjusted based on the response characteristics, and the prediction output of the adjusted early warning model is verified.

[0074] As a specific implementation method, the process of acquiring and correcting the basic information set under complex disaster scenarios, and constructing the prediction module of the early warning model by combining the quantitative expression of the coupling relationship includes:

[0075] The system acquires and initially classifies basic information sets under complex disaster scenarios to obtain classified data groups. Based on these data groups, a pre-established rule base is used for matching to determine key data subsets related to the disaster scenario. For these key data subsets, the system constructs the input layer structure of the early warning model, obtains the mapping relationship between the input layer and environmental conditions, and assesses the stability of the mapping relationship. If the stability of the mapping relationship is lower than a preset threshold, the key data subset is corrected through information processing to obtain a corrected data subset. Based on the corrected data subset and a quantification expression, the system constructs the prediction module of the early warning model.

[0076] In data processing and early warning model construction for complex disaster scenarios, acquiring multi-source data is the primary step. Multi-source data may include meteorological observation data, geographic information data, and historical disaster records, which are integrated to form a basic information set. For example, in a flood disaster scenario triggered by a rainstorm, meteorological data provides rainfall of 100 mm within 24 hours, geographic data provides topographic slope and river distribution, and historical records show that the area has experienced multiple floods. This initial integration of data forms an information set, providing a basis for subsequent classification.

[0077] For the initial classification of the information set, the data can be grouped by time, space, and disaster type. In terms of time, the data is divided into three groups: the past 24 hours, the past 7 days, and the past month; in terms of space, it is divided by administrative region or river basin; and in terms of disaster type, it is divided into rainstorms, floods, and secondary disasters. These categorized data groups lay the foundation for the subsequent matching rule base. For example, if a group shows that a certain river basin has experienced persistently high rainfall in the past 7 days, it may be related to flood risk.

[0078] When using a pre-established rule base for matching, a rule can be set to classify an area as high-risk if rainfall exceeds 80 mm and the terrain slope is greater than 15 degrees. The key data subset may include rainfall and terrain data for the watershed, used for further analysis. The rule base matching process can be continuously optimized based on historical disaster cases to ensure that the extracted data subset is highly relevant to actual disaster scenarios.

[0079] When constructing the input layer structure of an early warning model, the input layer may include variables such as rainfall, topographic features, and river flow. The mapping relationship is reflected in the association between these variables and environmental conditions such as soil moisture and vegetation cover. It is assumed that when soil moisture reaches 90%, the impact of rainfall on flooding is significantly enhanced; the stability of this mapping relationship needs to be verified using historical data. If the stability is insufficient, such as due to large data fluctuations, a subset of data needs to be corrected, such as by removing outliers or supplementing missing data.

[0080] When constructing the core prediction module based on the corrected subset of data and a quantified expression, rainfall and flow data can be substituted into the expression to generate flood risk levels. The logical framework might be designed as a hierarchical judgment: first, assess whether the rainfall reaches a threshold; then, combine flow data to determine the likelihood of flooding; and finally, output the risk distribution characteristics. This approach helps to accurately locate high-risk areas.

[0081] As a specific implementation method, the process of obtaining the response characteristics of the prediction module and adjusting the early warning model based on the response characteristics includes:

[0082] The process involves acquiring the response characteristics of the prediction module in the early warning model, extracting response values, determining whether the response values ​​exceed the specified range, and dynamically correcting the parameters of the core module based on response values ​​exceeding the threshold range. The magnitude of parameter adjustments is calibrated through information processing to determine the adjusted parameter configuration. The prediction module of the early warning model is updated with the adjusted parameter configuration until the stability of the response characteristic change trend meets the requirements, resulting in an optimized parameter set. Based on the optimized parameter set, the prediction function's operating environment is reloaded, and the performance of the prediction module on the updated response characteristics is obtained to determine whether the response values ​​have returned to the threshold range. If the response values ​​still exceed the threshold range, the triggering conditions are analyzed through information processing to obtain a condition correction scheme. For the condition correction scheme, the triggering condition settings of the early warning model are adjusted, and a support vector machine algorithm is used to model and analyze the fluctuation pattern of the response values, obtaining the distribution characteristics of the fluctuation pattern and determining whether the distribution characteristics meet the expected standards. The prediction output of the prediction module is updated based on the judgment results.

[0083] The determination of response characteristic distribution can be represented as a risk level map of different regions. When a region is shown as high-risk, flood control resources can be deployed in advance. This distribution characteristic provides an intuitive basis for disaster early warning and helps improve emergency response efficiency. Through the coordinated processing of the above multiple links, a complete chain is formed from data acquisition to model building, ensuring the reliability of the early warning system, while providing scientific support for disaster prevention and control and reducing potential losses.

[0084] When constructing an early warning model for complex disaster scenarios, the performance of the core module can be verified by simulating environmental conditions to obtain response characteristics and compare numerical data. Assuming a scenario involving a combined flood and landslide disaster, the response value of the core module is set to a range, such as a safe interval of 0.8 to 1.2. If a response value of 1.5 is detected, significantly exceeding the threshold range, the parameter adjustment process is initiated. Preliminary judgment indicates that the high value may be related to an excessively high weighting of the humidity factor in the input data. Therefore, the parameters need to be dynamically adjusted, reducing the weighting of the humidity factor from 0.6 to 0.4, and observing whether the response value returns to normal.

[0085] For the dynamic parameter correction of the adaptive mechanism, simulation calibration can be performed using historical disaster data. In one simulation test, the parameter adjustment range of the core module was calibrated to no more than 0.1 at a time to avoid excessive fluctuations that could lead to model instability. After adjustment, the response value decreased from 1.5 to 1.3, which, although an improvement, still did not reach a stable state. At this point, through a secondary optimization mechanism, the adjustment step size was further refined to 0.05, and combined with the wind speed factor for joint correction, ultimately bringing the value back to 1.1, which is within the threshold range. This approach can effectively improve the model's adaptability.

[0086] When updating the runtime environment of the prediction function, the changing trend of the response characteristics can be observed by loading the optimized parameter set. For example, if the fluctuation range of the response value gradually decreases over five consecutive tests, from an initial 0.3 to 0.1, it indicates that the trend is stabilizing. Judging this trend helps ensure the reliability of the model under different environments. Furthermore, if the triggering condition needs to be corrected, the reasons for the numerical deviations can be analyzed. For instance, if insufficient rainfall data collection frequency is found in a test leading to misjudgment, the triggering condition can be adjusted to update the data collection frequency to once per hour.

[0087] To model and analyze the fluctuation patterns of response values, a support vector machine (SVM) algorithm can be used to extract features from historical data. In one analysis, a periodic pattern was found, with a peak occurring every 6 hours, highly correlated with the rainfall cycle. Based on this distribution characteristic, the prediction output logic of the core module was updated to incorporate periodic fluctuations, ensuring the completeness and consistency of the output data. This method can significantly improve the accuracy of predictions, providing a more reliable basis for disaster early warning.

[0088] As a specific implementation method, the process of verifying the prediction output of the adjusted early warning model includes:

[0089] The prediction output of the early warning model is analyzed using uncertainty quantification methods to obtain fluctuation data and determine its distribution characteristics. Based on these characteristics, uncertainty factors are decomposed using information processing to obtain a set of influencing factors. It is then determined whether this set of factors meets a preset range standard. If the set exceeds the preset range standard, the adjustment results are corrected through information processing to obtain the corrected parameter configuration. Based on the corrected parameter configuration, the prediction module of the early warning model is updated, and the updated model performance data is obtained. Finally, it is determined whether the model performance data meets the timeliness and reliability requirements.

[0090] Based on the verified prediction output, early warning signal distributions under multiple scenarios are generated, and an early warning information mapping table is obtained. Based on the early warning information mapping table, early warning of combined rainstorm and flood disasters is realized.

[0091] As a specific implementation method, the process of generating early warning signal distributions in multiple scenarios based on the verified prediction output results and obtaining an early warning information mapping table is as follows:

[0092] Based on the verified prediction output results and the characteristics of the warning signals, a data integration tool is used to initially classify the signal distribution under multiple scenarios, resulting in classified signal distribution groups. According to the classified signal distribution groups, combined with environmental and adaptive conditions, the degree of matching between each group and the conditions is compared. If the matching degree does not reach a preset threshold, the signal distribution groups are reclassified to determine the adjusted distribution groups. Based on the adjusted distribution groups, feature extraction is performed on the signal distribution characteristics under different scenarios to obtain the extracted feature set. Based on the extracted feature set, combined with the warning signals and signal distribution, information processing is performed on the correlation between the feature set and the distribution map. If the feature set cannot be completely mapped to the distribution map, the feature set is supplemented to obtain a supplemented feature mapping. Based on the supplemented feature mapping, combined with the information mapping and the warning set, the mapping content is integrated using a data fusion tool to determine whether the integrated mapping content meets the condition matching requirements, obtaining an integrated warning information set. Based on the integrated warning information set, combined with condition matching and scenario analysis, an information verification tool is used to perform consistency checks on the set content. After passing the consistency check, combined with the signal distribution and distribution map, a warning information mapping table is obtained.

[0093] Specifically, based on the verification results and output data, and considering the characteristics of the warning signals, a data integration tool is used to initially classify the signal distribution under multiple scenarios, resulting in classified signal distribution groups. According to these classified signal distribution groups, and considering environmental conditions and adaptability requirements, the degree of matching between each group and the conditions is compared through information processing. If the matching degree does not reach a preset threshold, the signal distribution groups are reclassified to determine the adjusted distribution groups. Based on the adjusted distribution groups, and considering multiple scenario situations and scenario analysis, a support vector machine algorithm is used to extract features from the signal distribution characteristics under different scenarios, obtaining the extracted feature set. Based on the extracted feature set, and considering the warning signals and signal distribution, information processing is performed on the correlation between the feature set and the distribution map. If the feature set cannot be completely mapped to the distribution map, the feature set is supplemented to obtain a supplemented feature mapping. Based on the supplemented feature mapping, and combining the information mapping with the warning set, a data fusion tool is used to integrate the mapping content, determining whether the integrated mapping content meets the condition matching requirements, and obtaining the integrated warning information set. Based on the integrated set of early warning information, and combining condition matching and scenario analysis, an information verification tool is used to perform consistency checks on the content of the set. If the consistency check fails to pass a preset threshold, the information set is corrected to determine the final set of early warning information mappings. Based on the final set of early warning information mappings, and combining signal distribution and distribution diagrams, a data visualization tool is used to graphically process the mapping set to obtain early warning signal distribution views under multiple scenarios.

[0094] If the feature set cannot be completely mapped to the distribution diagram, for example, if some frequency features are missing from the diagram, supplementary processing is required. This can be done by backtracking through historical data to supplement missing frequency range data, such as adding signal records in the 40-50Hz range, thereby forming a complete feature mapping. This method ensures comprehensive information.

[0095] When integrating mapping content using data fusion tools, feature mappings can be combined with early warning information sets to determine if they meet matching requirements. For example, if the integration reveals that some mapping content lacks early warning information for low-temperature scenarios, the fusion strategy needs to be adjusted, increasing the weight of low-temperature scenarios to 30% to ensure information completeness. This integration improves the applicability of early warning information.

[0096] In the consistency check phase, if the consistency of the integrated warning information set fails to meet the standard in certain scenarios, such as a check result of 75% while the threshold is 85%, the information set needs to be corrected. One possible approach is to remove outlier data points and re-verify consistency to ensure the reliability of the final mapping set.

[0097] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for early warning of combined rainstorm and flood disasters based on meteorological big data, characterized in that, Includes the following steps: Long-term meteorological data is acquired and processed in layers. Based on the processed data, periodic characteristics and abnormal fluctuations of rainstorm frequency are obtained. The trend distribution of rainstorm frequency changes is obtained by combining time series methods. Based on the trend distribution of rainfall frequency variation and topographic and hydrological feature data, a correlation matrix between rainfall frequency and flood probability is obtained, and the initial parameter range for flood probability assessment is obtained based on the correlation matrix. Adaptive constraints for meteorological data analysis and environmental conditions are constructed. Based on the initial parameter range and adaptive constraints for flood probability assessment, and combined with dynamic correlation analysis, dynamic adjustment coefficients for probability assessment are obtained. Based on the dynamic adjustment coefficient of probability assessment and the trend distribution of rainstorm frequency changes, the coupling structure of rainstorm and flood is obtained and quantified, and the quantitative expression of the coupling relationship is obtained. Acquire and correct the basic information set under the complex disaster scenario, and construct the prediction module of the early warning model by combining the quantitative expression of the coupling relationship; Obtain the response characteristics of the prediction module, adjust the early warning model based on the response characteristics, and verify the prediction output of the adjusted early warning model; Based on the verified prediction output, early warning signal distributions under multiple scenarios are generated, and an early warning information mapping table is obtained. Based on the early warning information mapping table, early warning of combined rainstorm and flood disasters is realized.

2. The method for early warning of combined rainstorm and flood disasters based on meteorological big data according to claim 1, characterized in that, The process of obtaining the trend distribution of changes in the frequency of rainstorms includes: The long-term meteorological data is cleaned to obtain basic data. A hierarchical processing method is used to classify the basic data, dividing it according to time and geographical dimensions to obtain hierarchical data groups. Based on these hierarchical data groups, the periodic characteristics of rainstorm frequency are extracted, and the time series is decomposed using Fourier transform to obtain the periodic characteristics of rainstorm frequency. If the fluctuation amplitude of rainstorm frequency exceeds a preset threshold within a certain time period, it is marked as an abnormal fluctuation point, and the abnormal fluctuation distribution is obtained. After integrating the abnormal fluctuation distribution and periodic characteristics, the time series of frequency changes is smoothed using a moving average method. The trend distribution of rainstorm frequency changes is obtained based on the smoothed frequency change curve.

3. The method for early warning of combined rainstorm and flood disasters based on meteorological big data according to claim 1, characterized in that, The process of obtaining the initial parameter range for flood probability assessment includes: Correlation analysis was used to calculate the correlation strength between rainfall frequency and topographic and hydrological features, and a set of significantly correlated indicators was selected. Based on the set of significantly correlated indicators, a correlation matrix between rainfall frequency and flood probability was constructed, and the weight distribution of each indicator in the correlation matrix was obtained to obtain the weighted correlation relationship. Based on the weighted correlation relationship, key influencing factors were obtained. Based on the key influencing factors and information processing, the corresponding topographic and hydrological features were selected to obtain the distribution range of the key influencing factors. Combined with the assessment requirements of flood probability, the initial parameter range for flood probability assessment was obtained.

4. The method for early warning of combined rainstorm and flood disasters based on meteorological big data according to claim 1, characterized in that, The process of constructing adaptive constraints for meteorological data analysis and environmental conditions, based on the initial parameter range and adaptive constraints for flood probability assessment, and combined with dynamic correlation analysis, to obtain the dynamic adjustment coefficients for probability assessment includes: To meet the requirements of flood probability assessment, meteorological data and environmental conditions are integrated to obtain a comprehensive environmental dataset. Based on this dataset and dynamic correlation analysis, the interaction between meteorological data and environmental conditions is extracted to obtain the distribution characteristics of key triggering factors. If the correlation of a certain data point is lower than a preset threshold, that data point is removed, resulting in an optimized set of triggering factors. Based on the optimized set of triggering factors and the initial parameter range, a mapping relationship between flood probability and key triggering factors is constructed, and weights are assigned to this mapping relationship to obtain a weighted impact distribution. Based on the weighted impact distribution and adaptive constraints, the limiting relationship between environmental conditions and flood probability is obtained. If the weight of a certain adaptive constraint is lower than a preset threshold, its impact ratio is adjusted to obtain an optimized constraint distribution. For the optimized constraint distribution, a logistic regression model is used to quantify the dynamic adjustment of flood probability assessment, obtaining a candidate set of adjustment coefficients, which is then filtered to obtain the final coefficient range. Based on the final coefficient range and the actual needs of probability assessment, the dynamic adjustment coefficients are divided into intervals, and the applicability of each interval is judged. Based on the applicability results and integrated analysis, the dynamic adjustment coefficients for probability assessment are obtained.

5. The method for early warning of combined rainstorm and flood disasters based on meteorological big data according to claim 1, characterized in that, Based on the dynamic adjustment coefficient of probability assessment and the trend distribution of rainstorm frequency changes, the coupling structure of rainstorm and flood is obtained and quantified. The process of obtaining the quantitative expression of the coupling relationship includes: The dynamic adjustment coefficient is correlated and matched with the rainfall and trend distribution using a preset mapping rule to determine the initial coefficient adjustment range. Based on the coefficient adjustment range and the construction requirements of the coupling framework, a coupling structure between rainfall and flood is built by integrating flood correlation data. According to the coupling structure, multivariate regression analysis is applied to quantify the relationship between rainfall frequency, flood correlation and dynamic adjustment coefficient, and obtain a quantitative expression of the coupling relationship.

6. The method for early warning of combined rainstorm and flood disasters based on meteorological big data according to claim 1, characterized in that, The process of acquiring and correcting the basic information set under complex disaster scenarios, and then constructing the prediction module of the early warning model by combining the quantitative expression of the coupling relationship includes: The system acquires and initially classifies basic information sets under complex disaster scenarios to obtain classified data groups. Based on these data groups, a pre-established rule base is used for matching to determine key data subsets related to the disaster scenario. For these key data subsets, the system constructs the input layer structure of the early warning model, obtains the mapping relationship between the input layer and environmental conditions, and assesses the stability of the mapping relationship. If the stability of the mapping relationship is lower than a preset threshold, the key data subset is corrected through information processing to obtain a corrected data subset. Based on the corrected data subset and a quantification expression, the system constructs the prediction module of the early warning model.

7. The method for early warning of combined rainstorm and flood disasters based on meteorological big data according to claim 1, characterized in that, The process of obtaining the response features of the prediction module and adjusting the early warning model based on these features includes: The process involves acquiring the response characteristics of the prediction module in the early warning model, extracting response values, determining whether the response values ​​exceed the specified range, and dynamically correcting the parameters of the core module based on response values ​​exceeding the threshold range. The magnitude of parameter adjustments is calibrated through information processing to determine the adjusted parameter configuration. The prediction module of the early warning model is updated with the adjusted parameter configuration until the stability of the response characteristic change trend meets the requirements, resulting in an optimized parameter set. Based on the optimized parameter set, the prediction function's operating environment is reloaded, and the performance of the prediction module on the updated response characteristics is obtained to determine whether the response values ​​have returned to the threshold range. If the response values ​​still exceed the threshold range, the triggering conditions are analyzed through information processing to obtain a condition correction scheme. For the condition correction scheme, the triggering condition settings of the early warning model are adjusted, and a support vector machine algorithm is used to model and analyze the fluctuation pattern of the response values, obtaining the distribution characteristics of the fluctuation pattern and determining whether the distribution characteristics meet the expected standards. The prediction output of the prediction module is updated based on the judgment results.

8. The method for early warning of combined rainstorm and flood disasters based on meteorological big data according to claim 1, characterized in that, The process of validating the prediction output of the adjusted early warning model includes: The prediction output of the early warning model is analyzed using uncertainty quantification methods to obtain fluctuation data and determine its distribution characteristics. Based on these characteristics, uncertainty factors are decomposed using information processing to obtain a set of influencing factors. It is then determined whether this set of factors meets a preset range standard. If the set exceeds the preset range standard, the adjustment results are corrected through information processing to obtain the corrected parameter configuration. Based on the corrected parameter configuration, the prediction module of the early warning model is updated, and the updated model performance data is obtained. Finally, it is determined whether the model performance data meets the timeliness and reliability requirements.

9. The method for early warning of combined rainstorm and flood disasters based on meteorological big data according to claim 1, characterized in that, The process of generating early warning signal distributions across multiple scenarios and obtaining an early warning information mapping table based on the validated prediction output results: Based on the verified prediction output results and the characteristics of the warning signals, a data integration tool is used to initially classify the signal distribution under multiple scenarios, resulting in classified signal distribution groups. According to the classified signal distribution groups, combined with environmental and adaptive conditions, the degree of matching between each group and the conditions is compared. If the matching degree does not reach a preset threshold, the signal distribution groups are reclassified to determine the adjusted distribution groups. Based on the adjusted distribution groups, feature extraction is performed on the signal distribution characteristics under different scenarios to obtain the extracted feature set. Based on the extracted feature set, combined with the warning signals and signal distribution, information processing is performed on the correlation between the feature set and the distribution map. If the feature set cannot be completely mapped to the distribution map, the feature set is supplemented to obtain a supplemented feature mapping. Based on the supplemented feature mapping, combined with the information mapping and the warning set, the mapping content is integrated using a data fusion tool to determine whether the integrated mapping content meets the condition matching requirements, obtaining an integrated warning information set. Based on the integrated warning information set, combined with condition matching and scenario analysis, an information verification tool is used to perform consistency checks on the set content. After passing the consistency check, combined with the signal distribution and distribution map, a warning information mapping table is obtained.

Images