Flood disaster research and judgment method and device based on precipitation forecast grading correction

Abstract

The application relates to the technical field of flood prevention, and discloses a flood disaster research and judgment method and device based on graded correction of precipitation prediction, which comprises the following steps: obtaining historical typhoon data, historical rainwater condition data and real-time meteorological data of a target area; establishing an extreme precipitation correlation analysis model based on the historical typhoon data to obtain a medium and long-term prediction correction result; correcting short-term and imminent precipitation prediction by using a pre-constructed precipitation correction network based on the medium and long-term prediction correction result and the real-time meteorological data to obtain a full-cycle precipitation prediction correction result; performing similarity matching on the full-cycle precipitation prediction correction result and the historical rainwater condition data to screen out the most similar historical precipitation field; based on the most similar historical precipitation field, corresponding historical flood information and disaster information are retrieved to obtain a flood disaster research and judgment result. The application performs graded correction on medium and long-term and short-term and imminent precipitation prediction, effectively improves the accuracy of full-cycle precipitation prediction, and realizes rapid research and judgment of flood disasters.

Description

Technical Field

[0001] This invention relates to the field of flood control technology, specifically to a method and apparatus for assessing flood disasters based on graded correction of precipitation forecasts. Background Technology

[0002] In the field of flood disaster prevention, rapid and accurate flood disaster assessment is the foundation for flood control early warning and dispatch decisions. Existing flood disaster assessment and prediction technologies can be mainly divided into the following categories, but all of them have certain limitations.

[0003] The first category is assessment methods based on numerical precipitation forecasts and hydrodynamic models. For example, the paper "A Numerical Simulation Method for Community Rainwater Flooding Based on a Hydrological-Hydrodynamic Coupled Model" (publication number CN115510771A) proposes to achieve dynamic simulation of flood inundation range and depth by solving two-dimensional shallow water equations. Its advantages are clear physical mechanisms and detailed results; however, its disadvantages are that it is limited by incomplete understanding of the physical processes of rainstorms and the influence of complex terrain, resulting in insufficient accuracy of precipitation and location in medium- and long-term and short-term forecasts; and the hydrodynamic model is limited by computational efficiency and cannot achieve real-time calculation, thus failing to provide effective support for rapid disaster assessment and dispatch decisions.

[0004] The second category is similarity matching techniques based on historical flood characteristics. For example, the paper "A Flood Forecasting and Disaster Prevention Decision-Making Method and System" (publication number CN107609707B) proposes a method for similarity analysis using monitoring and sensing technology. Its advantage is that it can intuitively show the impact of disasters; however, its disadvantage is that it relies on measured rainfall or measured flow processes for synchronous analysis of flood disasters, and essentially lacks forecasting capabilities, failing to allow sufficient time for water conservancy project scheduling and emergency response. Summary of the Invention

[0005] This invention provides a flood disaster assessment method and apparatus based on precipitation forecast classification correction, which solves the problem that existing technologies are limited by insufficient accuracy of precipitation and local precipitation forecasts in medium- and long-term and short-term forecasts, as well as low computational efficiency of hydrodynamic models, making it difficult to quickly and accurately assess the disaster impact.

[0006] In a first aspect, the present invention provides a flood disaster assessment method based on precipitation forecast classification correction, the method comprising: Acquire historical typhoon data, historical rainfall data, and real-time meteorological data for the target area; An extreme precipitation correlation analysis model was established based on historical typhoon data, and the extreme precipitation correlation analysis model was used to correct medium- and long-term forecasts to obtain medium- and long-term forecast correction results. Based on the medium- and long-term forecast correction results and real-time meteorological data, a pre-constructed precipitation correction network is used to correct the short-term precipitation forecast, and the full-cycle precipitation forecast correction results are obtained. The full-cycle precipitation forecast correction results are matched with historical rainfall data to identify the most similar historical precipitation events. Based on the most similar historical precipitation events, corresponding historical flood information and disaster information are retrieved to obtain flood disaster assessment results.

[0007] This invention provides a flood disaster assessment method based on graded correction of precipitation forecasts. It acquires historical typhoon data, historical rainfall data, and real-time meteorological data for the target area; establishes an extreme precipitation correlation analysis model based on historical typhoon data to obtain medium- and long-term forecast correction results, enabling qualitative assessment of typhoon rainfall up to seven days in advance and quantitative evaluation up to three days in advance; and uses a pre-constructed precipitation correction network to correct short-term precipitation forecasts based on the medium- and long-term forecast correction results and real-time meteorological data, effectively reducing meteorological forecast uncertainty and obtaining full-cycle precipitation forecast correction results; thus achieving accurate forecasting. The invention performs similarity matching between the full-cycle precipitation forecast correction results and historical rainfall data to select the most similar historical precipitation events. Based on the most similar historical precipitation events, corresponding historical flood information and disaster information are retrieved to obtain flood disaster assessment results. This invention performs graded correction on medium- and long-term and short-term precipitation forecasts, effectively improving the accuracy of full-cycle precipitation forecasts, greatly extending the flood forecast lead time, and enabling rapid assessment of flood disasters. This provides sufficient time for water conservancy project scheduling and emergency response, and solves the problems of insufficient forecast accuracy and low assessment efficiency in existing technologies.

[0008] In one optional implementation, an extreme precipitation correlation analysis model is established based on historical typhoon data, including: Multi-dimensional characteristic indicators of typhoons affecting precipitation were extracted based on historical typhoon data. A statistical analysis method was used to establish a quantitative relationship between multi-dimensional typhoon characteristic indicators and precipitation in the target area; Based on historical rainfall data, the quantitative relationship was verified, and an extreme precipitation correlation analysis model was obtained.

[0009] In the above technical solution, multi-dimensional feature indicators are extracted from historical typhoon data, and a quantitative relationship between them and regional precipitation is established using statistical analysis. After verification with historical rainfall data, an extreme precipitation correlation analysis model is formed, which realizes a quantitative description of the medium- and long-term trend of extreme precipitation, improves the accuracy and reliability of medium- and long-term forecast correction, and provides a foundation for subsequent full-cycle precipitation forecast correction.

[0010] In one optional implementation, an extreme precipitation correlation analysis model is used to correct the medium- and long-term forecasts of extreme precipitation, resulting in corrected medium- and long-term forecasts, including: Extracting multi-dimensional feature indicators of the current weather system from real-time meteorological data; By inputting the multi-dimensional characteristic indicators of the current weather system into the extreme precipitation correlation analysis model, the medium- and long-term forecast of regional precipitation is calculated. Error verification and correction are performed on medium- and long-term forecasts based on historical rainfall and water conditions, resulting in corrected medium- and long-term forecasts.

[0011] In the above technical solution, multi-dimensional characteristic indicators of the current weather system are extracted from real-time meteorological data, input into the extreme precipitation correlation analysis model to calculate the medium- and long-term estimated value of regional precipitation, and the estimated value is verified and corrected based on historical rainfall and water conditions, thereby realizing the quantitative correction of medium- and long-term precipitation forecasts and improving the accuracy and reliability of forecast results.

[0012] In one alternative implementation, the pre-built precipitation correction network is implemented in the following manner: Obtain historical precipitation forecast data and historical meteorological station measured data at the same timestamp to construct a training dataset; Based on the training dataset, a deep learning algorithm was used to train the mapping relationship between historical precipitation forecast data and historical meteorological station measured data to obtain an initial correction network. The accuracy of the initial correction network was tested and the parameters were optimized using the validation dataset to obtain the precipitation correction network.

[0013] In the above technical solution, a training dataset is constructed by acquiring historical precipitation forecast data and measured data at the same time stamp. A deep learning algorithm is used to train the mapping relationship between the two to obtain an initial correction network. The network is then tested and its parameters are optimized using a validation dataset. This ensures that the precipitation correction network can identify and correct forecast biases, providing a reliable model foundation for accurate correction of short-term precipitation forecasts.

[0014] In one optional implementation, based on medium- and long-term forecast correction results and real-time meteorological data, a pre-constructed precipitation correction network is used to correct short-term precipitation forecasts to obtain full-cycle precipitation forecast correction results, including: Extract short-term precipitation forecast data for the current moment from real-time meteorological data; The medium- and long-term forecast correction results and short-term precipitation forecast data are used as inputs to the precipitation correction network to perform real-time correction of precipitation amount and location on the short-term precipitation forecast data, and output the corrected short-term precipitation forecast results. By sequentially linking the corrected medium- and long-term precipitation forecast results with the corrected short-term precipitation forecast results, a full-cycle precipitation forecast correction result from long-term to short-term is obtained.

[0015] In the above technical solution, by inputting the medium- and long-term forecast correction results and the current short-term precipitation forecast data into the precipitation correction network, the precipitation amount and location of the short-term forecast are corrected in real time, and the corrected short-term forecast results are sequentially linked with the medium- and long-term forecast correction results, thus realizing full-cycle precipitation forecast correction from long-term to short-term.

[0016] In one optional implementation, the full-cycle precipitation forecast correction results are matched with historical rainfall data to identify the most similar historical precipitation events, including: Based on historical rainfall data and full-cycle precipitation forecast correction results, multiple similarity algorithms are used to calculate the similarity between current precipitation and historical precipitation events in multiple dimensions, including surface average precipitation, spatial distribution, and overall error. These similarity algorithms include Nash coefficient, Pearson correlation coefficient, and root mean square error. The similarity across multiple dimensions is weighted and ranked, and the historical precipitation events most similar to the current precipitation are selected based on the ranking results.

[0017] In the above technical solution, three similarity algorithms, namely Nash coefficient, Pearson correlation coefficient and root mean square error, are used to calculate the similarity between the current precipitation and the historical precipitation from three dimensions: average precipitation, spatial distribution and overall error. The similarity of the three dimensions is weighted and sorted to achieve accurate screening of the most similar historical precipitation events, providing a reliable matching basis for subsequent retrieval of historical flood and disaster information.

[0018] In one optional implementation, based on the most similar historical precipitation events, corresponding historical flood information and disaster information are retrieved to obtain flood disaster assessment results, including: Using the most similar historical precipitation events as index identifiers, the corresponding flood process information is retrieved from the historical rainfall and water conditions database, and the corresponding disaster information is retrieved from the historical disaster information database, including disaster descriptions and data on the locations of affected points; By linking and integrating flood process information with disaster information, a flood disaster assessment result containing key flood characteristics and potential affected areas is generated.

[0019] In the above technical solution, the most similar historical precipitation events are used as index identifiers to retrieve corresponding flood process information from the historical rainfall and water conditions database. At the same time, disaster descriptions and disaster location data are retrieved from the historical disaster database. The flood process information and disaster information are linked and integrated to achieve rapid reproduction of historical disaster scenarios and fusion output of multi-source information, directly generating judgment results that include key flood characteristics and potential impact areas.

[0020] Secondly, the present invention provides a flood disaster assessment device based on precipitation forecast classification correction, the device comprising: The data acquisition module is used to acquire historical typhoon data, historical rainfall data, and real-time meteorological data for the target area. The medium- and long-term forecast module is used to establish an extreme precipitation correlation analysis model based on historical typhoon data, and to use the extreme precipitation correlation analysis model to correct the medium- and long-term forecasts, thereby obtaining the corrected medium- and long-term forecast results. The short-term correction module is used to correct short-term precipitation forecasts based on medium- and long-term forecast correction results and real-time meteorological data, using a pre-constructed precipitation correction network to obtain full-cycle precipitation forecast correction results. The similarity matching module is used to perform similarity matching between the full-cycle precipitation forecast correction results and historical rainfall data, and to select the most similar historical precipitation events. The analysis and output module is used to retrieve corresponding historical flood information and disaster information based on the most similar historical precipitation events to obtain flood disaster analysis results.

[0021] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the flood disaster assessment method based on precipitation forecast grading correction described in the first aspect or any corresponding embodiment.

[0022] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the flood disaster assessment method based on precipitation forecast grading correction described in the first aspect or any corresponding embodiment.

[0023] Fifthly, the present invention provides a computer program product, including computer instructions, which are used to cause a computer to execute the flood disaster assessment method based on precipitation forecast classification correction described in the first aspect or any corresponding embodiment above. Attached Figure Description

[0024] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of the first process of a flood disaster assessment method based on precipitation forecast classification correction according to an embodiment of the present invention; Figure 2This is a schematic diagram of the second process of the flood disaster assessment method based on precipitation forecast classification correction according to an embodiment of the present invention; Figure 3(a) is a schematic diagram of the simulation effect of the typhoon and rainstorm correlation analysis model according to an embodiment of the present invention; Figure 3(b) is a schematic diagram of the simulation effect of another storm-rainstorm correlation analysis model according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the cloud rain-to-ground rain correction network structure according to an embodiment of the present invention; Figure 5 This is a structural block diagram of a flood disaster assessment device based on precipitation forecast classification correction according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0028] According to an embodiment of the present invention, a method for assessing flood disasters based on precipitation forecast classification correction is provided. 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. Furthermore, 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.

[0029] In light of the problems in the background technology, there is an urgent need for a flood disaster assessment and prediction technology that can achieve both rapid assessment and high accuracy with a long lead time. Therefore, this invention innovatively proposes a typhoon-rainfall correlation analysis model, enabling long-term qualitative and quantitative assessment of precipitation; simultaneously, it constructs a "cloud-to-ground rainfall" correction technique, significantly reducing the uncertainty of weather forecasts. By incorporating high-precision precipitation data into historical similarity matching, it effectively solves the problems of low forecast accuracy, long computation time, and lack of lead time in existing technologies.

[0030] This embodiment provides a flood disaster assessment method based on precipitation forecast grading correction, which can be used in the aforementioned electronic equipment. Figure 1 This is a flowchart of a flood disaster assessment method based on precipitation forecast classification correction according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain historical typhoon data, historical rainfall data, and real-time meteorological data for the target area.

[0031] Historical typhoon data refers to relevant data recorded during the period when the target area was affected by typhoons in the past, including characteristic indicators such as the typhoon's origin, movement path, landfall location, intensity, wind speed, air pressure, movement speed, and water vapor supply, which are used to construct extreme precipitation correlation analysis models.

[0032] In this embodiment, a typhoon yearbook dataset is introduced to collect and statistically analyze the characteristic data of 20 typhoons that affected City A from 1952 to 2020, forming a typhoon database, as shown in Table 1 below: Table 1. Statistics on the Impact of Typhoon on City A

[0033] Historical rainfall and water conditions data refer to multi-source observation and recording data of precipitation events in the target area throughout history, including time-by-time measured precipitation, average precipitation, spatial distribution of precipitation, and information on flood processes (such as peak flow and peak time) and disaster information (such as disaster location and disaster description) generated by meteorological stations, which are used for model validation and similarity matching.

[0034] Real-time meteorological data refers to the current and short-term meteorological observation and forecast data of the target area, including meteorological numerical forecast products (such as short-term precipitation forecasts), real-time characteristic indicators of weather systems, and real-time monitoring data of meteorological stations, which are used as inputs for medium- and long-term forecast correction and short-term precipitation correction.

[0035] Specifically, historical typhoon data for the target area is obtained by collecting historical typhoon yearbook datasets, historical rainfall and water level data are obtained by organizing historical hydrological and meteorological observation records and water level reports, and real-time meteorological data is obtained by connecting to the numerical forecast products and real-time monitoring networks of meteorological departments, providing a data foundation for subsequent model building, forecast correction and similarity matching.

[0036] Step S102: Establish an extreme precipitation correlation analysis model based on historical typhoon data, and use the extreme precipitation correlation analysis model to perform medium- and long-term forecast correction to obtain medium- and long-term forecast correction results.

[0037] Among them, the extreme precipitation correlation analysis model refers to a mathematical model constructed based on historical typhoon and other weather system data to describe the quantitative relationship between weather system characteristics and precipitation in a target area. This model extracts multi-dimensional characteristic indicators affecting precipitation, uses statistical analysis to establish the correlation between these indicators and regional precipitation, and verifies the model based on historical rainfall and water level data, thereby achieving a quantitative description and forecast correction of the medium- and long-term trends in extreme precipitation.

[0038] Specifically, based on historical typhoon data, multi-dimensional characteristic indicators affecting precipitation are extracted. Through statistical analysis, a quantitative relationship between these indicators and regional precipitation is established and verified using historical rainfall data. An extreme precipitation correlation analysis model is constructed. This model is then applied to weather system characteristics extracted from current real-time meteorological data to calculate medium- and long-term precipitation forecasts. Error corrections are then performed using historical rainfall data, and finally, medium- and long-term forecast correction results are output.

[0039] Step S103: Based on the medium- and long-term forecast correction results and real-time meteorological data, a pre-constructed precipitation correction network is used to correct the short-term precipitation forecast to obtain the full-cycle precipitation forecast correction results.

[0040] The pre-constructed precipitation correction network refers to a mathematical model trained using deep learning algorithms based on historical precipitation forecast data and historical meteorological station measured data. This model is used to correct the precipitation amount and location in short-term precipitation forecasts in real time. By learning the mapping relationship between historical forecast data and measured data, this network can identify and correct systematic biases in numerical weather prediction products, thus achieving accurate correction of short-term precipitation forecasts.

[0041] Specifically, based on the medium- and long-term forecast correction results and the short-term precipitation forecast data in real-time meteorological data, the precipitation amount and location of the short-term forecast are corrected in real time through a pre-constructed precipitation correction network. The corrected short-term forecast results are output and then time-series linked with the medium- and long-term forecast correction results to finally form a full-cycle precipitation forecast correction result from long-term to short-term.

[0042] Step S104: Perform similarity matching between the full-cycle precipitation forecast correction results and historical rainfall data to select the most similar historical precipitation events.

[0043] Specifically, based on the full-cycle precipitation forecast correction results, three algorithms—Nash coefficient, Pearson correlation coefficient, and root mean square error—are used to calculate the similarity between the precipitation data and each precipitation event in historical rainfall data from three dimensions: surface average precipitation, spatial distribution, and overall error. The similarity of the three dimensions is then weighted and sorted to finally select the historical precipitation events most similar to the current precipitation scenario.

[0044] Step S105: Based on the most similar historical precipitation events, retrieve the corresponding historical flood information and disaster information to obtain the flood disaster assessment results.

[0045] Specifically, the most similar historical precipitation events are used as index identifiers to retrieve corresponding flood process information from the historical rainfall and water conditions database. At the same time, disaster descriptions and disaster location data are retrieved from the historical disaster database. The flood process information and disaster information are linked and integrated to generate flood disaster assessment results that include key flood characteristics and potential affected areas.

[0046] This embodiment provides a flood disaster assessment method based on graded correction of precipitation forecasts. It collects and statistically analyzes historical typhoon, rainfall, and disaster data using measured data from three lines of defense, forming a multi-source historical rainfall database. Utilizing a typhoon-rainfall correlation analysis model and a cloud-to-ground rainfall correction network, it performs full-cycle tracking correction of medium- and long-term and short-term precipitation forecasts. Based on the historical rainfall database and precipitation correction results, it employs historical similarity scene matching technology to select the most similar historical precipitation, displaying the basin flood information (peak flow, peak time, reservoir inflow, etc.), flash flood information (peak flow, flood process line, etc.), and disaster information (affected locations, disaster descriptions, etc.) for that precipitation event. The historical similarity scene matching technology simultaneously supports similarity matching using meteorological consultation result images as input. This research, through big data technology, achieves full-process coupling of cloud-to-ground rainfall and river water, simultaneously displaying disaster information and providing scientific reference for flood control scheduling decisions.

[0047] In an optional implementation, step S102 above, which involves establishing an extreme precipitation correlation analysis model based on historical typhoon data and using the extreme precipitation correlation analysis model to perform medium- and long-term forecast corrections to obtain medium- and long-term forecast correction results, includes the following specific steps: Step S1021: Extract multi-dimensional typhoon characteristic indicators that affect precipitation based on historical typhoon data; establish a quantitative relationship between the multi-dimensional typhoon characteristic indicators and precipitation in the target area using statistical analysis; verify the quantitative relationship based on historical rainfall data to obtain an extreme precipitation correlation analysis model.

[0048] The extreme precipitation correlation analysis model in this embodiment adopts the typhoon-rainfall correlation analysis model.

[0049] For example, to address the shortcomings in long-term forecasting capabilities for extreme rainstorms, a typhoon yearbook dataset was introduced to collect and statistically analyze the characteristic data of 20 typhoons that affected City A from 1952 to 2020, forming a typhoon database. Various statistical methods were used to clarify the changing trends of typhoons affecting City A. Qualitative analysis of the characteristics of typhoons affecting City A was conducted by statistically analyzing their formation origin, movement path, and landfall location. Simultaneously, satellite data was used to retrieve atmospheric precipitable water, and 24 numerical indicators were constructed from five aspects: typhoon intensity, movement speed, path, moisture supply, and distance from City A, as shown in Table 2 below. Table 2. Numerical Indicators of Typhoon Characteristics

[0050] Quantitative relationships were established between typhoon characteristic numerical indicators and the total rainfall and maximum daily rainfall intensity of typhoon in City A, as shown in Equations (1) and (2) below. Based on this, a typhoon-rainfall correlation analysis model was constructed, thereby enabling qualitative judgment of extreme precipitation such as typhoons seven days in advance and quantitative prediction three days in advance.

[0051] Average total precipitation (major precipitation process) = 12823 - 0.006 Typhoon duration - 0.54 Moving speed (km / h) + 0.44 Maximum wind speed (m / s) + 300 Rate of wind weakening (m / s / h) + 1.74 Minimum air pressure - 154 Rate of air pressure rise + 1.229 Duration after first landfall - 1.9 Wind speed at landfall + 0.7 Air pressure at landfall - 13.0 Wind speed at first landfall - 13.24 Air pressure at first landfall + 0.217 Distance from the center of City A / km + 11.8 Wind speed at the closest point to City A - 2.08 Air pressure at the closest point to City A (1); Maximum rainfall intensity = 11394 - 0.158 Typhoon duration - 0.88 Moving speed (km / h) - 0.10 Maximum wind speed (m / s) - 87 Rate of wind weakening (m / s / h) + 1.26 Minimum air pressure - 12 Rate of air pressure increase + 0.892 Duration after first landfall - 7.54 Wind speed at landfall - 4.65 Air pressure at landfall - 1.60 Wind speed at first landfall - 4.17 Air pressure at first landfall + 0.0624 Closest distance to city center A / km + 2.83 Wind speed at closest point to city A - 3.74 Air pressure at closest point to city A (2).

[0052] Step S1022: Extract multi-dimensional characteristic indicators of the current weather system from real-time meteorological data; input the multi-dimensional characteristic indicators of the current weather system into the extreme precipitation correlation analysis model to calculate the medium- and long-term forecast of regional precipitation; perform error verification and correction on the medium- and long-term forecast based on historical rainfall and water conditions data to obtain the medium- and long-term forecast correction results.

[0053] For example, based on the actual situation of "23.7", the average total precipitation and the maximum daily rainfall of the extreme rainstorm were quantitatively calculated by inputting the numerical indicators of typhoon characteristics. The actual test results of the extreme precipitation correlation analysis model are shown in Table 3 below: Table 3 Actual test results of “23.7”

[0054] Based on the above results, the relative error between the simulated value and the measured value is about 24%, and the absolute error is 79 mm. The results show that the model has a good application effect, as shown in Figure 3(a) and Figure 3(b).

[0055] In one alternative implementation, the pre-built precipitation correction network is implemented in the following manner: Historical precipitation forecast data and historical meteorological station measured data at the same time stamp are obtained to construct a training dataset. Based on the training dataset, a deep learning algorithm is used to train the mapping relationship between historical precipitation forecast data and historical meteorological station measured data to obtain an initial correction network. The correction accuracy of the initial correction network is tested and the parameters are optimized using a validation dataset to obtain a precipitation correction network.

[0056] In an optional implementation, step S103, based on the medium- and long-term forecast correction results and real-time meteorological data, uses a pre-constructed precipitation correction network to correct the short-term precipitation forecast, obtaining the full-cycle precipitation forecast correction results, and includes the following steps: Step S1031: Extract the short-term precipitation forecast data for the current moment from the real-time meteorological data; input the medium- and long-term forecast correction results and the short-term precipitation forecast data into the precipitation correction network, perform real-time correction on the precipitation amount and location of the short-term precipitation forecast data, and output the corrected short-term precipitation forecast results.

[0057] Specifically, short-term precipitation numerical forecast products for the current moment are obtained from real-time meteorological data, and their precipitation amount and location information are extracted as the input basis for subsequent corrections.

[0058] Step S1032: The medium- and long-term forecast correction results are time-series linked with the corrected short-term precipitation forecast results to obtain the full-cycle precipitation forecast correction results from long-term to short-term.

[0059] For example, the pre-constructed precipitation correction network adopts a cloud-to-ground precipitation correction network structure, such as... Figure 4As shown, the correction network employs an encoder-decoder architecture. The encoder extracts features and performs dimensionality reduction encoding on the input historical precipitation forecast data. The neural network layer learns the nonlinear mapping relationship between the forecast and observed data. The decoder reconstructs the corrected precipitation forecast from the encoded features. This network structure, trained through end-to-end deep learning, achieves accurate correction of short-term precipitation forecasts.

[0060] To address the issue of insufficient accuracy in short-term precipitation forecasts and their location, hourly precipitation forecasts and measured data from meteorological stations during the flood season from 2017 to 2021 were collected and compiled. A deep learning algorithm was used to construct a correction network of over 1300 cloud-to-ground precipitation lines. During network training, an encoding / decoding algorithm was employed to process the network input, thereby avoiding the significant skewness in precipitation distribution and the impact of different orders of magnitude on network learning efficiency and accuracy. Ultimately, real-time correction of cloud-to-ground precipitation was achieved across multiple time scales. Taking areas C and D of the "23.7" basin-wide catastrophic flood in River B of City A as examples, the RMSE (Root Mean Square Error) decreased by approximately 53% before and after correction, effectively correcting the forecasted precipitation.

[0061] In an optional implementation, step S104, which involves performing a similarity match between the full-cycle precipitation forecast correction results and historical rainfall data to select the most similar historical precipitation events, includes the following steps: Step S1041: Based on historical rainfall data and full-cycle precipitation forecast correction results, multiple similarity algorithms are used to calculate the similarity between the current precipitation and historical precipitation events in multiple dimensions, including surface average precipitation, spatial distribution, and overall error. The multiple similarity algorithms include Nash coefficient, Pearson correlation coefficient, and root mean square error.

[0062] Step S1042: Perform a weighted comprehensive ranking of the multi-dimensional similarity, and select the historical precipitation events most similar to the current precipitation based on the ranking results.

[0063] For example, the medium- and long-term forecasts of extreme precipitation such as typhoons are obtained from the typhoon-rainfall correlation analysis model, and the short-term precipitation forecast correction results are obtained from the cloud-to-ground precipitation correction network, thus obtaining the full-cycle precipitation forecast correction results. Historical rainfall and water conditions databases are formed by collecting and organizing hydrological briefing data from 2001 to 2024 and 5-minute measured precipitation data from meteorological stations. Based on the ground precipitation correction results, real-time similarity calculation and matching with historical precipitation are achieved. The similarity characteristics of precipitation are comprehensively considered from three perspectives: average precipitation, spatial distribution, and overall evaluation, to obtain a comprehensive similarity ranking, and the most similar historical precipitation events are selected.

[0064] The similarity between historical precipitation and measured corrected precipitation is calculated using the Nash coefficient (NSE). The similarity results tend to minimize the mean precipitation error, i.e., the surface average precipitation is optimal, which is used for long-term precipitation forecast coupling. The spatial distribution similarity between historical precipitation and measured corrected precipitation is calculated using the Pearson correlation coefficient. The similarity results tend towards the optimal spatial distribution, which can be used for coupling medium-term precipitation forecasts. The root mean square error (RMSE) is used to calculate the overall similarity between historical precipitation and measured corrected precipitation. The similarity results tend to minimize the overall error of all meteorological stations, i.e., the overall optimal result, which is used for short-term precipitation forecast coupling.

[0065] It should be noted that the historical similarity precipitation matching technology also supports similarity matching using meteorological consultation result images provided by meteorological departments as input. Utilizing the Java.awt package and OCR text recognition algorithms, through steps such as pixel assignment, text localization, and raster extraction, geographic information is imbued into the meteorological consultation result images, achieving fully automated digitization of meteorological consultation results within milliseconds with an accuracy rate of up to 95%, providing strong support for rapid assessment of flood disasters.

[0066] In an optional implementation, step S105 above, which retrieves corresponding historical flood information and disaster information based on the most similar historical precipitation events to obtain flood disaster assessment results, specifically includes the following steps: Step S1051: Using the most similar historical precipitation events as index identifiers, retrieve the corresponding flood process information from the historical rainfall and water conditions database, and retrieve the corresponding disaster information from the historical disaster database. The disaster information includes disaster descriptions and data on the locations of affected points.

[0067] Specifically, the most similar historical precipitation events selected are used as unique index identifiers to accurately match and retrieve the flood process information of the corresponding events from the historical rainfall and water situation database. At the same time, the disaster descriptions and disaster location data associated with the events are retrieved from the historical disaster situation database, so as to realize the rapid association and extraction of historical rainfall, water and disaster information.

[0068] Step S1052: The flood process information and disaster information are linked and integrated to generate flood disaster assessment results that include key flood characteristics and potential affected areas.

[0069] Specifically, the retrieved flood process information and disaster information are linked and integrated, and flood disaster assessment results are generated through data fusion, including key flood characteristics such as peak flow and peak time, as well as potential affected areas determined based on historical disaster locations.

[0070] The flood disaster assessment method based on graded correction of precipitation forecasts provided in this embodiment addresses the urgent needs of the current flood control system, achieving accurate full-cycle precipitation forecast correction and minute-level rapid assessment of flood disasters. The research results extend the lead time for typhoons and heavy rains, allowing more time for flood control deployment; through the cloud-to-landfall precipitation correction network, the accuracy of short-term meteorological precipitation forecasts is significantly improved, providing a foundation for accurate forecasting and assessment of basin floods, flash floods, and urban waterlogging; utilizing historical similarity scenario matching technology, and fully referencing historical disaster situations, the impact of disaster chains is considered using historical data, achieving minute-level assessment of rainfall, water levels, and disaster conditions. This provides new methods and ideas for rapid, accurate, and effective flood disaster prevention, and simultaneously achieves millisecond-level fully automated digitization of meteorological consultation results, providing a solid foundation for intelligent and unmanned assessment of flash flood risks in regional two-tier water resources systems.

[0071] As one or more specific application embodiments of the present invention, combined with Figures 2 to 4 The flood disaster assessment method based on precipitation forecast classification correction provided by this invention will be further described in detail, such as... Figure 2 As shown, the specific process is as follows: Step S1: Correct the medium- and long-term forecasts of extreme precipitation based on the typhoon-rainfall correlation analysis model.

[0072] The specific steps for correcting medium- and long-term extreme precipitation forecasts based on the typhoon-rainfall correlation analysis model include the following: Step S11: Construct a typhoon database for the study area.

[0073] In this embodiment, a typhoon yearbook dataset is introduced to collect and statistically analyze the characteristic data of 20 typhoons that affected City A from 1952 to 2020, forming a typhoon database, as shown in Table 1 above.

[0074] Step S12: Qualitatively assess extreme precipitation events such as typhoons in the study area.

[0075] In this embodiment, the characteristics of typhoons affecting City A are qualitatively analyzed by statistically analyzing their formation origin, movement path, and landfall location. Regarding the formation origin, typhoons affecting City A mainly formed between July 11th and August 10th, coinciding with City A's main flood season, accounting for 80% of the total. In terms of movement path, the main paths of typhoons affecting City A were northwestward and turning, accounting for 50% and 40% respectively. Regarding landfall location, the vast majority of typhoons affecting City A made landfall in Cities E and F, each accounting for 23% of the total number of typhoons affecting City A. At 12:00 on July 21, 2023, a certain typhoon formed in the Pacific Ocean. According to the forecast results, this typhoon matches the above three aspects and is highly likely to affect precipitation in City A. This allows for a qualitative assessment of the possibility of this typhoon affecting City A's precipitation seven days in advance.

[0076] Step S13: Quantitatively predict extreme precipitation such as typhoons in the study area.

[0077] In this embodiment, satellite data was used to retrieve atmospheric precipitable water. Twenty-four numerical indicators were constructed from five aspects: typhoon intensity, speed, path, water vapor supply, and distance from City A, as shown in Table 2 above. A stepwise regression method was used to establish quantitative relationships between the typhoon characteristic numerical indicators and the total typhoon rainfall and maximum daily rainfall intensity in the city, as shown in equations (1) and (2) above. On July 26, 2023, after a certain typhoon made landfall, based on the typhoon forecast information released by the Central Meteorological Observatory, 24 numerical indicators for the typhoon were simultaneously constructed and input into equations (1) and (2) to calculate the estimated total rainfall and maximum daily rainfall intensity in City A. The estimated results were close to the actual measured values ​​of "23·7", with relative errors of 23.87% and 5.67%, respectively. The actual verification results are shown in Table 3 above. This allows for a quantitative prediction of the impact of a certain typhoon on the precipitation in City A three days in advance.

[0078] Step S2: Correct the precipitation amount and landing area of ​​the short-term precipitation forecast based on the cloud rain-landing rain correction network.

[0079] The specific steps for correcting the precipitation amount and location in short-term precipitation forecasts based on the cloud-to-ground precipitation correction network include the following: Step S21: Construct a cloud-to-ground rainfall database for the study area.

[0080] In this embodiment, hourly precipitation forecast products from Ruisi and meteorological stations for the flood season in City A from 2017 to 2021 were collected and organized. Precipitation forecast products and measured data at the same time stamp were matched to form a database consisting of multiple cloud rain-ground rain data pairs.

[0081] Step S22: Construct the cloud rain-landing rain correction network.

[0082] In this embodiment, a cloud-to-ground rainfall correction network of over 1300 lines was constructed using deep learning algorithms. The network structure is as follows: Figure 4 As shown in Table 4, for the "23.7" basin-wide catastrophic flood in Areas C and D of River B in City A, the RMSE of short-term precipitation forecasts decreased by approximately 53% before and after correction, the mean absolute deviation decreased by 2.38, and the Pearson correlation coefficient increased by 0.13. The comparison of evaluation indicators before and after correction is shown in Table 4 below.

[0083] Table 4 Comparison of Evaluation Indicators Before and After Correction

[0084] Step S3: Real-time matching of forecast precipitation with historical precipitation using historical similarity scene matching technology.

[0085] The real-time matching of forecast precipitation with historical precipitation based on historical similarity scene matching technology specifically includes the following sub-steps: Step S31: Construct a historical rainfall and water situation database for the study area.

[0086] In this embodiment, water situation briefing data for City A from 2001 to 2024 and 5-minute measured precipitation data from meteorological stations were collected and organized to construct a historical rainfall and water situation database. Characteristic information such as rainfall, rainfall intensity, peak flow, and peak time of 140 floods were statistically analyzed, with a focus on verifying 116 floods during the flood season. For the flood season from 2017 to 2021, 5-minute measured data from 445 meteorological stations were further collected and organized, replacing the same events (44 events in total) in the historical rainfall and water situation database. Each precipitation event was numbered and used as a unique index identifier for the historical precipitation database.

[0087] Step S32: Real-time similarity matching between forecast precipitation and historical precipitation.

[0088] In this embodiment, step S201 obtains the medium- and long-term forecast results of extreme precipitation in City A, and step S202 obtains the short-term precipitation forecast correction results for City A, i.e., the full-cycle precipitation forecast correction results for City A. Based on the full-cycle precipitation correction results and the historical rainfall database constructed in step S2031, the real-time similarity calculation and matching between the groundfall precipitation correction results and historical precipitation are realized from three perspectives: average precipitation, spatial distribution, and overall evaluation.

[0089] The similarity between historical precipitation and measured corrected precipitation is calculated using the Nash coefficient (NSE). The similarity results tend to minimize the mean precipitation error, i.e., the surface average precipitation is optimal, which is used for long-term precipitation forecast coupling. The spatial distribution similarity between historical precipitation and measured corrected precipitation is calculated using the Pearson correlation coefficient. The similarity results tend towards the optimal spatial distribution, which can be used for coupling medium-term precipitation forecasts. The root mean square error (RMSE) is used to calculate the overall similarity between historical precipitation and measured corrected precipitation. The similarity results tend to minimize the overall error of all meteorological stations, i.e., the overall optimal result, which is used for short-term precipitation forecast coupling.

[0090] Step S33: Use the meteorological consultation result image as input to perform similarity calculation.

[0091] In this embodiment, the Java.awt package and OCR (Optical Character Recognition) text recognition algorithm are used to assign geographic information to meteorological consultation result images through steps such as pixel assignment, text positioning, and raster extraction. This achieves fully automatic digitization of meteorological consultation results in milliseconds with an accuracy rate of up to 95%, providing strong support for rapid assessment of flood disasters.

[0092] This embodiment provides a flood disaster assessment method based on graded correction of precipitation forecasts. It conducts trend changes and qualitative and quantitative analyses of five major precipitation influencing factors: typhoon intensity, path, and moisture supply. Based on the proposed spatiotemporal dual-scale typhoon rain quantitative identification method, regression equations for total typhoon rainfall and maximum daily rainfall in City A are coupled and established to construct a typhoon-rainfall correlation analysis model. The model can achieve qualitative assessment seven days in advance and quantitative prediction three days in advance; taking "23.7" as an example, the relative error is 23.87%. Utilizing encoding / decoding technology and deep learning algorithms, real-time correction of cloud-to-ground rainfall at multiple time scales is achieved, effectively improving the accuracy of the Ruisi precipitation numerical forecast product. Taking areas C and D of "23.7" as examples, the RMSE is reduced by approximately 53% after correction. Rainfall, water, and disaster data from 140 precipitation events between 2001 and 2024 have been collected and organized to form a historical rainfall and water situation database. It achieves real-time similarity matching between forecasted precipitation and historical precipitation, and simultaneously supports similarity calculation using meteorological consultation result images as input.

[0093] This embodiment also provides a flood disaster assessment device based on precipitation forecast grading correction. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" refers to a combination of software and / or hardware that can perform a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0094] This embodiment provides a flood disaster assessment device based on precipitation forecast classification correction, such as... Figure 5 As shown, it includes: The data acquisition module 501 is used to acquire historical typhoon data, historical rainfall data, and real-time meteorological data for the target area.

[0095] The medium- and long-term forecast module 502 is used to establish an extreme precipitation correlation analysis model based on historical typhoon data, and to use the extreme precipitation correlation analysis model to perform medium- and long-term forecast correction to obtain medium- and long-term forecast correction results.

[0096] The short-term correction module 503 is used to correct short-term precipitation forecasts based on medium- and long-term forecast correction results and real-time meteorological data, using a pre-constructed precipitation correction network to obtain full-cycle precipitation forecast correction results.

[0097] The similarity matching module 504 is used to perform similarity matching between the full-cycle precipitation forecast correction results and historical rainfall data, and to select the most similar historical precipitation events.

[0098] The analysis and output module 505 is used to retrieve corresponding historical flood information and disaster information based on the most similar historical precipitation events to obtain flood disaster analysis results.

[0099] In some optional implementations, the medium- and long-term forecast module 502 includes: The extreme precipitation correlation analysis model construction unit is used to extract multi-dimensional typhoon characteristic indicators that affect precipitation based on historical typhoon data; establish a quantitative relationship between the multi-dimensional typhoon characteristic indicators and the precipitation in the target area using statistical analysis methods; and verify the quantitative relationship based on historical rainfall and water conditions to obtain the extreme precipitation correlation analysis model.

[0100] In some optional implementations, the medium- and long-term forecast module 502 further includes: The medium- and long-term forecast correction unit is used to extract multi-dimensional characteristic indicators of the current weather system from real-time meteorological data; input the multi-dimensional characteristic indicators of the current weather system into the extreme precipitation correlation analysis model to calculate the medium- and long-term forecast of regional precipitation; and perform error verification and correction on the medium- and long-term forecast based on historical rainfall and water conditions data to obtain the medium- and long-term forecast correction results.

[0101] In some alternative implementations, the pre-built precipitation correction network is implemented in the following manner: Obtain historical precipitation forecast data and historical meteorological station measured data at the same timestamp to construct a training dataset; Based on the training dataset, a deep learning algorithm was used to train the mapping relationship between historical precipitation forecast data and historical meteorological station measured data to obtain an initial correction network. The accuracy of the initial correction network was tested and the parameters were optimized using the validation dataset to obtain the precipitation correction network.

[0102] In some alternative implementations, the short-term correction module 503 includes: The short-term precipitation forecast unit is used to extract short-term precipitation forecast data at the current moment from real-time meteorological data; it takes the medium- and long-term forecast correction results and the short-term precipitation forecast data as inputs to the precipitation correction network, performs real-time correction on the precipitation amount and location of the short-term precipitation forecast data, and outputs the corrected short-term precipitation forecast results.

[0103] The time-series connection unit is used to connect the medium- and long-term forecast correction results with the corrected short-term precipitation forecast results in a time series, so as to obtain the full-cycle precipitation forecast correction results from long-term to short-term.

[0104] In some alternative implementations, the similarity matching module 504 includes: The similarity calculation unit is used to calculate the similarity between current precipitation and historical precipitation events in multiple dimensions, including surface average precipitation, spatial distribution, and overall error, based on historical rainfall data and full-cycle precipitation forecast correction results. The similarity algorithms include Nash coefficient, Pearson correlation coefficient, and root mean square error.

[0105] The filtering unit is used to perform weighted comprehensive ranking of multi-dimensional similarity and filter out the historical precipitation events most similar to the current precipitation based on the ranking results.

[0106] In some optional implementations, the analysis output module 505 includes: The information retrieval unit is used to retrieve corresponding flood process information from the historical rainfall and water conditions database and corresponding disaster information from the historical disaster database, using the most similar historical precipitation events as index identifiers. The disaster information includes disaster descriptions and data on the locations of affected points.

[0107] The flood disaster assessment result generation unit is used to link and integrate flood process information with disaster information to generate flood disaster assessment results that include key flood characteristics and potential affected areas.

[0108] The flood disaster assessment device based on precipitation forecast classification correction provided in this embodiment of the invention can execute the flood disaster assessment method based on precipitation forecast classification correction provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0109] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0110] The following is a detailed reference. Figure 6This diagram illustrates a suitable structural design for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 601, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 602 or a program loaded from memory 608 into random access memory (RAM) 603. RAM 603 also stores various programs and data required for the operation of the electronic device. The processor 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0111] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 608 including, for example, magnetic tapes, hard disks, etc.; and communication devices 609. Communication device 609 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0112] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 609, or installed from a memory 608, or installed from a ROM 602. When the computer program is executed by the processor 601, it performs the functions defined in the flood disaster assessment method based on precipitation forecast grading correction of the embodiments of the present invention.

[0113] Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0114] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the flood disaster assessment method based on precipitation forecast grading correction shown in the above embodiments is implemented.

[0115] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0116] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A flood disaster assessment method based on graded correction of precipitation forecasts, characterized in that, The method includes: Acquire historical typhoon data, historical rainfall data, and real-time meteorological data for the target area; An extreme precipitation correlation analysis model was established based on the historical typhoon data, and the extreme precipitation correlation analysis model was used to perform medium- and long-term forecast correction to obtain medium- and long-term forecast correction results. Based on the medium- and long-term forecast correction results and real-time meteorological data, a pre-constructed precipitation correction network is used to correct the short-term precipitation forecast, resulting in a full-cycle precipitation forecast correction result. The full-cycle precipitation forecast correction results are matched with the historical rainfall data to select the most similar historical precipitation events. Based on the most similar historical precipitation events, corresponding historical flood information and disaster information are retrieved to obtain flood disaster assessment results.

2. The method according to claim 1, characterized in that, An extreme precipitation correlation analysis model was established based on the aforementioned historical typhoon data, including: Based on the historical typhoon data, multi-dimensional characteristic indicators of typhoons affecting precipitation are extracted. A statistical analysis method was used to establish a quantitative relationship between multi-dimensional typhoon characteristic indicators and precipitation in the target area; The quantitative relationship was verified based on the historical rainfall and water conditions to obtain an extreme precipitation correlation analysis model.

3. The method according to claim 1, characterized in that, The extreme precipitation correlation analysis model is used to correct the medium- and long-term forecasts of extreme precipitation, and the corrected medium- and long-term forecasts are obtained, including: Extract multi-dimensional feature indicators of the current weather system from the real-time meteorological data; The multi-dimensional characteristic indicators of the current weather system are input into the extreme precipitation correlation analysis model to calculate the medium- and long-term forecast of regional precipitation. Based on the historical rainfall and water level data, the medium- and long-term forecast values ​​are checked and corrected to obtain the medium- and long-term forecast correction results.

4. The method according to claim 1, characterized in that, The pre-built precipitation correction network is implemented in the following way: Obtain historical precipitation forecast data and historical meteorological station measured data at the same timestamp to construct a training dataset; Based on the training dataset, a deep learning algorithm is used to train the mapping relationship between historical precipitation forecast data and historical meteorological station measured data to obtain an initial correction network. The accuracy of the initial correction network was tested and the parameters were optimized using the validation dataset to obtain the precipitation correction network.

5. The method according to claim 1, characterized in that, Based on the aforementioned medium- and long-term forecast correction results and real-time meteorological data, a pre-constructed precipitation correction network is used to correct the short-term precipitation forecast, resulting in full-cycle precipitation forecast correction results, including: Extract short-term precipitation forecast data for the current moment from the real-time meteorological data; The medium- and long-term forecast correction results and the short-term precipitation forecast data are input into the precipitation correction network to perform real-time correction on the precipitation amount and location of the short-term precipitation forecast data, and output the corrected short-term precipitation forecast results. The medium- and long-term forecast correction results are combined with the corrected short-term precipitation forecast results to obtain full-cycle precipitation forecast correction results from long-term to short-term.

6. The method according to claim 1, characterized in that, The full-cycle precipitation forecast correction results are matched with the historical rainfall data to identify the most similar historical precipitation events, including: Based on the historical rainfall data and the full-cycle precipitation forecast correction results, multiple similarity algorithms are used to calculate the similarity between the current precipitation and historical precipitation events in multiple dimensions, including surface average precipitation, spatial distribution, and overall error. These multiple similarity algorithms include Nash coefficient, Pearson correlation coefficient, and root mean square error. The similarity across multiple dimensions is weighted and ranked, and the historical precipitation events most similar to the current precipitation are selected based on the ranking results.

7. The method according to claim 1, characterized in that, Based on the most similar historical precipitation events, corresponding historical flood information and disaster information are retrieved to obtain flood disaster assessment results, including: Using the most similar historical precipitation events as index identifiers, the corresponding flood process information is retrieved from the historical rainfall and water conditions database, and the corresponding disaster information is retrieved from the historical disaster information database, including disaster descriptions and data on the locations of affected points; The flood process information and disaster information are correlated and integrated to generate flood disaster assessment results that include key flood characteristics and potential affected areas.

8. A flood disaster assessment device based on precipitation forecast classification correction, characterized in that, The device includes: The data acquisition module is used to acquire historical typhoon data, historical rainfall data, and real-time meteorological data for the target area. The medium- and long-term forecast module is used to establish an extreme precipitation correlation analysis model based on the historical typhoon data, and to use the extreme precipitation correlation analysis model to perform medium- and long-term forecast correction to obtain the medium- and long-term forecast correction results. The short-term correction module is used to correct the short-term precipitation forecast based on the medium- and long-term forecast correction results and real-time meteorological data, using a pre-constructed precipitation correction network to obtain the full-cycle precipitation forecast correction results. The similarity matching module is used to perform similarity matching between the full-cycle precipitation forecast correction results and the historical rainfall data, and to filter out the most similar historical precipitation events. The analysis and output module is used to retrieve corresponding historical flood information and disaster information based on the most similar historical precipitation events to obtain flood disaster analysis results.

9. An electronic device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the flood disaster assessment method based on precipitation forecast grading correction as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the flood disaster assessment method based on precipitation forecast grading correction as described in any one of claims 1 to 7.

Images