Rainfall prediction method and system for Beijing, Tianjin and Hebei regions

Abstract

The invention discloses a rainfall prediction method and system for the Beijing, Tianjin and Hebei region, and relates to the technical field of rainfall prediction, and the method comprises the steps: obtaining multi-source data, carrying out the preprocessing of the data, extracting the core characteristics of the data, carrying out the dynamic rainfall modeling according to the core characteristics of the data, and simulating a rainfall evolution path; and constructing a rainfall prediction model, performing rainfall intensity prediction based on the rainfall evolution path, performing extreme weather early warning based on a prediction result, and generating a rainfall distribution diagram of a high-risk area. Through fusion of multi-source data, the generative adversarial network and spectral analysis, the spatial resolution of the data and the rainfall evolution prediction capability are improved, the real-time performance and the accuracy of rainfall prediction are improved in combination with dynamic rainfall modeling, a fractional order differential equation and a multiple regression model, and the prediction accuracy is improved. The rainfall prediction challenge under complex weather conditions can be effectively dealt with, and accurate prediction of rainfall intensity and extreme weather events is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of rainfall prediction, and in particular to a rainfall prediction method and system for the Beijing-Tianjin-Hebei region. Background Art

[0002] In recent years, with the frequent occurrence of climate change and extreme weather events, how to accurately predict and respond to rainfall and the secondary disasters caused by it has become an important topic in meteorological research and application. Traditional rainfall prediction methods mainly rely on observation data from meteorological stations and combine numerical weather forecast models to predict precipitation. However, when faced with complex terrain and changing climate conditions, traditional methods often have problems such as large prediction errors, poor real-time performance and low spatial resolution. In order to improve the accuracy and timeliness of rainfall prediction, in recent years, the application of multi-source data fusion and machine learning methods has gradually become an important direction for improving rainfall prediction accuracy. The diversification and real-time update of meteorological data provide more abundant input conditions for prediction models. At the same time, rainfall prediction methods based on big data analysis and artificial intelligence technology have also been gradually studied and applied. Despite this, existing rainfall prediction technologies still face several challenges, especially in large-scale regional applications. Existing technologies rely on traditional weather forecast models, and it is often difficult to make effective improvements in spatial resolution, prediction accuracy and real-time performance. The spatial coverage of meteorological station data is limited, and there is a large gap in spatial resolution between GNSS data and global meteorological data such as ERA5, which cannot provide sufficiently detailed prediction information. In addition, the dynamic evolution path of rainfall and its complexity often lead to inaccuracies in the traditional methods in rainfall intensity prediction and regional rainfall prediction. Especially in complex climate regions such as the Beijing-Tianjin-Hebei region, local precipitation and heavy rainfall events are difficult to be effectively warned and accurately predicted, which affects disaster prevention and emergency response. Summary of the invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides a rainfall prediction method and system for the Beijing-Tianjin-Hebei region to solve the problem that local precipitation and heavy rainfall events are difficult to be effectively warned and accurately predicted, thus affecting disaster prevention and emergency response, especially in complex climate regions such as the Beijing-Tianjin-Hebei region.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In a first aspect, the present invention provides a method for predicting rainfall in the Beijing-Tianjin-Hebei region, comprising:

[0007] Acquire multi-source data, pre-process the data, extract the core characteristics of the data, conduct dynamic rainfall modeling based on the core characteristics of the data, and simulate the rainfall evolution path;

[0008] Build a rainfall prediction model, predict rainfall intensity based on the rainfall evolution path, issue extreme weather warnings based on the prediction results, and generate rainfall distribution maps for high-risk areas;

[0009] Real-time rainfall forecast and warning information are visualized and stored in the database.

[0010] As a preferred solution of the rainfall prediction method for the Beijing-Tianjin-Hebei region described in the present invention, wherein: the acquisition of multi-source data and extraction of data core characteristics after preprocessing the data refers to collecting meteorological station data, GNSS data and ERA5 data, cleaning and standardizing the collected data, and using a spatial interpolation method to spatially align sparse observation data;

[0011] Use a generative adversarial network algorithm to increase the spatial resolution of ERA5 data to the kilometer level;

[0012] The spectrum analysis method is used to extract the time series of atmospheric precipitable water data in GNSS data. The time domain signal is converted into a frequency domain signal using fast Fourier transform, the frequency component is calculated, the amplitude and phase information of the frequency component is obtained, and the frequency point with an amplitude greater than the average amplitude is taken as the core periodic feature. The data change rate is calculated based on the extracted atmospheric precipitable water data.

[0013] As a preferred solution of the rainfall prediction method for the Beijing-Tianjin-Hebei region described in the present invention, wherein: the dynamic rainfall modeling is performed according to the core characteristics of the data, and the rainfall evolution path simulation refers to connecting the ERA5 data interface, selecting the Beijing-Tianjin-Hebei region and time range, extracting humidity, temperature and wind speed component data, integrating the temperature and humidity real-time observation data of multiple meteorological stations in the region, and storing the data as a CSV file with a unified format, each record containing a timestamp, longitude and latitude, temperature and humidity;

[0014] We traverse the collected data, mark missing values, fill in missing values ​​using time series linear interpolation, time-align the ERA5 data with the weather station data, and normalize the humidity, temperature, and wind speed components;

[0015] Get the humidity and temperature at the current time point as initial conditions, set the initial rainfall to zero and set the time step;

[0016] Fitting the evaporation constant from historical weather station data by nonlinear fitting method and the condensation constant , calculate the evaporation rate using the initial conditions at the first time step and condensation efficiency ;

[0017] The calculated results are used as the starting point of the next time step to iterate the evaporation rate and condensation efficiency to obtain the final evaporation rate. and condensation efficiency ;

[0018] Based on the calculated evaporation rate and condensation efficiency, a fractional differential equation is established to describe the change in rainfall:

[0019] ;

[0020] In the formula, is the fractional rate of change of rainfall, is the fractional order, which is set by experimental tuning;

[0021] Integrate the humidity, temperature, evaporation rate, condensation efficiency and rainfall at time t into the state variable matrix X(t);

[0022] Define the modulation function and use the modulation function method to fit the model parameters:

[0023] ;

[0024] In the formula, are the optimized model parameters, is the composite modulation function, is the state variable matrix, is the fractional derivative of the modulation function;

[0025] Initialize the basis functions of the Chebyshev polynomial, project the dynamic equations into the Chebyshev basis function space, and calculate the polynomial coefficients ;

[0026] The fractional differential equation is numerically solved using the Chebyshev polynomial method to obtain the rainfall time series solution. ;

[0027] Set the rainfall threshold Q, traverse the rainfall time series, if the current rainfall is greater than the threshold Q, mark the current time as the rainfall start time, if the current rainfall is less than or equal to the threshold Q and continues for n hours to meet the conditions, mark the current time as the rainfall stop time, get the rainfall time window according to the rainfall start time and stop time and record the initial coordinates of the current rainfall event;

[0028] Update the coordinates of the rainfall event based on the wind speed components and the initial coordinates of the rainfall event provided by ERA5:

[0029] ,

[0030] ;

[0031] In the formula, and are the coordinates of the current rainfall area, and are the coordinates of the rainfall area at the next time step, and is the wind speed component, is the time step;

[0032] The rainfall coordinates at each time step are recorded to form a dynamic rainfall path.

[0033] As a preferred solution of the rainfall prediction method for the Beijing-Tianjin-Hebei region described in the present invention, wherein: the construction of a rainfall prediction model and the prediction of rainfall intensity based on the rainfall evolution path refer to constructing a rainfall prediction model through a multivariate linear regression model, using state variable matrix data as training data to input into the multivariate linear regression model for iterative training, defining a loss function and an Adam optimizer to iteratively optimize model parameters, and when the loss of the multivariate linear regression model no longer decreases significantly during the continuous iteration process, the iteration is stopped to output the model parameters and update the multivariate linear regression model;

[0034] The real-time data within the coverage of the rainfall evolution path is input into the multivariate linear regression model to obtain the regional rainfall intensity forecast value. According to the rainfall time window range, the rainfall intensity forecast value is integrated and calculated to obtain the regional rainfall forecast value.

[0035] As a preferred solution of the rainfall prediction method for the Beijing-Tianjin-Hebei region described in the present invention, wherein: the extreme weather warning based on the prediction results and the generation of a rainfall distribution map for high-risk areas refer to setting rainfall thresholds A and B, and A>B. If the predicted rainfall is greater than or equal to the threshold A, it is extreme rainfall, an alarm is immediately issued, and emergency measures are taken. If the predicted rainfall is less than the threshold A and greater than or equal to the threshold B, it is moderate rainfall, there is a moderate impact, and it is recommended to take preventive measures. If the predicted rainfall is less than the threshold B, it is light rainfall, and there is no major risk.

[0036] Use GIS technology to generate a distribution map of high-risk areas, use red, yellow and blue to mark extreme risk areas, moderate risk areas and mild risk areas in sequence, superimpose regional administrative boundaries, and identify the affected cities and regions.

[0037] As a preferred solution of the rainfall prediction method for the Beijing-Tianjin-Hebei region described in the present invention, the visual display of real-time rainfall forecast and warning information refers to mapping the graded warning information onto the generated high-risk area distribution map, and drawing a curve graph of rainfall changes over time, superimposing the rainfall intensity distribution at different time points, and displaying the high-risk area distribution map and curve graph in real time through a visual interface.

[0038] As a preferred solution of the rainfall prediction method for the Beijing-Tianjin-Hebei region described in the present invention, the storage in the database refers to sorting the real-time rainfall forecast and warning information in chronological order and storing it in the central database. The database regularly performs integrity and security checks on the stored data and uploads it to the cloud for backup.

[0039] In a second aspect, the present invention provides a rainfall prediction system for the Beijing-Tianjin-Hebei region, comprising:

[0040] Data acquisition module, used to collect multi-source data and perform data cleaning and standardization, improve the spatial resolution of ERA5 data, and extract the core features of GNSS data;

[0041] The rainfall modeling module is used to construct the rainfall evolution path and simulate the time window and spatial variation of rainfall events;

[0042] Rainfall prediction module, used to predict rainfall intensity and regional rainfall based on rainfall evolution path;

[0043] Early warning module, used to classify risks based on predicted rainfall and generate distribution maps of high-risk areas;

[0044] The display storage module is used to generate real-time distribution maps of high-risk areas, store real-time rainfall forecasts and warning information, and regularly back up to the cloud.

[0045] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the rainfall prediction method for the Beijing-Tianjin-Hebei region as described in the first aspect of the present invention is implemented.

[0046] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the rainfall prediction method for the Beijing-Tianjin-Hebei region as described in the first aspect of the present invention.

[0047] The beneficial effects of the present invention are as follows: the present invention improves the spatial resolution of data and the ability to predict precipitation evolution by fusing multi-source data, adopting generative adversarial networks and spectral analysis, and combines dynamic rainfall modeling, fractional-order differential equations and multivariate regression models to improve the real-time and accuracy of rainfall prediction, and can effectively cope with the challenges of rainfall prediction under complex weather conditions, and achieve accurate prediction of rainfall intensity and extreme weather events. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0049] Figure 1 This is a flow chart of the rainfall prediction method for the Beijing-Tianjin-Hebei region in Example 1.

[0050] Figure 2 This is a structural diagram of the rainfall prediction system for the Beijing-Tianjin-Hebei region in Example 1. DETAILED DESCRIPTION

[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0054] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a rainfall prediction method for the Beijing-Tianjin-Hebei region, comprising the following steps:

[0055] S1. Obtain multi-source data and extract the core characteristics of the data after preprocessing. Perform dynamic rainfall modeling based on the core characteristics of the data to simulate the rainfall evolution path.

[0056] Specifically, obtaining multi-source data and extracting the core characteristics of the data after preprocessing the data refers to collecting meteorological station data, GNSS data, and ERA5 data, cleaning and standardizing the collected data, and using spatial interpolation methods to spatially align sparse observation data;

[0057] Use a generative adversarial network algorithm to increase the spatial resolution of ERA5 data to the kilometer level;

[0058] The spectrum analysis method is used to extract the time series of atmospheric precipitable water content data from GNSS data. The time domain signal is converted into a frequency domain signal using fast Fourier transform, the frequency component is calculated, the amplitude and phase information of the frequency component are obtained, and the frequency point with an amplitude greater than the average amplitude is taken as the core periodic feature. The data change rate is calculated based on the extracted atmospheric precipitable water content data;

[0059] The weather station data include temperature, humidity and rainfall data, the GNSS data include zenith wet delay and atmospheric precipitable water data, and the ERA5 data include humidity, temperature gradient and wind speed data.

[0060] By acquiring multi-source data (weather station, GNSS, ERA5) and extracting core features after preprocessing, the present invention can effectively improve the accuracy and spatial precision of rainfall forecasts. First, data cleaning and standardization ensure the consistency and quality of data from different sources, avoiding the impact of noise and inconsistency in the data on subsequent analysis. The use of spatial interpolation methods to spatially align sparse observation data effectively fills the gaps between meteorological stations and GNSS observation points, making the meteorological information of the entire forecast area more complete and continuous, especially in areas where the meteorological observation network is relatively sparse. Interpolation processing greatly improves the spatial coverage of the data. The application of the generative adversarial network (GAN) algorithm significantly improves the spatial details of low-resolution data by increasing the spatial resolution of ERA5 data to the kilometer level, thereby enhancing the model's ability to capture local precipitation changes, making rainfall forecasts more accurate and reliable. The combination of spectral analysis and fast Fourier transform can extract the periodic change characteristics of atmospheric water vapor from GNSS data. By calculating the amplitude and phase information of the frequency components, the key periodic factors affecting precipitation can be found, and based on this, the rate of change of atmospheric precipitable water data can be calculated, thereby providing more detailed spatiotemporal characteristics for rainfall intensity prediction.

[0061] Furthermore, dynamic rainfall modeling was performed based on the core characteristics of the data. The simulated rainfall evolution path refers to connecting to the ERA5 data interface, selecting the Beijing-Tianjin-Hebei region and time range, extracting humidity, temperature, and wind speed component data, integrating the real-time temperature and humidity observation data of multiple meteorological stations in the region, and storing the data as a CSV file with a unified format, where each record contains timestamp, longitude and latitude, temperature, and humidity;

[0062] We traverse the collected data, mark missing values, fill in missing values ​​using time series linear interpolation, time-align the ERA5 data with the weather station data, and normalize the humidity, temperature, and wind speed components;

[0063] By cleaning, standardizing and spatially interpolating the weather station data, GNSS data and ERA5 data, the noise and missing problems in the data can be effectively eliminated, and the quality and accuracy of the data can be improved. At the same time, the spectral analysis is used to extract the atmospheric precipitable water volume (PWV) characteristics in the GNSS data, which can identify the periodic characteristics of rainfall events and provide accurate basic data for subsequent modeling.

[0064] Get the humidity and temperature at the current time point as initial conditions, set the initial rainfall to zero and set the time step;

[0065] Fitting the evaporation constant from historical weather station data by nonlinear fitting method and the condensation constant , calculate the evaporation rate using the initial conditions at the first time step and condensation efficiency :

[0066] ,

[0067] ;

[0068] In the formula, h is humidity, T is temperature, is the reference temperature, obtained through statistical analysis of historical data;

[0069] The calculated results are used as the starting point of the next time step to iterate the evaporation rate and condensation efficiency to obtain the final evaporation rate. and condensation efficiency ;

[0070] Based on the calculated evaporation rate and condensation efficiency, a fractional differential equation is established to describe the change in rainfall:

[0071] ;

[0072] In the formula, is the fractional rate of change of rainfall, is the fractional order, which is set by experimental tuning;

[0073] Through dynamic modeling and rainfall evolution path simulation, the time window and spatial distribution of rainfall events can be accurately captured. The prediction of dynamic paths enables the model to track the development of rainfall events in real time, accurately predict the start and end time and range of rainfall, and provide more accurate rainfall prediction results.

[0074] Integrate the humidity, temperature, evaporation rate, condensation efficiency and rainfall at time t into the state variable matrix X(t);

[0075] Define the modulation function and use the modulation function method to fit the model parameters:

[0076] ,

[0077] ,

[0078] ;

[0079] In the formula, are the optimized model parameters, is the composite modulation function, is the state variable matrix, is the fractional derivative of the modulation function, Y is the total time span, is the Gamma function, is the historical time variable, is the fractional order, is the integer derivative of the modulation function, i is the index of the modulation function;

[0080] The nonlinear fitting method enables the model to better fit nonlinear parameters such as evaporation constant and condensation constant, thus improving the simulation accuracy. The fractional differential equation provides a more flexible and accurate mathematical description for rainfall changes, and is particularly suitable for dealing with the nonlinear characteristics of rainfall changes over time.

[0081] Initialize the basis functions of the Chebyshev polynomials:

[0082] ;

[0083] In the formula, is a Chebyshev polynomial of order k, and They are the first two order Chebyshev polynomials respectively;

[0084] Project the dynamic equations into the Chebyshev basis function space and calculate the polynomial coefficients :

[0085] ;

[0086] In the formula, is the rainfall rate data point at the mth time point, obtained by subtracting the condensation efficiency from the evaporation rate, and M is the total number of time points;

[0087] The fractional differential equation is numerically solved using the Chebyshev polynomial method to obtain the rainfall time series solution. :

[0088] ;

[0089] In the formula, are the coefficients of the Chebyshev polynomials, is the kth order Chebyshev polynomial, N is the highest order of the polynomial;

[0090] The use of Chebyshev polynomials can improve the accuracy and computational efficiency of numerical solutions, avoid numerical instability in solving high-order differential equations, and reduce computing time and resource consumption;

[0091] Set the rainfall threshold Q, traverse the rainfall time series, if the current rainfall is greater than the threshold Q, mark the current time as the rainfall start time, if the current rainfall is less than or equal to the threshold Q and continues for n hours to meet the conditions, mark the current time as the rainfall stop time, get the rainfall time window according to the rainfall start time and stop time and record the initial coordinates of the current rainfall event;

[0092] Update the coordinates of the rainfall event based on the wind speed components and the initial coordinates of the rainfall event provided by ERA5:

[0093] ,

[0094] ;

[0095] In the formula, and are the coordinates of the current rainfall area, and are the coordinates of the rainfall area at the next time step, and is the wind speed component, is the time step;

[0096] The rainfall coordinates at each time step are recorded to form a dynamic rainfall path.

[0097] Through multi-source data fusion, core feature extraction and dynamic modeling, combined with the innovative application of nonlinear fitting and fractional differential equations, the accuracy and timeliness of rainfall forecasting have been significantly improved. By capturing the rainfall evolution path in real time and accurately predicting rainfall intensity and spatial distribution, it can provide accurate decision-making basis for extreme weather warning. In addition, the Chebyshev polynomial method is used to optimize numerical solution and spatial interpolation technology, which effectively solves the problem of temporal and spatial inconsistency of data and improves the reliability and computational efficiency of prediction. Ultimately, this method provides scientific, timely and accurate support for meteorological warning, disaster management, flood control and drainage, and other fields, significantly enhancing the response capability and emergency response capability of regional rainfall forecasting.

[0098] S2. Build a rainfall prediction model, predict rainfall intensity based on the rainfall evolution path, issue extreme weather warnings based on the prediction results, and generate rainfall distribution maps for high-risk areas;

[0099] Specifically, constructing a rainfall prediction model and predicting rainfall intensity based on the rainfall evolution path means constructing a rainfall prediction model through a multivariate linear regression model, using state variable matrix data as training data to input into the multivariate linear regression model for iterative training, defining a loss function and an Adam optimizer for iterative optimization of model parameters, and stopping the iteration to output model parameters and update the multivariate linear regression model when the loss of the multivariate linear regression model no longer decreases significantly during continuous iterations;

[0100] The real-time data (meteorological station, GNSS, and ERA5 data) within the coverage of the rainfall evolution path are input into the multivariate linear regression model to obtain the regional rainfall intensity forecast value. According to the rainfall time window range, the rainfall intensity forecast value is integrated and calculated to obtain the regional rainfall forecast value.

[0101] The present invention can greatly improve the accuracy and reliability of rainfall intensity prediction by constructing a multivariate linear regression model, combining multidimensional meteorological data, and integrating data through a state variable matrix. In addition, the Adam optimizer is used to accelerate the training process, and the regional rainfall prediction value is obtained by integral calculation. Through these technical means, the present invention realizes the accurate prediction of rainfall intensity in the Beijing-Tianjin-Hebei region, and can provide effective support for extreme weather warning and disaster management, which has important application value.

[0102] Furthermore, issuing extreme weather warnings based on forecast results and generating rainfall distribution maps for high-risk areas means setting rainfall thresholds A and B by statistically analyzing historical rainfall data, and A>B. If the forecast rainfall is greater than or equal to threshold A, it is extreme rainfall, an alarm is immediately issued, and emergency measures are taken. If the forecast rainfall is less than threshold A and greater than or equal to threshold B, it is moderate rainfall with moderate impact, and preventive measures are recommended. If the forecast rainfall is less than threshold B, it is light rainfall with no major risk.

[0103] Use GIS technology to generate a distribution map of high-risk areas, use red, yellow and blue to mark extreme risk areas, moderate risk areas and mild risk areas in sequence, superimpose regional administrative boundaries, and identify the affected cities and regions.

[0104] By combining rainfall prediction, threshold setting and GIS technology, the accuracy and timeliness of rainfall warnings are not only improved, but also the visualization of rainfall and the timely identification of high-risk areas are realized, providing accurate data support for emergency responses to various disasters. Through visualization and regional rainfall distribution maps, the public can clearly see the intensity distribution of current rainfall events and understand the risk level of their respective regions. This not only helps to enhance the public's awareness of disaster warnings, but also promotes their disaster prevention preparations in daily life. For example, according to the alarm information of high-risk areas, preparations such as flood control and drainage and material reserves can be carried out in advance.

[0105] S3, visually displaying the real-time rainfall forecast and warning information and storing it in a database;

[0106] Specifically, visualizing the real-time rainfall forecast and warning information means mapping the graded warning information onto the generated high-risk area distribution map, drawing a curve graph of rainfall changes over time, superimposing the rainfall intensity distribution at different time points, and displaying the high-risk area distribution map and curve graph in real time through a visual interface.

[0107] By combining multi-dimensional rainfall forecast information (rainfall, rainfall intensity, time change, risk area), more comprehensive data support is provided for the meteorological forecast system, further improving the accuracy of the meteorological forecast system. In addition, by combining different meteorological data sources (weather stations, GNSS, ERA5, etc.), it can provide more reliable real-time data support for meteorological researchers, promote the further development of meteorological forecasting technology, and provide accurate warning information to the government, disaster emergency management departments and the public in a very short time by visualizing real-time rainfall forecasts and warning information. Decision makers can quickly determine which areas are at greater risk and take corresponding emergency measures (such as traffic control, flood prevention, emergency evacuation, etc.). This real-time and high efficiency greatly improves the speed and accuracy of emergency response and reduces the losses caused by extreme weather.

[0108] Furthermore, storing in a database means sorting the real-time rainfall forecast and warning information in chronological order and storing it in a central database, which regularly performs integrity and security checks on the stored data and uploads it to the cloud for backup.

[0109] By storing real-time rainfall forecast and warning information in a central database and performing regular integrity and security checks, combined with cloud backup technology, the efficiency of data storage and management has been significantly improved. This method can achieve real-time updates and reliable storage of meteorological warning data, provide an efficient data foundation for decision support, and ensure the stability and reliability of the meteorological warning system.

[0110] This embodiment also provides a rainfall prediction system for the Beijing-Tianjin-Hebei region, including:

[0111] Data acquisition module, used to collect multi-source data and perform data cleaning and standardization, improve the spatial resolution of ERA5 data, and extract the core features of GNSS data;

[0112] The rainfall modeling module is used to construct the rainfall evolution path and simulate the time window and spatial variation of rainfall events;

[0113] Rainfall prediction module, used to predict rainfall intensity and regional rainfall based on rainfall evolution path;

[0114] Early warning module, used to classify risks based on predicted rainfall and generate distribution maps of high-risk areas;

[0115] The display storage module is used to generate real-time distribution maps of high-risk areas, store real-time rainfall forecasts and warning information, and regularly back up to the cloud.

[0116] This embodiment also provides a computer device, which is suitable for the rainfall prediction method for the Beijing-Tianjin-Hebei region, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the rainfall prediction method for the Beijing-Tianjin-Hebei region proposed in the above embodiment.

[0117] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0118] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the rainfall prediction method for the Beijing-Tianjin-Hebei region as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0119] In summary, the present invention improves the spatial resolution of data and the ability to predict precipitation evolution by fusing multi-source data, adopting generative adversarial networks and spectral analysis. It also improves the real-time and accuracy of rainfall prediction by combining dynamic rainfall modeling, fractional-order differential equations and multivariate regression models. It can effectively cope with the challenges of rainfall prediction under complex weather conditions and achieve accurate prediction of rainfall intensity and extreme weather events.

[0120] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A rainfall prediction method for the Beijing-Tianjin-Hebei region, characterized by: include, Acquire multi-source data, pre-process the data, extract the core characteristics of the data, conduct dynamic rainfall modeling based on the core characteristics of the data, and simulate the rainfall evolution path; Build a rainfall prediction model, predict rainfall intensity based on the rainfall evolution path, issue extreme weather warnings based on the prediction results, and generate rainfall distribution maps for high-risk areas; Real-time rainfall forecast and warning information are visualized and stored in the database.

2. The method for predicting rainfall in the Beijing-Tianjin-Hebei region according to claim 1, characterized in that: The acquiring of multi-source data and extracting the core characteristics of the data after preprocessing the data refers to collecting weather station data, GNSS data and ERA5 data, cleaning and standardizing the collected data, and using a spatial interpolation method to spatially align the sparse observation data; Use a generative adversarial network algorithm to increase the spatial resolution of ERA5 data to the kilometer level; The spectrum analysis method is used to extract the time series of atmospheric precipitable water data in GNSS data. The time domain signal is converted into a frequency domain signal using fast Fourier transform, the frequency component is calculated, the amplitude and phase information of the frequency component is obtained, and the frequency point with an amplitude greater than the average amplitude is taken as the core periodic feature. The data change rate is calculated based on the extracted atmospheric precipitable water data.

3. The method for predicting rainfall in the Beijing-Tianjin-Hebei region according to claim 2, characterized in that: The dynamic rainfall modeling based on the core characteristics of the data and simulating the rainfall evolution path refers to connecting the ERA5 data interface, selecting the Beijing-Tianjin-Hebei region and time range, extracting humidity, temperature and wind speed component data, integrating the real-time temperature and humidity observation data of multiple meteorological stations in the region, and storing the data as a CSV file with a unified format, where each record contains a timestamp, longitude and latitude, temperature and humidity; We traverse the collected data, mark missing values, fill in missing values ​​using time series linear interpolation, time-align the ERA5 data with the weather station data, and normalize the humidity, temperature, and wind speed components; Get the humidity and temperature at the current time point as initial conditions, set the initial rainfall to zero and set the time step; Fitting the evaporation constant from historical weather station data by nonlinear fitting method and the condensation constant , calculate the evaporation rate using the initial conditions at the first time step and condensation efficiency ; The calculated results are used as the starting point of the next time step to iterate the evaporation rate and condensation efficiency to obtain the final evaporation rate. and condensation efficiency ; Based on the calculated evaporation rate and condensation efficiency, a fractional differential equation is established to describe the change in rainfall: ； In the formula, is the fractional rate of change of rainfall, is the fractional order, which is set by experimental tuning; Integrate the humidity, temperature, evaporation rate, condensation efficiency and rainfall at time t into the state variable matrix X(t); Define the modulation function and use the modulation function method to fit the model parameters: ； In the formula, are the optimized model parameters, is the composite modulation function, is the state variable matrix, is the fractional derivative of the modulation function; Initialize the basis functions of the Chebyshev polynomial, project the dynamic equations into the Chebyshev basis function space, and calculate the polynomial coefficients ; The fractional differential equation is numerically solved using the Chebyshev polynomial method to obtain the rainfall time series solution. ; Set the rainfall threshold Q, traverse the rainfall time series, if the current rainfall is greater than the threshold Q, mark the current time as the rainfall start time, if the current rainfall is less than or equal to the threshold Q and continues for n hours to meet the conditions, mark the current time as the rainfall stop time, get the rainfall time window according to the rainfall start time and stop time and record the initial coordinates of the current rainfall event; Update the coordinates of the rainfall event based on the wind speed components and the initial coordinates of the rainfall event provided by ERA5: ， ； In the formula, and are the coordinates of the current rainfall area, and are the coordinates of the rainfall area at the next time step, and is the wind speed component, is the time step; The rainfall coordinates at each time step are recorded to form a dynamic rainfall path.

4. The method for predicting rainfall in the Beijing-Tianjin-Hebei region according to claim 3, characterized in that: The construction of the rainfall prediction model and the prediction of rainfall intensity based on the rainfall evolution path refers to constructing the rainfall prediction model through a multivariate linear regression model, using the state variable matrix data as training data to input into the multivariate linear regression model for iterative training, defining the loss function and the Adam optimizer to iteratively optimize the model parameters, and stopping the iteration to output the model parameters and update the multivariate linear regression model when the loss of the multivariate linear regression model no longer decreases significantly during the continuous iteration process; The real-time data within the coverage of the rainfall evolution path is input into the multivariate linear regression model to obtain the regional rainfall intensity forecast value. According to the rainfall time window range, the rainfall intensity forecast value is integrated and calculated to obtain the regional rainfall forecast value.

5. The method for predicting rainfall in the Beijing-Tianjin-Hebei region according to claim 4, characterized in that: The extreme weather warning based on the forecast results and the generation of a rainfall distribution map for high-risk areas refer to setting rainfall thresholds A and B, and A>B. If the forecast rainfall is greater than or equal to threshold A, it is extreme rainfall, an alarm is immediately issued, and emergency measures are taken. If the forecast rainfall is less than threshold A and greater than or equal to threshold B, it is moderate rainfall with moderate impact, and preventive measures are recommended. If the forecast rainfall is less than threshold B, it is light rainfall with no major risk. Use GIS technology to generate a distribution map of high-risk areas, use red, yellow and blue to mark extreme risk areas, moderate risk areas and mild risk areas in sequence, superimpose regional administrative boundaries, and identify the affected cities and regions.

6. The method for predicting rainfall in the Beijing-Tianjin-Hebei region according to claim 5, characterized in that: The visual display of real-time rainfall forecast and warning information refers to mapping the graded warning information onto the generated high-risk area distribution map, drawing a curve graph of rainfall changes over time, superimposing the rainfall intensity distribution at different time points, and displaying the high-risk area distribution map and curve graph in real time through a visual interface.

7. The method for predicting rainfall in the Beijing-Tianjin-Hebei region according to claim 6, characterized in that: The storage in the database refers to sorting the real-time rainfall forecast and warning information in chronological order and storing it in a central database. The database regularly performs integrity and security checks on the stored data and uploads it to the cloud for backup.

8. A rainfall prediction system for the Beijing-Tianjin-Hebei region, based on the rainfall prediction method for the Beijing-Tianjin-Hebei region according to any one of claims 1 to 7, characterized in that: include, Data acquisition module, used to collect multi-source data and perform data cleaning and standardization, improve the spatial resolution of ERA5 data, and extract the core features of GNSS data; The rainfall modeling module is used to construct the rainfall evolution path and simulate the time window and spatial variation of rainfall events; Rainfall prediction module, used to predict rainfall intensity and regional rainfall based on rainfall evolution path; Early warning module, used to classify risks based on predicted rainfall and generate distribution maps of high-risk areas; The display storage module is used to generate real-time distribution maps of high-risk areas, store real-time rainfall forecasts and warning information, and regularly back up to the cloud.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the rainfall prediction method for the Beijing-Tianjin-Hebei region described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the rainfall prediction method for the Beijing-Tianjin-Hebei region described in any one of claims 1 to 7 are implemented.

Images