Heavy rainfall short-term and imminent forecasting method and system
By finely dividing the precipitation area and combining it with machine learning models, the problem of inaccurate forecasting of the rainfall center in traditional methods has been solved, achieving more accurate short-term forecasts of heavy rainfall and improving the accuracy and relevance of forecasts.
Patent Information
- Application Number
- CN202411028948.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-07-30
AI Technical Summary
Traditional short-term forecasting methods for heavy rainfall rely on single meteorological data, have poor resistance to interference, and are difficult to accurately predict the location and intensity of rainfall centers, resulting in large forecast errors and an inability to effectively cope with risks such as flash floods and geological disasters.
By dividing the precipitation area into multiple regularly shaped monitoring zones, analyzing precipitation trends using meteorological station monitoring data, determining the precipitation center area using machine learning models, and outputting short-term forecast results.
It improves the accuracy and relevance of short-term forecasts for heavy rainfall, enabling precise prediction of the location and intensity of precipitation centers, reducing losses caused by disasters, and meeting the needs for more refined weather forecasts.
Smart Images

Figure CN118962849A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of meteorological services, and in particular relates to a method and system for forecasting short-term heavy rainfall. Background Art
[0002] Heavy rainfall is a severe convective weather phenomenon with high rainfall intensity in a short period of time. It mainly occurs in the flood season every year and is often accompanied by severe weather phenomena such as strong winds, hail and thunderstorms. It is easy to cause natural disasters such as mountain torrents, mudslides, and urban floods. The short-term forecast of heavy rainfall generally refers to the forecast of heavy rainfall weather in the region within the next 0 to 12 hours. Improving the accuracy of short-term forecasts of heavy rainfall is one of the key research goals of meteorological researchers. However, due to the high suddenness of heavy rainfall and the high-dimensional nonlinearity of meteorological data, traditional short-term forecast methods are not effective. Moreover, traditional short-term forecast methods often only use a single meteorological data, and the model has weak anti-interference ability and is highly dependent on the accuracy of the data.
[0003] Normally, the 0-2 hour nowcast rainfall forecast is mainly based on extrapolation of radar or satellite observation data, while the 2-12 hour short-term rainfall forecast relies more on numerical weather forecast. According to the forecast method, short-term rainfall forecast is generally divided into statistical forecast method, numerical model forecast method and extrapolation-based forecast method.
[0004] At present, using radar data for extrapolation is one of the main methods for short-term rainfall forecasting. However, the traditional extrapolation method does not fully exploit the laws of historical observation data in terms of data utilization, and only relies on limited radar data for forecasting. The traditional forecast based on radar echo extrapolation has a fuzzy precipitation area and serious distortion, and there is a large error in estimating precipitation based on radar echoes. Therefore, the location and intensity forecast of the precipitation center are not very accurate. Although my country's ground meteorological automatic stations are densely distributed, it is still difficult for ground automatic stations to capture the precise location and intensity of the precipitation center due to the strong local characteristics of rainfall. Therefore, by using technical methods to improve the accuracy of the observation of the location and intensity of the center of heavy rainfall, more accurate observation data can be provided for short-term forecasting, which can effectively improve the central value and location accuracy of the heavy rainfall forecast, determine the area with the largest precipitation, provide heavy rainfall warnings for the public and industry departments, and avoid potential safety risks. For areas where flash floods and geological disasters occur frequently, accurate precipitation center forecasts can help take preventive measures in advance and reduce the losses caused by disasters. Therefore, it is urgent to propose a short-term forecast method for heavy rainfall and design related systems to improve the accuracy of short-term forecast of heavy rainfall. Summary of the invention
[0005] In order to solve the above problems, the present invention provides a method and system for forecasting short-term heavy rainfall, which improves the accuracy and pertinence of short-term heavy rainfall forecasting by finding the accurate precipitation center position and intensity.
[0006] In order to achieve the above object, the technical solution of the present invention is as follows:
[0007] On the one hand, a method for short-term forecasting of heavy rainfall is provided, comprising the following steps:
[0008] S1: monitoring the precipitation area, and when the precipitation area is monitored, dividing the precipitation area into a first monitoring area;
[0009] S2: Divide the first monitoring area with the meteorological monitoring station as the center to obtain at least three second monitoring areas with regular shapes, all of which are located in the first monitoring area, and the area of the second monitoring area is smaller than that of the first monitoring area;
[0010] S3: Calculating the precipitation intensity distribution at different locations in the second monitoring area to which the meteorological monitoring station belongs and the total precipitation in the second precipitation area through the precipitation data of the same time period monitored by each meteorological monitoring station;
[0011] S4: judging the precipitation change trend in the second monitoring area and the precipitation change trend between adjacent second monitoring areas based on the precipitation intensity distribution and total precipitation in each designated period corresponding to each second monitoring area;
[0012] S5: adjusting the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas, and obtaining the overlapping area of the second monitoring areas; when the precipitation increase path direction of each second monitoring area is toward the overlapping area of the second monitoring area, extracting the overlapping area as the precipitation center area;
[0013] S6: extracting precipitation-related features of the precipitation center area based on the location of the precipitation center area;
[0014] S7: Based on the trained machine learning model, the precipitation-related features of the precipitation center area are received and the short-term forecast results of heavy rainfall are output.
[0015] Furthermore, in S1, the precipitation area is divided into a first monitoring area, including: determining the monitoring purpose, collecting basic data, analyzing precipitation characteristics and determining the scope of the monitoring area.
[0016] Further, in S2, the first monitoring area is divided with the meteorological monitoring station as the center to obtain at least three second monitoring areas of regular shapes, including:
[0017] S201: Obtain the number and locations of meteorological monitoring stations within the first monitoring area;
[0018] S202: determining the shape of the second monitoring area based on the range of the first monitoring area to ensure that the shape distribution can cover the entire first monitoring area;
[0019] S203: based on the number and location of meteorological monitoring stations in the first monitoring area, dividing a second monitoring area of a regular shape centered on the meteorological monitoring station and determining the location of the second monitoring area;
[0020] S204: Determine the size of the second monitoring area based on the shape and size of the second monitoring area.
[0021] Further, S3: calculating the precipitation intensity distribution at different locations in the second monitoring area to which the meteorological monitoring station belongs and the total precipitation in the second precipitation area through the precipitation data of the same time period monitored by each meteorological monitoring station, including:
[0022] S301: performing radar detection based on the range of each second monitoring area through a meteorological monitoring station within the second monitoring area;
[0023] S302: performing radar detection based on the meteorological monitoring stations in each second monitoring area to obtain precipitation intensity and precipitation data for the second monitoring area;
[0024] S303: Analyze the precipitation intensity and precipitation amount data of each second detection area to obtain the precipitation intensity distribution in each second monitoring area and the total precipitation amount in the second precipitation area.
[0025] Furthermore, the precipitation data of each second monitoring area is analyzed to determine the trend of precipitation changes, including:
[0026] By analyzing the precipitation data in the second detection area, the precipitation change trend in the second monitoring area is obtained; by comparing the precipitation data of the adjacent second monitoring areas, the size of this group of precipitation data is judged, and the change trend of precipitation in the precipitation change area is obtained.
[0027] Further, S5: adjusting the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas, and obtaining the overlapping area of the second monitoring areas; when the precipitation increase path direction of each second monitoring area is toward the overlapping area of the second monitoring area, extracting the overlapping area as the precipitation center domain, including:
[0028] Based on comparing the precipitation change trends in adjacent second monitoring areas, the path of increasing or decreasing precipitation in the adjacent second monitoring areas is obtained. Through precipitation data analysis, the second monitoring area in the precipitation change area will move toward the area with higher precipitation data along the path direction of increasing precipitation, and the overlapping area of each second monitoring area is obtained based on the movement of each second monitoring area; when the direction of the precipitation increase path of each second monitoring area is toward the overlapping area of the second monitoring area, the overlapping area is extracted as the precipitation center area.
[0029] Further, in S5, the module for adjusting the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas to find the position of the precipitation center area includes:
[0030] S501: Collect precipitation data of corresponding meteorological stations in each second monitoring area;
[0031] S502: Collect and summarize precipitation data of meteorological stations within the second monitoring area within the first monitoring area;
[0032] S503: Compare the precipitation data of the weather stations collected to determine the precipitation trend and the path of precipitation increase or decrease;
[0033] S504: Determine whether the second monitoring area edge is on a route where precipitation increases, and change the size and position of each second monitoring area to obtain an overlapping area of each second monitoring area;
[0034] S505: Determine whether the direction of the precipitation increase path of each second monitoring area is toward the overlapping area of the second monitoring area. If not, return to the precipitation data comparison unit, compare the precipitation data in the direction of each precipitation increase path, and then output a larger precipitation increase path, and then adjust the second monitoring area until the direction of the precipitation increase path of each second monitoring area is toward the overlapping area of the second monitoring area, and then extract the overlapping area as the precipitation center area.
[0035] Further, S6: based on the location of the precipitation center area, extract precipitation-related characteristics of the precipitation center area, including: clarifying the specific location of the precipitation center area based on S1 to S5, collecting precipitation data of the precipitation center area, including historical data and real-time data; analyzing the main factors affecting the precipitation characteristics of the precipitation center area based on the data; and summarizing the main precipitation characteristics of the precipitation center area.
[0036] Further, in S7, the process of outputting a short-term heavy rainfall forecast result based on the received precipitation-related features of the precipitation center area based on the trained machine learning model is as follows: based on the main precipitation features of the precipitation center area, the features are input into the machine learning model that has been constructed and trained, and then the machine learning model can predict the rainfall in the precipitation center area at the next moment, thereby obtaining a future rainfall sequence for the predicted precipitation center area, and outputting the rainfall sequence can obtain the short-term rainfall forecast data for the precipitation center area.
[0037] On the other hand, a heavy rainfall short-term forecast system is provided, which is operated based on the heavy rainfall short-term forecast method according to any one of claims 1 to 9, and is characterized in that it includes:
[0038] A monitoring module, used for monitoring a precipitation area, and when a precipitation area is detected, dividing the precipitation area into a first monitoring area;
[0039] A division module is used to divide the first monitoring area with the meteorological monitoring station as the center to obtain at least three second monitoring areas of regular shapes, all of which are located in the first monitoring area, and the area of the second monitoring area is smaller than that of the first monitoring area;
[0040] A collection module, used to calculate the precipitation intensity distribution at different locations in the second monitoring area to which the meteorological monitoring station belongs and the total precipitation in the second precipitation area through the precipitation data of the same time period monitored by each meteorological monitoring station;
[0041] A judgment module, used for judging the precipitation change trend in the second monitoring area and the precipitation change trend between adjacent second monitoring areas based on the precipitation intensity distribution and total precipitation in each second monitoring area;
[0042] An adjustment module, used to adjust the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas, and simultaneously obtain the overlapping area of the second monitoring areas; when the precipitation increase path direction of each second monitoring area is toward the overlapping area of the second monitoring area, the overlapping area is extracted as the precipitation center area;
[0043] A feature extraction module, used for extracting precipitation-related features of the precipitation center area based on the location of the precipitation center area;
[0044] The output module is used to receive precipitation-related features in the precipitation center area based on the trained machine learning model and output short-term forecast results of heavy rainfall.
[0045] The above scheme has the following beneficial effects:
[0046] 1. The present invention, firstly, can reflect the precipitation characteristics of the entire target area by dividing the precipitation area into the first monitoring area, and the division of the second monitoring area helps to more accurately reflect the spatial variation of precipitation and provide a basis for the subsequent determination of the precipitation center area.
[0047] Compared with the existing technology, by comparing the precipitation data collected by the meteorological stations in each second monitoring area, the precipitation change trend of the adjacent second monitoring area is finally obtained, and the precipitation conditions of each location in the first monitoring area can be mastered, thereby obtaining the paths of precipitation increase or decrease of different sizes, so that the system will not miss any precipitation when monitoring the precipitation conditions, thereby improving the accuracy of this method. At the same time, by more accurately obtaining the precipitation center, a more accurate short-term forecast result of heavy rainfall can be output.
[0048] 2. The present invention adjusts the position of the second monitoring area by the precipitation change trend between adjacent second monitoring areas in the first monitoring area to find the position of the precipitation center area. The position of the precipitation center area of the first monitoring area can be accurately determined by the change in the size of the position of the second monitoring area, and some other precipitation points in the first monitoring area can be monitored, making the entire precipitation forecast more comprehensive.
[0049] 3. The present invention can improve the accuracy of forecast by predicting heavy rainfall by determining the precipitation center area. Currently, most traditional methods use radar data extrapolation to predict rainfall areas, while by determining the precipitation center area, the location of precipitation can be determined more accurately, which is more precise than traditional rainfall area forecasts. After determining the precipitation center area, the value of the precipitation center can be determined to achieve regular, quantitative and fixed-point forecasts, improve the accuracy and pertinence of precipitation forecasts, provide heavy rainfall warnings for the public and industry departments, avoid potential safety risks, and meet the country's urgent needs for refined weather forecasts.
[0050] 4. The present invention, through the location of the precipitation center area, extracts the relevant features of the precipitation center area, which is conducive to accurately identifying the location, intensity and range of the precipitation center, and can more accurately predict the occurrence and evolution of precipitation. By extracting and analyzing the characteristics of the precipitation center, a more sophisticated forecast model can be established to improve the spatiotemporal resolution and accuracy of the forecast. The precipitation center location is generally a high-incidence area for meteorological disasters such as heavy rain and floods. Extracting the characteristics of these areas helps to discover disaster signs earlier, thereby predicting the scope and extent of the disaster and reducing losses.
[0051] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] Figure 1 The present invention is a flowchart of an embodiment of a method for short-term forecasting of heavy rainfall.
[0053] Figure 2 It is a structural block diagram of an embodiment of a heavy rainfall short-term forecast system of the present invention. DETAILED DESCRIPTION
[0054] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0055] The following is further described in detail through specific implementation methods:
[0056] Embodiment 1:
[0057] Basically as attached Figure 1 A method for short-term forecasting of heavy rainfall is shown, comprising the following steps:
[0058] S1: Monitor the precipitation area. When the precipitation area is monitored, divide the precipitation area into a first monitoring area.
[0059] Among them, the precipitation area is divided into the first monitoring area, including: determining the purpose of monitoring, collecting basic data, analyzing precipitation characteristics and determining the scope of the monitoring area. First, it is necessary to clarify the specific purpose of monitoring the precipitation area. Different purposes may require different monitoring strategies and data accuracy. Then, by collecting basic geographic information of the area to be studied, historical meteorological data are obtained, including precipitation, precipitation frequency, precipitation type (such as rain, snowfall), etc. According to the historical meteorological data, the precipitation characteristics of the target area are analyzed, including the spatial distribution of precipitation, temporal changes, and extreme precipitation events. Finally, according to the analysis results, the precipitation area is determined and the precipitation area is divided into the first monitoring area.
[0060] S2: Divide the first monitoring area with the meteorological monitoring station as the center to obtain at least three second monitoring areas of regular shapes, all of which are located within the first monitoring area, and the area of the second monitoring area is smaller than that of the first monitoring area.
[0061] The first monitoring area is divided with the meteorological monitoring station as the center to obtain a plurality of second monitoring areas of regular shapes (regular shapes may be circular, fan-shaped, rectangular or triangular, etc., and a circular shape is selected in this embodiment) including:
[0062] S201: Calculate the number of meteorological monitoring stations, location distribution, and distances between adjacent meteorological monitoring stations in a first monitoring area using meteorological monitoring station data collected by a computer to obtain a distribution map of meteorological monitoring stations in the first monitoring area;
[0063] S202: dividing the first area into circles based on the distribution map of meteorological monitoring stations in the first monitoring area, so that the second detection area in the shape of a circle can cover the distribution map of meteorological monitoring stations in the first monitoring area;
[0064] S203: determining that each meteorological monitoring station is located at the center of the circle based on the fact that the second monitoring area in S202 is in a circular shape;
[0065] S204: Collecting basic data of the second monitoring area by a computer to determine the size of the second monitoring area;
[0066] S205: Based on S201-S204, obtain the distribution map of the first monitoring area and the second monitoring area.
[0067] S3: Calculate the precipitation intensity distribution at different locations within the second monitoring area to which the meteorological monitoring station belongs and the total amount of precipitation within the second precipitation area through the precipitation data of the same time period monitored by each meteorological monitoring station.
[0068] Specifically, it includes:
[0069] S301: Based on the range of each second monitoring area, precipitation observation and radar detection are performed through the meteorological monitoring station in the second monitoring area;
[0070] S302: obtaining precipitation intensity and precipitation data of the second monitoring area based on precipitation observation and radar detection by the meteorological monitoring stations in each second monitoring area;
[0071] S303: Analyze the precipitation intensity and precipitation amount data of each second detection area to obtain the precipitation intensity distribution in each second monitoring area and the total precipitation amount in the second precipitation area.
[0072] The precipitation intensity levels include:
[0073] Light rain: precipitation within 1 hour is less than or equal to 2.5 mm;
[0074] Moderate rain: precipitation within 1 hour is between 2.6 mm and 8.0 mm;
[0075] Heavy rain: precipitation within 1 hour ranges from 8.1 mm to 15.9 mm;
[0076] Heavy rain: precipitation reaches or exceeds 16 mm within 1 hour.
[0077] S4: Based on the precipitation intensity distribution and total precipitation in each second monitoring area, determine the precipitation change trend within the second monitoring area and the precipitation change trend between adjacent second monitoring areas.
[0078] Specifically, by analyzing the precipitation data in the second detection area, the precipitation change trend in the second monitoring area is obtained; by comparing the precipitation data of adjacent second monitoring areas, the size of this group of precipitation data is judged, and the change trend of precipitation in the precipitation change area is obtained.
[0079] S5: Adjust the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas, and obtain the overlapping area of the second monitoring area at the same time; when the precipitation increase path direction of each second monitoring area is toward the overlapping area of the second monitoring area, extract the overlapping area as the precipitation center area.
[0080] Specifically, the steps include:
[0081] S501: Collect precipitation data of corresponding meteorological stations in each second monitoring area;
[0082] S502: Collect and summarize precipitation data of meteorological stations within the second monitoring area within the first monitoring area;
[0083] S503: Compare the precipitation data of the weather stations collected to determine the precipitation trend and the path of precipitation increase or decrease;
[0084] S504: Determine whether the second monitoring area edge is on a route where precipitation increases, and change the size and position of each second monitoring area to obtain an overlapping area of each second monitoring area;
[0085] S505: Determine whether the direction of the precipitation increase path of each second monitoring area is toward the overlapping area of the second monitoring area. If not, return to the precipitation data comparison unit, compare the precipitation data in the direction of each precipitation increase path, and then output a larger precipitation increase path, and then adjust the second monitoring area until the direction of the precipitation increase path of each second monitoring area is toward the overlapping area of the second monitoring area, and then extract the overlapping area as the precipitation center area.
[0086] S6: Based on the location of the precipitation center area, extract precipitation-related features of the precipitation center area.
[0087] Specifically, based on S1 to S5, the specific location of the precipitation center area is clarified, and the precipitation data of the precipitation center area is collected, including historical data and real-time data; based on the data analysis, the main factors affecting the precipitation characteristics of the precipitation center area, such as topography, climate, monsoon, etc.; the main precipitation characteristics of the precipitation center area (precipitation intensity, precipitation form, precipitation, environmental conditions (such as wind field, humidity, temperature, etc.), terrain information, etc.) are summarized.
[0088] S7: Based on the trained machine learning model, the precipitation-related features of the precipitation center area are received and the short-term forecast results of heavy rainfall are output.
[0089] Specifically, based on the main precipitation characteristics of the precipitation center area, the characteristics are input into the machine learning model (support vector machine (SVM), random forest (Random Forest), deep neural network (DNN), etc.) that has been constructed and trained. Since the random forest (Random Forest) model has the advantages of high prediction accuracy in ensemble learning, processing nonlinear relationships and feature selection among the existing machine learning models that have been constructed, the present embodiment inputs the precipitation-related characteristics of the precipitation center area extracted in S6 into the random forest (Random Forest) model, and then the random forest (Random Forest) model can predict the rainfall map of the precipitation center area at the next moment, thereby obtaining the future rainfall sequence of the precipitation center area, and the rainfall sequence is output to obtain the short-term rainfall forecast map of the precipitation center area.
[0090] In this embodiment, the prediction of heavy rainfall by determining the precipitation center area can improve the accuracy of the forecast. At present, most traditional methods use radar data extrapolation to predict the rainfall area, while by determining the precipitation center area, the location of precipitation can be determined more accurately, which is more precise than the traditional rainfall area forecast. After the precipitation center area is determined, the value of the precipitation center can be determined to achieve a regular, quantitative and fixed-point forecast, provide heavy rainfall warnings for the public and industry departments, avoid potential safety risks, and meet the country's urgent needs for refined weather forecasts. By combining the determination of the precipitation center area with the existing Random Forest model, the accuracy of the prediction can be better improved, and large-scale data sets can be effectively processed and important features can be evaluated, thereby achieving the accuracy and pertinence of short-term forecasts of heavy rainfall.
[0091] Embodiment 2:
[0092] Basically as attached Figure 2 As shown: a heavy rainfall short-term forecast system, which operates based on the heavy rainfall short-term forecast method described in Example 1, includes: a division module, a collection module, a judgment module, an adjustment module, a feature extraction module, and an output module.
[0093] A monitoring module, used for monitoring a precipitation area, and when a precipitation area is detected, dividing the precipitation area into a first monitoring area;
[0094] A division module is used to divide the first monitoring area with the meteorological monitoring station as the center to obtain at least three second monitoring areas of regular shapes, all of which are located in the first monitoring area, and the area of the second monitoring area is smaller than that of the first monitoring area;
[0095] A collection module, used to calculate the precipitation intensity distribution at different locations in the second monitoring area to which the meteorological monitoring station belongs and the total precipitation in the second precipitation area through the precipitation data of the same time period monitored by each meteorological monitoring station;
[0096] A judgment module, used for judging the precipitation change trend in the second monitoring area and the precipitation change trend between adjacent second monitoring areas based on the precipitation intensity distribution and total precipitation in each second monitoring area;
[0097] An adjustment module, used to adjust the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas, and simultaneously obtain the overlapping area of the second monitoring areas; when the precipitation increase path direction of each second monitoring area is toward the overlapping area of the second monitoring area, the overlapping area is extracted as the precipitation center area;
[0098] A feature extraction module, used for extracting precipitation-related features of the precipitation center area based on the location of the precipitation center area;
[0099] The output module is used to receive precipitation-related features in the precipitation center area based on the trained machine learning model and output short-term forecast results of heavy rainfall.
[0100] Obviously, the above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the invention.
Claims
1. A method for short-term forecasting of heavy rainfall, characterized in that: The steps include: S1: monitoring the precipitation area, and when the precipitation area is monitored, dividing the precipitation area into a first monitoring area; S2: Divide the first monitoring area with the meteorological monitoring station as the center to obtain at least three second monitoring areas with regular shapes, all of which are located in the first monitoring area, and the area of the second monitoring area is smaller than that of the first monitoring area; S3: Calculating the precipitation intensity distribution at different locations in the second monitoring area to which the meteorological monitoring station belongs and the total precipitation in the second precipitation area through the precipitation data of the same time period monitored by each meteorological monitoring station; S4: judging the precipitation change trend in the second monitoring area and the precipitation change trend between adjacent second monitoring areas based on the precipitation intensity distribution and total precipitation in each second monitoring area; S5: adjusting the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas, and obtaining the overlapping area of the second monitoring areas; when the precipitation increase path direction of each second monitoring area is toward the overlapping area of the second monitoring area, extracting the overlapping area as the precipitation center area; S6: extracting precipitation-related features of the precipitation center area based on the location of the precipitation center area; S7: Based on the trained machine learning model, the precipitation-related features of the precipitation center area are received and the short-term forecast results of heavy rainfall are output.
2. The method for short-term forecasting of heavy rainfall according to claim 1, characterized in that: In S1, the precipitation area is divided into the first monitoring area, including: determining the monitoring purpose, collecting basic data, analyzing precipitation characteristics and determining the scope of the monitoring area.
3. The method for short-term forecasting of heavy rainfall according to claim 1, characterized in that: In S2, the first monitoring area is divided with the meteorological monitoring station as the center to obtain at least three second monitoring areas of regular shapes, including: S201: Obtain the number and locations of meteorological monitoring stations within the first monitoring area; S202: determining the shape of the second monitoring area based on the range of the first monitoring area to ensure that the shape distribution can cover the entire first monitoring area; S203: based on the number and location of meteorological monitoring stations in the first monitoring area, dividing a second monitoring area of a regular shape centered on the meteorological monitoring station and determining the location of the second monitoring area; S204: Determine the size of the second monitoring area based on the shape and size of the second monitoring area.
4. The method for short-term forecasting of heavy rainfall according to claim 1, characterized in that: S3: Calculating the precipitation intensity distribution at different locations in the second monitoring area to which the meteorological monitoring station belongs and the total precipitation in the second precipitation area through the precipitation data of the same time period monitored by each meteorological monitoring station, including: S301: performing radar detection based on the range of each second monitoring area through a meteorological monitoring station within the second monitoring area; S302: performing radar detection based on the meteorological monitoring stations in each second monitoring area to obtain precipitation intensity and precipitation data for the second monitoring area; S303: Analyze the precipitation intensity and precipitation amount data of each second detection area to obtain the precipitation intensity distribution in each second monitoring area and the total precipitation amount in the second precipitation area.
5. The method for short-term forecasting of heavy rainfall according to claim 4, characterized in that: Analyze the precipitation data of each second monitoring area to determine the trend of precipitation changes, including: By analyzing the precipitation data in the second detection area, the precipitation change trend in the second monitoring area is obtained; by comparing the precipitation data of the adjacent second monitoring areas, the size of this group of precipitation data is judged, and the change trend of precipitation in the precipitation change area is obtained.
6. The method for short-term forecasting of heavy rainfall according to claim 5, characterized in that: S5: adjusting the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas, and obtaining the overlapping area of the second monitoring area; When the precipitation increase paths of the second monitoring areas are all directed toward the overlapping area of the second monitoring areas, the overlapping area is extracted as the precipitation center domain, including: Based on comparing the precipitation change trends in adjacent second monitoring areas, the path of increasing or decreasing precipitation in the adjacent second monitoring areas is obtained. Through precipitation data analysis, the second monitoring area in the precipitation change area will move toward the area with higher precipitation data along the path direction of increasing precipitation, and the overlapping area of each second monitoring area is obtained based on the movement of each second monitoring area; when the direction of the precipitation increase path of each second monitoring area is toward the overlapping area of the second monitoring area, the overlapping area is extracted as the precipitation center area.
7. The method for short-term forecasting of heavy rainfall according to claim 6, characterized in that: In S5, adjusting the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas to find the position of the precipitation center area includes: S501: Collect precipitation data of corresponding meteorological stations in each second monitoring area; S502: Collect and summarize precipitation data of meteorological stations within the second monitoring area within the first monitoring area; S503: Compare the precipitation data of the weather stations collected to determine the precipitation trend and the path of precipitation increase or decrease; S504: Determine whether the second monitoring area edge is on a route where precipitation increases, and change the size and position of each second monitoring area to obtain an overlapping area of each second monitoring area; S505: Determine whether the direction of the precipitation increase path of each second monitoring area is toward the overlapping area of the second monitoring area. If not, return to the precipitation data comparison unit, compare the precipitation data in the direction of each precipitation increase path, and then output a larger precipitation increase path, and then adjust the second monitoring area until the direction of the precipitation increase path of each second monitoring area is toward the overlapping area of the second monitoring area, and then extract the overlapping area as the precipitation center area.
8. The method for short-term forecasting of heavy rainfall according to claim 1, characterized in that: S6: Based on the location of the precipitation center area, extract the precipitation-related characteristics of the precipitation center area, including: clarifying the specific location of the precipitation center area based on S1 to S5; collecting precipitation data of the precipitation center area, including historical data and real-time data; analyzing the main factors affecting the precipitation characteristics of the precipitation center area based on data; and summarizing the main precipitation characteristics of the precipitation center area.
9. The method for short-term forecasting of heavy rainfall according to claim 1, characterized in that: S7: Based on the trained machine learning model, the precipitation-related features of the precipitation center area are received to output the short-term forecast results of heavy rainfall. The process is as follows: based on the main precipitation features of the precipitation center area, the features are input into the machine learning model that has been constructed and trained. Then the machine learning model can predict the rainfall in the precipitation center area at the next moment, thereby obtaining the future rainfall sequence of the predicted precipitation center area. The rainfall sequence is output to obtain the short-term rainfall forecast data of the precipitation center area.
10. A heavy rainfall short-term forecast system, operating based on the heavy rainfall short-term forecast method according to any one of claims 1 to 9, characterized in that: include: A monitoring module, used for monitoring a precipitation area, and when a precipitation area is detected, dividing the precipitation area into a first monitoring area; A division module is used to divide the first monitoring area with the meteorological monitoring station as the center to obtain at least three second monitoring areas of regular shapes, all of which are located in the first monitoring area, and the area of the second monitoring area is smaller than that of the first monitoring area; A collection module, used to calculate the precipitation intensity distribution at different locations in the second monitoring area to which the meteorological monitoring station belongs and the total precipitation in the second precipitation area through the precipitation data of the same time period monitored by each meteorological monitoring station; A judgment module, used for judging the precipitation change trend in the second monitoring area and the precipitation change trend between adjacent second monitoring areas based on the precipitation intensity distribution and total precipitation in each second monitoring area; An adjustment module, used to adjust the position of the second monitoring area based on the precipitation change trend between adjacent second monitoring areas, and simultaneously obtain the overlapping area of the second monitoring area; When the precipitation increase paths of the second monitoring areas are all directed toward the overlapping area of the second monitoring areas, the overlapping area is extracted as the precipitation center area; A feature extraction module, used for extracting precipitation-related features of the precipitation center area based on the location of the precipitation center area; The output module is used to receive precipitation-related features in the precipitation center area based on the trained machine learning model and output short-term forecast results of heavy rainfall.
Citation Information
Patent Citations
Short-time heavy rainfall typing method based on feature cross fusion
CN116628626A
Geological disaster monitoring system and method based on big data
CN116758707A
Regional thunderstorm trend prediction system
CN116894161A
Rainfall measuring unit and method of attaching rain gauge
JP2011069644A
