Short-time heavy rainfall early warning management method based on meteorological indexes
Through the short-term heavy rainfall warning management method based on recurrent neural network, the meteorological data training model is used to predict the probability of short-term heavy rainfall, which solves the problem of unsatisfactory forecasting effect in the existing technology, and achieves higher forecast accuracy and timeliness , guarantees agriculture and life safety.
Patent Information
- Application Number
- CN202411189628.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-08-28
AI Technical Summary
Short-term heavy rainfall is sudden, local, uncertain and random, resulting in unsatisfactory results of existing meteorological forecasts and it is difficult to effectively warn and prevent disasters caused.
Using a short-term heavy rainfall warning management method based on recurrent neural network, we can collect and preprocess meteorological data, determine multiple types of forecast indicators and their weights, train neural network models, use real-time meteorological data to predict the probability of short-term heavy rainfall, and improve forecast accuracy rate and timeliness.
It has improved the accuracy and timeliness of short-term and near-term forecasts of heavy rainfall, protected grain production, safeguarded people's lives and property, and promoted local economic development and social stability.
Smart Images

Figure CN119001925A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of meteorological forecasting, and in particular to a short-term heavy rainfall early warning management method based on meteorological indicators. Background Art
[0002] Short-term heavy rainfall is one of the main meteorological disasters and one of the difficulties in weather forecasting. The characteristics of this weather phenomenon include short duration, heavy rain and strong locality. The heavy rainfall formed by the accumulation of short-term heavy rainfall can form flash floods in a short period of time, causing waterlogging in farmland, urban waterlogging, and even triggering geological disasters such as mudslides and mountain torrents, causing major economic losses and casualties. Short-term heavy rainfall has seriously affected urban life and transportation, and may also cause serious disasters such as mudslides or floods, causing serious harm to people’s lives and property. Therefore, it is of great significance to strengthen its mechanism research and near-term monitoring and early warning. However, due to the characteristics of suddenness, locality, uncertainty, randomness and short duration of short-term heavy rainfall, the forecast effect has not been very ideal for a long time. Summary of the invention
[0003] In order to solve the problems raised in the above background technology, the technical solution adopted by the present invention is:
[0004] A short-term heavy rainfall early warning management method based on meteorological indicators comprises the following steps:
[0005] S1. Collect local meteorological data, establish meteorological data sets, and classify and preprocess the meteorological data sets;
[0006] S2. By analyzing the characteristics of the meteorological data set, multiple forecast indicators for local short-term heavy rainfall weather are determined, and the weight of each forecast indicator is determined;
[0007] S3, based on recurrent neural network, a neural network model for objective forecasting of short-term heavy rainfall is obtained through training of meteorological data sets;
[0008] S4. When issuing a short-term heavy rainfall warning, real-time meteorological data is collected and input into a neural network model. In the neural network model, for the input meteorological data, its matching score with each type of forecast indicator is determined, and then all matching scores are weightedly added to obtain a comprehensive score, and then the predicted short-term heavy rainfall probability is obtained based on the comprehensive score.
[0009] In some embodiments, in step S1 , the collected meteorological data includes micaps data, automatic weather station data, and Doppler radar product data.
[0010] In some embodiments, in step S1, when collecting local meteorological data and establishing a meteorological data set, specifically:
[0011] Randomly select meteorological data from different periods and stations over several years;
[0012] In addition, the short-term heavy rainfall weather processes that occurred in the local area over the past few years were collected, and the regularity, regional and temporal characteristics of the short-term heavy rainfall were preliminarily analyzed. According to the site and time of the short-term heavy rainfall, the meteorological data for the period from the preset time before the short-term heavy rainfall began to the end of the short-term heavy rainfall were obtained;
[0013] A meteorological data set is established based on the collected meteorological data, and in the meteorological data set, different types of meteorological data in the same period are temporally associated.
[0014] In some embodiments, in step S1, when the meteorological data set is classified, specifically, the data in the meteorological data set is classified according to different types, sites and times;
[0015] When preprocessing the meteorological dataset, the size, scale, and center position of the images in the dataset are adjusted, and grayscale processing is performed.
[0016] In some embodiments, in step S2, when performing feature analysis on a meteorological data set, a feature detection algorithm is used through a combination of computer and manual processing to extract feature points or feature areas from the image to assist in image matching, recognition, and edge detection tasks, thereby identifying the small and medium-scale weather systems that cause local short-term heavy rainfall, and classifying them, determining the meteorological data features corresponding to small and medium-scale weather systems of different classifications as forecast indicators, determining the weight of each type of forecast indicator, and, for each type of forecast indicator at each site, selecting several groups of typical meteorological data for the site as evaluation criteria.
[0017] In some embodiments, in step S3, a neural network model corresponding to each site is established respectively, and when establishing the neural network model, a neural network model combining CNN and LSTM is used. During model training, different types of meteorological data of the site at the same time period are input, and characteristic information of different types of meteorological data is extracted through the attention mechanism, and matched with the typical meteorological data corresponding to each type of preset forecast indicator to obtain a matching score, and then all matching scores are weightedly added to obtain a comprehensive score, and then the probability of short-term heavy rainfall is obtained based on the comprehensive score.
[0018] In some embodiments, after step S3 and before step S4, a k-fold cross-validation method is also used to verify the accuracy of the neural network model. If the accuracy is greater than a preset threshold, step S4 is entered, otherwise, step S2 is returned to adjust the forecast indicators, the weights of the forecast indicators and / or the model parameters, and the model is retrained until the accuracy reaches the preset threshold.
[0019] Compared with the prior art, the present invention has the following beneficial effects:
[0020] The short-term heavy rainfall early warning management method based on meteorological indicators provided by the present invention can improve the accuracy and timeliness of short-term and near-term forecasts of short-term heavy rainfall, which is beneficial to protecting food production, and has very important practical and long-term significance for improving agricultural and industrial production, protecting people’s lives and property, and promoting local economic development and social stability. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Figure 1 A schematic diagram of the process of the short-term heavy rainfall early warning management method based on meteorological indicators provided by the present invention;
[0022] Figure 2-Figure 19 It is a meteorological data diagram in a specific embodiment. DETAILED DESCRIPTION
[0023] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the following further describes how the present invention is implemented in conjunction with the accompanying drawings and specific implementation methods.
[0024] Reference Figure 1 As shown, the present invention provides a short-term heavy rainfall early warning management method based on meteorological indicators, comprising the following steps:
[0025] S1. Collect local meteorological data, establish a meteorological data set, and classify and preprocess the meteorological data set.
[0026] S2. By analyzing the characteristics of the meteorological data set, multiple types of forecast indicators for local short-term heavy rainfall weather are determined, and the weight of each type of forecast indicator is determined.
[0027] S3. Based on the recurrent neural network, a neural network model for objective forecasting of short-term heavy rainfall is obtained through training with meteorological data sets.
[0028] S4. When issuing a short-term heavy rainfall warning, real-time meteorological data is collected and input into a neural network model. In the neural network model, for the input meteorological data, its matching score with each type of forecast indicator is determined, and then all matching scores are weightedly added to obtain a comprehensive score, and then the predicted short-term heavy rainfall probability is obtained based on the comprehensive score.
[0029] It is understandable that short-term heavy rainfall warning is highly regional, so data needs to be collected separately for different regions, and the judgment criteria may also be different. The following takes the situation in Tongliao, Inner Mongolia as an example to introduce a specific embodiment of the present invention.
[0030] In step S1, local meteorological data is collected, a meteorological data set is established, and the meteorological data set is classified and preprocessed.
[0031] In this embodiment, in step S1, meteorological data of Tongliao area in Inner Mongolia for several years (2008-2023 in this embodiment) are extracted, and the collected meteorological data include micaps data, automatic weather station data and Doppler radar product data.
[0032] In this embodiment, when collecting meteorological data, meteorological data of different time periods and different stations over several years are randomly selected; and short-term heavy rainfall weather processes occurring in the local area over several years are collected, and the regularity, regional and temporal characteristics of the occurrence of short-term heavy rainfall are preliminarily analyzed. According to the station and time where the short-term heavy rainfall occurs, meteorological data within the time period from a preset time (such as 24 hours) before the start of the short-term heavy rainfall to the end of the short-term heavy rainfall is obtained; a meteorological data set is established based on the collected meteorological data, and in the meteorological data set, meteorological data of different types in the same time period are temporally associated.
[0033] When collecting short-term heavy rainfall weather processes, search and filter the hourly precipitation to find out the short-term heavy rainfall processes that have occurred. When preliminarily analyzing the regularity and regional and temporal characteristics of short-term heavy rainfall, we can analyze and summarize from the aspects of spatial distribution, interannual changes, and daily changes, as follows.
[0034] (1) Spatial distribution
[0035] From the analysis of the number of short-term heavy rainfall in various places, the most is in Zarut Banner in the northern part of Tongliao, Inner Mongolia, followed by Naiman Banner (including Qinglongshan Station), then Kezuozhong Banner and Kulun Banner, and the least is in Horqin District and Holingol City: the short-term heavy rainfall in Tongliao is mainly distributed in the northern and southern mountainous areas, and relatively less in the plains. The reason why Holingol City is relatively small is that the area is smaller than that of other banners and counties and there are fewer regional automatic stations; the distribution characteristics of short-term heavy rainfall in various parts of the city are inseparable from the local terrain distribution characteristics of the rocky mountains and hills of the southern foot of the Greater Khingan Mountains in the north, the shallow mountains and loess hills on the edge of the western Liaoning mountains in the south, and the sandy alluvial plains in the middle of the Xiliaohe River Basin (Tongliao City has high terrain in the south and north, low and flat in the middle, and is saddle-shaped). The dynamic uplift of the terrain and the uneven thermal convection in the mountainous areas are important factors in the formation of multiple short-term heavy rainfall distribution.
[0036] (2) Interannual variation
[0037] If one or more of the national and regional stations in Tongliao experience short-term heavy rainfall in one day, it is defined as a short-term heavy rainfall day, and the average number of short-term heavy rainfall days in the studied years is 15.2. If a station has a short-term heavy rainfall on one day, it is counted as a station day, and the annual average is 40.6 station days. The study found that in 2009 and 2010, the number of short-term heavy rainfall days and station days showed a downward trend, while after 2011, the number of short-term heavy rainfall days and station days showed an upward trend, especially the station days showed a more obvious increase than the short-term heavy rainfall days. On the one hand, this is because many regional automatic stations have been newly deployed in Tongliao in recent years, and the spatial ability to detect short-term heavy rainfall has become more refined; on the other hand, it shows that with global warming, extreme weather and climate events are gradually becoming more frequent, and the scope of the impact of short-term heavy rainfall processes is gradually expanding.
[0038] (3) Seasonal changes
[0039] Tongliao area can experience short-term heavy rainfall from April to October, but it is mainly concentrated in June to September, accounting for 98% of the total short-term heavy rainfall days, of which June to July accounts for 65%. The earliest occurred on April 23 and the latest on October 10. July is the month with the most short-term heavy rainfall, followed by June. The peak of short-term heavy rainfall (June and July) is before the peak of rainfall in Tongliao (July and August). This is closely related to the formation characteristics of short-term heavy rainfall and the seasonal advance and retreat of subtropical high pressure.
[0040] (4) Daily variation
[0041] Statistics from all observation stations in Tongliao show that short-term heavy rainfall in Tongliao occurs at every moment, mainly concentrated between 11:00 and 23:00, accounting for 76% of the total, and short-term heavy rainfall between 12:00 and 20:00 accounts for 64% of the total. It can be seen that short-term heavy rainfall in Tongliao is highly concentrated in the afternoon and evening. This shows that in the period from afternoon to evening when the daily temperature change is the largest, it is the fact that the turbulence causes the instability of the atmospheric stratification, which is more likely to cause severe convective weather. The occurrence of short-term heavy rainfall has a close positive correlation with the occurrence of severe convective weather.
[0042] In addition, when classifying the meteorological data set, specifically, the data in the meteorological data set is classified according to different types, sites and times; when preprocessing the meteorological data set, the size, scale, and center position of the images in the data set are adjusted, and grayscale processing is performed to facilitate the subsequent establishment and training of the recurrent neural network.
[0043] In step S2, by performing feature analysis on the meteorological data set, multiple types of forecast indicators for local short-term heavy rainfall weather are determined, and the weight of each type of forecast indicator is determined.
[0044] In this embodiment, when performing feature analysis on a meteorological data set, a feature detection algorithm is used by combining a computer with manual processing to extract feature points or areas of certain importance or representativeness from the image to help perform tasks such as image matching, recognition, and edge detection, find out the small and medium-scale weather systems that cause local short-term heavy rainfall, and classify them, determine the meteorological data features corresponding to small and medium-scale weather systems of different classifications as forecast indicators, determine the weight of each type of forecast indicator, and, for each type of forecast indicator at each site, select several groups of typical meteorological data of the site as evaluation criteria.
[0045] In this embodiment, they are classified into: mesoscale convergence shear line, ground mesoscale convergence line, mesoscale low-vortex circulation, low-level jet and ultra-low-level jet, thunderstorm cell, thunderstorm group (including squall line), gust front, mesocyclone or mesoscale convergence zone (headwind zone), etc.
[0046] Most of the short-term heavy rainfall in Tongliao is directly caused by small and medium-scale systems embedded in the weather scale system under certain large-scale circulation situations. Since the encrypted wind field data of the automatic weather station and the Doppler radar product data have higher temporal and spatial resolution, the activities of small and medium-scale systems can be better monitored. Therefore, the small and medium-scale systems that cause short-term heavy rainfall are classified and summarized using the micaps data, the encrypted wind field data of the automatic weather station and the Doppler radar product data:
[0047] (1) Mesoscale convergence shear line
[0048] The long-term convergence of small and medium-scale wind fields enhances the convergence lift on the ground, and is also a mesoscale system and triggering condition for local large-scale precipitation. During the short-term heavy rainfall in Tongliao area, the discontinuity of the wind field or the maximum curvature of the cyclonic bending, i.e., the convergence line, is often analyzed in the middle and lower layers of the troposphere. Figure 2 As shown in the figure, it is the 850hPa convergence shear line at 20:00 on July 11, 2009, corresponding to the short-term heavy rainfall process in Holingol City; Figure 3 As shown in the figure, it is the 925hPa convergence shear line at 08:00 on July 13, 2009, corresponding to the short-term heavy rainfall process in Kezuozhong Banner. In addition to the convergence shear of wind direction, it also includes the convergence of wind speed, such as Figure 4 Shown is the 700hPa wind speed convergence line at 08:00 on July 23, 2008, corresponding to the short-term heavy rainfall process in Naiman Banner.
[0049] (2) Ground mesoscale convergence lines
[0050] By using the hourly wind field data of the encrypted automatic station, the effect of the ground wind field on short-term heavy rainfall was analyzed. The results showed that during the short-term heavy rainfall, the ground was always accompanied by the influence of mesoscale convergence lines with a length of tens to 200 km. The convergence lines showed both wind direction and wind speed convergence, which corresponded well to the heavy rainfall area. Figure 5 As shown in the figure, it is the ground convergence line at 20:00 on August 24, 2008, corresponding to the short-term heavy rainfall process in Kailu; Figure 6 As shown, this is the wind field convergence line of the automatic station at 19:00 on September 24, 2016, corresponding to the short-term heavy rainfall process in Kezuozhong Banner.
[0051] (3) Mesoscale low-vortex circulation
[0052] During the short-term heavy rainfall in Tongliao area, mesoscale low-vortex circulation is often analyzed in the wind field in the middle and lower layers of the troposphere. Figure 7 As shown in Figure 2, it is the 925hPa low vortex circulation at 08:00 on June 25, 2008, corresponding to the short-term heavy rainfall process in Horqin area; Figure 8 As shown, this is the 925hPa low vortex circulation at 08:00 on August 27, 2009, corresponding to the short-term heavy rainfall process in Kulun Banner.
[0053] (4) Low-level jets and ultra-low-level jets
[0054] During the short-term heavy rainfall in Tongliao area, low-level jets and ultra-low-level jets are often analyzed in the lower troposphere wind field. Fig. 9 As shown in the figure, it is the 850hPa low-level jet at 20:00 on June 27, 2009, corresponding to the short-term heavy rainfall process in Bayar, Zhalute Banner; Fig.10 As shown, this is the 925hPa ultra-low-level jet at 08:00 on July 16, 2010, corresponding to the short-term heavy rainfall process in Horqin District.
[0055] (5) Thunderstorm cells
[0056] like Fig.11 As shown in the figure, it is the reflectivity factor map at 14:07 on July 5, 2019, corresponding to the short-term heavy rainfall process in Zhalute Banner; Fig.12 As shown, this is the reflectivity factor map at 11:44 on August 21, 2015, corresponding to the short-term heavy rainfall process in Zhalute Banner.
[0057] (6) Thunderstorm groups (including squall lines)
[0058] A thunderstorm group is composed of multiple individual thunderstorms, such as Fig.13 As shown in the figure, it is the reflectivity factor map at 12:40 on August 20, 2015, corresponding to the short-term heavy rainfall process in Zhalute Banner; sometimes multiple thunderstorm cells are arranged linearly to form a squall line system, such as Fig.14As shown, this is the reflectivity factor map at 21:42 on July 25, 2019, corresponding to the short-term heavy rainfall process in Kezuohou Banner.
[0059] (7) Gust front
[0060] In many cases, strong echoes are accompanied by gust fronts, such as Fig.15 As shown in the figure, it is the reflectivity factor map at 21:58 on July 25, 2019, corresponding to the short-term heavy rainfall process in Kezuohou Banner; Fig.16 As shown, this is the reflectivity factor map at 17:10 on July 17, 2010, corresponding to the short-term heavy rainfall process in Kezuozhong Banner.
[0061] (8) Mesocyclone or mesoscale convergence zone (adverse wind zone)
[0062] Sometimes, obvious and strong convergence field features can be detected on the radial velocity diagram, reaching the characteristics of "mesocyclone", such as: Fig.17 As shown in the figure, it is the radial velocity diagram at 13:09 on August 1, 2011, corresponding to the short-term heavy rainfall process in Kezuozhong Banner; sometimes there will be "adverse wind areas", such as Fig.18 As shown in the figure, it is the radial velocity diagram at 18:18 on July 22, 2009, corresponding to the short-term heavy rainfall process in Naiman Banner; Fig.19 As shown, this is the radial velocity map at 12:02 on July 5, 2019, corresponding to the short-term heavy rainfall process in Kezuohou Banner.
[0063] The above pictures of different classifications can be stored as typical meteorological data after preprocessing and used as subsequent evaluation criteria.
[0064] In step S3, based on a recurrent neural network, a neural network model for objective forecasting of short-term heavy rainfall is obtained by training the meteorological data set.
[0065] In this embodiment, a neural network model corresponding to each site is established respectively, and when establishing the neural network model, a neural network model combining CNN and LSTM can be used. Moreover, when training the model, different types of meteorological data of the site at the same time period are input, and characteristic information of different types of meteorological data are extracted respectively through the attention mechanism, and matched with the typical meteorological data corresponding to each type of preset forecast indicator to obtain a matching score, and then all matching scores are weightedly added to obtain a comprehensive score, and then the probability of short-term heavy rainfall is obtained according to the comprehensive score.
[0066] In this embodiment, after step S3 and before step S4, a k-fold cross-validation method is also used to verify the accuracy of the neural network model. If the accuracy is greater than a preset threshold, step S4 is entered, otherwise the process returns to step S2, and after adjusting the forecast indicators, the weights of the forecast indicators and / or the model parameters, the model is retrained until the accuracy reaches the preset threshold.
[0067] When verifying the accuracy of the neural network model, the verification standard is as follows: In each verification of the k-fold cross-validation, for different types of meteorological data input at the same time period: when the meteorological data corresponds to the actual occurrence of short-term heavy rainfall, if the probability of short-term heavy rainfall calculated by the model is greater than 70%, then the model is considered to have accurately judged the short-term heavy rainfall, otherwise it is considered that the model has missed the judgment. When the meteorological data does not correspond to the actual occurrence of short-term heavy rainfall, if the probability of short-term heavy rainfall calculated by the model is greater than 70%, then the model is considered to have misjudged. When both the missed judgment rate and the misjudgment rate are less than the preset threshold (such as 10%), the accuracy of the model is considered to be qualified.
[0068] In this embodiment, in step S4, when a short-term heavy rainfall warning is issued, real-time meteorological data is collected and input into the neural network model corresponding to the site. In the neural network model, for the input meteorological data, its matching score with each type of forecast indicator is determined, and then all matching scores are weightedly added to obtain a comprehensive score, and then the predicted probability of short-term heavy rainfall is obtained based on the comprehensive score.
[0069] In summary, the short-term heavy rainfall warning management method based on meteorological indicators provided by the present invention can improve the accuracy and timeliness of short-term and near-term forecasts of short-term heavy rainfall, which is beneficial to protecting food production, and has very important practical and long-term significance for improving agricultural and industrial production, protecting people’s lives and property, and promoting local economic development and social stability.
[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution 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 solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.
Claims
1. A short-term heavy rainfall early warning management method based on meteorological indicators, characterized in that: The steps include: S1. Collect local meteorological data, establish meteorological data sets, and classify and preprocess the meteorological data sets; S2. By analyzing the characteristics of the meteorological data set, multiple forecast indicators for local short-term heavy rainfall weather are determined, and the weight of each forecast indicator is determined; S3, based on recurrent neural network, a neural network model for objective forecasting of short-term heavy rainfall is obtained through training of meteorological data sets; S4. When issuing a short-term heavy rainfall warning, real-time meteorological data is collected and input into a neural network model. In the neural network model, for the input meteorological data, its matching score with each type of forecast indicator is determined, and then all matching scores are weightedly added to obtain a comprehensive score, and then the predicted short-term heavy rainfall probability is obtained based on the comprehensive score.
2. The short-term heavy rainfall early warning management method based on meteorological indicators according to claim 1 is characterized in that: In step S1, the collected meteorological data include micaps data, automatic weather station data and Doppler radar product data.
3. The short-term heavy rainfall early warning management method based on meteorological indicators according to claim 2 is characterized in that: In step S1, local meteorological data is collected to establish a meteorological data set, specifically: Randomly select meteorological data from different periods and stations over several years; In addition, the short-term heavy rainfall weather processes that occurred in the local area over the past few years were collected, and the regularity, regional and temporal characteristics of the short-term heavy rainfall were preliminarily analyzed. According to the site and time of the short-term heavy rainfall, the meteorological data for the period from the preset time before the short-term heavy rainfall began to the end of the short-term heavy rainfall were obtained; A meteorological data set is established based on the collected meteorological data, and in the meteorological data set, different types of meteorological data in the same period are temporally associated.
4. The short-term heavy rainfall early warning management method based on meteorological indicators according to claim 3 is characterized in that: In step S1, when classifying the meteorological data set, specifically, the data in the meteorological data set is classified according to different types, sites and times; When preprocessing the meteorological dataset, the size, scale, and center position of the images in the dataset are adjusted, and grayscale processing is performed.
5. The short-term heavy rainfall early warning management method based on meteorological indicators according to claim 4 is characterized in that: In step S2, when the meteorological data set is subjected to feature analysis, a feature detection algorithm is used by combining computer processing with manual processing to extract feature points or feature areas from the image to help perform image matching, recognition, and edge detection tasks, find out the small and medium-scale weather systems that cause local short-term heavy rainfall, and classify them. The meteorological data features corresponding to the small and medium-scale weather systems of different classifications are determined as forecast indicators, and the weight of each type of forecast indicator is determined. In addition, for each type of forecast indicator at each station, several groups of typical meteorological data of the station are selected as evaluation criteria.
6. The short-term heavy rainfall early warning management method based on meteorological indicators according to claim 5 is characterized in that: In step S3, a neural network model corresponding to each site is established respectively, and when establishing the neural network model, a CNN combined with LSTM neural network model is adopted. During model training, different types of meteorological data of the site at the same time period are input, and characteristic information of different types of meteorological data is extracted through the attention mechanism, and matched with the typical meteorological data corresponding to each type of preset forecast indicator to obtain a matching score, and then all matching scores are weightedly added to obtain a comprehensive score, and then the probability of short-term heavy rainfall is obtained according to the comprehensive score.
7. The short-term heavy rainfall early warning management method based on meteorological indicators according to claim 6 is characterized in that: After step S3 and before step S4, the k-fold cross-validation method is also used to verify the accuracy of the neural network model. If the accuracy is greater than the preset threshold, step S4 is entered, otherwise, step S2 is returned to adjust the forecast indicators, the weights of the forecast indicators and / or the model parameters, and the model is retrained until the accuracy reaches the preset threshold.
Citation Information
Patent Citations
Interpretable short critical extreme rainfall prediction method based on attention mechanism
CN111815037A
Short-term approaching rainfall forecasting method based on double-flow convolution long short-term memory network
CN113610329A
Short-time rainfall prediction method, device and system based on bidirectional LSTM network
CN115907201A
Rainfall disaster early warning method and related equipment
CN116596131A
Short-time heavy rainfall probability forecast fusion method based on cyclic convolutional neural network
CN117290810A
Cited By
Radar reflectivity factor forecasting method based on FY-4A satellite
CN120370439A
