A method for identifying convection development based on networked wind profiler radar
Through the networked wind profile radar and deep neural network model, the development of strong convective weather is identified, and the problem of insufficient prediction of convective development in the existing technology is solved, and efficient monitoring and forecasting of convective weather is achieved.
Patent Information
- Application Number
- CN202210300833.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-25
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-03-25
AI Technical Summary
The existing technology lacks effective convective development prediction methods, making it difficult to accurately monitor and predict the occurrence and development of strong convective weather.
The identification method based on networked wind profile radar is adopted, and the atmospheric dynamic parameters are calculated through wind profile radar networking and finite element method, and the deep neural network model is constructed, and the functional mapping relationship of convective weather development is fitted to achieve the identification of convective development.
It improves the real-time monitoring and short-term forecasting capabilities of convective weather, can accurately capture the changes in the vertical dynamics of the atmosphere, and provides a detailed structural understanding and early warning reference for strong convective weather.
Smart Images

Figure CN114660610B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of meteorological forecasting, and in particular to a convection development identification method based on networked wind profile radar. Background Art
[0002] Severe convective weather is a weather phenomenon caused by strong vertical movement of the atmosphere. It is very destructive and is often accompanied by disasters such as heavy precipitation, strong winds and thunderstorms. For example, short-term heavy precipitation and convective strong winds are typical severe convective weather. Weather radars using Doppler and phased array technology can effectively monitor them. At present, combining professional theories and methods can analyze a large amount of valuable weather information such as the spatial location, moving speed and development trend of the storm, as well as the intensity of strong winds or precipitation. However, there is still a lack of effective prediction methods for the occurrence and development of severe convection. Summary of the invention
[0003] The purpose of the present invention is to solve the shortcomings existing in the prior art and to propose a convection development identification method based on networked wind profile radar.
[0004] In order to achieve the above object, the present invention adopts the following technical scheme: a method for identifying convection development based on a networked wind profiler radar, the method for identifying convection development based on a networked wind profiler radar comprising the following steps:
[0005] S1, wind profiler radar networking, first screen the wind profiler radar networking, the number of networked wind profiler radars is four or more, and the triangle method is used to determine the optimal networking site;
[0006] S2, calculates the data, and then uses the finite element method to traverse the triangle network, calculates the layered horizontal divergence and relative vorticity, generates the spatial distribution map of the horizontal divergence of each altitude layer, and generates the vertical profile map of the divergence and vorticity;
[0007] S3, data set construction, which includes divergence spatial distribution map, divergence and vorticity vertical profile map, wind profile map, and rainfall and wind force measured data of ground meteorological stations within the network range to form a data set;
[0008] S4, build a deep neural network model, build a deep neural network model based on convolutional neural network, and fit the functional mapping relationship between the image data sets such as the scatter diagram and profile diagram at the previous moment and the development of convective weather through iterative training of a large number of sample "data pairs";
[0009] S5, identify, apply and improve the function mapping relationship obtained by fitting to realize the identification of convection development.
[0010] In order to reduce the error, the present invention has the following improvements: in step S1, the triangle method is to let the sites form an equilateral triangle, which is the optimal networking form.
[0011] In order to accurately calculate, the present invention is improved in that, in step S2, the specific calculation method is: according to the determined longitude and latitude of the three vertex stations of the triangle and the average radius R of the earth, the components of the three sides of the triangle are calculated, and the horizontal divergence D and the relative vorticity ζ in the height triangle are calculated in combination with the horizontal wind data of the three stations in each layer;
[0012] Then, the networked triangles are traversed to obtain the horizontal divergence and relative vorticity values at each altitude layer above each triangle, and a spatial distribution diagram of the horizontal divergence is drawn. Then, the horizontal divergence and relative vorticity values calculated at each altitude layer are used to draw a vertical profile diagram of the divergence and vorticity.
[0013] In order to ensure the prediction effect of the deep neural network model, the present invention is improved in that, in step S4, the deep neural network model includes 2 convolutional layers, 2 maximum pooling layers and 2 fully connected layers.
[0014] In order to optimize the data set, the present invention has the following improvements: in step S4, the image data set can be enhanced and optimized by using geometric transformation.
[0015] In order to further optimize the prediction effect of the deep neural network model, the present invention has the following improvements: in step S5, the network topology structure is reasonably modified and configured to further enhance the model recognition capability.
[0016] Compared with the existing technology, the present invention utilizes wind profiler radar mesoscale network observations, and the horizontal wind at different heights can invert atmospheric dynamic parameters such as divergence, vorticity, wind shear and vertical velocity with higher temporal and spatial resolution. It can more accurately capture the characteristics of atmospheric vertical dynamic changes before convection is triggered, and can obtain real-time three-dimensional divergence and vorticity distribution, which helps meteorological forecasters understand the fine structure of the weather system and quickly find areas of strong upward motion for short-term and nowcasting of heavy precipitation, providing an important reference for severe convective weather monitoring and early warning and short-term and nowcasting, with high practicality and significant progress.
[0017] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.
[0018] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0020] Figure 1 A schematic flow chart of a method for identifying convection development based on a networked wind profiler radar is proposed for the present invention;
[0021] Figure 2 The present invention proposes a deep neural network model architecture diagram in a convective development identification method based on a networked wind profiler radar. DETAILED DESCRIPTION
[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.
[0023] Embodiment 1
[0024] See also Figure 1-2 A method for identifying convection development based on a networked wind profiler radar is provided. The method for identifying convection development based on a networked wind profiler radar comprises the following steps:
[0025] S1, wind profiler radar networking, first screen the wind profiler radar networking, the number of networked wind profiler radars is four or more, and the triangle method is used to determine the optimal networking site;
[0026] S2, calculates the data, and then uses the finite element method to traverse the triangle network, calculates the layered horizontal divergence and relative vorticity, generates the spatial distribution map of the horizontal divergence of each altitude layer, and generates the vertical profile map of the divergence and vorticity;
[0027] S3, data set construction, which includes divergence spatial distribution map, divergence and vorticity vertical profile map, wind profile map, and rainfall and wind force measured data of ground meteorological stations within the network range to form a data set;
[0028] S4, build a deep neural network model, build a deep neural network model based on convolutional neural network, and fit the functional mapping relationship between the image data sets such as the scatter diagram and profile diagram at the previous moment and the development of convective weather through iterative training of a large number of sample "data pairs";
[0029] S5, identify, apply and improve the function mapping relationship obtained by fitting to realize the identification of convection development.
[0030] In this embodiment, in step S1, the triangle method is to let the sites form an equilateral triangle. The equilateral triangle is the optimal networking form because when the maximum internal angle of the triangle increases, the data and shape correlation error will also increase, so the equilateral triangle is the optimal networking form.
[0031] In this embodiment, in step S2, the specific calculation method is: according to the determined longitude and latitude of the three vertex stations of the triangle and the average radius R of the earth, the components of the three sides of the triangle are calculated, and the horizontal divergence D and the relative vorticity ζ in the height triangle are calculated in combination with the horizontal wind data of the three stations in each layer;
[0032] Then, the networked triangles are traversed to obtain the horizontal divergence and relative vorticity values at each altitude layer above each triangle, and a spatial distribution diagram of the horizontal divergence is drawn. Then, the horizontal divergence and relative vorticity values calculated at each altitude layer are used to draw a vertical profile diagram of the divergence and vorticity.
[0033] In this embodiment, in step S4, the deep neural network model includes 2 convolutional layers, 2 maximum pooling layers and 2 fully connected layers;
[0034] A deep neural network model suitable for convective development identification was built, where the input is a three-dimensional vector with a length and width of 100 pixels, consisting of three images (a horizontal divergence spatial distribution image, a divergence and vorticity vertical profile image, and a wind profile image);
[0035] After two convolutions and two poolings, it is finally sent to two fully connected layers. The activation functions used in the convolutional layer and the fully connected layer are both ReLU functions. In addition, to avoid overfitting, the fully connected layer makes 20% of the neurons dormant.
[0036] In this embodiment, in step S4, the image data set can be enhanced and optimized by using geometric transformation. Geometric transformation refers to a mapping from a set with a geometric structure to itself or other such sets. By making certain transformations to the graphics, it will be helpful to discover the implicit conditions of the problem.
[0037] In this embodiment, in step S5, the network topology structure is reasonably modified and configured to further enhance the model recognition capability and improve the model recognition accuracy.
[0038] It can be seen from the above embodiments that the present invention utilizes wind profiler radar mesoscale network observations, and horizontal winds at different heights can invert atmospheric dynamic parameters such as divergence, vorticity, wind shear and vertical velocity with higher temporal and spatial resolution, which can more accurately capture the characteristics of atmospheric vertical dynamic changes before convection is triggered, and can obtain real-time three-dimensional divergence and vorticity distribution, which helps meteorological forecasters understand the fine structure of the weather system and quickly find areas of strong upward motion for short-term nowcasting of heavy precipitation, providing an important reference for severe convective weather monitoring and early warning and short-term nowcasting, and has high practicality and significant progress.
[0039] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A method for identifying convection development based on networked wind profiler radars. It is characterized in that The convection development identification method based on networked wind profiler radar comprises the following steps: S1, wind profiler radar networking, first screen the wind profiler radar networking, the number of wind profiler radars in the network is more than four, and the triangle method is used to determine the optimal networking site; S2, calculates the data, and then uses the finite element method to traverse the triangle network, calculates the layered horizontal divergence and relative vorticity, generates the spatial distribution map of the horizontal divergence of each altitude layer, and generates the vertical profile map of the divergence and vorticity; S3, data set construction, which includes divergence spatial distribution map, divergence and vorticity vertical profile map, wind profile map, and rainfall and wind force measured data of ground meteorological stations within the network range to form a data set; S4, build a deep neural network model, build a deep neural network model based on convolutional neural network, and fit the functional mapping relationship between the scatter diagram and profile diagram image data set at the previous moment and the development of convective weather through iterative training of a large number of sample "data pairs"; S5, identify, apply and improve the function mapping relationship obtained by fitting to realize the identification of convection development; In step S2, the specific calculation method is: according to the longitude and latitude of the three vertex stations of the triangle and the average radius R of the earth, the components of the three sides of the triangle are calculated, and the horizontal divergence D and relative vorticity ζ in the height triangle are calculated in combination with the horizontal wind data of the three stations in each layer; Then, the networked triangles are traversed to obtain the horizontal divergence and relative vorticity values at each altitude layer above each triangle, and a spatial distribution diagram of the horizontal divergence is drawn. Then, the horizontal divergence and relative vorticity values calculated at each altitude layer are used to draw a vertical profile diagram of the divergence and vorticity.
2. According to the method for identifying convection development based on networked wind profiler radar according to claim 1, Features: In step S1, the triangle method is to make the sites form an equilateral triangle, which is the optimal networking form.
3. According to the method for identifying convection development based on networked wind profiler radar according to claim 1, Features: In step S4, the deep neural network model includes 2 convolutional layers, 2 maximum pooling layers and 2 fully connected layers.
4. The method for identifying convection development based on networked wind profiler radar according to claim 1, Features: In step S4, the image data set is enhanced and optimized by using a geometric transformation method.
5. The method for identifying convection development based on networked wind profiler radar according to claim 1, Features: In step S5, the network topology structure is reasonably modified and configured to further improve the model recognition capability.
