Wind field inversion method based on X-band radar and numerical mode

Abstract

The invention provides a wind field inversion method based on an X-band radar and a numerical mode, and the method comprises the following steps: firstly, taking a three-dimensional wind field predicted based on a 3-kilometer numerical mode of a coarse resolution as an initial field of wind field inversion in a wind field inversion process; then, high-resolution S-band and X-band radar observation data are introduced, the radar reflectivity is subjected to quality control, and radial wind is subjected to speed deblurring processing; then combining the coarse resolution mode data and the high resolution radar data to perform data fusion; introducing a conservation equation as a constraint condition and calculating a loss function in a mode of solving the vertical velocity in the high-precision three-dimensional wind field; and finally, iteratively solving the optimal solution of the loss function based on an optimization algorithm to obtain an inversion wind field. According to the technical scheme, the problem of lack of 1-2 kilometer low-altitude wind field monitoring of the S-band radar is solved.

Description

Technical Field

[0001] This application belongs to the fields of atmospheric science and remote sensing technology, and relates to a wind field inversion method based on X-band radar and numerical models. Background Technology

[0002] In recent years, with the continuous development of detection technology, people's ability to monitor the Earth's atmosphere has been constantly improving. For example, in precipitation detection, the reflectivity provided by S-band radar, which is updated every 6 minutes, has been widely used in convection monitoring, precipitation monitoring inversion, and precipitation forecasting. It is also widely used in the inversion of low-level wind fields, playing an indispensable role in the monitoring of severe weather.

[0003] Unlike direct observation products, radar observations provide indirect measurements of particle quantity, morphology, and advection, such as reflectivity and radial wind speed. However, the detection range of a single, fixed ground-based radar is limited and insufficient to cover complex and changing weather systems. In addition, my country's terrain is extremely complex, and most operational radars generally experience varying degrees of obstruction at an elevation angle of 0.5 degrees, which limits the ability of the new generation of weather radar networks to detect low-level meteorological targets.

[0004] Doppler radar observations utilize changes in the power and phase of electromagnetic waves after scattering and reflection at particle surfaces to reflect movement. In some areas, radar observations occur above the water vapor condensation level (within the range of 2.5–4 km). On one hand, the quantity and morphology of particles detected by radar are used to determine rainfall intensity, a common principle in quantitative precipitation estimation. On the other hand, radar can monitor wind speed. While the particles detected by reflectivity do not directly determine wind speed, they act as tracers in the wind field, reflecting wind direction and intensity to some extent. Therefore, in traditional S-band radar wind field inversion processes, wind speeds above 2.5 km can typically only be retrieved in most areas. Thus, when using a single or multiple S-band radars, precipitation and wind field monitoring at 1–2 km altitudes has always been a "missing piece" in low-altitude monitoring. Figure 1 The diagram illustrates a blind zone in the volume scan region when using S-band radar to retrieve wind fields in existing technologies. (Reference) Figure 1 The volume scan area of ​​the S-band radar is an approximately gray semi-circular area, and the white double-arrow area is a blind zone that the S-band radar cannot detect. Summary of the Invention

[0005] This application provides a wind field inversion method based on X-band radar and numerical models to solve the problem of the lack of monitoring of precipitation and wind fields at low altitudes of 1-2 km using S-band radar.

[0006] In a first aspect, this application provides a wind field inversion method based on X-band radar and numerical models, comprising the following steps: during the wind field inversion process, observation data from X-band radar is introduced to establish an inversion wind field region covering low elevation angle space; the wind field predicted based on a 3-kilometer numerical model is used as the initial field for wind field inversion; the observation data from S-band radar and the X-band radar are combined with the initial field for data fusion to correct the initial field; the conservation equation is added as a constraint condition to the loss function, and the loss function is iteratively solved based on an optimization algorithm to obtain the inverted wind field.

[0007] In this application, during the wind field inversion process, observational data from X-band radar in the low-altitude wind field region is introduced to compensate for the limitations of single or multiple S-band radars, which can only invert wind speeds above 2.5 km and precipitation and wind field monitoring at 1-2 km due to altitude and observation angle restrictions. The X-band radar network observation data increases the accuracy of observations on the horizontal plane and compensates for the low elevation angle of the S-band radar on the vertical plane, achieving minute-by-minute temporal resolution. Furthermore, during the wind field inversion process, a forecast field from a numerical model (3 km) is incorporated as the initial field for wind field inversion to achieve real-time updates of the radar's three-dimensional variational assimilation.

[0008] In one implementation of the first aspect, the introduction of observation data from X-band radar to establish an inverted wind field region covering low elevation angle space includes: determining the spatial range of the inverted wind field region; determining the intersection regions between any two radars in the effective scanning areas of all S-band and X-band radars within the spatial range; and determining the center point of the inverted wind field region based on the coverage of all these intersection regions.

[0009] In one implementation of the first aspect, using the wind field predicted by the 3-kilometer numerical model as the initial field for wind field inversion includes: establishing a three-dimensional space in a Cartesian coordinate system based on the geopotential height field of isobaric surfaces in the model data. This three-dimensional space includes vertical height, latitude, and longitude, wherein the vertical height of the isobaric surfaces is as follows:

[0010]

[0011] in, This represents the potential height at time t. This represents the current altitude field. Represents the reference state in the mode integral;

[0012] Establish an initial three-dimensional wind field space in Cartesian coordinates based on the latitude and longitude coordinates in the model:

[0013]

[0014] Based on the center point of the inverted wind field region, a refined initial wind field tensor matrix is ​​established:

[0015]

[0016] in, Indicates the center point of the inverted wind field region. , and z represents the resolution along the tensor's X-axis, Y-axis, and vertical Z-axis, respectively. Represents the grid points on the X-axis of the fine-resolution tensor. Represents the grid points on the Y-axis of the tensor;

[0017] The velocity constraints for grid points in the numerical model are determined using the following formula:

[0018]

[0019] in, This represents the interpolation function, which uses nearest-neighbor interpolation to obtain a coarse-resolution wind field. Interpolation to fine resolution ,in This represents the initial field, i.e., the coordinates in the 3km numerical model. The value of the wind field.

[0020] In one implementation of the first aspect, the loss function is defined as:

[0021]

[0022] in, This indicates the wind field observed by S-band radar and X-band radar; The loss function representing the conservation of mass; The loss function representing vorticity conservation; The loss function representing a smooth linear constraint; This indicates a 3-kilometer numerical model constraint.

[0023] In one implementation of the first aspect, the wind field observed by the S-band radar and the X-band radar The specific formula is as follows:

[0024] in, x, y, and z represent the horizontal and vertical distances of each point in the Cartesian coordinate system from the radar station.

[0025] In one implementation of the first aspect, based on the law of conservation of mass

[0026]

[0027]

[0028] Define loss function The specific formula is as follows:

[0029]

[0030] in, For density.

[0031] In one implementation of the first aspect, the vertical vorticity conservation relation is:

[0032]

[0033] Calculate the loss function for vorticity conservation based on the stated vertical vorticity conservation relationship. The specific formula is as follows:

[0034]

[0035] in, .

[0036] In one implementation of the first aspect, the loss function of the smooth linear constraint The specific formula is as follows:

[0037]

[0038] in, W is the weighting coefficient.

[0039] In one implementation of the first aspect, the 3-kilometer numerical model constraint The specific formula is as follows:

[0040]

[0041]

[0042] Among them, the model wind field is Time and observation time Translation between them; To interpolate the wind field to the grid points of the forecast area, Translation position at any time The positional relationship with the whole hour is as follows:

[0043]

[0044] The wind field is:

[0045] in, , , The three-dimensional wind field is interpolated to the grid points of the forecast area.

[0046] In one implementation of the first aspect, solving the loss function includes:

[0047] in,

[0048]

[0049] right Solve the differential and calculate the multidimensional gradient from the tensor. scalar These are weighting coefficients used to determine the weighting of each cost function relative to the total cost function. relative contribution

[0050] Secondly, this application provides an electronic device, the electronic device comprising: a memory storing a computer program thereon; and a processor communicatively connected to the memory for executing the computer program in the above-described wind field inversion method based on X-band radar and numerical models.

[0051] Thirdly, this application provides a computer-readable storage medium storing a computer program that, when executed by an electronic device, implements the aforementioned wind field inversion method based on X-band radar and numerical models.

[0052] As described above, the wind field inversion method based on X-band radar and numerical models described in this application has the following beneficial effects:

[0053] In the wind field inversion process, observational data from X-band radar in the low-altitude wind field region is introduced to compensate for the limitations of single or multiple S-band radars, which can only invert wind speeds above 2.5 km and precipitation and wind field monitoring at 1-2 km due to altitude and observation angle restrictions. X-band radar network observation data increases the accuracy of observations on the horizontal plane and compensates for the low elevation angle of the S-band radar at the vertical layer, achieving minute-by-minute temporal resolution. Furthermore, during the wind field inversion process, a forecast field from a numerical model (3 km) is incorporated as the initial field to achieve real-time updates of the radar's three-dimensional variational assimilation. Attached Figure Description

[0054] Figure 1 This diagram illustrates the existing technology that utilizes the volume scan area of ​​an S-band radar at the lowest elevation angle to create a blind zone.

[0055] Figure 2 The diagram shown is a flowchart illustrating a wind field inversion method based on X-band radar and numerical models as described in an embodiment of this application.

[0056] Figure 3 The diagram shows a module flow diagram of a wind field inversion method based on X-band radar and a 3-kilometer mode as described in an embodiment of this application.

[0057] Figure 4 The images shown are wind field inversion diagrams obtained using existing technologies and the wind field inversion method described in this patent embodiment, taking the S-band radar and three X-band radars in Dinghai area of ​​Zhoushan, Zhejiang Province as examples. Detailed Implementation

[0058] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0059] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0060] The technical solutions in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0061] like Figure 2 As shown, this embodiment provides a flowchart illustrating a wind field inversion method based on X-band radar and numerical models. (Reference) Figure 2 The wind field inversion method includes the following steps:

[0062] Step S101: During the wind field inversion process, X-band radar observation data is introduced to establish an inverted wind field area covering the low elevation angle space.

[0063] Step S102: Use the wind field predicted by the 3-kilometer numerical model as the initial field for wind field inversion.

[0064] Step S103: Combine the observation data from the S-band radar and the X-band radar with the initial field to perform data fusion, so as to correct the initial field.

[0065] Step S104: The conservation equation is added as a constraint to the loss function, and the loss function is solved iteratively based on the optimization algorithm to obtain the inverted wind field.

[0066] In step S101, during the wind field inversion process, since the S-band radar has an observation blind zone at the lowest elevation angle, in order to solve this problem, this embodiment introduces the observation data of the X-band radar, which increases the accuracy of observation on the horizontal plane, and can make up for the deficiency of the S-band radar in the low elevation angle region on the vertical plane.

[0067] Specifically, this step includes:

[0068] Step S1011: Determine the spatial range of the inverted wind field area.

[0069] Step S1012: Determine the intersection area between any two of the S-band and X-band radars within the space range based on their respective effective scanning areas.

[0070] Step S1013: Determine the center point of the inverted wind field region based on all these intersecting regions covering the inverted wind field region.

[0071] In this embodiment, the spatial range of the inverted wind field area is determined according to the inverted target (e.g., the area of ​​concern for typhoons). Then, all S-band and X-band radars within the determined spatial range are selected. These radars are set up at different locations within the spatial range, and each radar has its own effective scanning area. If there is an intersection between the effective scanning areas of two radars (including radars of the same band or radars of different bands), the intersection is taken as part of the inverted wind field area. Thus, the entire inverted wind field area can be covered by the intersection between all these radars.

[0072] Unlike existing technologies, this embodiment incorporates observation data from X-band radar, which increases observation accuracy on the horizontal plane and compensates for the low-elevation blind zone of S-band radar on the vertical plane. Furthermore, in terms of temporal resolution, S-band radar can achieve an update frequency of every 6 minutes, while X-band radar can achieve an update frequency of every 1 minute.

[0073] From a mathematical perspective, based on S-band radar observations, X-band radars are introduced at multiple locations within the observation area for observation. The volume scan areas covered by X-band radars and / or S-band radars at different locations will overlap. The center point of the intersection area of ​​multiple radar volume scans is taken as the center point for establishing the inverted wind field area with local high resolution coverage of low elevation angle space. Then, the maximum intersection is solved based on the following formula (1-1):

[0074]

[0075] in, This represents a point in the inverted wind field region. R represents the center position of the i-th radar (which can be an X-band or S-band radar), and R represents the effective radius of the effective scanning area. In practical applications, due to the rapid attenuation of X-band radar, the effective radius R is usually chosen to be 50 kilometers.

[0076] In step S102, the wind field predicted based on the 3-kilometer numerical model is used as the initial field for wind field inversion.

[0077] Step S1021: Based on the geopotential height field of isobaric surfaces in the model data, establish a three-dimensional space in a Cartesian coordinate system. The three-dimensional space includes vertical height, latitude, and longitude. The vertical height of the isobaric surfaces is as follows:

[0078]

[0079] in, This represents the potential height at time t. This represents the current altitude field. This represents the base state in the mode integral.

[0080] Step S1022: Establish the initial three-dimensional wind field space in Cartesian coordinates based on the latitude and longitude coordinates in the model:

[0081]

[0082] Based on the center point of the inverted wind field region, a refined initial wind field tensor matrix is ​​established:

[0083]

[0084] in, Indicates the center point of the inverted wind field region. , and z represents the resolution along the tensor's X-axis, Y-axis, and vertical Z-axis, respectively. Represents the grid points on the X-axis of the fine-resolution tensor. This represents a grid point on the Y-axis of the tensor.

[0085] Those skilled in the art will understand that, in the wind field inversion process, the unknowns to be solved (e.g., zonal wind field u, meridional wind field v, and vertical wind field w) are a high-order tensor (multidimensional array), which yields a complete analysis field on a fixed three-dimensional forecast area grid. In the mathematical formulation of the optimization algorithm, this massive tensor containing all unknowns is flattened into a one-dimensional super column vector, denoted as X (state vector), and the state tensor corresponds to the vector X acted upon by the matrix in the mathematical equation.

[0086] Step S1023: Determine the velocity constraints of the grid points in the numerical model according to the following formula:

[0087]

[0088] in, This represents the interpolation function, which uses nearest-neighbor interpolation to obtain a coarse-resolution wind field. Interpolation to fine resolution ,in This represents the initial field, specifically the coordinates in the 3km numerical model. The value of the wind field.

[0089] In this embodiment, during the wind field inversion process, a 3-kilometer numerical model is used to provide the initial guess field or physical constraints required for the inversion. This initial field is a complete but coarse-resolution wind field. The model simulates a physically consistent and spatiotemporally complete three-dimensional atmospheric state (including wind, temperature, pressure, humidity, etc.) based on a fixed region; this state is called the initial field (or background field). Interpolation is used to calculate the zonal wind field u, meridional wind field v, and vertical wind field w at uniform grid points with a spatial resolution of 1 kilometer and a vertical resolution of 500 meters. To avoid data loss and improve computational efficiency, nearest neighbor interpolation or linear interpolation is generally used.

[0090] In step S103, the observation data from the S-band radar and the X-band radar are combined with the initial field to perform data fusion, so as to correct the initial field.

[0091] Specifically, in the wind field inversion process, data fusion refers to the assimilation of sparse but precise radar observation data (real-time) with the initial field (complete but potentially biased model background). In this embodiment, observation data (including radial wind data and reflectivity) from multiple S-band and X-band radars can be combined with the initial field for data fusion. That is, in the three-dimensional grid tensor constructed from the initial field, the difference between the observation data from the S-band or X-band radar and the 3-kilometer numerical model background value at the same grid point is compared. Based on the reliability of both (observation error and model error), an optimal correction value is calculated to correct the initial field, generating a completely new analytical field that integrates all information.

[0092] In step S104, the conservation equation is added as a constraint to the loss function, and the loss function is iteratively solved based on the optimization algorithm to obtain the inverted wind field.

[0093] Specifically, the loss function described in this step is defined as follows:

[0094]

[0095] in, This indicates the wind field observed by S-band radar and X-band radar; The loss function representing the conservation of mass; The loss function representing vorticity conservation; The loss function representing a smooth linear constraint; This indicates a 3-kilometer numerical model constraint.

[0096] Unlike existing methods, this embodiment employs a 3km numerical model in the loss function. As a pattern constraint, and The wind field observed by X-band radar has been added.

[0097] In practical applications, the wind fields observed by the S-band radar and X-band radar The specific formula is as follows:

[0098]

[0099] in, x, y, and z represent the horizontal and vertical distances of each point in the Cartesian coordinate system from the radar station.

[0100] According to the law of conservation of mass:

[0101]

[0102]

[0103] Define loss function The specific formula is as follows:

[0104]

[0105] in, For density.

[0106] Vertical vorticity conservation relation:

[0107]

[0108] Calculate the loss function for vorticity conservation based on the vertical vorticity conservation relation. The specific formula is as follows:

[0109]

[0110] in, .

[0111] The loss function of the smooth linear constraint The specific formula is as follows:

[0112]

[0113] in, W is the weighting coefficient.

[0114] The 3-kilometer numerical model constraint The specific formula is as follows:

[0115]

[0116]

[0117] Among them, the model wind field is Time and observation time Translation between them; To interpolate the wind field to the grid points of the forecast area, Translation position at any time The positional relationship with the whole hour is as follows:

[0118]

[0119] The wind field is:

[0120] in, , , The three-dimensional wind field is interpolated to the grid points of the forecast area.

[0121] Solving the loss function includes:

[0122]

[0123] in,

[0124]

[0125] right Solve the differential and calculate the multidimensional gradient from the tensor. Among them, scalar These are weighting coefficients used to determine the weighting of each cost function relative to the total cost function. The relative contribution. Generally, the default is... , , , , .

[0126] By employing a machine learning framework and utilizing the GPU of a mainframe to calculate the gradient of the loss function on multidimensional tensors, computation time is reduced and timeliness is improved.

[0127] At any given moment or during real-time system operation, reflectivity and radial wind are extracted from multiple X-band radars closest to the current moment, and quality control is performed simultaneously. The quality control scheme includes: cross-correlation coefficient greater than 0.8, basic reflectivity greater than 5, and velocity deambiguity.

[0128] This technical solution introduces wind field inversion data from X-band radar, updated every 1 minute, into the existing 6-minute S-band radar observation data. By more finely characterizing the spatiotemporal resolution of radar reflectivity and radial wind, and adding a real-time updated 3-kilometer regional mode, it can compensate for the observation blind zone of S-band at low altitudes of 1-2 kilometers in real time, thereby improving the spatiotemporal resolution of wind field monitoring.

[0129] like Figure 3 The diagram shown is a schematic flowchart of a wind field inversion method based on X-band radar and a 3-kilometer mode, as described in an embodiment of this application.

[0130] refer to Figure 3 Compared to the existing wind field inversion process using S-band radar observation data (solid line module box), this embodiment adds a dashed line module box. Specifically, by fusing radial wind data from X-band radar observation data, velocity de-ambiguity can be achieved, and reflectivity data can be used for clutter removal. In the wind field inversion process, besides calculating mass conservation... vorticity conservation After smoothing constraints In addition to the loss function, the initial field is established using forecast data from the 3-kilometer numerical model as a model constraint. Establishing a unified tensor field and using the gradient of the tensor to calculate the inverted wind field can improve computational efficiency.

[0131] Using the wind field inversion method of this technical solution, taking the S-band radar and three X-band radars in Dinghai area of ​​Zhoushan, Zhejiang Province as examples. The center point latitude and longitude is (122.2, 30.11), and the time is 17:20 on July 29, 2025. The wind speed characterization in Zhoushan area is compared between the traditional solution and this solution.

[0132] like Figure 4 As shown in (a) and (b), the S-band and X-band radars, in a cross-section at 29.3 degrees North latitude, have radial winds at the ordinate of 0-15 km. Figure 4In (a), the traditional scheme inverts a wide wind field range, but the resolution is low, and there is a significant gap in the description of the wind field in the near-surface layer. Compared to the S-band, the X-band radar has a smaller radial wind range, but it effectively compensates for the shortcomings of the S-band in the lower layers. The radial wind speed ranges from 0 to 24 m / s, and the inverted zonal and meridional wind speeds both reach 20 m / s. The emergence of wind speeds from zero indicates that this technical scheme can further invert the near-surface wind field.

[0133] The scope of protection for the wind field inversion method based on X-band radar and numerical model described in this application is not limited to the order of steps listed in this embodiment. Any solution implemented by adding, subtracting, or replacing steps in the prior art based on the principles of this application is included within the scope of protection of this application.

[0134] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0135] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0136] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0137] This application also provides an electronic device, which includes: a memory storing a computer program thereon; and a processor communicatively connected to the memory for executing the computer program to implement the above-described wind field inversion method based on X-band radar and numerical models.

[0138] This application also provides a computer-readable storage medium storing a computer program that, when executed by an electronic device, implements the aforementioned wind field inversion method based on X-band radar and numerical models. Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing a processor. The program can be stored in a computer-readable storage medium, which is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disc, and any combination thereof. The aforementioned storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video optical disc), or a semiconductor medium (e.g., solid-state drive), etc.

[0139] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0140] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A wind field inversion method based on X-band radar and numerical models, characterized in that, The method includes the following steps: In the wind field inversion process, observation data from X-band radar is introduced to establish an inverted wind field area covering low elevation angle space; The wind field predicted by the 3-kilometer numerical model was used as the initial field for wind field inversion; The initial field is corrected by combining the observation data from the S-band radar and the X-band radar with the data from the initial field. The conservation equation is added as a constraint to the loss function, and the loss function is iteratively solved based on the optimization algorithm to obtain the inverted wind field.

2. The method according to claim 1, characterized in that, The introduction of X-band radar observation data to establish an inverted wind field region covering low elevation angle space includes: Determine the spatial extent of the inverted wind field area; The intersection area between any two S-band and X-band radars within the aforementioned spatial range is determined based on their respective effective scanning areas. Based on all these intersecting regions covering the inverted wind field area, the center point of the inverted wind field area is determined.

3. The method according to claim 2, characterized in that, The use of wind fields predicted by a 3-kilometer numerical model as the initial field for wind field inversion includes: Based on the geopotential height field of isobaric surfaces in model data, a three-dimensional space in Cartesian coordinates is established. This three-dimensional space includes vertical height, latitude, and longitude. The vertical height of the isobaric surfaces is as follows: in, This represents the potential height at time t. This represents the current altitude field. Represents the reference state in the mode integral; Establish an initial three-dimensional wind field space in Cartesian coordinates based on the latitude and longitude coordinates in the model: Based on the center point of the inverted wind field region, a refined initial wind field tensor matrix is ​​established: in, Indicates the center point of the inverted wind field region. , and z represents the resolution along the tensor's X-axis, Y-axis, and vertical Z-axis, respectively. Represents the grid points on the X-axis of the fine-resolution tensor. Represents the grid points on the Y-axis of the tensor; The velocity constraints for grid points in the numerical model are determined using the following formula: in, This represents the interpolation function, which uses nearest-neighbor interpolation to obtain a coarse-resolution wind field. Interpolation to fine resolution ,in This represents the initial field, i.e., the coordinates in the 3km numerical model. The value of the wind field.

4. The method according to claim 1, characterized in that, The loss function is defined as follows: in, This indicates the wind field observed by S-band radar and X-band radar; The loss function representing the conservation of mass; The loss function representing vorticity conservation; The loss function representing a smooth linear constraint; This indicates a 3-kilometer numerical model constraint.

5. The method according to claim 4, characterized in that, The wind field observed by the S-band and X-band radars The specific formula is as follows: in, x, y, and z represent the horizontal and vertical distances of each point in the Cartesian coordinate system from the radar station.

6. The method according to claim 4, characterized in that, According to the law of conservation of mass Define loss function The specific formula is as follows: in, For density.

7. The method according to claim 4, characterized in that, Vertical vorticity conservation relation: Calculate the loss function for vorticity conservation based on the stated vertical vorticity conservation relationship. The specific formula is as follows: in, .

8. The method according to claim 4, characterized in that, The loss function of the smooth linear constraint The specific formula is as follows: in, W is the weighting coefficient.

9. The method according to claim 4, characterized in that, The 3-kilometer numerical model constraint The specific formula is as follows: Among them, the model wind field is Time and observation time Translation between them; To interpolate the wind field to the grid points of the forecast area, Translation position at any time The positional relationship with the whole hour is as follows: The wind field is: in, , , The three-dimensional wind field is interpolated to the grid points of the forecast area.

10. The method according to claim 4, characterized in that, Solving the loss function includes: in, right Solve the differential and calculate the multidimensional gradient from the tensor. scalar These are weighting coefficients used to determine the weighting of each cost function relative to the total cost function. The relative contribution.

11. An electronic device, characterized in that, The electronic device includes: a memory storing a computer program thereon; and a processor communicatively connected to the memory for executing the computer program to implement the wind field inversion method based on X-band radar and numerical models as described in any one of claims 1 to 10.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by an electronic device, the program implements the wind field inversion method based on X-band radar and numerical model as described in any one of claims 1 to 10.

Images