CN-122017849-A - Wind field inversion method based on X-band radar and numerical mode

CN122017849ACN 122017849 ACN122017849 ACN 122017849ACN-122017849-A

Abstract

The application provides a wind field inversion method based on an X-band radar and a numerical mode, which comprises the following steps of firstly taking a three-dimensional wind field predicted based on a 3 km numerical mode with coarse resolution as an initial wind field for wind field inversion, then introducing high-resolution S-band and X-band radar observation data, performing quality control on radar reflectivity, performing speed deblurring on radial wind, then combining the coarse resolution mode data and the high-resolution radar data, performing data fusion, then introducing a conservation equation as a constraint condition in a mode of solving vertical speed in the high-precision three-dimensional wind field, calculating a loss function, and finally iteratively solving an 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 monitoring missing of the S-band radar 1-2 km low-altitude wind field is solved.

Inventors

DAI XIANGLIN
CHEN YEFENG
CHEN FENG
Yu Zhenshou
QIU JINJING
CHEN MEITING
CHEN SHUQIN

Assignees

浙江省气象科学研究所

Dates

Publication Date: 20260512
Application Date: 20260408

Claims (12)

1. The wind field inversion method based on the X-band radar and the numerical mode is characterized by comprising the following steps of: in the wind field inversion process, introducing observation data of an X-band radar to establish an inversion wind field area covering a low elevation angle space; taking a wind field predicted based on a 3 km numerical mode as an initial field for wind field inversion; combining the observation data of the S-band radar and the X-band radar with the initial field to perform data fusion so as to correct the initial field; And adding a conservation equation as a constraint condition into a loss function, and iteratively solving the loss function based on an optimization algorithm to obtain the inversion wind field.
2. The method of claim 1, wherein the introducing observation data for an X-band radar to establish an inverted wind field region covering a low elevation space comprises: Determining the space range of an inversion wind field area; Determining a crossing area between every two radars based on the effective scanning areas of any two radars in all S-band radars and X-band radars in the space range; and covering the inversion wind field area according to all the intersection areas and determining the central point of the inversion wind field area.
3. The method of claim 2, wherein taking a wind field based on a 3 km numerical model forecast as an initial field for wind field inversion comprises: according to potential height field of the medium pressure surface in the mode data, establishing a three-dimensional space of a Cartesian coordinate system, wherein the three-dimensional space comprises vertical height, latitude and longitude, and the vertical height of the medium pressure surface is as follows: Wherein, the The potential height at time t is indicated, The height field representing the current moment in time, Representing a reference state in the mode integration; establishing an initial field three-dimensional wind field space under a Cartesian coordinate system based on longitude and latitude coordinates in the mode: based on the center point of the inversion wind field area, establishing a refined wind field initial field tensor matrix: Wherein, the Representing the center point of the inverted wind field region, 、 And Z represents the resolution in the tensor X-axis, Y-axis and vertical direction Z respectively, The grid points on the X-axis of the refined resolution tensor are represented, Representing grid points on the tensor Y-axis; Determining a speed constraint for a grid point in a numerical mode according to the following formula: Wherein, the Representing interpolation functions, i.e. wind fields with coarse resolution using adjacent interpolation Interpolation to fine resolution Wherein Representing the initial field, i.e. the coordinates in 3 km numerical mode Is a wind field value of (a).
4. The method of claim 1, wherein the loss function is defined as: Wherein, the The wind field observed by the S-band radar and the X-band radar is represented; a loss function representing conservation of mass; A loss function representing conservation of vorticity; a loss function representing a smooth linear constraint; Representing a 3 km numerical mode constraint.
5. The method of claim 4, wherein the S-band radar and the X-band radar observe wind fields The specific formula of (2) is as follows: Wherein, the And x, y and z are the horizontal distance and the vertical distance from each point to the radar site in the rectangular coordinate system.
6. The method of claim 4, wherein the mass conservation formula is based on Defining a loss function The specific formula of (2) is as follows: Wherein, the Is the density.
7. The method of claim 4, wherein the vertical vorticity conservation relationship: calculating the loss function of the vorticity conservation according to the vertical vorticity conservation relation The specific formula is as follows: Wherein, the 。
8. The method of claim 4, wherein the smooth linear constraint loss function The specific formula of (2) is as follows: Wherein, the W is a weight coefficient.
9. The method of claim 4, wherein the 3 km value mode constraint The specific formula of (2) is as follows: Wherein the mode wind field is in Time of day and observation time Translation between; for interpolation to wind fields at grid points of the forecasted area, Time of day translation position The positional relationship with the whole hour is: The wind field is as follows: Wherein, the , , For a three-dimensional wind field interpolated to the grid points of the forecasted area.
10. The method of claim 4, wherein solving the loss function comprises: Wherein, the For a pair of Solving the differentiation, calculating multidimensional gradient according to tensors Scalar quantity Is a weight coefficient for determining the total cost function of each cost function Is a relative contribution of (c).
11. An electronic device comprising a memory having a computer program stored thereon, and a processor communicatively coupled to the memory for executing the computer program to implement the X-band radar and numerical mode based wind park inversion method of any of claims 1-10.
12. A computer-readable storage medium, on which a computer program is stored, characterized in that the program, when executed by an electronic device, implements the X-band radar and numerical mode based wind park inversion method according to any one of claims 1 to 10.

Description

Wind field inversion method based on X-band radar and numerical mode Technical Field The application belongs to the technical fields of atmospheric science and remote sensing, and relates to a wind field inversion method based on an X-band radar and a numerical mode. Background In recent years, with the continuous development of detection technology, the ability of monitoring the earth's atmosphere is continuously improved. For example, in the detection of precipitation, the reflectivity updated every 6 minutes provided by the S-band radar has been widely used in convection monitoring, inversion of precipitation monitoring, and precipitation prediction, and also widely used in inversion of low-altitude wind fields, and plays an indispensable role in monitoring disaster weather. Unlike the direct measurement observation products, radar observation provides indirect measurements of particle number, morphology and advection in terms of reflectivity, radial wind speed, etc. However, the single radar with fixed ground is limited in detection range, is insufficient to cover a weather system with complex changes, is very complex in terrain conditions in China, and is widely shielded by most business radars at 0.5-degree elevation angles to limit the detection capability of a new-generation weather radar network on low-layer meteorological targets. Doppler radar observation is to reflect movement by using changes in power after scattering and reflection of electromagnetic waves on the particle surface. In some areas, the radar observation height is above the water condensation height (in the range of 2.5 to 4 km). On the one hand, the rain intensity is judged according to the quantity and the form of particles monitored by a radar, which is a common principle of quantitative precipitation estimation. Radar, on the other hand, is able to monitor wind speed. The particles whose reflectivity is monitored do not directly determine the wind speed, but they act as trace particles in the wind field, reflecting to some extent the direction and intensity of the wind. Therefore, in the conventional S-band radar inversion wind field process, wind speeds of more than 2.5 km can be inverted in most areas. Therefore, when a single-section S-band radar or a multi-section S-band radar is used, precipitation and wind-field monitoring of 1 to 2 km is always "missing puzzle" of low-altitude monitoring. Fig. 1 is a schematic diagram of an observation blind area existing in a body scanning area in a wind field inverted by using an S-band radar in the prior art. Referring to fig. 1, the body sweep area of the S-band radar is an approximately gray semicircular area, and the white double-arrow area is a blind area which cannot be detected by the S-band radar. Disclosure of Invention The application provides a wind field inversion method based on an X-band radar and a numerical mode, which is used for solving the problem that precipitation and wind fields which are 1-2 kilometers lower than the air are always lost when an S-band radar is adopted for monitoring. According to the wind field inversion method based on the X-band radar and the numerical mode, the method comprises the steps of introducing observation data of the X-band radar in a wind field inversion process to establish an inversion wind field area covering a low elevation angle space, taking a wind field predicted based on the 3 km numerical mode as an initial field of wind field inversion, combining the observation data of the S-band radar and the observation data of the X-band radar with the initial field to conduct data fusion so as to correct the initial field, adding a conservation equation as a constraint condition into a loss function, and iteratively solving the loss function based on an optimization algorithm to obtain an inversion wind field. In the application, in the wind field inversion process, the defect that the wind speed of more than 2.5 km and the rainfall of 1-2 km and the wind field monitoring can only be inverted due to the limitation of the height and the observation angle of a single or a plurality of S-band radars are overcome by introducing the observation data of the X-band radars in a low-altitude wind field area. On one hand, the X-band radar networking observation data increases the observation precision on the horizontal plane, and on the other hand, the defect of low elevation angle of the S-band is overcome on the vertical layer, and the time resolution can be up to every minute. Further, in the wind field inversion process, a forecasting field in a numerical mode (3 km) is added as an initial field of wind field inversion to realize radar three-position variation assimilation real-time update. In one implementation manner of the first aspect, the method for introducing the observation data of the X-band radar to establish the inversion wind field area covering the low elevation angle space includes determining a spatial range of the inversion