Water vapor chromatography modeling method and device based on GNSS-MET

Abstract

The invention discloses a water vapor chromatography modeling method and device based on GNSS-MET, and belongs to the technical field of atmospheric water vapor information inversion. According to the method, satellite signals and meteorological parameters are obtained through a GNSS-MET receiver, and the oblique path water vapor content (SWV) is calculated to serve as chromatography input; densely and uniformly distributed observation stations are selected to construct a chromatography area, and voxel grids are divided; a horizontal constraint is constructed by using Gaussian distance weighting, a vertical constraint is constructed in combination with a water vapor vertical index distribution characteristic, and ERA5 reanalysis data is introduced as a prior constraint, so that the limitation of a traditional sounding data constraint is solved; and resolving the tomographic equation through a singular value decomposition (SVD) method to obtain three-dimensional water vapor density distribution. According to the method, the high temporal-spatial resolution characteristic of ERA5 data is utilized, the precision and reliability of water vapor chromatography are remarkably improved, and the method is suitable for the fields of extreme weather prediction, meteorological monitoring and the like.

Description

Technical Field

[0001] This invention belongs to the field of atmospheric water vapor information inversion technology, specifically relating to a water vapor tomography modeling method and apparatus based on GNSS-MET. Background Technology

[0002] The troposphere contains over 90% of atmospheric water vapor, and its variations are closely related to extreme weather events and meteorological disasters. High spatiotemporal resolution water vapor monitoring is of great significance for weather forecasting, climate research, and disaster early warning. Traditional water vapor detection methods mainly include radiosondes, microwave radiometers, weather radar, and optical radiometers, but these methods have significant limitations: radiosondes provide sparse observations and are expensive; microwave radiometers are susceptible to cloud and rain interference; weather radar has difficulty detecting non-precipitating cloud data; and optical radiometers cannot operate in all weather conditions.

[0003] In recent years, water vapor inversion technology based on the Global Navigation Satellite System (GNSS) has been widely used due to its advantages such as all-weather operation, low cost, and high accuracy. GNSS signals experience delays as they pass through the atmosphere; these delays can be calculated to invert atmospheric water vapor content. GNSS-MET receivers, combined with Meteorological Observatory (MET) systems, can not only acquire satellite signals but also simultaneously measure meteorological parameters such as temperature, humidity, and pressure, providing more comprehensive data support for water vapor inversion. However, traditional GNSS water vapor inversion mainly obtains precipitable water (PWV) at the zenith, reflecting only the total vertically integrated water vapor volume above the station and failing to reveal the three-dimensional spatial distribution characteristics of water vapor.

[0004] To address this issue, water vapor tomography emerged. This technique establishes a mathematical relationship between the slant path water vapor content (SWV) of GNSS signals and a three-dimensional voxel grid, constructing tomographic equations to invert the three-dimensional water vapor density distribution within a region. However, due to the non-uniform distribution of GNSS signal rays, the tomographic equations often suffer from rank deficiency, leading to unstable solutions. Existing methods typically improve the solution performance by introducing horizontal, vertical, and prior constraints. Among these, prior constraints are crucial to the accuracy of the tomographic results. Traditional methods often use radiosonde data as prior constraints, but radiosonde data is only valid at specific stations, failing to reflect the horizontal changes in water vapor within the tomographic region, and has low spatiotemporal resolution, making it difficult to meet the requirements of high-precision tomography.

[0005] The ERA5 reanalysis data provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) offers a new approach to solving the aforementioned problems. ERA5 boasts high spatiotemporal resolution (1-hour temporal resolution, 37 vertical pressure layers) and provides gridded data of meteorological parameters such as temperature, humidity, and pressure globally, enabling a more accurate description of the spatial distribution characteristics of water vapor. Compared to radiosonde data, ERA5 is more suitable as a priori constraint background field for water vapor tomography. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a GNSS-MET-based water vapor tomography modeling method and apparatus. It utilizes ERA5 data to construct prior constraints, optimizes the tomographic equations by combining horizontal and vertical constraints, and solves the three-dimensional water vapor density distribution using singular value decomposition (SVD). This method fully leverages the spatiotemporal resolution advantages of ERA5 data, significantly improving the accuracy and spatial resolution of water vapor tomography, and providing more reliable technical support for meteorological research and disaster early warning.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A GNSS-MET-based water vapor tomography modeling method, the method comprising:

[0009] Step S1: Based on the GNSS-MET receiver, acquire satellite signals and meteorological parameters respectively, and calculate the slant path water vapor content (SWV).

[0010] Step S2: Obtain historical observation data and perform three-dimensional voxel mesh generation. Convert the slant path water vapor content (SWV) into an integral form corresponding to the voxel mesh and construct the observation equation of the tomographic model.

[0011] Step S3: Apply horizontal constraints, vertical constraints, and prior constraints to the observation equations of the tomography model to obtain the joint observation equations; wherein, the horizontal constraints are based on the Gaussian distance weighting method, the vertical constraints are based on the exponential distribution characteristics of water vapor, and the prior constraints are based on the water vapor density background field calculated from the reanalysis dataset.

[0012] Step S4: Solve the joint observation equations using the singular value decomposition method to obtain the three-dimensional water vapor density distribution.

[0013] On the other hand, the present invention provides a GNSS-MET-based water vapor tomography modeling device, comprising a solution module, a voxel partitioning module, a constraint module, and an output module, wherein...

[0014] The calculation module is used to acquire satellite signals and meteorological parameters based on the GNSS-MET receiver, and calculate the slant path water vapor content (SWV).

[0015] The voxel partitioning module is used to acquire historical observation data and perform three-dimensional voxel mesh partitioning, convert the slant path water vapor content (SWV) into an integral form corresponding to the voxel mesh, and construct the observation equation of the tomographic model.

[0016] The constraint module is used to apply horizontal, vertical and prior constraints to the observation equations of the tomographic model to obtain the joint observation equations. The horizontal constraints are based on the Gaussian distance weighting method, the vertical constraints are based on the exponential distribution characteristics of water vapor, and the prior constraints are based on the water vapor density background field calculated from the reanalysis dataset.

[0017] The output module is used to solve the joint observation equations using the singular value decomposition method to obtain the three-dimensional water vapor density distribution.

[0018] Thirdly, the present invention provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned GNSS-MET-based water vapor tomography modeling method.

[0019] Fourthly, the present invention provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enable the processor to implement the aforementioned GNSS-MET-based water vapor tomography modeling method.

[0020] The beneficial effects of this invention are as follows:

[0021] This invention significantly improves the accuracy and reliability of water vapor tomography by introducing high spatiotemporal resolution ERA5 reanalysis data as a priori constraints, compared to traditional radiosonde data constraint methods, and can more accurately reflect the three-dimensional distribution characteristics of water vapor. Secondly, by employing GNSS-MET receiver network observations combined with an intelligent constraint system design, the rank deficiency problem of the tomographic equations is effectively solved, making the inversion results more stable.

[0022] In terms of practicality, this invention achieves water vapor monitoring with a 1-hour time resolution, greatly improving the timeliness of observations and providing strong support for short-term weather forecasting. Simultaneously, based on the widely deployed GNSS-MET network and free ERA5 data, it significantly reduces observation costs and improves the convenience of data acquisition.

[0023] Furthermore, this invention is not limited by weather conditions, enabling continuous monitoring around the clock. The obtained three-dimensional water vapor products can be directly applied to multiple fields such as numerical weather prediction and disaster early warning, possessing significant operational application value. Cross-validation using multi-source data ensures the reliability of the inversion results, providing high-quality data support for meteorological research and operational applications. Attached Figure Description

[0024] Figure 1 This is a flowchart of a water vapor tomography modeling method based on GNSS-MET according to the present invention;

[0025] Figure 2 This is a three-dimensional water vapor density distribution map;

[0026] Figure 3 This is a comparison chart of the water vapor profile results obtained by this method and the water vapor profile of a weather balloon. Detailed Implementation

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0028] This invention provides a GNSS-MET-based water vapor tomography modeling method, which extracts high-precision water vapor parameter information; performs network observations to obtain basic observation equations; constructs complete tomography equations based on radiosonde and reanalysis data and determines the weight matrix; and calculates water vapor density to generate water vapor profiles, etc. As can be seen, this invention realizes the modeling of three-dimensional water vapor information through water vapor parameter information, and further demonstrates the water vapor distribution in three-dimensional space.

[0029] like Figure 1 As shown, the GNSS-MET-based water vapor tomography modeling method of the present invention specifically includes the following steps:

[0030] Step S1: Based on the GNSS-MET receiver, acquire satellite signals and meteorological parameters respectively, and calculate the slant path water vapor content (SWV).

[0031] To obtain current and historical water vapor parameter estimates, specifically, the system receives and processes navigation satellite system signals, saves the information received by the station as observation and meteorological files, acquires the station's temperature, humidity, and pressure data, and inverts the east-west and north-south horizontal gradients of the zenith tropospheric wet delay. Furthermore, using the wet projection function and gradient values, the obtained zenith wet delay (ZWD) is converted into the oblique path wet delay (SWD).

[0032] (1)

[0033] In the formula, This represents the wet mapping function, where ele and azi represent the satellite cutoff elevation angle and azimuth angle, respectively. and This is due to atmospheric anisotropy, resulting in a wetted horizontal gradient term for water vapor in the north-south and east-west directions. This represents the post-hoc residual correction of the observed values.

[0034] Slant path wet delay (SWD) is further converted into slant path water vapor content (SWV) as the chromatography input:

[0035] (2)

[0036] In the formula, Atmospheric water vapor conversion coefficient:

[0037] (3)

[0038] In the formula, Let be the specific gas constant of water vapor. ; The density of liquid water; The atmospheric weighted average temperature; The adjusted wet refractive index coefficient ; and for calculation of and All are atmospheric refractive index constants; , , ; and Molar mass of water vapor and dry air molecules, respectively , .

[0039] Step S2: Select a uniformly distributed GNSS-MET receiver network, acquire its historical observation data and perform three-dimensional voxel grid division, convert the slant path water vapor content (SWV) into an integral form corresponding to the voxel grid, and construct the observation equation of the tomographic model.

[0040] First, single-point observations from GNSS-MET receiver stations are acquired to obtain satellite and meteorological observation information. When performing three-dimensional water vapor tomography, some stations are selected for network observation to construct a network observation tomography area that is divided into voxel combinations of a certain size, in order to construct the tomography model equations for subsequent purposes.

[0041] Specifically, a certain number of densely and uniformly distributed GNSS-MET receiver stations are selected to acquire observational information. At least one radiosonde station is included for vertical constraint and subsequent accuracy verification. A tomographic region encompassing these stations is constructed. During voxel-based 3D tomography, the tomographic region is first divided into voxel blocks in a specific manner. The final water vapor density or wet refractive index obtained from the tomography corresponds to these voxel blocks. The horizontal division of voxels, i.e., the horizontal resolution of the tomography, can be adjusted according to the distribution density of the stations within the tomographic region. The vertical resolution uses a non-uniform stratification method. This is because ERA5 is introduced as a priori constraint, but the ERA5 stratification data only covers approximately 37 altitude layers. Furthermore, considering that water vapor distribution should conform to the principle of high water vapor content and high vertical resolution at lower levels, and low water vapor content and lower vertical resolution at higher levels, the vertical division is based on these two points. A total of n voxels are obtained, of which... , , and These represent the number of rows, columns, and layers of the voxel block within the chromatography region, respectively.

[0042] In this example, satellite and meteorological observation information for the Hong Kong area was acquired. Three-dimensional tomography was performed on 12 stations and one radiosonde station (45004). The tomographic region was divided into voxel blocks according to a specific method. The final water vapor density or wet refractive index obtained from the tomography corresponds to these voxel blocks. The tomographic range was selected along the latitudinal direction. Step size is Longitude direction Step size is Considering the characteristic of water vapor density being denser at the bottom and sparser at the top in the vertical direction, height intervals of 300, 300, 300, 300, 500, 800, 1200, 1700, 2200, and 2800 m were set, and the study area was ultimately divided into... A tomographic grid. The tomography was conducted in May 2013, as shown below. Figure 2 and Figure 3 The figures show the three-dimensional water vapor density distribution at 00:00 on May 1st and the water vapor profile compared to the sounding station.

[0043] The slant path water vapor content (SWV) is the integral of the water vapor density along the signal ray path, which can be expressed as the sum of the products of the intercept of the satellite signal through each grid and the atmospheric water vapor density within that grid:

[0044] (4)

[0045] In the formula, Indicates that the ray is in Intercept within the grid, express The estimated water vapor density value within the grid, here These represent the sequence numbers in the three dimensions of latitude, longitude, and elevation, respectively. By representing all signal rays within the study area in the above form, the observation equations of the tomography model can be obtained:

[0046] (5)

[0047] In the formula, y is a column vector composed of water vapor content (SWV) on the rays passing through the top of the study area, A is the coefficient matrix of the observation equation, generated by the grid number and intercept of the signal ray, and x is a column matrix composed of the unknown parameter water vapor density.

[0048] Step S3: Apply horizontal, vertical, and prior constraints to the observation equations of the tomographic model. Horizontal constraints are constructed based on the voxel partitioning method using Gaussian distance weighting. Vertical constraints are constructed based on the exponential distribution of water vapor in the vertical direction. Prior constraints are applied by calculating the water vapor density of each tomographic voxel block using an appropriate ERA5 time period as the background field. Furthermore, the water vapor density at the top of the layer is calculated from historical ERA5 data as a boundary constraint and integrated into the vertical constraints, ultimately constructing the tomographic model equations within the networked tomographic region.

[0049] Water vapor density can be calculated from historical reanalysis data such as ERA5 or radiosonde data. The difference is that radiosonde data directly provides water vapor pressure and temperature, while ERA5 requires obtaining water vapor pressure from temperature and relative humidity first. The unit is or :

[0050] (6)

[0051] In the formula, Temperature is expressed in units of . Then pressurize the saturated water vapor. Calculate water vapor pressure based on relative humidity:

[0052] (7)

[0053] In the formula, Indicates relative humidity. This represents the water vapor pressure. Only with water vapor pressure and temperature can the water vapor density of each layer be obtained.

[0054] (8)

[0055] In the formula The water vapor density at each height point. The water vapor pressure at each floor level, in units of... or .

[0056] Based on the characteristic that water vapor density is correlated in the horizontal direction and the correlation is stronger the closer the distance, the Gaussian weighted function method is used to determine the coefficient values, with a larger weight for closer distances. The horizontal constraint equation is as follows:

[0057] (9)

[0058] In the formula, and These are the coefficient and value matrices for the horizontal constraints, respectively. Based on the characteristic that water vapor density exhibits a negative exponential correlation in the vertical direction, the coefficient values ​​are determined according to the water vapor density relationship between adjacent mesh layers. The vertical constraint condition equation is then applied as follows:

[0059] (10)

[0060] In the formula, and These are the coefficient and value matrices for the vertical constraints, respectively. Based on the radiosonde data from the previous few days of tomography, boundary constraints can be applied to the top layer of the tomographic region, using the water vapor density and coefficients obtained from radiosonde data to replace the last layer in the original vertical constraints. The water vapor density near each layer is calculated using ERA5, yielding the PWV results for any past ERA5 time. These results are then matched with GNSS-MET station data to select several discontinuous PWV times with the closest water vapor content. The a priori constraint equation is applied using the average water vapor density of the ERA5 layers at these times:

[0061] (11)

[0062] In the formula, I is the identity matrix of prior constraint coefficients, and C is the water vapor density value obtained from ERA5. Therefore, the joint observation equation after applying all constraints is:

[0063] (12)

[0064] Step S4: Use the Singular Value Decomposition (SVD) method to solve the three-dimensional water vapor density of the tomographic region of the joint observation equation, and analyze its root mean square error, correlation coefficient, water vapor profile, etc. compared with reanalysis data and radiosonde data.

[0065] The tomographic model equations are solved using the SVD method. Specifically, the coefficient matrix B on the left side of the joint observation equations is first decomposed using SVD as follows:

[0066] (13)

[0067] In the formula, It is a matrix containing singular values. The coefficient matrix of the observation equation The square root of the eigenvalue, Let B be the rank of the coefficient matrix B of the joint observation equation. and Let B be the left singular and right singular vectors, respectively. The generalized inverse of B is defined as:

[0068] (14)

[0069] The column matrix x, composed of the unknown parameters water vapor density obtained from the tomography, combined with B and the observation matrix L on the right side of the joint observation equation, can be expressed as:

[0070] (15)

[0071] The RMS and STD at different altitudes of a given latitude and longitude point were compared and verified with radiosonde and ERA5 data respectively. The calculation methods for both are as follows:

[0072] (16)

[0073] In the formula, This refers to the number of chromatography layers. This represents the water vapor density value obtained from the chromatographic inversion. This represents the water vapor density value calculated from radiosonde data or ERA5. Within the voxel block where the radiosonde station is located, i.e., at the station's latitude and longitude, the water vapor profiles of the three parameters are compared along the altitude. Since ERA5 provides the water vapor density for each tomographic voxel block, the correlation between the tomographic and ERA5 results can be analyzed. Using May 1, 2013 as the selected date, the water vapor density values ​​for all tomographic grids on May 1, 2013, are obtained.

[0074] On the other hand, this invention provides a GNSS-MET-based water vapor tomography modeling device, which includes various modules capable of implementing the steps of the aforementioned method. Specifically, it includes a solution module for acquiring satellite signals and meteorological parameters based on a GNSS-MET receiver and calculating the slant path water vapor content (SWV); a voxel partitioning module for acquiring historical observation data and performing three-dimensional voxel mesh partitioning, converting the slant path water vapor content into an integral form corresponding to the voxel mesh to construct the observation equation of the tomography model; a constraint module for applying horizontal constraints, vertical constraints, and prior constraints to the observation equation to generate a joint observation equation; wherein the horizontal constraint is based on the Gaussian distance weighting method, the vertical constraint is based on the exponential distribution characteristics of water vapor, and the prior constraint is based on the water vapor density background field calculated from ERA5 reanalysis data; and an output module for solving the joint observation equation using the singular value decomposition method to obtain the three-dimensional water vapor density distribution result.

[0075] Thirdly, the present invention provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned GNSS-MET-based water vapor tomography modeling method.

[0076] Fourthly, the present invention provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enable the processor to implement the aforementioned GNSS-MET-based water vapor tomography modeling method.

[0077] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A water vapor tomography modeling method based on GNSS-MET, characterized in that, The method includes: Step S1: Based on the GNSS-MET receiver, acquire satellite signals and meteorological parameters respectively, and calculate the slant path water vapor content (SWV). Step S2: Obtain historical observation data and perform three-dimensional voxel mesh generation. Convert the slant path water vapor content (SWV) into an integral form corresponding to the voxel mesh and construct the observation equation of the tomographic model. Step S3: Apply horizontal constraints, vertical constraints, and prior constraints to the observation equations of the tomography model to obtain the joint observation equations; wherein, the horizontal constraints are based on the Gaussian distance weighting method, the vertical constraints are based on the exponential distribution characteristics of water vapor, and the prior constraints are based on the water vapor density background field calculated from the reanalysis dataset. Step S4: Solve the joint observation equations using the singular value decomposition method to obtain the three-dimensional water vapor density distribution.

2. The water vapor tomography modeling method based on GNSS-MET according to claim 1, characterized in that, The step S1 of calculating the slant path moisture content (SWV) specifically includes: converting the zenith wet delay (ZWD) to the slant path wet delay (SWD) using a wet projection function, and then converting the SWD to SWV by combining the atmospheric moisture conversion coefficient.

3. The water vapor tomography modeling method based on GNSS-MET according to claim 1, characterized in that, The three-dimensional voxel mesh division in step S2 is specifically as follows: the resolution in the horizontal direction is adjusted according to the distribution density of the stations, and a non-uniform layering method with denser bottom and sparser top is adopted in the vertical direction.

4. The water vapor tomography modeling method based on GNSS-MET according to claim 1, characterized in that, The specific implementation of the horizontal constraint in step S3 is as follows: based on the correlation of water vapor density in the horizontal direction, a Gaussian weighting function is used to determine the constraint coefficient, and the closer the grid points are, the greater the weight.

5. The water vapor tomography modeling method based on GNSS-MET according to claim 1, characterized in that, The specific implementation of the vertical constraint in step S3 is as follows: based on the characteristic that water vapor has a negative exponential distribution in the vertical direction, a relationship equation for water vapor density between adjacent layers is established, and radiosonde data is introduced as the top-level boundary constraint.

6. The water vapor tomography modeling method based on GNSS-MET according to claim 1, characterized in that, The specific implementation method of the prior constraint in step S3 is as follows: select multiple ERA5 time data that are closest to the water vapor content of the observation period, and calculate their average layered water vapor density as background field constraints.

7. The water vapor tomography modeling method based on GNSS-MET according to claim 1, characterized in that, Step S4 further includes comparing and verifying the obtained three-dimensional water vapor density distribution with the reanalysis dataset and radiosonde data, calculating the root mean square error and correlation coefficient, and drawing the water vapor profile.

8. A water vapor tomography modeling device based on GNSS-MET, characterized in that, include: The calculation module is used to acquire satellite signals and meteorological parameters based on the GNSS-MET receiver, and calculate the slant path water vapor content (SWV). The voxel partitioning module is used to acquire historical observation data and perform three-dimensional voxel mesh partitioning, convert the slant path water vapor content (SWV) into an integral form corresponding to the voxel mesh, and construct the observation equation of the tomographic model. The constraint module is used to apply horizontal, vertical and prior constraints to the observation equations of the tomographic model to obtain the joint observation equations. The horizontal constraints are based on the Gaussian distance weighting method, the vertical constraints are based on the exponential distribution characteristics of water vapor, and the prior constraints are based on the water vapor density background field calculated from the reanalysis dataset. The output module is used to solve the joint observation equations using the singular value decomposition method to obtain the three-dimensional water vapor density distribution.

9. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When one or more programs are executed by the one or more processors, the one or more processors implement the GNSS-MET-based water vapor tomography modeling method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores executable instructions that, when executed by a processor, enable the processor to implement the GNSS-MET-based water vapor tomography modeling method as described in any one of claims 1-7.

Images