GNSS vertical displacement-based earth surface mass load inversion method and computer system

By combining GNSS vertical displacement and weighting functions and boundary constraints of multi-source geographic data, the minimum and smoothest model is constructed, which solves the signal leakage and uncertainty of GNSS inversion results under sparse observation network, and improves the changes in surface quality Inversion accuracy and reliability have promoted the application of GNSS in regional climate monitoring and surface quality research.

CN119004616AInactive Publication Date: 2024-11-22WUHAN UNIV
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411105323.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-11-22
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

When the prior art uses GNSS data to invert surface quality changes, especially in the case of a sparsely distributed GNSS observation network in large areas, there are problems such as signal leakage, large uncertainty and low spatial resolution, making it difficult to obtain reliable inversion results.

Method used

Using the minimum model and the smoothest model based on geophysical inversion theory, combined with GNSS vertical displacement and other multi-source geologic data (such as satellite gravity data, rainfall data, and hydrological model data), jointly inversion is constructed through weighting functions and boundary constraints. Method to improve the accuracy and reliability of the inversion result.

Benefits of technology

It significantly improves the inversion accuracy and reliability of surface quality changes, reduces signal leakage errors, enhances the accuracy of inversion results under sparse GNSS observation network conditions, and promotes the research on regional climate monitoring and surface quality migration of GNSS application.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119004616A_ABST
    Figure CN119004616A_ABST
Patent Text Reader

Abstract

The invention discloses a GNSS (Global Navigation Satellite System) vertical displacement-based earth surface mass load inversion method and a computer system. According to the method, an inversion model is constructed based on a minimum model and a most gentle model in a geophysical inversion theory, a GNSS elevation time sequence is used as observation data, various prior information can be taken as a reference model to be included in an inversion process, and meanwhile, extra earth scientific data can be included in a solving process to be taken as quantitative constraints, so that the method is suitable for a global navigation satellite system. Multi-source data joint inversion is realized, and the precision and reliability of an inversion result are greatly improved. The invention provides an effective new method for researching the earth surface mass migration under the condition of a non-dense GNSS observation network in the current region, and promotes the application of the GNSS in the fields of near-real-time region climate monitoring, earth surface mass migration research and the like.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of geodesy and geodynamics, and is a new method for inverting surface mass load by using GNSS vertical displacement based on a mathematical model. Background Art

[0002] The redistribution of surface mass will cause changes in the gravity field and elastic deformation of the solid earth. Satellite gravity technology (such as GRACE and GRACE-FO) provides a global high-precision time-varying gravity field, providing a new means for large-scale surface mass migration and redistribution. However, GRACE / GRACE-FO observation technology still has difficult-to-overcome technical obstacles, such as limited temporal resolution (10 days or one month) and spatial resolution (~300km), large uncertainties (instrument measurement errors, signal leakage, background model bias, etc.) and high data missing rate. The Global Navigation Satellite System is another space geodetic technology that can accurately measure the displacement of the earth's surface in an all-weather, continuous, and high temporal and spatial resolution manner, including the elastic displacement caused by the mass load of surface and underground fluids on the crust, as well as the load deformation caused by changes in the atmosphere, oceans, and terrestrial water reserves (Van dam et al. 2001; Larochelle et al. 2018). At present, the methods for inverting surface mass changes based on GNSS vertical and horizontal motion are mainly spherical harmonics method and Green's function method.

[0003] After GNSS accurately measures the crustal deformation caused by surface loads, the observed data can be globally expanded using spherical harmonics, and then the surface mass changes can be effectively estimated using load theory. In areas with limited space, due to the small amount of observation data, there is a problem that the surface spherical harmonics do not satisfy orthogonality. The Slepian basis function can be introduced to improve the precision and accuracy of non-global modeling applications under limited observation bandwidth. However, the Slepian basis function requires truncation of the spherical harmonic coefficients by high-order terms. The loss of the truncated basis function on the signal will be reflected in the spatial resolution, resulting in serious signal leakage; on the other hand, the optimal number of truncation terms of the basis function depends on experience, and it is difficult to ensure that the optimal spectral smoothing solution is obtained within the local area. As the study area expands, the sharp increase in the parameters to be estimated for the Slepian basis function brings uncertainty to the inversion results.

[0004] The inversion method based on the load Green's function is suitable for areas with dense GNSS stations. Since the surface mass load covers the entire study area, inverting the regional surface mass changes through discrete GNSS observation data is a typical ill-posed problem. In order to solve the ill-posedness problem of the inversion equation and improve the reliability of the inversion results, regularization theory is usually used to constrain the inversion results. The traditional load Green's function inversion model uses the Laplacian matrix as the regularization constraint matrix. However, the Laplacian matrix is ​​an isotropic second-order differential operator and is extremely sensitive to noise. When the GNSS stations are sparse, the spatial resolution of the inversion results is low, and at the boundary of the inversion area, the signal leakage error outside the area will bring significant uncertainty to the final result. At the same time, the signal-to-noise ratio of the GNSS vertical displacement is low, and the surface mass changes inverted using this method have certain uncertainties. Summary of the invention

[0005] Therefore, there is an urgent need to further develop inversion methods for obtaining reliable surface mass change inversion methods using GNSS data, especially in the case of sparsely distributed GNSS observation networks in large areas. It is necessary to invent surface mass migration inversion methods that integrate multi-source observation data for constraints, so as to provide technical support for further expanding the application of GNSS in terrestrial water cycle, extreme climate monitoring and related geological research.

[0006] The present invention proposes a method for jointly inverting surface quality changes by utilizing GNSS vertical displacement and other multi-source geoscientific data (including but not limited to: satellite gravity data, rainfall data and hydrological model data).

[0007] The present invention constructs a joint inversion method for surface mass load based on the minimum model and the most gentle model in geophysical inversion theory. The method uses GNSS elevation time series as observation data, and can incorporate a variety of prior information as reference models into the inversion process. At the same time, additional geoscience data can be incorporated into the solution process as quantitative constraints, realizing the joint inversion of multi-source data, greatly improving the accuracy and reliability of the inversion results. The present invention provides an effective new method for the study of surface mass migration under the current regional non-dense GNSS observation network conditions, and promotes the application of GNSS in near-real-time regional climate monitoring, surface mass migration research and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A flowchart of a method for performing joint inversion of surface mass load using GNSS elevation time series is provided in accordance with an embodiment of the present invention.

[0009] Figure 2 A schematic diagram of a chessboard model and random site distribution provided in one embodiment of the present invention.

[0010] Figure 3 A schematic diagram of an unconstrained inversion result and its residual provided by an embodiment of the present invention.

[0011] Figure 4 A schematic diagram of the inversion results and residuals of a reference model applied to a specific grid provided in one embodiment of the present invention.

[0012] Figure 5 A schematic diagram of an inversion result and a residual thereof with boundary constraints imposed on a specific grid provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0013] Figure 1 A flow chart showing a joint inversion of surface mass load using GNSS elevation time series is shown. Figure 1 As shown, the present invention integrates GNSS vertical displacement and other multi-source geological data (prior reference model, which can be InSAR data, hydrological model data, satellite gravity data) to perform joint inversion of surface mass changes.

[0014] Step 1: discretize the inversion region.

[0015] The study area is divided into a series of grid units (the inversion will be performed in grid units below). Each grid uses a disk model, and the Green function of the vertical displacement of the surface caused by it can be expressed as:

[0016]

[0017] Among them, h n is the Love number of vertical displacement load; G is the gravitational constant; R is the radius of the earth; g is the gravitational acceleration; P n is the Legendre function; θ is the angular distance between the GNSS station and the mass load (disc load); n is the order; Γ n The following recursive formula can be used for calculation:

[0018]

[0019] Figure 2 An example of grid division in the process of discretizing the inversion area is given, such as Figure 2 The chessboard is divided into 1°×1° grids. According to the grid positions and station positions (black circles are GNSS stations), the Green’s function between each grid and each station can be calculated using formula (1).

[0020] Step 2: Construct the objective function of the inversion model based on the minimum and flattest models.

[0021] In the inversion process, it is crucial to choose a suitable model objective function. Due to the limited number of observations, the model objective function can stabilize the inversion effect. At the same time, a series of prior information needs to be introduced into the inversion process to constrain the inversion results. According to the geophysical inversion evaluation theory, the present invention proposes an inversion model that combines the minimum model and the most gentle model:

[0022]

[0023] Where, the first term corresponds to the minimum model, the second and third terms correspond to the flattest models in the longitude and latitude directions; S is the disk area, h0 is the reference model; h is the surface mass change, expressed in equivalent water height; r is the radius of the earth; w s , w λ is a spatially independent weighting function; w(θ) is a weighting function of the angular distance between the GNSS station and the disk payload; α s , α λ Adjust w s , w λ Weighting coefficients of the three-term weighting function.

[0024] Studies have shown that the displacement caused by the disk load decays rapidly as it moves away from the center of the disk. When using the minimum structural model for inversion, it is necessary to consider the weight of the displacement caused by the disk model around the GNSS station. To address this problem, an angular distance weighting function in the spherical coordinate system is proposed, which can improve this problem to a certain extent:

[0025]

[0026] Where n is the order, N is the number of GNSS stations, and θ ji Represents the angular distance between the jth site and the i-th grid.

[0027] Step 3: Construct a data fitting difference function.

[0028] The elastic response of the stations in the study area to the mass load, that is, the vertical displacement at the station, is used as the observation data. For the model results constructed by inversion, it is necessary to measure the difference between the results and the observed data. The common sum of squared errors is used to measure the degree of fit between the inversion model and the observed data:

[0029]

[0030] Among them, h pre and d obe are the predicted data and observed data respectively, σ j is the standard deviation of the data, W dis a diagonal matrix with diagonal elements equal to 1 / σ j .

[0031] Step 4: Construct the inversion objective function.

[0032] The data fit is poor and meets the condition In the case of , that is, when the data fitting error is less than the set limit error, To solve the inversion model objective function Φ m The minimum value of . Using the first-order finite difference approximation, the inversion objective function can be discretized into a matrix form. By introducing the regularization parameter, this constrained optimization problem is transformed into an unconstrained optimization problem, that is, solving the minimum value of the following inversion objective function:

[0033]

[0034] Among them, μ is the regularization parameter, W m It is the matrix form of each weight coefficient in formula (2).

[0035] Step 5: Add boundary constraints.

[0036] In order to meet the requirements of prior information, the discretized grid can be quantitatively constrained in equation (2), and the inversion objective function (5) becomes the constraint equation (6), that is, the inversion result must meet some mandatory constraints:

[0037]

[0038] Among them, h1 and h2 are the lower and upper limits of the inequality constraints respectively.

[0039] The Lagrange multiplier and quadratic penalty function are introduced into the above formula, and the final inversion objective function with boundary constraints is:

[0040]

[0041] Among them, C0, C1, C2 represent equality constraints, lower bound constraints, and upper bound constraints respectively:

[0042]

[0043] Among them, λ is the Lagrange multiplier, M is the penalty factor, and F represents the constraint information at each grid. If there is constraint information, the value is 1, otherwise it is 0. Y1 and Y2 are slack variables introduced by the lower and upper bound constraints. M0, M2, and M3 represent the penalty factors for equality constraints, lower bound constraints, and upper bound constraints. F0, F1, and F2 represent the constraint information for equality constraints, lower bound constraints, and upper bound constraints.

[0044] Step 6: Conjugate gradient solution;

[0045] In the unconstrained case, the derivative of formula (5) is expanded as follows:

[0046]

[0047] Set the reference model h0 to the zero model and use the conjugate gradient algorithm to solve the above formula to obtain the inversion result under unconstrained conditions ( Figure 3 ) and its difference with the original model ( Figure 2 ) is the residual between .

[0048] Before inversion, the above unconstrained inversion results ( Figure 3 ) is used as the reference model value and assigned to h0. At this time, the conjugate gradient algorithm is used for inversion. The result and its residual are as follows: Figure 4 Given the reference model information for a specific area, the inversion results in the five blocks are more consistent with the original model ( Figure 2 ).

[0049] because Figure 2 The range of the chessboard model is 0-500mm. Without setting a reference model, we set upper and lower limits of 0-500mm for all grids. Figure 3 The grid in the medium and large squares is set with upper and lower limits of 450-500 mm, and then formula (7) is expanded and solved:

[0050]

[0051] In the formula, h0 is the reference model, which can be the InSAR data, hydrological model data, or satellite gravity data in the figure.

[0052] The inversion results are as follows Figure 5 As shown, after boundary constraints (450-500 mm) were imposed on the specific five blocks, the inversion results were confined within the specified boundaries, the inversion results fit the original model well, and the residuals were greatly reduced, indicating that imposing additional boundary constraints can greatly improve the reliability and accuracy of the inversion results.

[0053] The present invention also provides an embodiment of a computer system. The computer system includes a processor and a memory. The memory is used to store non-temporary computer-readable instructions (e.g., one or more computer program modules). The processor is used to run the non-temporary computer-readable instructions, and when the non-temporary computer-readable instructions are run by the processor, one or more steps in the method for joint inversion of surface mass load using GNSS elevation time series described above can be executed. The memory and the processor can be interconnected via a bus system and / or other forms of connection mechanisms.

[0054] For example, the processor may be a central processing unit (CPU), a graphics processing unit (GPU), or other forms of processing units with data processing capabilities and / or program execution capabilities. For example, the central processing unit (CPU) may be an X86 or ARM architecture, etc. The processor may be a general-purpose processor or a special-purpose processor, and may control other components in the computer to perform desired functions.

[0055] For example, the memory may include any combination of one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, erasable programmable read-only memory (EPROM), compact disc read-only memory (CD-ROM), USB memory, flash memory, etc. One or more computer program modules may be stored on the computer-readable storage medium, and the processor may run one or more computer program modules to implement various functions of the computer.

[0056] The present invention also provides an embodiment of a computer-readable storage medium, which is used to store non-transitory computer-readable instructions. When the non-transitory computer-readable instructions are executed by a computer, one or more steps in the above-mentioned method for joint inversion of surface mass load using GNSS elevation time series can be implemented. When the joint inversion of surface mass load using GNSS elevation time series provided by the embodiment of the present invention is implemented in the form of software and sold or used as an independent product, it can be stored in a computer-readable storage medium. For the relevant description of the storage medium, please refer to the corresponding description of the memory in the computer above, which will not be repeated here.

Claims

1. A method for inverting surface mass load, characterized in that: include: The study area is divided into several grids, and the inversion is performed in grid units, with disks used to represent the grids; The objective function of constructing the inversion model based on the minimum model and the flattest model is: Where the first term corresponds to the minimum model, the second and third terms correspond to the flattest models in the longitude and latitude directions, S is the disk area, h0 is the reference model, h is the surface mass change expressed in equivalent water height, r is the radius of the earth, and w is s , w λ is a spatially independent weighting function; w(θ) is a weighting function of the angular distance between the GNSS station and the disk payload, α s , α λ Adjust w s , w λ The weighting coefficients of the three weighting functions; The vertical displacement at the GNSS station in the study area is used as the observation data, and the sum of squared errors is used to measure the degree of fit between the inversion model and the observation data: Where N is the number of GNSS stations, j represents the jth GNSS station; h pre and dob e are the inversion model prediction data and GNSS observation data, respectively, j is the standard deviation of the data, W d is a diagonal matrix with diagonal elements equal to 1 / σ j ; When the data fitting error is less than the set limit error, the objective function Φ of the inversion model is solved. m Minimum value of: Where μ is the regularization parameter; Derivative expansion of formula (3): The conjugate gradient algorithm is used to solve formula (4) and the inversion result of surface mass load under unconstrained conditions is obtained.

2. The method for inverting surface mass load according to claim 1, characterized in that: The reference model h0 is InSAR data, hydrological model data or satellite gravity data.

3. The method for inverting surface mass load according to claim 1, characterized in that: The weighting function w(θ) is designed as shown in (5) below to improve the problem that the displacement caused by the disk load decays rapidly as it moves away from the center of the disk: Where n is the order, P n is the Legendre function, N is the number of GNSS stations, θ ji Represents the angular distance between the jth site and the i-th grid.

4. The method for inverting surface mass load according to claim 1, characterized in that: In equation (1), the discretized grid is quantitatively constrained, and the inversion objective function (3) becomes the constraint equation (6): Where h1 and h2 are the lower and upper limits of the inequality constraints respectively.

5. The method for inverting surface mass load according to claim 4, characterized in that: Introducing Lagrange multipliers and quadratic penalty functions into equation (6), we obtain the inversion objective function with boundary constraints: Among them, C0, C1, C2 represent equality constraints, lower bound constraints, and upper bound constraints respectively: Among them, λ is the Lagrange multiplier, M is the penalty factor, F represents the constraint information at each grid. If there is constraint information, the value is 1, otherwise it is 0, Y1 and Y2 are slack variables introduced by the lower bound constraint and the upper bound constraint, M1, M2, and M3 represent the penalty factors in the three cases of equality constraint, lower bound constraint, and upper bound constraint, and F1, F2, and F3 represent the constraint information in the three cases of equality constraint, lower bound constraint, and upper bound constraint.

6. The method for inverting surface mass load according to claim 5, characterized in that: Expand and solve formula (7): The conjugate gradient algorithm is used to solve the above formula (9) and obtain the inversion result of the surface mass load under the constraint condition.

7. A computer system, characterized in that: include: processor; a memory including one or more computer program modules; The one or more computer program modules are stored in the memory and configured to be executed by the processor, and the one or more computer program modules include instructions for implementing the method for inverting surface mass load as described in any one of claims 1-6.

8. A computer-readable storage medium for storing non-transitory computer-readable instructions, characterized in that: When the non-transitory computer-readable instructions are executed by a computer, the method for inverting surface mass load according to claims 1 to 6 can be implemented.

Citation Information

Cited By

  • Method for inverting land water reserves based on Slepian basis function combined with GNSS three-dimensional deformation and related equipment

    CN121632286A

  • Elastic load earth surface quality inversion method and system based on spatial structure constraint

    CN122616166A