Method, device, electronic equipment and medium for correcting gravity wave drag at the top of the intermediate layer

Abstract

The present application relates to the field of space exploration technology, and more particularly to a method, apparatus, electronic device, and medium for correcting gravity wave drag at the mesosphere. The method comprises: obtaining an ultradense gravity wave momentum flux (MF) of a target step length; fitting the ultradense MF of the target step length to obtain the gravity wave drag and residual at the mesosphere, estimating a first error of the gravity wave drag and a confidence interval of the residual under a target confidence level; removing outliers from the ultradense MF based on the confidence interval of the residual, fitting the ultradense MF after removing the outliers to obtain a corrected gravity wave drag at the mesosphere, estimating a second error of the gravity wave drag under a target confidence level, and calculating an error reduction rate based on the first error and the second error. This method solves the problem that gravity wave drag obtained based on meteor radar observations often has systematically large values ​​and, under the same confidence level, its error interval is also more divergent.

Description

Technical Field

[0001] The present application relates to the field of space exploration technology, and in particular to a method, device, electronic equipment and medium for correcting gravity wave drag at the top of an intermediate layer. Background Art

[0002] The mesopause (generally considered to be around 80 km in altitude) is the lowest point on Earth's surface, located at the boundary between the mesosphere and the lower thermosphere. Due to a lack of solar heating and intense radiative cooling from carbon dioxide, the mesopause is the coldest region on Earth, reaching temperatures as low as -100°C. The mesopause is the transition zone between the neutral atmosphere and the thermosphere / ionosphere, where important physical and chemical processes occur and where the exchange of matter and energy within the atmosphere is crucial.

[0003] Gravity waves (GWs) are one of the most important fluctuations in atmospheric dynamics, active at the mesopause. It is generally believed that most GWs originate in the lower atmosphere and propagate upward. Due to the decrease in atmospheric density, the amplitude of GWs increases exponentially, exceeding a critical level of instability at the mesopause, where they break and deposit momentum and energy into the background atmosphere. This drives the structural evolution of the local space environment and, consequently, creates vertical coupling between the atmosphere at lower altitudes and the mesopause. Gravity waves can also cause adiabatic cooling of the polar mesopause in summer and adiabatic heating in winter, leading to residual meridional circulation at this altitude.

[0004] Gravity wave momentum flux (MF) and its evolution at altitude can be used to quantitatively characterize wind disturbances caused by gravity waves and provide a crucial parameter, gravity wave drag, that characterizes the impact of gravity waves on the background atmosphere. Meteor radar, which inverts atmospheric dynamic parameters in the mesopause region based on meteor detection, is unaffected by day and night and weather conditions. It can continuously detect meteors (with an altitude resolution of 0.1 km) in meteor-dense regions of the mesopause (generally 80–90 km). It is a widely deployed advanced device for measuring atmospheric wind fields and high-frequency, small-scale fluctuations.

[0005] Currently, there are mature technologies for inverting atmospheric gravity wave momentum flux and gravity wave drag in the mesopause region based on meteor radar. However, studies both domestically and internationally have found that compared with other observational methods, the gravity wave drag derived from meteor radar observations often has systematically larger values, and under the same confidence level, its error range is also more divergent. Summary of the Invention

[0006] The present application provides a method, device, electronic device and medium for correcting gravity wave drag at the top of the mesosphere to address the problem that gravity wave drag obtained based on meteor radar observations often has systematically large values ​​and, under the same confidence level, its error range is also more divergent.

[0007] A first aspect of the present application provides a method for correcting gravity wave drag at the top of an intermediate layer, comprising the following steps: obtaining an ultra-dense gravity wave momentum flux MF of a target step size; fitting the ultra-dense MF of the target step size to obtain the gravity wave drag and residual at the top of the intermediate layer, and estimating a first error of the gravity wave drag and a confidence interval of the residual under a target confidence level; removing outliers from the ultra-dense MF according to the confidence interval of the residual, fitting the ultra-dense MF after removing the outliers to obtain a corrected gravity wave drag at the top of the intermediate layer, estimating a second error of the gravity wave drag under a target confidence level, and calculating an error reduction rate based on the first error and the second error.

[0008] Optionally, obtaining the ultra-dense gravity wave momentum flux MF of the target step size includes: obtaining the height resolution of the meteor detection; determining the target step size according to the height resolution of the meteor detection, wherein the target step size is greater than or equal to the height resolution; obtaining the height matrix of the mesosphere top region, and determining the ultra-dense MF of the target step size by inverting from the height matrix according to the target step size within the target height range centered on the current height.

[0009] Optionally, fitting an overdense MF of a target step size to obtain the gravity wave drag and residual at the top of the mesosphere includes: obtaining an all-one matrix corresponding to the altitude matrix; constructing a data matrix based on the altitude matrix and the all-one matrix in series; determining a divergence parameter vector based on the data matrix and the overdense matrix corresponding to the overdense MF; calculating the gravity wave drag at the top of the mesosphere based on the divergence parameter vector and the average atmospheric density in the area of ​​the top of the mesosphere, and calculating the residual corresponding to the overdense matrix based on the divergence parameter vector, the altitude matrix, the all-one matrix and the overdense matrix corresponding to the overdense MF.

[0010] Optionally, the calculation formula of the first error is:

[0011]

[0012] in, is the average atmospheric density in the upper mesosphere region of 82–90 km; v1 is the matrix (A Ht T A Ht ) -1 The first diagonal element of Ht The data matrix constructed by the height matrix Ht and the corresponding all-one matrix One is calculated as follows:

[0013]

[0014] For the standard normal distribution Quantiles are calculated as follows:

[0015] f(a) is the probability density function of the standard normal distribution,

[0016]

[0017] F(a) is the cumulative distribution function corresponding to f(a),

[0018]

[0019] for When , the value of a;

[0020] To estimate the standard deviation, the calculation formula is:

[0021]

[0022] Optionally, removing outliers in the ultra-dense MF according to the confidence interval of the residual includes: locating outlier numbers in the ultra-dense MF according to the confidence interval of the residual; and removing all outliers in the ultra-dense MF according to the outlier numbers.

[0023] Optionally, the gravity wave drag force after correction is:

[0024]

[0025] Among them, GWD new is the corrected GW drag; b 0,new is the corrected divergence parameter vector b new The first dollar of

[0026] The second error is:

[0027]

[0028] Among them, v 1，new is the matrix (A Ht,new T A Ht,new ) -1 The first diagonal element of is the estimated standard deviation after correction, specifically:

[0029] Let the ultra-dense MF with the corresponding step length after removing the outliers be Fp new , and Fp new The corresponding residual is r new ,but

[0030] r new=Fp new -(b 0，new *Ht new +b 1,new *One new )

[0031] =[r 1，new r 2,new … r x，new …r N,new ] T

[0032] =[r′1 r′2 r′3…r′ N-n-1 r′ N-n ] T ,

[0033] Among them, r x，new Represents the same as Fp new The corresponding residual matrix r new The subscript in the residual element is x, which is the same as Fp x Correspondingly, the above residual matrix r new The total number of residual elements in is Nn, and Among them, [x1 x2 ... x n ] is the number of the outlier; [r′1 r′2 r′3 ... r′ N-n-1 r′ N-n ] T For [r 1，new r 2,new ... r x，new ... r N,new ] T The method of renumbering in sequence according to the subscript;

[0034]

[0035] A second aspect of the present application provides a device for correcting gravity wave drag at the top of the intermediate layer, including: an acquisition module for acquiring an ultra-dense gravity wave momentum flux MF of a target step length; a first estimation module for fitting the ultra-dense MF of the target step length to obtain the gravity wave drag and residual at the top of the intermediate layer, and estimating a first error of the gravity wave drag and a confidence interval of the residual under a target confidence level; a second estimation module for eliminating outliers in the ultra-dense MF according to the confidence interval of the residual, fitting the ultra-dense MF after eliminating the outliers to obtain a corrected gravity wave drag at the top of the intermediate layer, estimating a second error of the gravity wave drag under a target confidence level, and calculating an error reduction rate based on the first error and the second error.

[0036] The third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and runnable on the processor, wherein the processor executes the program to implement the intermediate layer top gravity wave drag correction method of the first aspect.

[0037] The fourth embodiment of the present application provides a computer-readable storage medium having a computer program or instruction stored thereon. When the computer program or instruction is executed, the method for correcting the gravity wave drag of the intermediate layer top according to the first aspect is implemented.

[0038] The fifth embodiment of the present application provides a computer program product, including a computer program or instructions, which, when executed, implements the intermediate layer top gravity wave drag correction method of the first aspect.

[0039] Therefore, this application has the following beneficial effects:

[0040] The present embodiment obtains and fits the ultradense gravity wave momentum flux MF of a target step size to obtain the gravity wave drag and residual at the mesosphere top. The first error of the gravity wave drag and the confidence interval of the residual are estimated under a target confidence level. Outliers are removed from the ultradense MF. The ultradense MF after outlier removal is fitted to obtain a corrected gravity wave drag at the mesosphere top. The second error of the gravity wave drag is estimated under a target confidence level. The error reduction rate is calculated based on the first and second errors. This corrects the systematically overstated gravity wave drag values ​​obtained based on meteor radar observations and the more divergent error intervals under the same confidence level, significantly converging the values. This solves the problem that the gravity wave drag obtained based on meteor radar observations often has systematically overstated values ​​and, under the same confidence level, has a more divergent error interval.

[0041] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0043] Figure 1 This is a flow chart of a method for correcting gravity wave drag at the top of the intermediate layer according to an embodiment of the present application;

[0044] Figure 2 A schematic diagram of GW drag and GW drag error fitting at a confidence level of 90% and an intuitive summary of statistical results under different step-size conditions provided in accordance with one embodiment of the present application;

[0045] Figure 3A schematic diagram of locating outliers in an ultra-dense MF based on residual confidence intervals according to one embodiment of the present application;

[0046] Figure 4 A schematic diagram showing a comparison of the GW drag and the GW drag error at a confidence level of 90% before and after outlier removal according to an embodiment of the present application;

[0047] Figure 5 This is a block diagram of a device for correcting gravity wave drag at the top of the intermediate layer according to an embodiment of the present application;

[0048] Figure 6 A schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0049] The following describes in detail embodiments of the present application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0050] The following describes, with reference to the accompanying drawings, a method, apparatus, electronic device, and medium for correcting gravity wave drag at the mesosphere according to an embodiment of the present application. To address the problem mentioned in the background art above that gravity wave drag obtained based on meteor radar observations often has systematically overstated values ​​and, under the same confidence level, a more divergent error range, the present application provides a method for correcting gravity wave drag at the mesosphere. In this method, an ultradense gravity wave momentum flux (MF) of a target step size is acquired and fitted to obtain the gravity wave drag and residual at the mesosphere. A first error of the gravity wave drag and a confidence range for the residual are estimated under the target confidence level. Outliers in the ultradense MF are removed, and the ultradense MF after removing the outliers is fitted to obtain a corrected gravity wave drag at the mesosphere. A second error of the gravity wave drag is estimated under the target confidence level. An error reduction rate is calculated based on the first and second errors. This method corrects the systematically overstated values ​​of the gravity wave drag obtained based on meteor radar observations and the more divergent error range under the same confidence level, significantly converging the gravity wave drag. This solves the problem that the gravity wave drag force obtained based on meteor radar observations often has a systematically large value, and under the same confidence level conditions, its error range is also more divergent.

[0051] Specifically, Figure 1 A schematic flow chart of a method for correcting gravity wave drag at the top of an intermediate layer provided in an embodiment of the present application.

[0052] like Figure 1 As shown, the method for correcting the gravity wave drag at the top of the intermediate layer includes the following steps:

[0053] In step S101 , the ultra-dense gravity wave momentum flux MF of the target step length is obtained.

[0054] The target step size is the sliding step size of the ultra-dense MF determined according to the height resolution of meteor detection, which is expressed as Step and is not specifically limited here. The method of obtaining the ultra-dense gravity wave momentum flux MF of the target step size will be described in detail below and will not be repeated here.

[0055] It can be understood that the embodiment of the present application determines the target step length according to the height resolution of meteor detection, and obtains the ultra-dense gravity wave momentum flux MF of the target step length.

[0056] In an embodiment of the present application, obtaining the ultra-dense gravity wave momentum flux MF of the target step size includes: obtaining the height resolution of the meteor detection; determining the target step size according to the height resolution of the meteor detection, wherein the target step size is greater than or equal to the height resolution; obtaining the height matrix of the mesosphere top region, and determining the ultra-dense MF of the target step size by inverting from the height matrix according to the target step size within the target height range centered on the current height.

[0057] Among them, the height matrix of the middle layer top area Ht=[Ht1 Hr2 ... Ht x ... Ht N ] T , Ht x represents the xth height element in the height matrix Ht, N represents the total number of height elements in the height matrix Ht, and Ht x -Ht x-1 =Step, the target altitude range centered on the current altitude is represented by RGE, which is set according to actual needs and is not specifically limited here; Fp is used to represent the super-dense MF of the target step length, and the formula Fp=Func_MF(Step,Ht,RGE)=[Fp1 Fp2 ... Fp x ... Fp N ] T A super-dense MF is implemented by inverting the target step size from the height matrix to determine the target step size.

[0058] It can be understood that the embodiment of the present application determines the target step size that is greater than the resolution by obtaining the height resolution of meteor detection, and obtains the height matrix of the top area of ​​the middle layer and the target height range centered on the current height. By substituting the three into the corresponding formula, the ultra-dense MF of the target step size can be inverted.

[0059] In step S102, the ultra-dense MF of the target step size is fitted to obtain the gravity wave drag and residual of the intermediate layer top, and the first error of the gravity wave drag and the confidence interval of the residual are estimated under the target confidence level condition.

[0060] Among them, the gravity wave drag force and residual of the top of the intermediate layer are obtained by fitting the ultra-dense MF with the target step size through the least squares method.

[0061] It can be understood that the embodiment of the present application can illustrate the least squares method for fitting the ultra-dense MF of the target step size to obtain the gravity wave drag and residual at the top of the intermediate layer, and estimate the first error of the gravity wave drag and the confidence interval of the residual under the target confidence level condition. The specific implementation method will be described in detail below and will not be repeated here.

[0062] In an embodiment of the present application, an ultra-dense MF of a target step size is fitted to obtain the gravity wave drag and residual at the top of the mesosphere, including: obtaining an all-one matrix corresponding to the altitude matrix; constructing a data matrix based on the altitude matrix and the all-one matrix in series; determining a divergence parameter vector based on the data matrix and the ultra-dense matrix corresponding to the ultra-dense MF; calculating the gravity wave drag at the top of the mesosphere based on the divergence parameter vector and the average atmospheric density in the area of ​​the top of the mesosphere, and calculating the residual corresponding to the ultra-dense matrix based on the divergence parameter vector, the altitude matrix, the all-one matrix and the ultra-dense matrix corresponding to the ultra-dense MF.

[0063] Among them, the all-one matrix corresponding to the height matrix is ​​a matrix with the same dimension as the height matrix but all elements in the matrix are 1, which is represented by One; the data matrix is ​​constructed by connecting the height matrix and the all-one matrix in series, that is, connecting the two arrays of the height matrix and the all-one matrix in series, which can be obtained by the formula Implementation, where A Ht is the constructed data matrix, Ht is the height matrix, One is the all-one matrix; through the data matrix A Ht The super-dense matrix Fp corresponding to the super-dense MF determines the divergence parameter vector b, which can be determined by the formula b = [b0 b1] T =(A Ht T A Ht ) -1 A Ht T Fp is realized, where b0 is the first element of the divergence parameter vector and b1 is the second element of the divergence parameter vector; according to the first element b0 of the divergence parameter vector and the average atmospheric density of the mesosphere top area The gravity wave drag GWD at the top of the mesosphere can be calculated by the formula Implementation: According to the divergence parameter vectors b0 and b1, the height matrix Ht, the all-one matrix One and the super-dense matrix Fp corresponding to the super-dense MF, the residual r corresponding to the super-dense matrix can be calculated by the formula r=Fp-(b0*Ht+b1*One)=[r1 r2 ... r x ... r N ] T accomplish.

[0064] It can be understood that after obtaining the ultra-dense matrix corresponding to the ultra-dense MF, the embodiment of the present application can coordinate the data matrix obtained by concatenating the height matrix and the corresponding all-one matrix to calculate and determine the divergence parameter vector. By substituting the average atmospheric density and the divergence parameter vector b in the top of the mesosphere according to the formula, the gravity wave drag force of the top of the mesosphere can be obtained; at the same time, the divergence parameter vector, the height matrix, the all-one matrix and the ultra-dense matrix corresponding to the ultra-dense MF are substituted into the corresponding formula to calculate the residual corresponding to the ultra-dense matrix.

[0065] In the embodiment of the present application, the calculation formula of the first error is:

[0066]

[0067] in, is the average atmospheric density in the upper part of the mesosphere; v1 is the matrix (A Ht T A Ht ) -1 The first diagonal element of Ht The data matrix constructed by the height matrix Ht and the corresponding all-one matrix One is calculated as follows:

[0068]

[0069] For the standard normal distribution Quantiles are calculated as follows:

[0070] f(a) is the probability density function of the standard normal distribution,

[0071]

[0072] F(a) is the cumulative distribution function corresponding to f(a),

[0073]

[0074] for When , the value of a;

[0075] To estimate the standard deviation, the calculation formula is:

[0076]

[0077] It is understandable that the embodiment of the present application obtains the average atmospheric density of the mesosphere top region. Construct a data matrix A by concatenating the height matrix and the corresponding all-one matrix Ht , calculate the matrix (A Ht TA Ht ) -1 The first diagonal element v1 of the standard normal distribution Quantile and the estimated standard deviation Substituting into the calculation formula of the first error, the first error can be calculated.

[0078] In step S103, outliers in the overdense MF are eliminated according to the confidence interval of the residual, and the overdense MF after outliers are eliminated is fitted to obtain the corrected gravity wave drag at the top of the intermediate layer. The second error of the gravity wave drag is estimated under the target confidence level condition, and the error reduction rate is calculated based on the first error and the second error.

[0079] Among them, the confidence interval of the residual is [rint1,rint2], as follows:

[0080] rint1=[rint11 rint12 ... rint1 x ... rint1 N ] T

[0081] rint2=[rint21 rint22 ... rint2 x ... rint2 N ] T

[0082] Among them, rint1 x (rint2 x ) represent the xth lower (upper) limit element in the residual confidence interval lower (upper) limit matrix rint1 (rint2), and r x Correspondingly, N represents the total number of lower (upper) limit elements in the above lower (upper) limit matrix rint1 (rint2); taking the xth lower (upper) limit element as an example, the calculation method is as follows:

[0083]

[0084] in, is a t-distribution with ν degrees of freedom Quantiles, specifically:

[0085] v=N-2

[0086] Suppose the probability density function of a t-distribution with degrees of freedom ν is g(a), which is as follows:

[0087]

[0088] The cumulative distribution function corresponding to g(a) is G(a), which is as follows:

[0089]

[0090] but for When , the value of a; std() is a mature function for solving standard deviation; r Student,x is the xth Studentized residual matrix, which is calculated by removing the xth element from Ht, One, and Fp. The results after removal are Ht Student,x 、One Student,x 、Fp Student,x , Ht Student,x and One Student,x Combined into the xth Studentized data matrix A Ht,x , re-use the least squares fit to obtain the xth Studentized residual r Student,x , let the xth Studentized divergence parameter vector be b x Among them, b 0,x and b 1,x are the first and second elements of the Studentized divergence parameter vector, respectively; all lower (upper) limit elements from 1 to N are processed according to the above process until the N×2 residual confidence interval [rint1, rint2] is filled; after obtaining the N×2 residual confidence interval [rint1, rint2], the outliers in the ultra-dense MF are located and removed. After removing the outliers in the ultra-dense MF, the least squares method is used again to fit the corrected GW drag. The GW drag error is estimated under the condition of a confidence level of 1-α, and compared with the previous results. The error reduction rate is calculated based on the first error and the second error.

[0091] It can be understood that the embodiment of the present application can calculate the confidence interval of the residual through the above method, locate and eliminate outliers in the ultra-dense MF, and fit the ultra-dense MF after eliminating the outliers to obtain the corrected gravity wave drag at the top of the intermediate layer, estimate the second error of the gravity wave drag under the target confidence level condition, calculate the error reduction rate based on the first error and the second error, and the method of estimating the second error of the gravity wave drag and calculating the error reduction rate based on the first error and the second error will be described in detail below and will not be repeated here.

[0092] In the embodiment of the present application, removing outliers in the ultra-dense MF according to the confidence interval of the residual includes: locating the outlier numbers in the ultra-dense MF according to the confidence interval of the residual; and removing all outliers in the ultra-dense MF according to the outlier numbers.

[0093] Among them, the xth lower limit and upper limit tuple [rint1 x ,rint2x ] as an example, if rint1 x <0, and rint2 x > 0, then in the plane Fp-Ht with Fp as the horizontal axis and Ht as the vertical axis, the point (Fp x , Ht x ) is the cluster point, and x is the number of the cluster point; otherwise, in the plane of Fp-Ht, point (Fp x , Ht x ) is an outlier point, and x is the number of the outlier point. All lower limit and upper limit tuples from 1 to N are processed according to the above process until every point in (Fp, Ht) is classified. Suppose a total of n outliers are obtained, and the numbers of the eliminated outliers are set as [x1 x2 ... x n ], and remove all outliers in the super-dense MF according to the outlier numbers.

[0094] It can be understood that the embodiment of the present application determines whether a point is a convergent point or an outlier by comparing the positive and negative values ​​of the lower limit element and the upper limit element in a lower limit element and an upper limit element tuple, and numbers the obtained outliers, and determines and eliminates all outliers in the super-dense MF by numbering.

[0095] In the embodiment of the present application, the corrected gravity wave drag force is:

[0096]

[0097] Among them, GWD new is the corrected GW drag; b 0,new is the corrected divergence parameter vector b new The first dollar of

[0098] The second error is:

[0099]

[0100] Among them, v 1,new is the matrix (A Ht,new T A Ht,new ) -1 The first diagonal element of is the estimated standard deviation after correction, specifically:

[0101] Let the ultra-dense MF with the corresponding step length after removing the outliers be Fp new , and Fp new The corresponding residual is r new ,but

[0102] r new =Fp new -(b 0，new *Htnew +b 1，new *One new )

[0103] =[r 1，new r 2,new ...r x，new ... r N,new ] T

[0104] =[r′1 r′2 r′3... r′ N-n-1 r′ N-n ] T ,

[0105] Among them, r x,new Represents the same as Fp new The corresponding residual matrix r new The subscript in the residual element is x, which is the same as Fp x Correspondingly, the above residual matrix r new The total number of residual elements in is Nn, and Among them, [x1x2...x n ] is the number of the outlier; [r′1 r′2 r′3 ... r′ N-n-1 r′ N-n ] T For [r 1,new r 2,new ... r x,new ... r N,new ] T The method of renumbering in sequence according to the subscript;

[0106]

[0107] It is understandable that the embodiment of the present application can calculate the corrected divergence parameter vector b after removing the outliers. new The average atmospheric density of the mesosphere top region The corrected gravity wave drag force is calculated; at the same time, similar to the calculation of the first error, the embodiment of the present application obtains the average atmospheric density of the mesosphere top area Get the height matrix Ht after removing outliers new And the corresponding all-one matrix One new , concatenated into a data matrix A Ht,new , calculate the matrix (A Ht,new T A Ht,new ) -1 The first diagonal element v 1,new , standard normal distribution Quantile and the estimated standard deviation Substituting into the calculation formula of the second error, the second error can be calculated.

[0108] According to the method for correcting the gravity wave drag at the top of the mesosphere proposed in an embodiment of the present application, the gravity wave drag and residual at the top of the mesosphere are obtained by acquiring the ultra-dense gravity wave momentum flux MF of a target step size and fitting it, the first error of the gravity wave drag and the confidence interval of the residual are estimated under the target confidence level, and outliers in the ultra-dense MF are removed. The ultra-dense MF after removing the outliers is fitted to obtain the corrected gravity wave drag at the top of the mesosphere, the second error of the gravity wave drag is estimated under the target confidence level, and the error reduction rate is calculated based on the first error and the second error. The systematically large values ​​of the gravity wave drag obtained based on meteor radar observations and the more divergent error intervals under the same confidence level are corrected, so that the intervals are greatly converged.

[0109] The following is a further description of the method for correcting gravity wave drag at the mesopause through a specific example. This example is based on observation data and atmospheric density data from the meteor radar at the YC station at 40°N in August 2020. The correction of GW drag in the mesopause region based on ultra-dense MF outlier diagnosis is performed. The specific steps are as follows:

[0110] Step 1: According to the height resolution of meteor detection, input the sliding step size of the super-dense MF (the step size is not less than the resolution), and gradually use the mature inversion method to calculate within a certain height range centered on the current height to obtain the super-dense MF of the corresponding step size, specifically:

[0111] Assume that the height resolution of meteor detection is Met_Resolution, Met_Resolution = 0.1km;

[0112] Assume that the sliding step of the ultra-dense MF is Step, then: Step ≥ Met_Resolution;

[0113] In this embodiment, Step∈[0.1 0.5 1 2], unit: km;

[0114] Assume that: under the condition of step length Step, the height matrix of the mesosphere top region 82-90 km is Ht, which is as follows:

[0115] Ht=[Ht1 Ht2 ... Ht x ... Ht N ] T

[0116] Among them, Ht x represents the xth height element in the height matrix Ht of 82-90 km, N represents the total number of height elements in the height matrix Ht, and Htx -Ht x-1 =Step;

[0117] Assume that a certain altitude range centered on the current altitude is RGE, RGE = 2km;

[0118] Assume that the super dense MF corresponding to the step size is Fp, specifically:

[0119] Fp=Func_MF(Step,Ht,RGE)=[Fp1 Fp2 ... Fp x ... Fp N ] T

[0120] Among them, Func_MF() is a mature inversion method for ultra-dense MF matrix, Fp x represents the xth super-dense MF element in the super-dense MF matrix Fp calculated using the mature inversion method, and Ht x Correspondingly, N represents the total number of super-dense MF elements in the above super-dense MF matrix Fp;

[0121] Step 2: Based on the overdense MF obtained in step 1, the least squares method is used to fit the GW drag and residual corresponding to the overdense MF. The confidence intervals of the GW drag error and residual are estimated under the condition of a confidence level of 1-α (α = 0.1), and compared with the fitting results under the original step size conditions. Specifically:

[0122] Let the all-one matrix corresponding to the height matrix Ht be One, specifically:

[0123] One=[1 1 ... 1 ... 1] T

[0124] Among them, all elements of One are 1, and there are N elements in total, which is consistent with the height matrix Ht;

[0125] Construct a data matrix by connecting Ht and One in series. Let the constructed data matrix be A Ht , specifically:

[0126]

[0127] Based on the super-dense MF obtained in step 1, the least squares method is used to fit the GW drag and residual corresponding to the super-dense MF. The GW drag error and residual confidence interval are estimated under the condition of confidence level 1-α, and compared with the fitting results under the original step size conditions. Specifically:

[0128] Let the divergence parameter vector be b, specifically:

[0129] b=[b0 b1] T =(A Ht T A Ht ) -1 A Ht T Fp

[0130] Among them, b0 is the first element of the divergence parameter vector, b1 is the second element of the divergence parameter vector;

[0131] Assume that the average atmospheric density in the upper mesosphere region of 82-90 km is The drag force of gravity waves is GWD, then

[0132]

[0133] Assume that the residual error corresponding to Fp obtained by fitting is r, then

[0134] r=Fp-(b0*Ht+b1*One)=[r1 r2 ... r x ... r N ] T

[0135] Among them, r x Represents the xth residual element in the residual matrix r corresponding to Fp, which is the same as Fp x Correspondingly, N represents the total number of residual elements in the above residual matrix r;

[0136] Assume that the GW drag error (confidence level is 1-α) is GWD_Error, then

[0137]

[0138] in, is a standard normal distribution Quantiles, specifically:

[0139] Suppose the probability density function of a standard normal distribution is f(a), then we have

[0140]

[0141] Then the cumulative distribution function corresponding to f(a) is F(a), then we have

[0142]

[0143] but for When , the value of a;

[0144] v1 is the matrix (A Ht T A Ht) -1 The first diagonal element of ;

[0145] To estimate the standard deviation, specifically:

[0146]

[0147] is the average atmospheric density in the upper region of the mesosphere at 82-90 km;

[0148] Assume that the residual confidence interval (confidence level is 1-α) is [rint1, rint2], specifically:

[0149] rint1=[rint11 rint12 ... rint1 x ... rint1 N ] T

[0150] rint2=[rint21 rint22 ... rint2 x ... rint2 N ] T

[0151] Among them, rint1 x (rint2 x ) represent the xth lower (upper) limit element in the residual confidence interval lower (upper) limit matrix rint1 (rint2), and r x Correspondingly, N represents the total number of lower (upper) limit elements in the above lower (upper) limit matrix rint1 (rint2);

[0152] Taking the xth lower (upper) limit element as an example, the calculation method is as follows:

[0153]

[0154] in, is a t-distributed with v degrees of freedom Quantiles, specifically:

[0155] v=N-2

[0156] Assume that the probability density function of a t-distribution with a degree of freedom of ν is g(a), then we have

[0157]

[0158] The cumulative distribution function corresponding to g(a) is G(a), then we have

[0159]

[0160] but for When , the value of a;

[0161] std() is a mature function for solving standard deviation;

[0162] Assume r Student，x is the xth Studentized residual matrix, specifically:

[0163] Eliminate the xth element in Ht, One, and Fp (the results after elimination are Ht Student,x 、One Student,x 、Fp Student,x ), Ht Student,x and One Student,x Combined into the xth Studentized data matrix A Ht,x , re-use the least squares fit to obtain the xth Studentized residual r Student,x , specifically:

[0164] Ht Student,x =[Ht1 Ht2 ... Ht x-1 Ht x+1 ... Ht N ] T

[0165] One Student,x =[1 1 ... 1 1 ... 1]

[0166]

[0167] Fp Student,x =[Fp1 Fp2 ... Fp x-1 Fp x+1 ... Fp N ] T

[0168] Let the xth Studentized divergence parameter vector be b x , specifically:

[0169] b x =[b 0，x b 1,x ] T =(A Ht，x T A Ht,x ) -1 A Ht,x T Fp Student,x

[0170] Among them, b0，x and b 1，x are the first and second elements of the Studentized divergence parameter vector respectively;

[0171] r Student,x =Fp Student,x -(b 0，x *Ht Student，x +b 1，x *One Student,x )

[0172] For all lower (upper) limit elements from 1 to N, the above process is followed until the N×2 residual confidence interval [rint1, rint2] is filled;

[0173] Adjust the sliding step size Step of the super-dense MF (in this example, Step∈[0.1 0.5 1 2], unit: km), repeat the above steps, and compare the GW drag and GW drag error (confidence level is 1-α) corresponding to the super-dense MF under different step size conditions (including the original step size, in this case, the original step size is 2 km);

[0174] Figure 2 The following is a schematic diagram of the fitting and intuitive statistical results of the GW drag and GW drag error (with a confidence level of 90%) under different step length conditions (2 km, 1 km, 0.5 km, and 0.1 km, respectively, with 2 km being the original step length). Figure 2 As shown:

[0175] When Step=2km, GWD=194.8m / s / d, GWD_Error=77.7m / s / d;

[0176] When Step = 1 km, GWD = 159.2 m / s / d, GWD_Error = 52 m / s / d;

[0177] When Step=0.5km, GWD=141.5m / s / d, GWD_Error=30.7m / s / d;

[0178] When Step = 0.1 km, GWD = 111 m / s / d, GWD_Error = 16.8 m / s / d;

[0179] Depend on Figure 2 Analysis shows that as Step decreases, the value of GWD gradually decreases, and the error range of its 90% confidence level gradually converges;

[0180] Step 3: Based on the residual confidence interval obtained in step 2, locate the outliers in the super-dense MF, specifically:

[0181] Take the xth lower limit and upper limit tuple [rint1 x , rint2 x ] as an example:

[0182] If rint1 x <0, and rint2 x > 0, then in the plane of Fp-Ht (with Fp as the horizontal axis and Ht as the vertical axis), the point (Fp x , Ht x ) is the cluster point, and x is the number of the cluster point; otherwise, in the plane of Fp-Ht, point (Fp x , Ht x ) is an outlier, and x is the number of the outlier;

[0183] For all lower and upper bound tuples from 1 to N, the above process is followed until every point in (Fp, Ht) is classified;

[0184] Assume that a total of n outliers are obtained in this step, and the number of the removed outliers is set to [x1 x2 ... x n ];

[0185] In this embodiment, Figure 3 The schematic diagram of locating outliers in super-dense MF according to residual confidence intervals, where solid points represent convergent points, hollow points represent outliers, the horizontal axis is the subscript number of the convergent / outlier point, the vertical axis is the residual value, and the error bar represents the confidence interval with a confidence level of 90%, as shown in Figure 2. Figure 3 As shown:

[0186] n=7

[0187] [x1 x2 ... x n ]=[1 4 8 9 10 25 29]

[0188] Step 4: Based on the ultra-dense MF obtained in step 1, remove the outliers obtained in step 3 and re-fit the calibrated GW drag using the least squares method. Estimate the GW drag error under the condition of a confidence level of 1-α and compare it with the previous results. Specifically:

[0189] As mentioned in step 3, the n outliers are numbered as [x1 x2 ... x n ];

[0190] Put the Fp and Ht obtained in step 1, the [x1 x2 ... x n ] elements are all eliminated (the results after elimination are Fp new 、Ht new、One new ), and construct the correction data matrix A Ht,new , using least squares fitting, the corrected GW drag is obtained, and the GW drag error is estimated under the condition of a confidence level of 1-α, specifically:

[0191] Ht new =[Ht1 Ht2 ... Ht x ... Ht N ] T

[0192] Among them, Ht x The height element with the subscript x in the height matrix Ht of 82-90 km is represented, and here:

[0193]

[0194] With Ht new Correspondingly,

[0195] One new =[1 1 ... 1 ... 1] T

[0196]

[0197] Fp new =[Fp1 Fp2 ... Fp x ... Fp N ] T

[0198] Similarly, here

[0199] Let the corrected divergence parameter vector be b new , specifically:

[0200] b new =[b 0，new b 1，new ] T =(A Ht,new T A Ht,new ) -1 A Ht,new T Fp new

[0201] Among them, b 0，new is the first element of the corrected divergence parameter vector, b 1，new is the second element of the corrected divergence parameter vector;

[0202] Let the corrected GW drag be GWD new, GW drag error (confidence level is 1_α) is GWD_Error new ,but:

[0203]

[0204] in, is the average atmospheric density in the mesosphere top region of 82-90 km (consistent with the value in step 2), is a standard normal distribution Quantile (same as the value in step 2), v 1，new is the matrix (A Ht,new T A Ht,new ) -1 The first diagonal element of is the estimated standard deviation after correction, specifically:

[0205] Assume that after correction, new The corresponding residual is r new ,but

[0206] r new =Fp new -(b 0，new *Ht new +b 1，new *One new )

[0207] =[r 1，new r 2,new ... r x，new ... r N,new ] T

[0208] =[r′1 r′2 r′3 ... r′ N-n-1 r′ N-n ] T

[0209] Among them, r x，new Represents the same as Fp new The corresponding residual matrix r new The subscript in the residual element is x, which is the same as Fp x Correspondingly, the above residual matrix r new The total number of residual elements in is Nn, and

[0210] In addition, [r′1 r′2 r′3 ... r′ N-n-1 r′ N-n ] T To [r 1，new r 2,new ... rx,new ... r N，new ] T The subscripts of are renumbered in order;

[0211]

[0212] The obtained GWD new 、GWD_Error new Compare the results with those in step 2;

[0213] In this embodiment, if Figure 4 As shown:

[0214] GWD new =120.3m / s / d

[0215] GWD_Error new =11.9m / s / d;

[0216] Figure 4 The figure shows the comparison of GW drag and GW drag error (confidence level is 90%) before and after removing outliers. The solid points represent the integrated points, the hollow points represent the outliers, and the dotted line represents the least squares fitting result. Figure 4 The result analysis shows that the error range of 90% confidence level of GWD has further converged;

[0217] Combine Figure 2 The GW drag and drag error under the original step length (2 km) condition show that the new estimated GW drag is 74.5 m / s / d lower than the original method, and the error at the 90% confidence level is 65.8 m / s / d lower.

[0218] In summary, the present method in the embodiment can effectively correct the systematically large values ​​of gravity wave drag obtained based on meteor radar observations, and effectively correct the more divergent error intervals under the same confidence level conditions, making them converge significantly.

[0219] Next, the intermediate layer top gravity wave drag correction device proposed according to the embodiment of the present application is described with reference to the accompanying drawings.

[0220] Figure 5 It is a block diagram of the intermediate layer top gravity wave drag correction device of an embodiment of the present application.

[0221] like Figure 5 As shown, the intermediate layer top gravity wave drag correction device 10 includes: an acquisition module 201, an estimation module 202 and a calculation module 203.

[0222] Among them, the acquisition module 201 is used to obtain the ultra-dense gravity wave momentum flux MF of the target step length; the estimation module 202 is used to fit the ultra-dense MF of the target step length to obtain the gravity wave drag and residual at the top of the intermediate layer, and estimate the first error of the gravity wave drag and the confidence interval of the residual under the target confidence level condition; the calculation module 203 is used to eliminate outliers in the ultra-dense MF according to the confidence interval of the residual, fit the ultra-dense MF after eliminating the outliers to obtain the corrected gravity wave drag at the top of the intermediate layer, estimate the second error of the gravity wave drag under the target confidence level condition, and calculate the error reduction rate based on the first error and the second error.

[0223] In an embodiment of the present application, the acquisition module 201 is further used to: obtain the height resolution of meteor detection; determine the target step size based on the height resolution of meteor detection, wherein the target step size is greater than or equal to the height resolution; obtain the height matrix of the top area of ​​the intermediate layer, and determine the ultra-dense MF of the target step size by inverting from the height matrix based on the target step size within the target height range centered on the current height.

[0224] In an embodiment of the present application, the first estimation module 202 is further used to: obtain an all-one matrix corresponding to the altitude matrix; construct a data matrix based on the series connection of the altitude matrix and the all-one matrix; determine a divergence parameter vector based on the data matrix and the super-dense matrix corresponding to the super-dense MF; calculate the gravity wave drag at the top of the mesosphere based on the divergence parameter vector and the average atmospheric density in the mesosphere top area, and calculate the residual corresponding to the super-dense matrix based on the divergence parameter vector, the altitude matrix, the all-one matrix and the super-dense matrix corresponding to the super-dense MF.

[0225] In the embodiment of the present application, the calculation formula of the first error is:

[0226]

[0227] in, is the average atmospheric density in the upper mesosphere region of 82-90 km; v1 is the matrix (A Ht T A Ht ) -1 The first diagonal element of Ht The data matrix constructed by the height matrix Ht and the corresponding all-one matrix One is calculated as follows:

[0228]

[0229] For the standard normal distribution Quantiles are calculated as follows:

[0230] f(a) is the probability density function of the standard normal distribution,

[0231]

[0232] F(a) is the cumulative distribution function corresponding to f(a),

[0233]

[0234] for When , the value of a;

[0235] To estimate the standard deviation, the calculation formula is:

[0236]

[0237] In the embodiment of the present application, the second estimation module 203 is further configured to: locate outlier numbers in the ultra-dense MF according to the confidence interval of the residual; and remove all outliers in the ultra-dense MF according to the outlier numbers.

[0238] In the embodiment of the present application, the corrected gravity wave drag force is:

[0239]

[0240] Among them, GWD new is the corrected GW drag; b 0，new is the corrected divergence parameter vector b new The first dollar of

[0241] The second error is:

[0242]

[0243] Among them, v 1，new is the matrix (A Ht，new T A Ht,new ) -1 The first diagonal element of is the estimated standard deviation after correction, specifically:

[0244] Let the ultra-dense MF with the corresponding step length after removing the outliers be Fp new , and Fp new The corresponding residual is r new ,but

[0245] r new =Fp new -(b 0，new *Ht new +b 1，new *One new )

[0246] =[r 1，new r2，new ... r x，new ... r N，new ] T

[0247] =[r′1 r′2 r′3 ... r′ N-n-1 r′ N-n ] T ,

[0248] Among them, r x，new Represents the same as Fp new The corresponding residual matrix r new The subscript in the residual element is x, which is the same as Fp x Correspondingly, the above residual matrix r new The total number of residual elements in is Nn, and Among them, [x1 x2 ... x n ] is the number of the outlier; [r′1 r′2 r′3 ... r′ N-n-1 r′ N-n ] T For [r 1，new r 2，new ... r x，new ... r N，new ] T The method of renumbering in sequence according to the subscript;

[0249]

[0250] It should be noted that the above explanation of the embodiment of the intermediate layer top gravity wave drag correction method is also applicable to the intermediate layer top gravity wave drag correction device of this embodiment, and will not be repeated here.

[0251] According to the intermediate layer top gravity wave drag correction device proposed in the embodiment of the present application, through the coordinated action of the acquisition module, the first estimation module and the second estimation module, it is possible to obtain the ultra-dense gravity wave momentum flux MF of the target step size and perform fitting to obtain the gravity wave drag and residual of the intermediate layer top, estimate the first error of the gravity wave drag and the confidence interval of the residual under the target confidence level condition, remove outliers in the ultra-dense MF, fit the ultra-dense MF after removing the outliers to obtain the corrected gravity wave drag of the intermediate layer top, estimate the second error of the gravity wave drag under the target confidence level condition, calculate the error reduction rate based on the first error and the second error, correct the systematically large value of the gravity wave drag obtained based on the meteor radar observation and the more divergent error interval under the same confidence level condition, and make the interval converge significantly.

[0252] Figure 6This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0253] Memory 301 , processor 302 , and computer programs stored in the memory 301 and executable on the processor 302 .

[0254] When the processor 302 executes the program, the intermediate layer top gravity wave drag correction method provided in the above embodiment is implemented.

[0255] Furthermore, the electronic device further includes:

[0256] The communication interface 303 is used for communication between the memory 301 and the processor 302 .

[0257] The memory 301 is used to store computer programs that can be run on the processor 302 .

[0258] The memory 301 may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile memory, such as at least one disk memory.

[0259] If the memory 301, processor 302, and communication interface 303 are implemented independently, the communication interface 303, memory 301, and processor 302 can be connected to each other via a bus and communicate with each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0260] Optionally, in a specific implementation, if the memory 301, the processor 302 and the communication interface 303 are integrated on a chip, the memory 301, the processor 302 and the communication interface 303 can communicate with each other through an internal interface.

[0261] The processor 302 may be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0262] An embodiment of the present application further provides a computer-readable storage medium having a computer program or instruction stored thereon. When the computer program or instruction is executed, the above-mentioned method for correcting gravity wave drag at the top of the intermediate layer is implemented.

[0263] An embodiment of the present application also provides a computer program product, including a computer program or instructions, which, when executed, implements the above-mentioned method for correcting the gravity wave drag force at the top of the intermediate layer.

[0264] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0265] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "N" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0266] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing a custom logical function or process step, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed in a different order than shown or discussed, including performing functions in a substantially simultaneous manner or in a reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application pertain.

[0267] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, the steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement the method: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array, a field programmable gate array, etc.

[0268] A person skilled in the art may understand that all or part of the steps carried out in the method for implementing the above-mentioned embodiment may be completed by instructing the relevant hardware through a program, and the above-mentioned program may be stored in a computer-readable storage medium, which, when executed, includes one of the steps of the method embodiment or a combination thereof.

[0269] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for correcting gravity wave drag at the top of the intermediate layer, characterized in that: The following steps are involved: Obtain the ultra-dense gravity wave momentum flux MF of the target step length; Fitting the ultra-dense MF of the target step size to obtain the gravity wave drag and residual at the top of the intermediate layer, and estimating the first error of the gravity wave drag and the confidence interval of the residual under the target confidence level condition; Outliers in the overdense MF are eliminated according to the confidence interval of the residual, the overdense MF after the outliers are eliminated is fitted to obtain a corrected gravity wave drag at the top of the intermediate layer, a second error of the gravity wave drag is estimated under the target confidence level condition, and an error reduction rate is calculated based on the first error and the second error.

2. The method for correcting the gravity wave drag force of the intermediate layer top according to claim 1, characterized in that: The step of obtaining the ultra-dense gravity wave momentum flux MF of the target step size includes: Get the high resolution of meteor detection; Determining the target step size according to the height resolution of the meteor detection, wherein the target step size is greater than or equal to the height resolution; The height matrix of the intermediate layer top region is obtained, and within the target height range centered at the current height, the ultra-dense MF of the target step length is determined by inverting from the height matrix according to the target step length.

3. The method for correcting the gravity wave drag force of the intermediate layer top according to claim 1, characterized in that: The ultra-dense MF fitting the target step size to obtain the gravity wave drag and residual at the top of the intermediate layer includes: Get the all-one matrix corresponding to the height matrix; Constructing a data matrix by serially connecting the height matrix and the all-ones matrix; Determine a divergence parameter vector according to the data matrix and the super-dense matrix corresponding to the super-dense MF; The gravity wave drag of the mesosphere top is calculated according to the divergence parameter vector and the average atmospheric density in the mesosphere top region, and the residual corresponding to the super-dense matrix is ​​calculated according to the divergence parameter vector, the height matrix, the all-one matrix and the super-dense matrix corresponding to the super-dense MF.

4. The method for correcting the gravity wave drag force of the intermediate layer top according to claim 3, characterized in that: The calculation formula of the first error is: ， in, is the average atmospheric density in the mesosphere top region; is a matrix The first diagonal element of , where is the height matrix And the corresponding all-one matrix The constructed data matrix is ​​calculated as follows: ， in, Represents the height matrix The Height units, Represents the height matrix The total number of medium and high level elements; For the standard normal distribution Quantiles are calculated as follows: is the probability density function of the standard normal distribution, ， For The corresponding cumulative distribution function, ， for hour, The value of To estimate the standard deviation, the calculation formula is: ， in, represents the residual matrix corresponding to the ultra-dense MF of the corresponding step size The Residual element, is the residual matrix corresponding to the ultra-dense MF of the corresponding step size The total number of residual elements in .

5. The method for correcting the gravity wave drag force of the intermediate layer top according to claim 1, characterized in that: Eliminating outliers in the super-dense MF according to the confidence interval of the residual includes: Locating the outlier number in the super-dense MF according to the confidence interval of the residual; All outliers in the super-dense MF are removed according to the outlier numbers.

6. The method for correcting the gravity wave drag force of the intermediate layer top according to claim 4, characterized in that: The corrected gravity wave drag force is: in, is the corrected GW drag; is the corrected divergence parameter vector The first dollar; The second error is: ， in, is a matrix The first diagonal element of is the height matrix after removing outliers And the corresponding all-one matrix The data matrix is ​​concatenated. is the estimated standard deviation after correction, specifically: Assume that the ultra-dense MF with the corresponding step length after removing the outliers is ,and The corresponding residual is ,but in, Represents The corresponding residual matrix The subscript in The residual element of Correspondingly, the residual matrix The total number of residual elements in , where n is the number of outliers, is the residual matrix corresponding to the ultra-dense MF of the corresponding step size The total number of residual elements in , and ,in, is the number of the outlier point; for The method of renumbering in sequence according to the subscript; is the corrected divergence parameter vector The second element of ， in, Represents the residual matrix corresponding to the ultra-dense MF of the corresponding step size after removing the outliers The Residual element.

7. An intermediate layer top gravity wave drag correction device, characterized in that: include: An acquisition module is used to obtain the ultra-dense gravity wave momentum flux MF of the target step length; a first estimation module, configured to fit the ultra-dense MF of the target step size to obtain the gravity wave drag and residual at the top of the intermediate layer, and estimate the first error of the gravity wave drag and the confidence interval of the residual under a target confidence level condition; The second estimation module is configured to remove outliers from the overdense MF according to a confidence interval of the residual, fit the overdense MF after removing the outliers to obtain a corrected gravity wave drag at the top of the intermediate layer, estimate a second error of the gravity wave drag under the target confidence level, and calculate an error reduction rate based on the first error and the second error.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for correcting gravity wave drag at the top of the intermediate layer according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed, the method for correcting the gravity wave drag at the top of the intermediate layer according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed, the method for correcting the gravity wave drag at the top of the intermediate layer according to any one of claims 1 to 6 is implemented.

Images