Atmospheric waveguide monitoring and diagnostic method and system

Abstract

The application discloses an atmospheric waveguide monitoring and diagnosing method and system, and the method comprises the following steps: acquiring meteorological data of a FY-4 satellite, wherein the meteorological data comprises water vapor pressure, air pressure and temperature; pre-processing the meteorological data to obtain a first data set; processing the first data set based on a horizontal high-resolution construction algorithm to obtain a high-resolution second data set; obtaining atmospheric refractive index according to the second data set; and obtaining atmospheric waveguide data according to the atmospheric refractive index. The meteorological data of the meteorological satellite is processed based on the horizontal high-resolution construction algorithm to obtain high-resolution atmospheric refractive index; the atmospheric waveguide is generated due to atmospheric refraction effect, and therefore, the parameters of the atmospheric waveguide can be obtained according to the atmospheric refractive index and its profile, so that the atmospheric waveguide in a sea area range can be monitored.

Description

Technical Field

[0001] This invention relates to the field of atmospheric waveguide monitoring technology, specifically to an atmospheric waveguide monitoring and diagnostic method and system. Background Technology

[0002] Atmospheric waveguides, as an important component of the marine environment, have a significant impact on shipborne and shore-based electronic equipment such as radar, communications, and electronic countermeasures. When the refraction of electromagnetic waves exceeds the curvature of the Earth, the electromagnetic waves will be trapped in a certain space, forming an atmospheric waveguide phenomenon. Based on their formation and altitude, atmospheric waveguides can be divided into three main categories: evaporation waveguides, surface waveguides, and suspended waveguides.

[0003] Traditional atmospheric waveguide monitoring methods face challenges such as insufficient real-time monitoring capabilities over large sea areas and weak monitoring capabilities for various types of atmospheric waveguides. Furthermore, considering factors such as foreign restrictions on the export of waveguide monitoring equipment and related hardware and software, there is an urgent need to design and develop methods for atmospheric waveguide monitoring utilizing domestically developed technologies and resources.

[0004] Currently, there are three main types of methods for real-time monitoring of atmospheric waveguides both domestically and internationally: direct measurement techniques, model diagnostic techniques, and inversion monitoring techniques. Direct measurement techniques involve directly measuring the temperature, humidity, and air pressure of each pressure layer, then calculating the atmospheric refractive index of each layer to obtain the atmospheric refractive index profile. Model diagnostic techniques utilize the principle of similarity, using a small amount of meteorological and hydrological data as input, and employing appropriate algorithms to obtain the atmospheric refractive index profile and related parameters of the atmospheric waveguide. Inversion monitoring techniques analyze the intensity of radar, communication, and laser signals affected by atmospheric waveguides, and then invert to obtain the atmospheric refractive index profile in the pressure layer, thereby obtaining relevant atmospheric waveguide data.

[0005] The aforementioned technologies still have significant gaps in terms of wide-area coverage and real-time monitoring capabilities before practical application. For example, using Raman lidar to monitor atmospheric waveguides is limited by the large size and weight of Raman lidar, which allows for single-point measurements of atmospheric waveguides on land. It cannot be easily installed on shipboard platforms, thus preventing the measurement of atmospheric waveguide data at sea. Furthermore, to measure atmospheric waveguides within a specific area using this method, it is necessary to move the measurement equipment from one location to another, severely limiting the measurement area. Most existing atmospheric waveguide monitoring methods are only applied to scientific research; large-scale operational capabilities have not yet been established, failing to meet the support needs of both military and civilian applications for atmospheric waveguides. This severely restricts the availability of early warning detection and navigation support equipment within my country's maritime areas.

[0006] With the continuous advancement of satellite remote sensing technology and computer hardware and software capabilities, monitoring the temporal and spatial variations of atmospheric refractive index using satellite remote sensing and low-level spatial modeling techniques has become an important research direction. For example, the Fengyun-4 satellite can collect meteorological data over a large area of ​​sea in real time; however, there is currently a lack of technology to monitor atmospheric waveguides using meteorological data from meteorological satellites. Therefore, there is an urgent need for an atmospheric waveguide monitoring technology based on meteorological satellites that meets these requirements. Summary of the Invention

[0007] To address the aforementioned technical problems in the existing technology, this invention provides an atmospheric waveguide monitoring and diagnostic method and system, which obtains atmospheric waveguides based on meteorological data collected by the Fengyun-4 satellite, and realizes atmospheric waveguide monitoring over a large sea area.

[0008] This invention discloses an atmospheric waveguide monitoring and diagnostic method, the method comprising: acquiring meteorological data from the Fengyun-4 satellite, the meteorological data including water vapor pressure, air pressure, and temperature; preprocessing the meteorological data to obtain a first dataset; processing the first dataset based on a horizontal high-resolution construction algorithm to obtain a high-resolution second dataset; obtaining the atmospheric refractive index based on the second dataset; and obtaining atmospheric waveguide data based on the atmospheric refractive index.

[0009] Preferred high-resolution horizontal construction algorithms include:

[0010] Based on the known points in the first dataset, obtain the estimated points;

[0011] Based on known points and estimated points, a high-resolution second dataset is obtained;

[0012] The estimated point is represented as:

[0013]

[0014] in, It is the estimated point, λ i Represented as weighting coefficients, z i Represented as a known point;

[0015] λ is obtained by inverting the following matrix. i :

[0016]

[0017] Where, r ij The variance is semivariance, and φ is the Lagrange multiplier.

[0018]

[0019] E[z] represents the unbiased estimator, z i and zj These are represented as different known points.

[0020] Preferably, the semivariance is solved using the following formula:

[0021] r ij =r(z) i -z j )

[0022] Among them, z i -z j It is expressed as the distance d between two points. ij r() represents d ij and r ij The fitting function;

[0023] The weighting coefficients must satisfy the following constraints:

[0024]

[0025] Preferably, the atmospheric refractive index is corrected:

[0026] The formula for correcting the atmospheric refractive index is selected from any of the following formulas:

[0027] N r =p1N 3 +p2N 2 +p3N+p4 (21)

[0028]

[0029] N r =a0+a2cos(wN)+b2sin(wN) (23)

[0030] Where N represents the atmospheric refractive index, N r This represents the corrected atmospheric refractive index, where p1, p2, p3, p4, a1, b1, c1, a0, a2, b2, and w are coefficients.

[0031] Preferably, the correction formula is selected based on minimizing the root mean square error;

[0032] Based on the selected correction formula, the corresponding coefficients are solved.

[0033] Preferably, the root mean square error corrected by Formula 23 is the smallest in the northern sea area with a pressure of 700-800 hPa. The coefficients of Formula 23 are: a0 = 226.3, a2 = 2.456, b2 = 18.65, w = 0.05497.

[0034] In the eastern sea area with a pressure of 700-800 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 220.5, a2 = 30.45, b2 = -18.34, w = 0.04256.

[0035] In the southern sea area with a pressure of 700-800 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 217.4, a2 = 42.89, b2 = -31.77, w = 0.01879.

[0036] In the northern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 271.5, a2 = 37.59, b2 = 33.99, w = 0.02007.

[0037] In the eastern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 22 is the smallest. The coefficients of Formula 22 are: a1 = 342.8, b1 = 413.7, c1 = 292.2.

[0038] In the southern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 21 is the smallest. The coefficients of Formula 21 are: p1 = -1.832e-0.5, p2 = 0.01296, p3 = -2.057, p4 = 241.3.

[0039] In the northern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 286.8, a2 = -13.14, b2 = 18.12, w = 0.046.

[0040] In the eastern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 306.3, a2 = 27.24, b2 = 21.84, w = 0.037.

[0041] In the southern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 336.4, a2 = 47.32, b2 = 40.96, w = 0.01612.

[0042] In the northern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 314.3, a2 = 0.2996, b2 = 17.35, w = 0.05992.

[0043] In the eastern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by formulas 21 and 23 is the smallest. The coefficients of formula 21 are: p1 = -0.0001241, p2 = 0.1175, p3 = -36.05, p4 = 3896. The coefficients of formula 23 are: a0 = 315.5, a2 = -19.89, b2 = 29.53, w = 0.028.

[0044] In the southern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by Formula 22 is the smallest. The coefficients of Formula 22 are: a1 = 468.9, b1 = 562.2, c1 = 394.8.

[0045] Preferably, the root mean square error corrected by Formula 23 is the smallest in the northern sea area with a pressure of 700-800 hPa. The coefficients of Formula 23 are: a0 = 226.3, a2 = 2.456, b2 = 18.65, w = 0.05497.

[0046] In the eastern sea area with a pressure of 700-800 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 220.5, a2 = 30.45, b2 = -18.34, w = 0.04256.

[0047] In the southern sea area with a pressure of 700-800 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 217.4, a2 = 42.89, b2 = -31.77, w = 0.01879.

[0048] In the northern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 271.5, a2 = 37.59, b2 = 33.99, w = 0.02007.

[0049] In the eastern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 22 is the smallest. The coefficients of Formula 22 are: a1 = 342.8, b1 = 413.7, c1 = 292.2.

[0050] In the southern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 21 is the smallest. The coefficients of Formula 21 are: p1 = -1.832e-0.5, p2 = 0.01296, p3 = -2.057, p4 = 241.3.

[0051] In the northern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by formula 2.3 is the smallest. The coefficients of formula 2.3 are: a0 = 286.8, a2 = -13.14, b2 = 18.12, w = 0.046.

[0052] In the eastern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 306.3, a2 = 27.24, b2 = 21.84, w = 0.037.

[0053] In the southern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 336.4, a2 = 47.32, b2 = 40.96, w = 0.01612.

[0054] In the northern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 314.3, a2 = 0.2996, b2 = 17.35, w = 0.05992.

[0055] In the eastern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by formulas 21 and 23 is the smallest. The coefficients of formula 21 are: p1 = -0.0001241, p2 = 0.1175, p3 = -36.05, p4 = 3896. The coefficients of formula 23 are: a0 = 315.5, a2 = -19.89, b2 = 29.53, w = 0.028.

[0056] In the southern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by Formula 22 is the smallest. The coefficients of Formula 22 are: a1 = 468.9, b1 = 562.2, c1 = 394.8.

[0057] Preferably, it is determined whether the meteorological data meets the following conditions:

[0058] or

[0059] Where X represents the parameters of the meteorological data. The mean is expressed as the average value, and STD is expressed as the standard deviation of the deviation.

[0060] If the conditions are met, the meteorological data is considered abnormal. The abnormal data in the meteorological data is then removed to obtain the first dataset.

[0061] Preferably, the data for the atmospheric waveguide includes an atmospheric refractive index profile;

[0062] The atmospheric pressure ranges for the meteorological data are: 700hPa-800hPa, 800hPa-900hPa, 900hPa-1000hPa, and 1000hPa-1100hPa, where the atmospheric pressure range is used to describe the altitude range.

[0063] Preferably, the atmospheric refractive index is calculated using the following method:

[0064]

[0065] Where N represents the atmospheric refractive index, T represents the atmospheric temperature, P represents the atmospheric pressure, e represents the water vapor pressure, q represents the specific humidity, and D and E are constants.

[0066] This invention also provides a system for implementing the above-mentioned atmospheric waveguide monitoring and diagnostic method, comprising an acquisition module, a preprocessing module, a horizontal high-resolution module, and an atmospheric waveguide analysis module.

[0067] The acquisition module is used to acquire meteorological data from meteorological satellites;

[0068] The preprocessing module is used to preprocess the meteorological data to obtain a first dataset;

[0069] The horizontal high-resolution module is used to process the first dataset based on the horizontal high-resolution construction algorithm to obtain a high-resolution second dataset.

[0070] The atmospheric waveguide analysis module is used to obtain the atmospheric refractive index based on the second dataset; and to obtain atmospheric waveguide data based on the atmospheric refractive index.

[0071] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0072] The meteorological data from meteorological satellites is processed using a horizontal high-resolution construction algorithm to obtain a high-resolution atmospheric refractive index. Atmospheric waveguides are generated due to atmospheric refraction effects, so the parameters of atmospheric waveguides can be obtained based on the atmospheric refractive index and its profile, enabling the monitoring of atmospheric waveguides over a large sea area. Attached Figure Description

[0073] Figure 1 This is a flowchart of the atmospheric waveguide monitoring method of the present invention;

[0074] Figure 2 This is a corrected scatter plot of atmospheric refractive index;

[0075] Figure 3 It is a graph showing the average error and root mean square error of atmospheric refractive index in the northern, eastern and southern sea areas at the same time.

[0076] Figure 4 This is a graph showing the average error and root mean square error of atmospheric refractive index at different times in the northern sea area.

[0077] Figure 5 This is a graph showing the average error and root mean square error of atmospheric refractive index at different times in the eastern sea area.

[0078] Figure 6 This is a graph showing the average error and root mean square error of atmospheric refractive index at different times in the southern sea area.

[0079] Figure 7 This is the system logic block diagram of the present invention. Detailed Implementation

[0080] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0081] The present invention will now be described in further detail with reference to the accompanying drawings:

[0082] An atmospheric waveguide monitoring and diagnostic method, such as Figure 1 As shown, the method includes:

[0083] Step 101: Obtain meteorological data from the Fengyun-4 satellite, including water vapor pressure, air pressure, and temperature.

[0084] Step 102: Preprocess the meteorological data to obtain the first dataset. For example, remove outlier data.

[0085] Step 103: Based on the horizontal high-resolution construction algorithm, process the first dataset to obtain a high-resolution second dataset.

[0086] The Fengyun-4 satellite has a horizontal resolution of only 16 km, while assessing radar performance in atmospheric duct environments requires meteorological data with a horizontal grid spacing of at least 1 km or even meters. Therefore, it is necessary to improve the resolution of the dataset.

[0087] Step 104: Obtain the atmospheric refractive index based on the second dataset.

[0088] The atmospheric refractive index can be calculated using the following method:

[0089]

[0090]

[0091] Where N represents the atmospheric refractive index, T represents the atmospheric temperature, P represents the atmospheric pressure, e represents the water vapor pressure, q represents the specific humidity, and D and E are constants that can be determined experimentally. In one specific embodiment, D = 77.6 K·hPa -1Given E = 4810 K and ε = 0.622, and considering the atmospheric pressure P, temperature T, and specific humidity q, the atmospheric refractive index N can be calculated.

[0092] Formula 1 can also be written as:

[0093] Step 105: Obtain the atmospheric refractive index profile and the data / parameters of the atmospheric waveguide based on the atmospheric refractive index.

[0094] Based on the horizontal high-resolution construction algorithm, meteorological data from the Fengyun-4 satellite are processed to obtain the atmospheric refractive index. Atmospheric waveguides are generated due to atmospheric refraction effects. Therefore, the parameters of atmospheric waveguides can be obtained based on the atmospheric refractive index and its profile, enabling the monitoring and diagnosis of atmospheric waveguides over a large sea area.

[0095] The preprocessing in step 102 includes removing outlier data from the meteorological data.

[0096] Data is considered abnormal if it meets the following conditions:

[0097] or

[0098] Where X represents the parameters of the meteorological data. denoted as the mean, and STD as the standard deviation of the deviation.

[0099] The atmospheric refractive index can also be corrected; and atmospheric waveguide data can be obtained based on the corrected atmospheric refractive index. The formula for correcting the atmospheric refractive index is selected from any of the following formulas:

[0100] N r =p1N 3 +p2N 2 +p3N+p4 (21)

[0101]

[0102] N r =a0+a2cos(wN)+b2sin(wN) (23)

[0103] Where N represents the atmospheric refractive index, Nr represents the corrected atmospheric refractive index, and p1, p2, p3, p4, a1, b1, c1, a0, a2, b2 and w are coefficients.

[0104] The above-mentioned correction formulas can be screened by minimizing the root mean square error; based on the screened correction formulas, the corresponding coefficients can be solved.

[0105] Example 1

[0106] Step 301: Obtain meteorological data from the Fengyun-4 satellite via the Fengyun Remote Sensing Satellite Data Service Network. Specifically, obtain regional atmospheric profile products from the GIIRS channel matching dataset, and select atmospheric temperature and specific humidity profile products, including air pressure, temperature, and specific humidity information for 101 layers in the 0hPa-1100hPa range.

[0107] Obtain radiosonde data, including meteorological elements such as temperature, dew point temperature, geopotential height, wind speed and direction from the near-surface to the upper atmosphere, to assess the reliability of atmospheric waveguide monitoring based on meteorological data.

[0108] Step 302: Select satellite sampling points near the radiosonde data points and obtain meteorological data for those points. Match the radiosonde data and meteorological data spatiotemporally to facilitate the verification of atmospheric waveguide monitoring data based on meteorological data using the radiosonde data. The selected satellite sampling points and radiosonde points are matched horizontally by latitude and longitude; vertically, they are matched using nearby pressure layers. If no corresponding satellite data is available for the radiosonde data, the data for that pressure layer is discarded.

[0109] Step 303: Preprocessing, removing outlier data, and obtaining the first dataset.

[0110] After removing outliers, the mean temperature deviation of each pressure layer decreased, the root mean square error decreased significantly, and the mean specific humidity deviation of each pressure layer decreased significantly.

[0111] In the specific data analysis, the first dataset was grouped and analyzed according to four pressure layers: 700hPa-800hPa, 800hPa-900hPa, 900hPa-1000hPa, and 1000hPa-1100hPa. The satellite data was also divided into northern, eastern, and southern sea areas. For example, the atmospheric refractive index was calculated using Formula 1. The pressure range is used to describe the altitude range.

[0112] Step 304: Based on the horizontal high-resolution construction algorithm, process the first dataset to obtain a high-resolution second dataset. The horizontal high-resolution construction algorithm includes: obtaining estimated points based on known points in the first dataset; and obtaining a high-resolution second dataset based on the known points and the estimated points.

[0113] The estimated point is represented as:

[0114]

[0115] in, It is the estimated point, λ i Represented as weighting coefficients, z iLet (x0, y0) be a known point. The estimated point at (x0, y0) is... The difference from the true value z0 is minimized, that is...

[0116]

[0117] Simultaneously satisfying unbiased estimation

[0118] For any point (x, y) within the region, its expected value c and variance σ 2 They must all be the same. That is, for any point (x, y), we have:

[0119] E[z(x,y)]=E[z]=c (303);

[0120]

[0121] Therefore, the value z(x, y) at any point within the region is composed of the average value c of the region and the random error R(x, y) at that point, i.e.

[0122] z(x,y)=E[z(x,y)]+R(x,y)=c+R(x,y) (305),

[0123] Where R(x,y) represents the deviation at point (x,y), and its variance is a constant.

[0124] Var[R(x,y)]=σ 2 (306).

[0125] Unbiased constraints Will Substituting it in, we get

[0126]

[0127] Since for any z we have E[z] = c, then

[0128]

[0129] Right now

[0130]

[0131] Formula 309 is λ i One of the constraints.

[0132] Introducing the cost function:

[0133] Then there is

[0134]

[0135] Let C ij =Cov(z) i ,z j )=Cov(R i ,R j ), here R i =z i -c, meaning point (x) i ,y i The deviation of the attribute value at point () from the average attribute value of the region is then:

[0136]

[0137] Define the semivariance function r ij =σ 2 -C ij Its equivalent is

[0138]

[0139] According to the definition above, R i =z i -c, with z i -z j =R i -R j Then there is

[0140]

[0141] And because

[0142]

[0143] E[R i R j ] = E[(z i -c)(z j -c)]=Cov(z i ,z j ) = C ij (315);

[0144] therefore

[0145]

[0146] And because

[0147]

[0148] E[R i R j ] = E[(z i -c)(z j -c)]=Cov(z i ,z j ) = Cij (317)

[0149] therefore

[0150]

[0151] Then only calculation is needed. It represents the similarity of attributes; spatial correlation is represented by distance, defined as the geometric distance between i and j.

[0152]

[0153] Assume r ij With d ij There exists a certain functional relationship, which can be arbitrary. To determine which functional relationship best meets our requirements, we need to process the existing dataset as follows:

[0154] {z(x1,y1),z(x2,y2),……,z(x n-1 ,y n-1 ),z(x n ,y n )}

[0155] Calculate the distance between any two points and semivariance At this point, we can obtain n. 2 (d) ij ,r ij The set of data pairs is used to fit the set of data pairs to a curve that best meets our requirements, which is used to fit the relationship between d and r. After calculation, the following functional relationship can be obtained:

[0156] r=r(d) (320)

[0157] That is, r ij =r(z) i -z j )

[0158] Then for any two points (x) i ,y i ),(x j ,y j First, the distance d can be calculated. ij Then, based on the functional relationship between them, the semivariance r of these two points is obtained. ij .

[0159] Substituting the defined semivariance function back into the cost function J, we have:

[0160]

[0161] because

[0162]

[0163] Where J is λ i To minimize J, we can subtract J from λ. i Find the partial derivative and set it to 0, that is...

[0164]

[0165] There is an additional condition: the calculated λ must be guaranteed. i Satisfy the formula For problems involving constraints and seeking optimal solutions, we can construct a new function as follows:

[0166]

[0167] Where φ is the Lagrange multiplier, if the parameters φ,λ1,λ2,…,λ that satisfy the requirements are solved... n The value of can be found such that When it is true, the minimum value of J is, i.e.

[0168]

[0169]

[0170] Because of C ij =Conv(z i , z j ) = C ji Therefore, r ij =r ji So there are

[0171]

[0172] This expression can be written as a system of linear equations as follows:

[0173]

[0174] Written in matrix form:

[0175]

[0176] λ is obtained by inverting the matrix. i Substitute The estimated point at (x0, y0) can then be obtained. Where r... ij φ is the semivariance, and φ is the Lagrange multiplier.

[0177] Right now

[0178] r ij =r(z) i -z j )

[0179] E[z] represents the unbiased estimator, z i and z j Let z represent different known points respectively. i -z j It is expressed as the distance d between two points. ij r() represents d ij and r ij The functional relationship.

[0180] Adaptability analysis of horizontal high-resolution construction algorithms:

[0181] Data from the northern, eastern, and southern sea areas were selected and tested in stratified layers from 700 hPa to 1100 hPa. 50%, 60%, 70%, and 80% of the data from each layer were taken as the original data for interpolation, and the remaining 50%, 40%, 30%, and 20% of the data were used as validation data. The mean error and root mean square error were used as standards to measure the optimal interpolation region. Validation was performed on each region size, from 3*3 to 9*9 grids.

[0182] In the northern sea area, as the percentage of input data increases, the average error of the N value does not change significantly, remaining roughly stable between -0.5N and 0.5N. Similarly, the root mean square error (RMSE) of the N value also does not change significantly, remaining roughly stable between 0.5N and 4.5N. However, the RMSE exceeds 3N in some areas for 4x4 and 5x5 grids. Considering the average error, the range of variation is small, and the average error differences across grid sizes are not significant. Therefore, the focus is primarily on the RMSE of each grid size. From the perspective of RMSE, grids exhibiting severe "shaking" in different pressure layers should be excluded. The 4x4 and 5x5 grids show obvious severe "shaking" with prominent "peaks," making them unsuitable for selecting grid areas of this size. Overall, as the grid size increases, the average error and root mean square error obtained by interpolation using each grid size do not show a linear relationship. However, the average error and root mean square error of 7*7 grid, 8*8 grid, and 9*9 grid are slightly better than those of other grid sizes. The average error and root mean square error of these grids change less with the change of pressure layer. Therefore, 7*7 grid, 8*8 grid, and 9*9 grid can be selected as the interpolation area in the northern sea area.

[0183] In the eastern sea area, except for the 3x3 grid, the average error of the N value does not change much with the increase of the percentage of input data, remaining roughly stable between -0.3N and 0.2N. The average error of the 3x3 grid is less stable than that of other grid sizes across all pressure layers, while the average errors of other grid sizes show less "fluctuation" across all pressure layers. With the increase of the percentage of input data, the root mean square error of the N value does not change much, remaining roughly stable between 0N and 4N. As the pressure layer increases from 700 hPa to 1100 hPa, the graphs for each percentage of input data show an upward trend in the root mean square error as the pressure layer increases. From the perspective of average error, except for the 3*3 grid, the average error of the other grids does not vary much, but the difference in average error between grid sizes is less than 0.2N. Therefore, the root mean square error of each grid size is mainly examined. From the perspective of root mean square error, grids with severe "shaking" in each pressure layer are excluded. Among them, the root mean square error of the 5*5 grid shows obvious "shaking" and "peaks" when the input data is 70% and 80%, while the root mean square error of the 3*3 grid is significantly smaller than that of other grid sizes when the input data is 80%. Moreover, when the input data is 50%, 60%, and 70%, there is no significant difference between the 3*3 grid and other grid sizes, thus eliminating the random error caused by the small amount of data in the 3*3 grid. Overall, as the grid size increases, the average error and root mean square error obtained by interpolation using each grid size do not show a linear relationship. Although the average error of the 3*3 grid is larger than that of other grids, the base is very small. However, in terms of root mean square error, it is significantly better than other grid sizes. The root mean square error of the 3*3 network also changes less with the change of the pressure layer. Therefore, the 3*3 grid can be selected as the interpolation area in the eastern sea area.

[0184] In the southern sea area, as the percentage of input data increases, the average error of the N value does not change significantly, remaining roughly stable between -0.6N and 1.5N. The average error of 3*3 and 4*5 grids shows better stability across all pressure layers than other grid sizes, while the average error of 5*5 and 6*6 grids exhibits greater "jitter" across all pressure layers. As the percentage of input data increases, the root mean square error (RMSE) of the N value changes significantly, generally ranging from 0N to 20N. Except for the 3*3 grid, the RMS error of the other grids exceeds 5N at all pressure layers. From the perspective of average error, except for the 5*5 grid, the difference in average error among all grid sizes is within 1N. Therefore, the RMS error of each grid size is the primary focus. From the perspective of RMS error, grids exhibiting severe "jitter" across pressure layers should be excluded. The 5*5, 6*6, and 7*7 grids show significant "jitter" with prominent "peaks," making them unsuitable for selection in this grid size region. Overall, as the grid size increases, the average error and root mean square error obtained by interpolation using each grid size do not show a linear relationship. However, the average error and root mean square error of the 3*3 grid are significantly better than those of other grid sizes. With the change of the pressure layer, the average error and root mean square error of the 3*3 grid change less. Therefore, it is more appropriate to use the 3*3 grid as the interpolation area in the southern sea area.

[0185] Based on the combined performance of average error and root mean square error across different grid sizes in various pressure layers of the northern, eastern, and southern sea areas, the 3x3 grid region significantly outperforms other grid sizes in the eastern and southern sea areas. However, in the northern sea area, the 7x7, 8x8, and 9x9 grid regions show significantly better performance. The percentage of input data has little impact on the average error and root mean square error; more than 50% of the input data is sufficient to meet the requirements for improved resolution.

[0186] Atmospheric refractive indices calculated from Fengyun-4 satellite data at 08:00, 14:00, and 20:00 were selected as datasets for three regions. The optimal interpolation regions for the northern sea area were randomly selected as 7*7, 8*8, and 9*9 grids, while the optimal interpolation regions for the eastern and southern sea areas were selected as 3*3 grids. Figure 3 The average error and root mean square error of atmospheric refractive index at the same time are shown. From the root mean square error, the root mean square error of atmospheric refractive index in the northern and eastern sea areas is less than 4N, while in the southern sea area, a small portion in the 1000hPa-1050hPa range is greater than 4N, but still less than 6.5N. In comparison, the accuracy of atmospheric refractive index of RS80 radiosonde can only reach ±2.0N. Considering the error of actual measurement and calculation, the best accuracy does not exceed ±6.5N. The actual horizontal high-resolution second dataset shows better results. Therefore, the horizontal high-resolution construction algorithm is applicable to all sea areas.

[0187] Figure 4 The average error and root mean square error of atmospheric refractive index at different times in the northern sea area are shown. The average error of N value does not vary much, and is roughly stable between -0.4N and 0.2N. The root mean square error ranges roughly between 0.2N and 3.2N. Figure 5 The average error and root mean square error of atmospheric refractive index at different times in the eastern sea area are shown. The average error of N value does not vary much, and is roughly stable between -0.5N and 0.52N. The root mean square error ranges roughly between 0.1N and 2.8N. Figure 6 The average error and root mean square error of atmospheric refractive index at different times in the southern sea area are shown. The average error of N value does not vary much and is roughly stable between -0.5N and 0.78N, while the root mean square error ranges roughly between 0N and 3.5N.

[0188] The average error of N values ​​in the morning, noon, and evening does not vary much across different sea areas. In the northern sea area, the average error of N values ​​in the morning, noon, and evening shows no strong correlation. In the eastern sea area, the average error of N values ​​in the morning, noon, and evening shows significant differences between 850 hPa and 1100 hPa, with the morning data being significantly larger, the noon data showing no significant deviation, and the evening data being significantly smaller. In the southern sea area, the average error of N values ​​in the morning, noon, and evening shows a strong correlation with other pressure layers except for the high-pressure layer.

[0189] From the perspective of root mean square error, the average error of N values ​​in the early, middle, and late periods varies little across different sea areas, generally ranging from 0 to 3.5 N. In the northern sea area, the root mean square error data of N values ​​in the early, middle, and late periods show greater fluctuations but no obvious temporal correlation. In the eastern sea area, the average error of N values ​​in the early, middle, and late periods fluctuates less in the low-pressure layer but fluctuates significantly more in the high-pressure layer. In the southern sea area, the average error of N values ​​in the early, middle, and late periods shows strong correlations with other pressure layers, except for the 700 hPa pressure layer.

[0190] It is evident that the horizontal high-resolution modeling algorithm can be used for data processing at any time in the three sea areas. However, there is still a certain error between the atmospheric refractive index processed using the horizontal high-resolution modeling algorithm and the radiosonde data.

[0191] Step 305: Correct the atmospheric refractive index.

[0192] Figure 2 The atmospheric refractive index scatter plot and correction curves corrected by Equation 21 are shown. The correction results for different sea areas and different pressure levels are shown in Table 1.

[0193] Table 1

[0194]

[0195]

[0196] As shown in Table 1, a better correction formula can be selected for each sea area and each pressure layer, and the root mean square error of each correction formula can meet the monitoring requirements. Among them, the root mean square error corrected by Formula 23 is smallest in the northern sea area with a pressure of 700-800 hPa, with the coefficients of Formula 23 being: a0 = 226.3, a2 = 2.456, b2 = 18.65, w = 0.05497; the root mean square error corrected by Formula 23 is smallest in the eastern sea area with a pressure of 700-800 hPa, with the coefficients of Formula 23 being: a0 = 220.5, a2 = 30.45, b2 = -18.34, w = 0.04256; the root mean square error corrected by Formula 23 is smallest in the southern sea area with a pressure of 700-800 hPa, with the coefficients of Formula 23 being: a0 = 217.4, a2 = 42.89, b2 = -31.77, w = 0.01879. (800-900 hPa) In the northern sea area at 800-900 hPa, the root mean square error corrected by Formula 23 is the smallest, with coefficients of a0 = 271.5, a2 = 37.59, b2 = 33.99, w = 0.02007. In the eastern sea area at 800-900 hPa, the root mean square error corrected by Formula 22 is the smallest, with coefficients of a1 = 342.8, b1 = 413.7, c1 = 292.2. In the southern sea area at 800-900 hPa, the root mean square error corrected by Formula 21 is the smallest, with coefficients of p1 = -1.832e-0.5, p2 = 0.01296, p3 = -2.057, p4 = 241.3. In the northern sea area at 900-1000 hPa, Formula 2... The formula with the smallest corrected root mean square error (RMSE) is: a0 = 286.8, a2 = -13.14, b2 = 18.12, w = 0.046. In the eastern sea area at 900-1000 hPa, the formula with the smallest corrected RMSE is: a0 = 306.3, a2 = 27.24, b2 = 21.84, w = 0.037. In the southern sea area at 900-1000 hPa, the formula with the smallest corrected RMSE is: a0 = 336.4, a2 = 47.32, b2 = 40.96, w = 0.01612.In the northern sea area at 1000-1100 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 314.3, a2 = 0.2996, b2 = 17.35, w = 0.05992. In the eastern sea area at 1000-1100 hPa, the root mean square error corrected by Formulas 21 and 23 is the smallest. The coefficients of Formula 21 are: p1 = -0.0001241. p2 = 0.1175, p3 = -36.05, p4 = 3896, the coefficients of formula 23 are: a0 = 315.5, a2 = -19.89, b2 = 29.53, w = 0.028; in the southern sea area with 1000-1100 hPa, the root mean square error corrected by formula 22 is the smallest, the coefficients of formula 22 are: a1 = 468.9, b1 = 562.2, c1 = 394.8.

[0197] Based on the calculation results of the correction formula and coefficients, the correction formula and corresponding coefficients for each sea area and pressure layer can be quickly found; through meteorological satellite data, the corrected atmospheric refractive index and related atmospheric waveguide parameters can be accurately calculated.

[0198] Example 2

[0199] This embodiment provides a system for implementing the above-described atmospheric waveguide monitoring and diagnostic method, such as... Figure 7 As shown, it includes 1. an acquisition module, 2. a preprocessing module, 3. a horizontal high-resolution module, and 4. an atmospheric waveguide analysis module.

[0200] Acquisition module 1 is used to acquire meteorological data from meteorological satellites;

[0201] Preprocessing module 2 is used to preprocess the meteorological data to obtain a first dataset;

[0202] The horizontal high-resolution module 3 is used to process the first dataset based on the horizontal high-resolution construction algorithm to obtain a high-resolution second dataset;

[0203] The atmospheric waveguide analysis module 4 is used to obtain the atmospheric refractive index based on the second dataset; and to obtain atmospheric waveguide data based on the atmospheric refractive index.

[0204] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An atmospheric waveguide monitoring and diagnostic method, characterized in that, The method includes: Acquire meteorological data from the Fengyun-4 satellite, including water vapor pressure, air pressure, and temperature; The meteorological data is preprocessed to obtain the first dataset; Based on a horizontal high-resolution construction algorithm, the first dataset is processed to obtain a high-resolution second dataset; wherein the horizontal high-resolution construction algorithm includes: Based on the known points in the first dataset, obtain the estimated points; Based on known points and estimated points, a high-resolution second dataset is obtained; The estimated point is represented as: in, It is the estimated point, λ i Represented as weighting coefficients, z i Represented as a known point; λ is obtained by inverting the following matrix. i : Where, r ij The variance is semivariance, and φ is the Lagrange multiplier. E[z] represents the unbiased estimator, z i and z j These are respectively represented as different known points; The semivariance is calculated using the following formula: r ij =r(z i -z j ) Among them, z i -z j It is expressed as the distance d between two points. ij r() represents d ij and r ij The fitting function; The weighting coefficients must satisfy the following constraints: Based on the second dataset, the atmospheric refractive index is obtained; Data for the atmospheric waveguide is obtained based on the atmospheric refractive index.

2. The atmospheric waveguide monitoring and diagnostic method according to claim 1, characterized in that, The atmospheric refractive index is also corrected: The formula for correcting the atmospheric refractive index is selected from any of the following formulas: N r =p1N 3 +p2N 2 +p3N+p4 (21) N r =a0+a2cos(wN)+b2sin(wN) (23) Where N represents the atmospheric refractive index, N r This represents the corrected atmospheric refractive index, where p1, p2, p3, p4, a1, b1, c1, a0, a2, b2, and w are coefficients.

3. The atmospheric waveguide monitoring and diagnostic method according to claim 2, characterized in that, Based on minimizing the root mean square error, the modified formulas are selected. Based on the selected correction formula, the corresponding coefficients are solved.

4. The atmospheric waveguide monitoring and diagnostic method according to claim 2, characterized in that, In the northern sea area with a pressure of 700-800 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 226.3, a2 = 2.456, b2 = 18.65, w = 0.05497. In the eastern sea area with a pressure of 700-800 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 220.5, a2 = 30.45, b2 = -18.34, w = 0.04256. In the southern sea area with a pressure of 700-800 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 217.4, a2 = 42.89, b2 = -31.77, w = 0.01879. In the northern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 271.5, a2 = 37.59, b2 = 33.99, w = 0.02007. In the eastern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 22 is the smallest. The coefficients of Formula 22 are: a1 = 342.8, b1 = 413.7, c1 = 292.

2. In the southern sea area with a pressure of 800-900 hPa, the root mean square error corrected by Formula 21 is the smallest. The coefficients of Formula 21 are: p1 = -1.832e-0.5, p2 = 0.01296, p3 = -2.057, p4 = 241.

3. In the northern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by formula 2.3 is the smallest. The coefficients of formula 2.3 are: a0 = 286.8, a2 = -13.14, b2 = 18.12, w = 0.

046. In the eastern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 306.3, a2 = 27.24, b2 = 21.84, w = 0.

037. In the southern sea area with a pressure of 900-1000 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 336.4, a2 = 47.32, b2 = 40.96, w = 0.01612. In the northern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by Formula 23 is the smallest. The coefficients of Formula 23 are: a0 = 314.3, a2 = 0.2996, b2 = 17.35, w = 0.05992. In the eastern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by formulas 21 and 23 is the smallest. The coefficients of formula 21 are: p1 = -0.0001241, p2 = 0.1175, p3 = -36.05, p4 = 3896. The coefficients of formula 23 are: a0 = 315.5, a2 = -19.89, b2 = 29.53, w = 0.

028. In the southern sea area with a pressure of 1000-1100 hPa, the root mean square error corrected by Formula 22 is the smallest. The coefficients of Formula 22 are: a1 = 468.9, b1 = 562.2, c1 = 394.

8.

5. The atmospheric waveguide monitoring and diagnostic method according to claim 1, characterized in that, The preprocessing includes: Determine whether the meteorological data meets the following conditions: or Where X represents the parameters of the meteorological data. The mean is expressed as the average value, and STD is expressed as the standard deviation of the deviation. If the conditions are met, the meteorological data is considered abnormal. The abnormal data in the meteorological data is then removed to obtain the first dataset.

6. The atmospheric waveguide monitoring and diagnostic method according to claim 1, characterized in that, The data for the atmospheric waveguide includes atmospheric refractive index profiles; The atmospheric pressure ranges for the meteorological data are: 700hPa-800hPa, 800hPa-900hPa, 900hPa-1000hPa, and 1000hPa-1100hPa, where the atmospheric pressure range is used to describe the altitude range.

7. The atmospheric waveguide monitoring and diagnostic method according to claim 1, characterized in that, The preprocessing includes: the method for calculating the atmospheric refractive index includes: Where N represents the atmospheric refractive index, T represents the atmospheric temperature, P represents the atmospheric pressure, e represents the water vapor pressure, q represents the specific humidity, and D and E are constants.

8. A system for implementing the atmospheric waveguide monitoring and diagnostic method as described in any one of claims 1-7, characterized in that, It includes an acquisition module, a preprocessing module, a horizontal high-resolution module, and an atmospheric waveguide analysis module. The acquisition module is used to acquire meteorological data from meteorological satellites; The preprocessing module is used to preprocess the meteorological data to obtain a first dataset; The horizontal high-resolution module is used to process the first dataset based on the horizontal high-resolution construction algorithm to obtain a high-resolution second dataset. The atmospheric waveguide analysis module is used to obtain the atmospheric refractive index based on the second dataset; and to obtain atmospheric waveguide data based on the atmospheric refractive index.

Images