Simulation and forecasting capability evaluation method for occultation constellation observation system

Abstract

The invention provides an occultation constellation observation system simulation and forecast capability evaluation method, which comprises the following steps of: simulating and calculating an atmospheric refractive index by using a GNSS (Global Navigation Satellite System) radio occultation refractive index observation operator; simulating and calculating a one-dimensional bending angle by using a GNSS radio occultation one-dimensional bending angle observation operator; simulating and calculating a two-dimensional bending angle by using a GNSS radio occultation two-dimensional bending angle observation operator; correcting the bending angle of the L2 frequency on the basis of the difference between the bending angles of the L1 and L2 frequencies; and completing quality control of occultation observation data based on the atmospheric refractive index, the one-dimensional bending angle, the two-dimensional bending angle, the corrected bending angle of the L2 frequency and the actually collected occultation observation data, wherein the atmospheric refractive index, the one-dimensional bending angle and the two-dimensional bending angle are calculated through simulation. By establishing the error evaluation model for the autonomous occultation constellation observation data, the assimilation application effect of the autonomous occultation constellation observation data in the numerical weather forecasting system can be optimized, and direct technology and experience can be provided for business operation of future observation data in the numerical weather forecasting system.

Description

Technical Field

[0001] The present invention relates to the technical field of observation data error evaluation, and particularly to a method for evaluating the simulation and prediction capabilities of a occultation constellation observation system. Background Technique

[0002] Meteorological satellites emerged in 1960. Spaceborne microwave imager data emerged in 1992. Advanced Microwave Sounding Unit / Microwave Humidity Sounder (AMSU / MHS) data emerged in 1998. High-Resolution Infrared Radiation Sounder (HIRS) data was improved in 1998. High-spectral infrared sounding data emerged in 2002. Occultation data emerged in 2006, etc.

[0003] However, at that time, meteorological satellites had not yet realized their true value until they encountered numerical weather prediction (NWP). Since the physical equations in numerical weather prediction are partial differential equations, an initial condition is required for numerical solution. The better the initial condition, the higher the prediction accuracy. The process of determining the optimal initial condition is to perform an optimal combination and information fusion of the background fields (prior information) of the atmospheric (oceanic) states at a series of past moments and the atmospheric (oceanic) observations (new information) at a series of different moments using a certain algorithm to calculate the weight ratio, resulting in a new atmospheric (oceanic) state analysis field (posterior information). This process is called data assimilation (or data assimilation, and data assimilation is further advanced than data assimilation).

[0004] The operational implementation of the four-dimensional variational (4D-Var) data assimilation method is a milestone in global operational numerical weather prediction. The four-dimensional variational data assimilation method was operational at the European Centre for Medium-Range Weather Forecasts (ECMWF) in 1997. After more than 20 years of development, innovative scientific research results have continuously improved its main components. This includes combining the forecast model with an efficient radiative transfer model for more fully utilizing satellite radiance data, enabling a large amount of satellite remote sensing data to be fully applied in numerical weather prediction, and the accuracy of numerical prediction has also advanced rapidly.

[0005] There are generally two methods to evaluate the contribution of different satellite observation data to numerical weather prediction. The first one is the Observation System Experiments (OSEs), which is a traditional method to assess the impact of a specific observation network on numerical weather prediction. The Observation System Experiments consist of two experiments covering the same time period. The first experiment (control experiment) assimilates all the observations used in the operational process, and the second experiment (comparison experiment) systematically adds (or removes) selected observation datasets from the assimilation process (if it is a new observation addition experiment, a relatively poor data assimilation system is required; for removing observations, it is in a complete assimilation system, all to make the experimental effect more obvious) to evaluate the improvement (or decay) degree of the numerical prediction quality when this type of observation is allowed (or rejected).

[0006] The first application of OSEs lies in the evaluation of the effect of assimilating new observation data. When new observation data enters the data assimilation system, first, it is necessary to evaluate the compatibility degree of the new observation data with all previous types of observation data and the analysis field. The second step is to evaluate the impact of the new observation data on the prediction. Since evaluating the impact of a new type of observation data cannot rely solely on case studies or short-term experiments, OSEs require a long-term and stable operation. Generally, monthly average statistical test experiments in different seasons are adopted, and the computational workload is relatively large. The second application of OSEs is to study the impact of the most basic observations, mainly by running the most basic observation system to study the basic operation mechanism of the four-dimensional variational assimilation method. For example, studying the impact of GNSS occultation observations on the variational bias correction of radiance data in four-dimensional variational assimilation enables us to understand the main mechanism that the wind increment from geostationary satellite radiance observations is very sensitive to humidity, because after adding GNSS occultation observations in the comparison experiment, the bias value and correction time of the radiance data (AMSU-A channel 9) are greatly reduced.

[0007] Another one is the adjoint-based evaluation method, called the Forecast Sensitivity to Observation Impact (FSOI) method. The FSOI method uses the basic operator in the four-dimensional variational data assimilation system to quantify the sensitivity of a certain type of observation to the analysis field or prediction error. The most attractive part of FSOI is that it evaluates the impact of a certain type of observation when all other observations are in the data assimilation system, while OSEs need to remove this type of observation. When evaluating a certain type of observation, FSOI does not need to run experiments as long as OSEs to obtain a stable result. Therefore, FSOI would be a very good means to monitor the contribution of a certain type of observation. In addition, the modification of the observation system in OSEs will continuously accumulate the impact on the analysis and observation, but the adjoint system adopted by FSOI runs independently within the system.

[0008] In fact, OSEs and FSOI will show similar results. However, due to the differences introduced above, they are not the same type of method. Therefore, we can consider them as complementary evaluation methods, and it is very necessary to use them in combination.

[0009] Since the observational data of the already launched meteorological satellites can be quantitatively evaluated through the above two methods, we can quantitatively evaluate the contribution of the payload data of the yet-to-be-launched meteorological satellites to the numerical weather prediction system. This is the Observing System Simulation Experiment (OSSEs).

[0010] The Observing System Simulation Experiment has the word "simulation" more than the Observing System Experiment. This means that in the Observing System Simulation Experiment, simulated observations are used instead of real observational data as used in the Observing System Experiment. Since the satellite payload has not been launched yet, OSSEs is an extension of OSEs. The Observing System Simulation Experiment has three functions. First, it quantitatively evaluates whether a new observing system will have a positive effect on numerical weather prediction in the future and how large the positive effect will be. Second, it makes design and optimization decisions for a new type of observing system. Third, it studies the performance of the data assimilation system when the true value is known.

[0011] Since OSSEs is to evaluate the impact of simulated observations, a "true value" atmospheric state field is needed to simulate observations and verify the assimilation effect. However, we do not know the true value in nature, so an idealized atmospheric state field is needed to represent the "true value", which is called the Nature Run in OSSEs; Although OSSEs is aimed at simulated observations, the various subsystems and components used in its numerical weather prediction must be the systems and components actually used in the operation; During the process of observing simulation, the orbit, error, etc. must be carried out according to real parameters. Due to the above characteristics, the OSSEs system must be synchronized with the actual numerical weather prediction operation system. Summary of the Invention

[0012] In order to overcome the deficiencies of the prior art, the purpose of the present invention is to provide a method for evaluating the simulation and prediction capabilities of a occultation constellation observing system.

[0013] To achieve the above purpose, the present invention provides the following solution:

[0014] A method for evaluating the simulation and prediction capabilities of a occultation constellation observing system, comprising:

[0015] Using the GNSS radio occultation refractive index observation operator to simulate and calculate the atmospheric refractive index;

[0016] Using the GNSS radio occultation one-dimensional bending angle observation operator to simulate and calculate the one-dimensional bending angle;

[0017] Simulate and calculate the two-dimensional bending angle using the GNSS radio occultation two-dimensional bending angle observation operator;

[0018] Calculate the difference in bending angles between L1 and L2 frequencies using the geometric optics method;

[0019] Correct the bending angle of the L2 frequency based on the difference in bending angles between L1 and L2 frequencies;

[0020] Based on the simulated atmospheric refractive index, one-dimensional bending angle, two-dimensional bending angle, corrected L2 frequency bending angle, and actual collected occultation observation data, complete the quality control of the occultation observation data.

[0021] Preferably, the simulation and calculation of the atmospheric refractive index using the GNSS radio occultation refractive index observation operator includes:

[0022] Adopt the formula:

[0023] n = 10 -6 N LOC +1

[0024] Simulate and calculate the atmospheric refractive index; where n is the atmospheric refractive index and N LOC is the local refractive index.

[0025] Preferably, the simulation and calculation of the one-dimensional bending angle using the GNSS radio occultation one-dimensional bending angle observation operator includes:

[0026] Adopt the formula:

[0027]

[0028] Simulate and calculate the one-dimensional bending angle; where α represents the one-dimensional bending angle, a represents the influencing factor, x = nr, n represents the atmospheric refractive index, and r represents the radius of the earth.

[0029] Preferably, the simulation and calculation of the two-dimensional bending angle using the GNSS radio occultation two-dimensional bending angle observation operator includes:

[0030] Adopt the formula:

[0031]

[0032] Simulate and calculate the two-dimensional bending angle; where s is the distance along the ray path, n is the atmospheric refractive index, φ is the angle between the local radius vector and the tangent of the ray path, r represents the radius of the earth, and θ represents the two-dimensional bending angle.

[0033] Preferably, the calculation formula for the difference in bending angles between L1 and L2 frequencies is:

[0034]

[0035] Among them, α 2 (a)-α 1 (a) represents the difference in bending angle between L1 and L2 frequencies, a represents the influence factor, r 0 is the height of the ion density extreme, k 4 =40.3m 3 s -2 , TEC = ∫n e dr,n e is the ion density, r is the radius of the Earth, and f 1 and f 2 Indicates the carrier frequency of the L1 and L2 frequencies.

[0036] Preferably, based on the simulated atmospheric refractive index, one-dimensional bending angle, two-dimensional bending angle and bending angle of the corrected L2 frequency and the actually collected occultation observation data, the quality control of the occultation observation data is completed, including:

[0037] Perform range checks on the simulated atmospheric refractive index, one-dimensional bending angle, two-dimensional bending angle, bending angle of the corrected L2 frequency, and the actual collected occultation observation data, and remove data that are not within the set range;

[0038] Calculate the double-weighted mean and double-weighted standard deviation of each data point in the actual collected occultation observation data;

[0039] The consistency of occultation observation data was checked based on double-weighted mean and double-weighted standard deviation;

[0040] Remove data corresponding to consistency check values ​​that are not within the preset range;

[0041] The occultation observation data that have completed the consistency check are subjected to compatibility check and symmetry check, and the data that fail the check are removed to complete the quality control of the occultation observation data.

[0042] Preferably, the calculation formulas for the double-weighted mean and double-weighted standard deviation are:

[0043]

[0044] Where M is the median of the occultation observation data, w i is the weight function, MA D stands for median absolute deviation, i.e. |X i -Median of M|,X i represents the i-th data sample in the occultation observation data, represents the bi-weighted mean, and BSTD(X) represents the bi-weighted standard deviation.

[0045] The present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor. The transceiver, the memory, and the processor are connected through the bus. It is characterized in that when the computer program is executed by the processor, the steps in the above-mentioned method for evaluating the simulation and prediction capabilities of an occultation constellation observation system are implemented.

[0046] The present invention also provides a computer-readable storage medium, on which a computer program is stored. It is characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for evaluating the simulation and prediction capabilities of an occultation constellation observation system are implemented.

[0047] The beneficial effects of the method for evaluating the simulation and prediction capabilities of an occultation constellation observation system provided by the present invention are as follows: Compared with the prior art, through the observation system simulation test and the systematic optimization application test with infrared hyperspectral observation data, the present invention can quantitatively evaluate the potential impact of the observation and detection elements of unknown observation and detection satellites on the weather and climate predicted by the existing operational system before launch and deployment, and adjust the deployment and operation of the observation and detection satellites to maximize the impact of the observation and detection elements, which can provide direct technology and experience for the future operational use of observation and detection data in the numerical weather prediction system. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0049] Figure 1 is a flowchart of the method for evaluating the simulation and prediction capabilities of an occultation constellation observation system provided by the embodiment;

[0050] Figure 2 is a schematic diagram of the principle of a GPS occultation event provided by the embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0052] References to "embodiments" in this specification mean that specific features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of the present application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0053] The terms "first", "second", "third", "fourth", etc. in the specification, claims, and drawings of the present application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a series of steps, processes, methods, etc. included do not limit to the listed steps, but optionally further include steps not listed, or optionally further include other step elements inherent to these processes, methods, products, or devices.

[0054] To make the above objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0055] Please refer to Figure 1 , a method for evaluating the simulation and prediction capabilities of a satellite occultation constellation observation system, comprising:

[0056] Step 1: Use the GNSS radio occultation refractive index observation operator to simulate and calculate the atmospheric refractive index;

[0057] In Step 1, the formula is adopted:

[0058] n = 10 -6 N LOC +1

[0059]

[0060] Simulate and calculate the atmospheric refractive index; where the local refractive index N LOC is a function of the ionospheric density function N e and the atmospheric temperature T, dry air pressure P d and water vapor pressure P w of water.

[0061] Step 2: Use the GNSS radio occultation one-dimensional bending angle observation operator to simulate and calculate the one-dimensional bending angle;

[0062] In Step 2, the formula is adopted:

[0063]

[0064] Simulate and calculate the one-dimensional bending angle; where x = nr.

[0065] Step 3: Use the GNSS radio occultation two-dimensional bending angle observation operator to simulate and calculate the two-dimensional bending angle;

[0066] In Step 3, the formula is adopted:

[0067]

[0068] Simulate and calculate the two-dimensional bending angle.

[0069] Step 4: Use the geometric optics method to calculate the difference in bending angles between L1 and L2 frequencies;

[0070] Among them, the calculation formula for the difference in bending angles between L1 and L2 frequencies is:

[0071]

[0072] Among them, α 2 (a) - α 1 (a) represents the difference in bending angles between L1 and L2 frequencies.

[0073] Step 5: Correct the bending angle of L2 frequency based on the difference in bending angles between L1 and L2 frequencies;

[0074] Step 6: Based on the simulated atmospheric refractive index, one-dimensional bending angle, two-dimensional bending angle, and the corrected bending angle of L2 frequency and the actually collected occultation observation data, complete the quality control of the occultation observation data.

[0075] Among them, Step 6 includes:

[0076] Check the data of each vertical profile. If α < 0 or N < 0, then eliminate this data;

[0077] Calculate the biweighted mean and biweighted standard deviation of each data point in the occultation observation data;

[0078] Among them, the calculation formulas for the biweighted mean and biweighted standard deviation are:

[0079]

[0080] Among them, M is the median, w i is the weight function, and the median absolute deviation is represented by MAD.

[0081] Perform a consistency test on the data based on the biweighted mean and biweighted standard deviation;

[0082] Remove the data corresponding to the consistency test values outside the preset range;

[0083] Perform compatibility tests and symmetry tests on occultation observation data, and remove data outside the set range to complete the quality control of occultation observation data.

[0084] The following further describes the method for evaluating errors in occultation constellation observation data of the present invention in combination with specific embodiments:

[0085] The present invention mainly focuses on in - research and future occultation constellation satellite systems, and conducts research on simulation technologies for autonomous occultation constellation observation systems, natural operation data of meteorological numerical forecasts, error statistics and evaluation technologies for autonomous occultation constellation observation data, and data assimilation and numerical forecasting technologies for autonomous occultation observations. This is to support a rigorous and quantitative assessment of the potential impact of new satellite observation and detection systems and the optimization of observation and detection strategies for replacing existing systems. At the same time, it also prepares for the application of data assimilation of new types of data in numerical forecasting systems.

[0086] 1) High - resolution global natural operation atmospheric state field data.

[0087] The vertical resolution of high - resolution global natural operation atmospheric state field data should be sufficient to simulate real boundary layer and tropopause processes. Complete physical parameterizations of cloud, hydrometeor, and radiation interactions in the atmosphere are included in the current atmospheric model. To simulate observations, it is very important to have realistic natural operation cloud parameters, the true occurrence types in space and time of natural occurrences, and the accuracy of simulating infrared and visible radiation and other data types affected by clouds for obtaining true coverage. In addition, at higher resolutions, in the global high - resolution natural operation system, surface momentum, energy, and mass exchanges will be coupled with ocean and land surface models that include realistic seasonal variations in sea ice, snow, and vegetation cover. Realistic surface conditions are necessary for simulating infrared and microwave radiation with the correct coverage and accuracy affected by the surface. Similarly, changes in clouds, hydrometeors, and the land surface on the order of several meters to dozens of meters will have an important impact on real observation systems.

[0088] 2) High - precision simulation technology for GNSS radio occultation observations

[0089] 1) GNSS radio occultation refractive index observation operator

[0090] The original data of GPS radio occultation technology are the optical path lengths L1 and L2. They are the line integrals of the refractive index n along the ray path:

[0091]

[0092] Here, LEO and GPS refer to the paths from the satellite to the receiver, n(f 1 ) and n(f 2)Refers to the atmospheric refractive index at each position along the ray path s. The atmospheric refractive index n is determined by the local refractive index N of the atmosphere LOC as follows:

[0093] n = 10 -6 N LOC + 1 (2)

[0094] The local refractive index N LOC is a function of the ionospheric density function N e and the atmospheric temperature T, the dry air pressure P d and the water vapor pressure P w as follows:

[0095]

[0096] where k 1 , k 2 and k 3 are all coefficients that adjust the refraction effects of different parts of the atmosphere and the ionosphere. k 1 is the influence coefficient of dry air on the refractive index, k 2 is the influence coefficient of water vapor on the refractive index, k 3 is the influence coefficient of the ionosphere on the refractive index, Z d is the refraction factor of the dry air part, Z w is the refraction factor of the water vapor part.

[0097] 2) GNSS radio occultation one-dimensional bending angle observation operator

[0098] The one-dimensional bending angle observation operator is as follows:

[0099] α = H α F I H N (T, p, q) (4)

[0100] where H N (T, p, q) can complete the calculation of the non-spherical symmetric refractive index at the model grid points, and F I interpolates the refractive index at the model grid points to the near-surface tangent point position of the ray; H α represents the calculation process of obtaining the bending angle using the Abel transform based on the refractive index at the near-surface tangent point position.

[0101] Assuming that the state of the Earth's atmosphere is (approximately) spherically symmetrically distributed, there is a one-to-one functional relationship between the bending angle (as a function of the influence factor a) and the refractive index exponent (as a function of the radius r), where r t is the tangent height:

[0102]

[0103] Substituting \(x = nr\) into Equation (5), we can obtain:

[0104]

[0105] 3) GNSS Radio Occultation Two-Dimensional Bending Angle Observation Operator

[0106] The observed geometric structure is shown in Figure 2 The two-dimensional ray tracing method is based on solving the differential equation that defines the ray path. It uses the horizontal grid introduced above. Additionally, since the step size is adjusted as part of the model, the refractive index \(N\) and radius \(r\) are calculated in the full model layer and do not need to be inserted into the 250 m vertical grid.

[0107] The calculation of the bending angle is based on the numerical solution of the combined equation that defines the ray path in circular polar coordinates.

[0108]

[0109]

[0110] where \(s\) is the distance along the ray path, \(n\) is the refractive index exponent, and \(\varphi\) is the angle between the local radial vector and the tangent of the ray path. In fact, Equation (9) can be simplified by assuming At the tangent point where the ray bending is maximum, \(\varphi=\pi / 2\), so Physically, since the ray path is almost parallel to near the tangent point, this term is very small. The ray bending is caused by the refractive index gradient perpendicular to the path. Therefore, Equation (9) can be simplified to:

[0111]

[0112] The total bending angle is obtained by adding the bending angles along the path segments between the tangent point and the LEO and between the tangent point and the GPS. The ray tracing observation operator shown above, and the ray tracing method is called "theoretically the best observation operator". The GNSS radio occultation refractive index observation operator, the GNSS radio occultation one-dimensional bending angle observation operator, and the GNSS radio occultation two-dimensional bending angle observation operator provide important information for subsequent data assimilation, error correction, and quality control, and provide basic observation data and error models for the entire assimilation system. The results calculated through these operators can help optimize the prediction model, reduce errors, and improve the accuracy of numerical weather prediction.

[0113] 3) Evaluation of the Observation Error of Autonomous Occultation Data.

[0114] The error simulation process applied to GNSS radio occultation observations is based on a statistical comparison of models and observational data for two different seasons. Statistics are performed separately for the Northern Hemisphere (NH, latitudes higher than 20N), the Tropics (TR, latitudes between 20S and 20N), and the Southern Hemisphere (SH, latitudes lower than 20S), and also separately for NH winter and NH summer. The data are restricted to horizontal layers of 1 km. The mean difference between the observed and model first guess (a 6-hour forecast) of the refractive index and the standard deviation of these differences are presented as percentages of the refractive index. The NH winter case shows data above 40 km because the altitude of the refractive index profile during this period is increased to 60 km, while the NH summer case does not.

[0115] The main differences between the two seasons, winter and summer, are above 25 km and below 10 km in altitude. Between 10 km and 25 km, there is no obvious seasonal dependence. It is found that the standard deviation of the difference in the refractive index is only a function of latitude. In fact, a plot of the standard deviation of the difference between the observed and model parts as a function of latitude shows good symmetry around the equator. Thus, a cosine function can be used to fit the standard deviation function. GPS RO observations have high accuracy in the altitude range from 10 km to 25 km, and a cut-off value (or threshold) for removing an observation is set to 3 times the standard deviation value.

[0116] The details of the quality control structures for four different altitudes are summarized in Table 1. In the actual code, the change in the altitude range is smoothed, so the cut-off value is a continuous function of height.

[0117] Table 1 New assimilation system suitability standard deviation and cut-off value settings

[0118]

[0119] The most significant problem in introducing relevant errors lies in the insufficient understanding of the correlation matrix of combined observations and forward model errors. In the stratosphere, the correlation is mainly determined by the errors introduced during the bending angle processing. The bending angle is obtained by taking the time derivative of the phase delay observations. The finite difference approximation and the phase delay in this process can introduce a significant negative correlation. However, it should be noted that the differentiation is performed on 50Hz phase data, with a vertical sampling data every about 10 meters. The typical vertical stratification of the bending angle assimilated in the system is about 200m. If vertically sparsified, the distance may be even greater. In principle, if the phase delay is smoothed within a vertical range less than 200m, the negative bending angle error correlation can be reduced, but this may increase the noise level. In the troposphere, this correlation may be determined by the relevant forward model errors. One of the largest errors in the one-dimensional bending angle operator is using the influencing factors provided by the observations to obtain the tangent point height. In regions with strong horizontal gradients, this approximation will introduce positive correlation errors.

[0120] 4) Quantitative analysis technology for multi-observation collaborative optimization of the autonomous occultation constellation

[0121] In the application of data assimilation systems, multi-observation collaborative optimization means that, on the premise of ensuring that each type of observation data makes a positive contribution to the numerical prediction system after separate assimilation, multiple types of observation data are assimilated simultaneously, and the contribution to the numerical prediction system can produce an effect of 1+1≥2. This requires systematic analysis and parameter optimization of multiple observations simultaneously. Especially, the assimilation process of infrared hyperspectral data is relatively complex, and its key technologies such as cloud detection determine the assimilation effect of infrared hyperspectral observation data. At the same time, the assimilation effect of the autonomous occultation constellation observation system in the numerical prediction system also needs further adjustment and optimization. Therefore, breaking through the multi-observation collaborative optimization and quantitative analysis technology of the autonomous occultation constellation is a very crucial step. Due to the unbiased characteristics of occultation data, the autonomous occultation constellation data can be used to provide a bias correction standard for infrared hyperspectral data, which is usually called "Anchor". At the same time, the autonomous occultation constellation data can also make up for the data missing problem of infrared hyperspectral data in cloud and rain areas, and can also make up for the missing part of infrared hyperspectral data in the vertical spatial distribution.

[0122] 5) Error correction and quality control technology for autonomous occultation data

[0123] According to the error characteristics of GNOS from the occultation observation data of FY-3C and FY-3D, it is necessary to correct the tracking signal of L2. The existing processing algorithms for L2 signals will lead to large errors. Therefore, a new extrapolation algorithm for L2 signals is needed. One of the most important assumptions is that the total bending angle can be expressed as the linear sum of the neutral atmosphere bending angle and the frequency-dependent ionospheric bending angle. Therefore, subtracting the L1 bending angle value from the L2 bending angle value at a common influencing factor removes the influence of the neutral atmosphere bending angle. The extrapolation method is based on a thin ionospheric shell model, assuming that the ionosphere is approximated by a Delta function at a specific height. For a vertically local refraction region, just above the tangent point, the ionospheric contribution to the bending angle α at frequency f can be expressed as:

[0124]

[0125] where x = nr is the product of the refractive index and the Earth's radius, and a is the influencing factor, and at the same time n e is the ion density. Usually, the ion density can be expressed as the vertical integral of the number of ions, TEC, i.e., TEC = ∫n e dr. Assuming a very thin ionospheric shell, then the above equation can be written as,

[0126]

[0127] where r 0 is the height of the ion density extremum, which is assumed to be 300 km above the ground.

[0128] Then the difference in bending angles between L1 and L2 frequencies can be expressed as:

[0129]

[0130] where, if we define then we can get,

[0131]

[0132] Here, at 20 km vertically upward from the lowest available L2 bending angle value, we calculate x so using the least squares method based on the difference in L1 and L2 bending angles observed by the geometric optics method. The maximum height of the vertical interval is limited to 70 km.

[0133] Outliers to be identified in quality control can be distinguished by first studying the statistical characteristics and statistical distribution structures of a large number of observational data themselves and the differences between the observational data and model simulations, then finding reasonable theoretical bases for confirmation, and finally designing practical and feasible solutions. A commonly used method in quality control is to identify so-called outliers, that is, those data with relatively large distances from the mean value. Therefore, outliers are usually determined according to the distance of the data from the mean value of the data sample (Xi, i = 1, 2,...) based on the value of the standard deviation. However, outliers themselves can have a great impact on the mean value and standard deviation of the data, thus affecting the identification of outliers. To reduce this possibility, the identification of outliers can utilize the so-called biweight mean and biweight standard deviation BSTD:

[0134]

[0135] where M is the median and w i is the weight function. Using MAD to represent the median absolute deviation (i.e., the median of |X i - M), then the weight function w i is defined as

[0136]

[0137] The quality control of occultation observational data can be completed by the following four steps.

[0138] Step 1: Range check

[0139] Since the ionospheric error is opposite to that of the upper troposphere, a range check is first performed, that is, the data of each vertical profile are checked as follows. If α < 0 or N < 0, then the data are excluded.

[0140] Step 2: Data consistency test

[0141] Calculate the Z-score of each data:[[]]

[0142]

[0143] where X i represents the bending angle α i or the refractive index Ni, and the subscript "i" represents the i-th observational data. It is double weighted average, and BSTD(X) is double weighted mean square deviation. Suspected abnormal data (i.e., outlier data) are determined respectively according to the criteria that the Z-score is greater than 3, 4, and 5. After comparing the amount of suspected abnormal data and the corresponding observed values with the model simulation values calculated from the large-scale analysis field temperature and water vapor, it is found that there are more outlier data in the refractive index than in the bending angle. Data with a large difference between the observed value and the model simulation value are basically removed in the second step, which is a very ideal result. Because the quality control formula (1.23) in this step is for the data itself and does not input any model simulation values.

[0144] Step 3: Compatibility test with the background field

[0145] Calculate the Z-score of the following variables:

[0146]

[0147] where, α obs and α model represent the bending angles in the observation and the model respectively, N obs and N model represent the atmospheric refractive indices in the observation and the model respectively. Data with a Z-score greater than 4 are regarded as suspected abnormal data.

[0148] Step 4: Symmetry test

[0149] There may be negative systematic errors in the occultation observation data at the bottom layer of the troposphere. Therefore, a symmetry test is conducted on the data errors passing through the above three steps. The symmetry test is only carried out on the data below 4 km. Denote the sets

[0150] XL α ={α:α obs -α model <0}, XG α ={α:α obs -α model >0} (20)

[0151] According to the sets XL α and XG α , construct the following two data sets:

[0152] YL α ={-α:α∈XL α}, YG α ={-α:α∈XG α} (21)

[0153] Then the data set XL α ∪YL α and the data set XG α ∪YG αBoth are symmetric. Calculate the biweight standard deviations BSTD of these two data sets respectively L and BSTD G , and take

[0154] BSTD = min{BSTD L , BSTD G} (22)

[0155] is the standard deviation of the original data set. Finally, calculate the Z-score for each piece of data. Data with a Z-score greater than 3 is defined as potentially incorrect data.

[0156] According to the embodiments provided by the present invention, the present invention discloses the following technical effects:

[0157] (1) Optimize the application efficiency and level of autonomous occultation constellation observation data and geostationary satellite infrared observation data in the numerical weather prediction system. By establishing an error assessment model for autonomous occultation constellation observation data, the assimilation application effect of autonomous occultation constellation observation data in the numerical weather prediction system can be optimized;

[0158] (2) Quantitatively evaluate the potential impact of the observation and detection elements of future observation and detection satellites on the numerical prediction system. Through observation system simulation experiments and systematic optimization application experiments with infrared hyperspectral observation data, the potential impact of the observation and detection elements of unknown observation and detection satellites on the weather and climate predicted by the existing operational system can be quantitatively evaluated before launch and deployment, and the deployment and operation of the observation and detection satellites can be adjusted to maximize the impact of the observation and detection elements, which can provide direct technology and experience for the future operation of observation and detection data in the numerical weather prediction system.

[0159] (3) Quantitatively evaluate the cost and utility of the observation and detection satellite system, and provide a reliable basis for optimal design and final decision-making. The construction and deployment of observation and detection satellites will consume a large amount of human, financial and other resources. The effective application of its products in the numerical weather prediction system is the key link to exert its application efficiency. The observation system simulation experiment system can quantitatively evaluate the cost and expense of the observation and detection satellite system, and can provide an optimal design for what elements should be observed at what location and time to obtain the best effect, and provide a reliable basis for design and decision-making. Most importantly, the observation system prediction ability assessment experiment can be carried out before the actual physical construction of the observation network.

[0160] The present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor. The transceiver, the memory, and the processor are connected through the bus. It is characterized in that when the computer program is executed by the processor, the steps in the above method for evaluating the simulation and prediction capabilities of an occultation constellation observation system are implemented. Compared with the prior art, the beneficial effects of the electronic device provided by the present invention are the same as those of the method for evaluating the simulation and prediction capabilities of an occultation constellation observation system described in the above technical solution, and will not be elaborated here.

[0161] The present invention also provides a computer-readable storage medium, on which a computer program is stored. It is characterized in that when the computer program is executed by a processor, the steps in the above method for evaluating the simulation and prediction capabilities of an occultation constellation observation system are implemented. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present invention are the same as those of the method for evaluating the simulation and prediction capabilities of an occultation constellation observation system described in the above technical solution, and will not be elaborated here.

[0162] In the present specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the methods disclosed in the embodiments, since they correspond to the devices disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description of the device part.

[0163] In this article, specific examples are used to elaborate on the principles and implementation manners of the present invention. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present invention. At the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for simulating and evaluating the prediction capability of an occultation constellation observation system, characterized in that: include: The atmospheric refractive index is simulated and calculated using the GNSS radio occultation refractive index observation operator; The one-dimensional bending angle is calculated by using the GNSS radio occultation one-dimensional bending angle observation operator to simulate; The 2D bending angle is simulated and calculated using the GNSS radio occultation 2D bending angle observation operator; The difference in bending angles between L1 and L2 frequencies was calculated using the geometrical optics method; Correcting the bending angle of the L2 frequency based on the difference between the bending angles of the L1 and L2 frequencies; The quality control of the occultation observation data is completed based on the simulated atmospheric refractive index, one-dimensional bending angle, two-dimensional bending angle and bending angle of the corrected L2 frequency and the actually collected occultation observation data.

2. The method for simulating and predicting the occultation constellation observation system according to claim 1, characterized in that: The method of using the GNSS radio occultation refractive index observation operator to simulate and calculate the atmospheric refractive index includes: Using the formula: n=10 -6 N LOC +1 Simulate and calculate the atmospheric refractive index; where n is the atmospheric refractive index, N LOC is the local refractive index.

3. The method for simulating and predicting the occultation constellation observation system according to claim 2, characterized in that: The method of using the GNSS radio occultation one-dimensional bending angle observation operator to simulate and calculate the one-dimensional bending angle comprises: Using the formula: The one-dimensional bending angle is simulated and calculated; wherein α represents the one-dimensional bending angle, a represents the influence factor, x=nr, n represents the atmospheric refractive index, and r represents the radius of the earth.

4. The method for simulating and predicting the occultation constellation observation system according to claim 3, characterized in that: The GNSS radio occultation two-dimensional bending angle observation operator simulates and calculates the two-dimensional bending angle, including: Using the formula: The simulation calculates the 2D bending angle; where s is the distance along the ray path, n is the atmospheric refractive index, φ is the angle between the local vector and the tangent of the ray path, r represents the radius of the Earth, and θ represents the 2D bending angle.

5. The method for simulating and predicting the occultation constellation observation system according to claim 4, characterized in that: The difference in bending angles between L1 and L2 frequencies is calculated as: Where, α2(a)-α1(a) represents the difference in bending angles between L1 and L2 frequencies, a represents the influence factor, r0 is the height of the ion density extreme, and k4 = 40.3 m 3 s -2 , TEC = ∫n e dr,n e is the ion density, r represents the radius of the Earth, and f1 and f2 represent the carrier frequencies of the L1 and L2 frequencies.

6. A system for constructing a machine learning satellite observation operator according to claim 5, characterized in that: Based on the simulated atmospheric refractive index, one-dimensional bending angle, two-dimensional bending angle and the bending angle of the corrected L2 frequency and the actual collected occultation observation data, the quality control of the occultation observation data is completed, including: Perform range checks on the simulated atmospheric refractive index, one-dimensional bending angle, two-dimensional bending angle, bending angle of the corrected L2 frequency, and the actual collected occultation observation data, and remove data that are not within the set range; Calculate the double-weighted mean and double-weighted standard deviation of each data point in the actual collected occultation observation data; The consistency of occultation observation data was checked based on double-weighted mean and double-weighted standard deviation; Remove data corresponding to consistency check values ​​that are not within the preset range; The occultation observation data that have completed the consistency check are subjected to compatibility check and symmetry check, and the data that fail the check are removed to complete the quality control of the occultation observation data.

7. A system for constructing a machine learning satellite observation operator according to claim 6, characterized in that: The calculation formulas for the double-weighted mean and double-weighted standard deviation are: Where M is the median of the occultation observation data, w i is the weight function, MA D stands for median absolute deviation, i.e. |X i -Median of M|,X i represents the i-th data sample in the occultation observation data, represents the double-weighted mean, and BSTD(X) represents the double-weighted standard deviation.

8. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps in the method for simulating and evaluating the prediction capability of an occultation constellation observation system as described in any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method for simulating and evaluating the prediction capability of an occultation constellation observation system as described in any one of claims 1 to 7 are implemented.

Images