Near space multi-source observation data assimilation method and electronic device

Abstract

The application provides a near space multi-source observation data assimilation method and an electronic device, and relates to the technical field of assimilation processing. The method comprises the following steps: preprocessing a plurality of observation data of the near space to obtain a plurality of initial data; performing conversion processing on each initial data to obtain a plurality of observation sequences in a unified format; and performing assimilation processing on the plurality of observation sequences in a constructed initial background field based on an assimilation algorithm to obtain a target analysis field. The application is aimed at the observation data collected by a plurality of observation devices in the near space, the observation data having different spatial and temporal resolutions and different data formats. The observation data is preprocessed to improve the effectiveness of the data. The plurality of data is uniformly converted to obtain observation sequences in a unified format for subsequent assimilation processing. The plurality of observation sequences are assimilated in the initial background field by using the assimilation algorithm to obtain a target analysis field with high precision and reliability, thereby meeting the use requirements of various scenes.

Description

Technical Field

[0001] This application relates to the field of assimilation processing technology, and more specifically, to a method and electronic device for assimilating near-space multi-source observation data. Background Technology

[0002] Currently, weather forecasting or atmospheric reanalysis is usually performed based on analytical fields assimilated from near-space atmospheric data. Existing technologies typically use atmospheric dynamic equations and physical parameterization schemes, combined with observational data, to adjust model state variables to obtain the analytical field. Alternatively, statistical relationships can be established using historical observational data, and data fusion can be performed through regression, interpolation, or machine learning methods to obtain the analytical field.

[0003] However, current near-space atmospheric data assimilation methods primarily rely on single data sources or limited types of observational data, leading to the following problems: Dispersed data sources: Near-space observational data comes from various instruments, with inconsistent data formats and spatiotemporal resolutions, making unified processing difficult. Insufficient quality control: Existing methods lack a systematic approach to identifying and handling outliers, affecting the reliability of assimilation results. Poor compatibility of assimilation systems: Most assimilation systems struggle to directly support the access and fusion of multi-source heterogeneous data. Limited initial field accuracy: Insufficient data fusion results in low accuracy of model initial fields, impacting the quality of forecasts and reanalysis products. Therefore, current near-space atmospheric data assimilation methods exhibit poor assimilation effects, failing to yield high-precision and reliable analytical fields, and thus failing to meet practical application requirements. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a near-space multi-source observation data assimilation method and electronic device to improve the problem of poor analysis field quality in the prior art.

[0005] To address the aforementioned problems, in a first aspect, embodiments of this application provide a near-space multi-source observation data assimilation method, the method comprising: Multiple observation data from near space are preprocessed separately to obtain multiple initial data; Each type of initial data is transformed to obtain multiple observation sequences in a uniform format; Based on the assimilation algorithm, the target analysis field is obtained by assimilating multiple observation sequences in the constructed initial background field.

[0006] In the above implementation process, considering the diverse spatiotemporal resolutions and data formats of observation data collected by various observation devices in near space, and to improve data effectiveness, preprocessing can be performed on each type of observation data separately to obtain various types of initial data. Since the initial data have different formats, a unified format conversion process can be performed on all data to obtain a unified format observation sequence for subsequent assimilation processing. Assimilation algorithms can be used to assimilate multiple observation sequences in the initial background field, effectively improving the assimilation effect of near-space atmospheric data and obtaining a high-precision and reliable target analysis field, meeting the needs of various data analysis application scenarios.

[0007] Optionally, the transformation process performed on each type of initial data to obtain multiple observation sequences in a uniform format includes: Determine the mapping function between each type of initial data and the uniform target format; Based on the mapping function, the initial data is transformed to obtain the observation sequence in the target format; The observation sequence includes: observation point, observation value, three-dimensional location information, observation time, observation type identifier, and observation error variance.

[0008] In the above implementation process, due to the differences between various observation data, there are also differences among the various initial data obtained from preprocessing. To achieve format conversion, a mapping function can be determined to convert each type of initial data to a unified target format. The initial data is then converted using this mapping function to obtain multiple observation sequences in a unified target format. These observation sequences can include various types of observation parameters. A mapping function can be established from the mode state variables of the initial function to the observations in the target format, thereby achieving data format conversion through mapping, ensuring consistency between spatiotemporal and physical dimensions, and obtaining multiple corresponding observation sequences for subsequent processing.

[0009] Optionally, the assimilation algorithm-based process of assimilating multiple observation sequences in a constructed initial background field to obtain the target analysis field includes: The initial background field for constructing a full atmospheric community climate model; Based on the assimilation algorithm, the initial analysis field is obtained by assimilation processing of multiple observation sequences in the initial background field; wherein, the assimilation algorithm includes: ensemble Kalman filter algorithm or localized particle filter algorithm; The initial analysis field is optimized to obtain the target analysis field.

[0010] In the above implementation process, during assimilation, an initial background field of a full-atmosphere community climate model can be constructed to ensure better physical balance and spatiotemporal continuity of the assimilation process, thereby providing more reliable start-up conditions for upper-atmosphere numerical models. To overcome the biases and uncertainties caused by assimilation from a single data source, an assimilation algorithm can be used to assimilate multiple observation sequences within the initial background field to obtain a corresponding initial analysis field. This effectively reduces analysis errors and significantly improves the analysis accuracy of key atmospheric parameters such as temperature, wind field, and density in near-space. Furthermore, the initial analysis field can be further optimized based on actual conditions to obtain a target analysis field with higher accuracy and effectiveness. This innovative approach assimilates observational information from multiple sources to obtain a high-precision, high-effectiveness, and high-reliability target analysis field.

[0011] Optionally, the initial background field for constructing the full atmospheric community climate model includes: Determine the current status data; wherein, the status data includes: time information and location information; Based on the state data, extract the corresponding historical data from the full atmospheric community climate model; Based on the historical data, the initial background field is constructed.

[0012] In the above implementation process, the actual state data of the current near space can be determined. Based on the state data, historical data with high similarity to the current state can be extracted from the full-atmosphere community climate model, thereby constructing the initial background field of the full-atmosphere community climate model based on historical data. This allows for the construction of a more relevant initial background field based on historical conditions, providing reliable initiation conditions for assimilation processing.

[0013] Optionally, the assimilation process based on the assimilation algorithm, performing assimilation processing on the initial background field according to multiple observation sequences to obtain the initial analysis field, includes: The multiple observation sequences are input into the data assimilation research test platform; Based on the assimilation algorithm set in the data assimilation research and testing platform, multiple observation sequences are fused with the initial background field to obtain an analysis field set; The initial analysis field is obtained by processing the set of analysis fields based on the spatial localization mechanism and the covariance expansion mechanism.

[0014] In the above implementation process, multiple observation sequences are input into a data assimilation research test platform. Using various assimilation algorithms set within the platform, these sequences are fused with the initial background field to obtain a corresponding set of analytical fields. Considering the impact of localization and covariance inflation on the analytical field set, spatial localization and covariance inflation mechanisms are used to process the set, yielding the corresponding initial analytical field. The ability to fuse data and the background field using the configured assimilation algorithms effectively improves the data assimilation effect.

[0015] Optionally, optimizing the initial analysis field to obtain the target analysis field includes: Determine the influencing factors that affect the initial analysis field; wherein the influencing factors include at least one of the following: weight information determined based on the type and accuracy requirements of the observation data, assimilation influence radius adjusted based on atmospheric variability characteristics at different levels in the near space, time decay factor, and adaptive assimilation residual error model; Based on the aforementioned influencing factors, the initial analysis field is adjusted to obtain the target analysis field.

[0016] In the aforementioned implementation process, during optimization, various influencing factors affecting the initial analysis field can be identified from multiple perspectives, such as observation quality, spatial coverage, and temporal consistency. These factors are then used to dynamically adjust the analysis field, ultimately yielding the final target analysis field. This ability to identify and adjust the initial analysis field from multiple angles further improves the accuracy and reliability of the target analysis field.

[0017] Optionally, the method further includes: The atmospheric analysis field is obtained by averaging the target analysis field. Based on the atmospheric analysis field, atmospheric analysis and processing are performed to obtain weather forecast data for the observation area corresponding to the near space.

[0018] In the above implementation process, after obtaining the target analysis field, since the target analysis field is a collection of multiple analysis fields, it can be averaged to obtain a high-precision atmospheric analysis field. Based on this atmospheric analysis field, atmospheric analysis can be performed to obtain weather forecast data for the corresponding observation area in the adjacent space. The ability to perform weather forecasting based on the target analysis field effectively improves the accuracy of weather forecast data.

[0019] Optionally, the various observation data include at least one of the following: meteor radar data from different observation points, mesosphere-stratospheric-tropospheric radar data, lidar data, and satellite data; The preprocessing of various observation data in the near space yields various initial data, including: The meteor radar data is subjected to least squares inversion, effective event screening, error estimation, and spatiotemporal consistency verification to obtain the first initial data. And / or, perform sliding window smoothing, outlier removal, and beam consistency verification on the mesosphere-stratospheric-tropospheric radar data to obtain second initial data; And / or, perform data filtering and constraint processing on the lidar data to obtain third initial data; And / or, perform spatial matching and error processing on the satellite data to obtain fourth initial data.

[0020] In the above implementation process, for different types of observation data collected by different devices, differentiated preprocessing methods can be set according to the actual situation of each type of observation data to obtain various types of initial data, such as first initial data, second initial data, third initial data, and fourth initial data. This ability to perform differentiated preprocessing on different types of observation data effectively improves the effectiveness of various types of initial data.

[0021] Secondly, embodiments of this application also provide an electronic device, which includes a memory and a processor. The memory stores program instructions, and when the processor executes the program instructions, it performs the steps in any of the methods described in the first aspect above.

[0022] Thirdly, embodiments of this application also provide a computer program product, the computer program product including a computer program / instruction, which, when executed by a processor, implements the steps in the method described in any of the first aspects above.

[0023] In summary, the embodiments of this application provide a near-space multi-source observation data assimilation method and electronic device. This method preprocesses observation data with varying spatiotemporal resolutions and data formats acquired by multiple observation devices in near space to improve data effectiveness. It also performs a unified format conversion process on the various data to obtain a unified format observation sequence for subsequent assimilation. Furthermore, an assimilation algorithm is used to assimilate multiple observation sequences in the initial background field, resulting in a highly accurate and reliable target analysis field that meets the needs of various scenarios. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A block diagram illustrating an electronic device provided in an embodiment of this application; Figure 2 A flowchart illustrating a near-space multi-source observation data assimilation method provided in an embodiment of this application; Figure 3 A detailed flowchart of step S300 provided for an embodiment of this application; Figure 4 A detailed flowchart of step S400 provided for an embodiment of this application; Figure 5 A detailed flowchart of step S410 provided for an embodiment of this application; Figure 6 A detailed flowchart of step S420 provided for an embodiment of this application; Figure 7 A detailed flowchart of step S430 provided for an embodiment of this application; Figure 8 A flowchart illustrating another near-space multi-source observation data assimilation method provided in this application embodiment; Figure 9 A detailed flowchart of step S200 provided for an embodiment of this application; Figure 10 A schematic diagram of temperature assimilation increment provided in an embodiment of this application; Figure 11 This is a schematic diagram of zonal wind assimilation increment provided in an embodiment of this application.

[0026] Icons: 100 - Electronic device; 111 - Memory; 112 - Memory controller; 113 - Processor; 114 - Peripheral interface; 115 - Input / output unit; 116 - Display unit. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of the embodiments of this application.

[0028] Currently, near-space atmospheric data assimilation mainly relies on a single data source or a limited number of observational data, which presents the following problems: Dispersed data sources: Near-space observational data comes from various instruments, with inconsistent data formats and spatiotemporal resolutions, making unified processing difficult. Insufficient quality control: Existing methods lack a systematic approach to identifying and handling outliers, affecting the reliability of assimilation results. Poor compatibility of assimilation systems: Most assimilation systems cannot directly support the access and fusion of multi-source heterogeneous data. Limited initial field accuracy: Due to insufficient data fusion, the accuracy of model initial fields is low, affecting the quality of forecasts and reanalysis products. Therefore, current near-space atmospheric data assimilation methods have poor assimilation effects, failing to obtain high-precision and reliable analytical fields, and thus failing to meet practical application requirements.

[0029] To address the aforementioned issues, this application provides a near-space multi-source observation data assimilation method and electronic device. This method preprocesses observation data with varying spatiotemporal resolutions and formats acquired by multiple observation devices in near space to improve data effectiveness. It also performs a unified format conversion on the various data sets to obtain a unified format observation sequence for subsequent assimilation. An assimilation algorithm is then used to assimilate multiple observation sequences in the initial background field, resulting in a highly accurate and reliable target analysis field that meets the needs of various scenarios.

[0030] This application provides a near-space multi-source observation data assimilation method, which can be applied to electronic devices such as servers, personal computers (PCs), tablets, smartphones, personal digital assistants (PDAs), cloud devices, and other electronic devices with logical computing functions.

[0031] Optionally, please refer to Figure 1 , Figure 1 This is a block diagram illustrating an electronic device according to an embodiment of this application. The electronic device 100 may include a memory 111, a memory controller 112, a processor 113, a peripheral interface 114, an input / output unit 115, and a display unit 116. Those skilled in the art will understand that... Figure 1The structure shown is for illustrative purposes only and does not limit the structure of the electronic device 100. For example, the electronic device 100 may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0032] The aforementioned memory 111, memory controller 112, processor 113, peripheral interface 114, input / output unit 115, and display unit 116 are electrically connected directly or indirectly to each other to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines. The aforementioned processor 113 is used to execute executable modules stored in the memory.

[0033] The memory 111 can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory 111 stores programs. After receiving execution instructions, the processor 113 executes the programs. The methods executed by the electronic device 100 as defined in any embodiment of this application can be applied to the processor 113, or implemented by the processor 113.

[0034] The aforementioned processor 113 may be an integrated circuit chip with signal processing capabilities. The processor 113 may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it may also be a digital signal processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor.

[0035] The peripheral interface 114 described above couples various input / output devices to the processor 113 and the memory 111. In some embodiments, the peripheral interface 114, the processor 113, and the memory controller 112 can be implemented on a single chip. In other instances, they can be implemented on separate chips.

[0036] The input / output unit 115 described above is used to provide user input data. The input / output unit 115 can be, but is not limited to, a mouse and keyboard, etc.

[0037] The aforementioned display unit 116 provides an interactive interface (e.g., a user interface) between the electronic device 100 and the user, or displays image data for the user's reference. In this embodiment, the display unit can be a liquid crystal display (LCD) or a touch display. If it is a touch display, it can be a capacitive touchscreen or a resistive touchscreen that supports single-point and multi-point touch operations. Supporting single-point and multi-point touch operations means that the touch display can sense touch operations generated simultaneously from one or more locations on the touch display and pass the sensed touch operations to the processor for calculation and processing. In this embodiment, the display unit 116 can display various data such as the final target analysis field.

[0038] The electronic device in this embodiment can be used to execute various steps in the near-space multi-source observation data assimilation methods provided in the embodiments of this application. The implementation process of the near-space multi-source observation data assimilation methods is described in detail below through several embodiments.

[0039] Please see Figure 2 , Figure 2 This is a flowchart illustrating a near-space multi-source observation data assimilation method provided in an embodiment of this application. The method may include steps S200-S400.

[0040] Step S200: Preprocess the various observation data in the near space to obtain various initial data.

[0041] In particular, for observation data with different spatiotemporal resolutions and data formats collected by various observation devices in near space, in order to improve the effectiveness of the data and considering the differences between various observation data, the various observation data can be preprocessed separately to obtain various types of initial data.

[0042] Optionally, the various observation data can be data collected using various types of observation equipment, including different types of radar and satellites. Near space can include multiple observation points, and one or more observation devices can be deployed for data collection based on the actual conditions at each observation point.

[0043] Step S300: Transform each type of initial data to obtain multiple observation sequences in a uniform format.

[0044] Since various initial data have different formats, a unified format conversion process can be performed on the various data to obtain a unified format observation sequence for subsequent assimilation processing.

[0045] Optionally, the unified format is the target format, and the specific data format of the target format can be selected and set based on the actual situation of the assimilation system such as DART (Data Assimilation Research Testbed).

[0046] Optionally, given the diverse range of near-space observation devices, heterogeneous data formats, and varying spatiotemporal resolutions, this application can automatically identify and parse the data structure of observation data from different sources, implement targeted preprocessing (i.e., quality control processes, such as valid event screening, outlier removal, and signal-to-noise ratio threshold judgment), and uniformly convert the data into a unified target format (such as obs_seq) that can be identified by an assimilation system (such as DART), ensuring the consistency of data in time, space, and physical dimensions, and providing high-quality input for subsequent fusion.

[0047] Step S400: Based on the assimilation algorithm, the target analysis field is obtained by assimilation processing of multiple observation sequences in the constructed initial background field.

[0048] Among them, the assimilation algorithm can be used to assimilate multiple observation sequences in the initial background field, which effectively improves the assimilation effect of near-space atmospheric data and obtains a target analysis field with high accuracy and reliability.

[0049] It should be noted that the assimilation algorithm is set up in the data assimilation research test platform and is able to fuse the observation sequence into the initial background field.

[0050] exist Figure 2 In the illustrated embodiment, various observation data with different spatiotemporal resolutions and data formats collected by multiple observation devices in near space are preprocessed to improve the effectiveness of the data. The data are also converted to a unified format to obtain a unified observation sequence for subsequent assimilation. The assimilation algorithm is used to assimilate multiple observation sequences in the initial background field to obtain a target analysis field with high accuracy and reliability, which meets the needs of various scenarios.

[0051] Optionally, please refer to Figure 3 , Figure 3This is a detailed flowchart of step S300 provided in an embodiment of the present application. Step S300 may include steps S310-S320.

[0052] Step S310: Determine the mapping function between each type of initial data and the uniform target format.

[0053] Since there are differences between various observation data, there are also differences between the various initial data obtained from preprocessing. In order to achieve format conversion, a mapping function can be determined to convert each initial data to a unified target format.

[0054] Optionally, the target format can be the standard input format obs_seq of the DART system, and the mapping function can be a function that can characterize the data mapping relationship between the pattern state variables of the initial data and the observation parameters of the target format.

[0055] Step S320: Based on the mapping function, the initial data is transformed to obtain the observation sequence in the target format.

[0056] The initial data can be transformed using a mapping function to obtain multiple observation sequences in a unified target format. These observation sequences can include various types of observation parameters.

[0057] Optionally, the observation sequence may include various types of observation parameters such as observation points, observation values, three-dimensional location information, observation time, observation type identifier, and observation error variance.

[0058] For example, an observation sequence may include: observation points (such as observation point number, relative position, etc.), observation values ​​(such as temperature, wind speed, refractive index, etc.), three-dimensional location information (longitude, latitude, altitude / air pressure), observation time (UTC accurate to the second), observation type identifier (DART internal code, such as RADIOSONDE_U_WIND), and observation error variance.

[0059] Optionally, a linked list structure can be used to store multiple observation sequences to support efficient read / write and random access in the future.

[0060] exist Figure 3 In the illustrated embodiment, a mapping function can be established from the mode state variables of the initial function to the observations of the target format, so as to realize the data format conversion through mapping, ensure the consistency of spatiotemporal and physical dimensions, and obtain multiple observation sequences for subsequent processing.

[0061] Optionally, please refer to Figure 4 , Figure 4The following is a detailed flowchart of step S400 provided in an embodiment of this application. Step S400 may include steps S410-S430.

[0062] Step S410: Construct the initial background field for the full atmospheric community climate model.

[0063] In the process of assimilation, an initial background field for a full atmospheric community climate model can be constructed to enable the assimilation process to have better physical balance and spatiotemporal continuity, thereby providing more reliable start-up conditions for upper atmospheric numerical models and other models.

[0064] Optionally, historical data can be extracted from historical simulations or climate states of the WACCM model, i.e., the whole atmospheric community climate model, to construct the initial background field.

[0065] For example, the initial background field can be a meteorological background field, i.e. a dynamically constrained background field, which can gradually pull back the model wind, temperature, and surface pressure for reanalysis to improve the realism of mid-to-low-level weather.

[0066] Step S420: Based on the assimilation algorithm, the initial analysis field is obtained by assimilation processing of multiple observation sequences in the initial background field.

[0067] In order to overcome the bias and uncertainty caused by assimilation from a single data source, an assimilation algorithm can be used to assimilate multiple observation sequences in the initial background field to obtain the corresponding initial analysis field, which effectively reduces analysis error and significantly improves the analysis accuracy of key atmospheric parameters such as temperature, wind field, and density in near space.

[0068] Optionally, the assimilation algorithm may include: the Ensemble Adjustment Kalman Filter (EAKF) algorithm, or the Localized Particle Filter algorithm, and other algorithms that can achieve data fusion.

[0069] Optionally, the assimilation algorithm can update the state of the set members by minimizing the analysis error covariance, thereby obtaining the corresponding initial analysis field.

[0070] Step S430: Optimize the initial analysis field to obtain the target analysis field.

[0071] Furthermore, the initial analysis field can be further optimized to make further adjustments based on the actual situation, resulting in a target analysis field with higher accuracy and effectiveness.

[0072] Optionally, the optimization process may include: performing dynamic weight adjustment, dynamically calculating and assigning different assimilation weights to each observation data based on the type of observation data, instrument calibration error, spatial coverage density of the observation, and data freshness, etc.

[0073] exist Figure 4 The illustrated embodiment innovatively assimilates observation information from multiple sources to obtain a target analysis field with high precision, high effectiveness, and high reliability.

[0074] Optionally, please refer to Figure 5 , Figure 5 The following is a detailed flowchart of step S410 provided in an embodiment of this application. Step S410 may include steps S411-S413.

[0075] Step S411: Determine the current state data.

[0076] This can determine the actual state data of the current near space. The state data can include various data that can characterize the actual situation of the near space, such as time information and location information.

[0077] Step S412: Based on the state data, extract the corresponding historical data from the full atmospheric community climate model.

[0078] Among these, historical data with high similarity to the current state can be extracted from the overall atmospheric community climate model based on the state data.

[0079] Optionally, historical data may include various types of data such as historical simulation models of the WACCM model or historical climate data.

[0080] Step S413: Based on historical data, construct the initial background field.

[0081] Among them, the initial background field of the whole atmospheric community climate model can be constructed based on historical data.

[0082] exist Figure 5 In the illustrated embodiment, a more relevant initial background field can be constructed based on historical data, providing reliable initiation conditions for assimilation processing.

[0083] Optionally, please refer to Figure 6 , Figure 6 The following is a detailed flowchart of step S420 provided in an embodiment of this application. Step S420 may include steps S421-S423.

[0084] Step S421: Input multiple observation sequences into the data assimilation research test platform.

[0085] Step S422: Based on the assimilation algorithm set in the data assimilation research test platform, multiple observation sequences are fused with the initial background field to obtain the analysis field set.

[0086] Multiple observation sequences can be input into the data assimilation research test platform. Based on the various types of assimilation algorithms set in the data assimilation research test platform, the multiple observation sequences are fused with the initial background field to obtain multiple corresponding analysis fields as a set of analysis fields.

[0087] Step S423: The analysis field set is processed based on the spatial localization mechanism and the covariance expansion mechanism to obtain the initial analysis field.

[0088] In particular, considering the impact of localization and expansion on the set of analytical fields, the set of analytical fields can be processed based on the spatial localization mechanism and the covariance expansion mechanism to obtain the corresponding initial analytical fields.

[0089] Alternatively, spatial localization can be implemented based on a spatial localization mechanism to reduce long-range spurious correlations in the analysis field, and covariance expansion can be used to prevent the adverse effects of set divergence.

[0090] exist Figure 6 In the illustrated embodiment, the set assimilation algorithm can be used to fuse the data and background field, effectively improving the data assimilation effect.

[0091] Optionally, please refer to Figure 7 , Figure 7 The following is a detailed flowchart of step S430 provided in an embodiment of this application. Step S430 may include steps S431-S432.

[0092] Step S431: Determine the influencing factors that affect the initial analysis field.

[0093] During optimization, various influencing factors that affect the initial analysis field can be identified from multiple perspectives, such as observation quality, spatial coverage, and temporal consistency.

[0094] Optionally, the influencing factors may include one or more of the following: weight information determined based on the type and accuracy requirements of the observation data, assimilation influence radius adjusted based on atmospheric variability characteristics at different levels in the near space, time decay factor, and adaptive assimilation residual error model.

[0095] Optionally, the adjustment of weight information may include: assigning higher weights to high-precision observation data types (such as lidar temperature). The adjustment of the assimilation influence radius may include: adjusting the assimilation influence radius according to the atmospheric variability characteristics at different levels in the near space. The adjustment of the time decay factor may include: introducing a time decay factor into non-real-time observation data to reduce its assimilation weight. The update method for the adaptive assimilation residual error model may include: updating the observation error model in real time based on the assimilation residual (observation increment).

[0096] Step S432: Based on the influence factor, adjust the initial analysis field to obtain the target analysis field.

[0097] Among these, the analysis field can be dynamically adjusted through various influencing factors to obtain the final target analysis field.

[0098] Optionally, the dynamic adjustment method may include: dynamically adjusting the assimilation weight based on observation quality, spatial coverage, and temporal consistency.

[0099] exist Figure 7 In the illustrated embodiment, multiple different types of influencing factors can be identified from multiple different perspectives to adjust the initial analysis field, further improving the accuracy and reliability of the target analysis field.

[0100] Optionally, please refer to Figure 8 , Figure 8 This is a flowchart illustrating another near-space multi-source observation data assimilation method provided in an embodiment of this application. The method may further include steps S510-S520.

[0101] Step S510: Perform mean processing based on the target analysis field to obtain the atmospheric analysis field.

[0102] After obtaining the target analysis field, since the target analysis field is a collection of multiple analysis fields, the target analysis field can be averaged to obtain a high-precision atmospheric analysis field.

[0103] Step S520: Based on the atmospheric analysis field, perform atmospheric analysis and processing to obtain weather forecast data for the observation area corresponding to the near space.

[0104] Among these methods, atmospheric analysis and processing can be performed based on the atmospheric analysis field to obtain weather forecast data for the observation area corresponding to the adjacent space of the observation data.

[0105] Optionally, starting from the current atmospheric analysis field, atmospheric dynamics and thermodynamic equations can be used to extrapolate the future evolution of the atmosphere through numerical models, and then weather forecast data within the observation area can be output in combination with release technology.

[0106] exist Figure 8 In the illustrated embodiment, weather forecasting can be performed based on the target analysis field, effectively improving the accuracy of weather forecast data.

[0107] It should be noted that the various observation devices may include: meteor radar, meso-stratospheric-tropospheric radar, lidar, GNSS (Global Navigation Satellite System) occultation and satellite devices, etc., and the various observation data may include: meteor radar data, meso-stratospheric-tropospheric radar data, lidar data, and satellite data from different observation points, one or more of these.

[0108] Optionally, please refer to Figure 9 , Figure 9 The following is a detailed flowchart of step S200 provided in an embodiment of this application. Step S200 may include steps S210-S240.

[0109] Step S210: Perform least squares inversion, effective event screening, error estimation and spatiotemporal consistency verification on the meteor radar data to obtain the first initial data.

[0110] Optionally, the horizontal wind field can be inverted from the arrival angle and radial velocity of meteor trails using the least squares method. Effective event selection may include retaining only meteor events with a signal-to-noise ratio (SNR) greater than a preset threshold and power within a reasonable range. Error estimation may include quantifying the uncertainty of the inversion results using the number of meteors involved in the calculation and wind speed error information. Spatiotemporal consistency verification may include identifying and eliminating anomalous abrupt change points by combining wind field data from adjacent height layers and time points.

[0111] And / or, in step S220, the meso-stratospheric-tropospheric radar data is subjected to sliding window smoothing, outlier removal, and beam consistency verification to obtain the second initial data.

[0112] Optionally, considering the susceptibility of mesosphere-stratospheric-tropospheric radar data to clutter interference, the sliding window smoothing process may include: averaging the wind field profile at 5 centers to construct a reference profile. Outlier removal may include: a 3σ outlier removal method, i.e., calculating the deviation between the original data and the reference value, and removing data points exceeding the mean ± 3 times the standard deviation. Beam consistency verification may include: using multi-beam observation results for cross-validation to improve data reliability.

[0113] And / or, in step S230, perform data filtering and constraint processing on the lidar data to obtain the third initial data.

[0114] Optionally, for lower stratospheric temperature data such as lidar data, data filtering may include: removing layers with low signal-to-noise ratios based on the statistical characteristics of photon number in nitrogen molecule Raman scattering. Constraint processing may include vertical continuity constraints and using regularization methods to smooth the temperature profile and avoid non-physical oscillations.

[0115] And / or, in step S240, spatial matching and error processing are performed on the satellite data to obtain the fourth initial data.

[0116] Optionally, spatial matching for satellite data may include: resampling the satellite's orbital data to a regular grid and aligning it with the pattern grid points. Error model construction may include: assigning corresponding error variances to different observation data based on instrument calibration and historical statistics.

[0117] exist Figure 9 In the illustrated embodiment, for different types of observation data collected by different devices, differentiated preprocessing methods can be set based on the actual situation of each type of observation data, such as physical characteristics and error sources, to obtain various types of initial data, such as first initial data, second initial data, third initial data, and fourth initial data. This ability to perform differentiated preprocessing on different types of observation data effectively improves the effectiveness of various types of initial data.

[0118] For example, taking the Mohe station of the project at noon on August 1, 2023, as an example, the observation data includes meteor radar data, mesosphere-stratospheric-tropospheric radar data, and lidar data: Data Acquisition: Acquire various L2-level radar products and satellite-aided data. Meteor radar data was processed using least squares inversion and effective event filtering; mesosphere-stratospheric-tropospheric radar data was processed using 5-point smoothing and 3σ anomaly removal. Format Conversion: All initial data were uniformly converted to obs_seq format, with time aligned to the UTC hour. Assimilation Process: Input the initial background field in WACCM mode, run the DART filter module, and the assimilation period was 6 hours. Output Analysis Field: Generate 3D analysis products of temperature and wind field; verification showed that the RMSE was reduced by 25% compared to single-data assimilation.

[0119] Optionally, please refer to Figure 10 , Figure 10 This diagram illustrates a temperature assimilation increment as provided in an embodiment of this application, where the temperature assimilation increments on August 1, 2023 are (a) 20 km, (b) 30 km, (c) 40 km, (d) 50 km, (e) 60 km, and (f) 70 km. Figure 10The figure shows the temperature increments at different altitudes before and after assimilation. As can be seen from the figure, the temperature increment is relatively small at altitudes of 20–50 km, approximately ±1 K, while it is larger at altitudes of 60–70 km, approximately ±3 K. This phenomenon is a reasonable result determined by the physical characteristics of the mesosphere, the distribution of observation systems, and model uncertainties. In the lower stratosphere (20–30 km), the observation systems are most densely packed, with global GPS occultation coverage, high vertical resolution, and dense AIRS / MLS satellite observations, resulting in smaller errors. However, in the mesosphere (60–70 km), observations are sparse only by a few dedicated satellites such as SABER / TIMED, with low horizontal resolution and increased instrument errors to ±2–3 K. Furthermore, the temperature variability in the mesosphere increases significantly with altitude. The daily temperature variation at 20–30 km is only 1–2 K, with a seasonal variation of 5–10 K, while the daily temperature variation at 60–70 km can reach 20–30 K, significantly influenced by the solar cycle. This increase in natural variability directly leads to a larger increment.

[0120] Optionally, please refer to Figure 11 , Figure 11 This illustration provides a schematic diagram of zonal wind assimilation increments for embodiments of this application. The zonal wind assimilation increments on August 1, 2023 are (a) 20km, (b) 30km, (c) 40km, (d) 50km, (e) 60km, and (f) 70km. Looking at the zonal wind (U) increments at different altitudes, the 20km-30km altitude reflects the U-wind field increment in the lower and middle layers of the atmosphere. The overall increment is relatively small, approximately ±1m / s, mainly in yellow and light blue, indicating a relatively low adjustment range of the assimilation system to the background wind speed. The increments gradually increase from 60km to 70km, with the central value exceeding ±6m / s, indicating more drastic wind speed adjustments. This is mainly due to the vertical gradient changes in mid-atmosphere observation constraints, the complexity of dynamic processes, and model system bias. The small increments at 20-30km indicate dense observations (sounding, satellite winds), strong dynamic constraints (quasi-geostrophic equilibrium dominates), and relatively mature model simulation of stratospheric circulation with small system biases. The sharp increase in altitude increments of 60–70 km reflects the superposition of multiple challenges near the mesopause: extremely sparse observations (covered by only a few satellites and radars), imperfect parameterization of complex physical processes such as gravity wave fragmentation and molecular diffusion transition, strong diurnal variation and tidal influences (solar radiation cycles cause diurnal temperature variations of 20–30 K, which in turn amplifies wind field adjustments through thermo-wind relationships), and systematic biases in the model regarding meso-layer climatology (such as underestimating the gravity wave drag effect and the intensity of the polar night jet stream). The Chinese region, located in the mid-latitudes, is in a dynamically unstable region transitioning from the stratospheric jet stream to the meso-layer circulation, making it exceptionally sensitive to gravity wave uplift and dissipation processes, further amplifying the significant corrections required by the assimilation system to reconcile sparse observations with the biased background field.

[0121] This application also provides a computer program product, which includes a computer program / instructions. When the computer program / instructions are executed by a processor, they implement the steps in any of the near-space multi-source observation data assimilation methods described above.

[0122] In summary, this application provides a near-space multi-source observation data assimilation method and electronic equipment, which can improve the accuracy of temperature and wind field assimilation analysis by about 20%~30%, achieve an outlier data removal rate of over 95%, improve data reliability, support real-time data stream access, have a processing latency of less than 10 minutes, have strong scalability, and support rapid access to various new observation devices.

[0123] In the several embodiments provided in this application, it should be understood that the disclosed device can also be implemented in other ways. The device embodiments described above are merely illustrative; for example, the block diagrams in the accompanying drawings illustrate the possible architecture, functions, and operations of the device according to various embodiments of this application. In this regard, each block in the block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram, and combinations of block diagrams, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0124] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0125] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0126] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application. It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0127] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

[0128] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A near-space multi-source observation data assimilation method, characterized in that, The method includes: Multiple observation data from near space are preprocessed separately to obtain multiple initial data; Each type of initial data is transformed to obtain multiple observation sequences in a uniform format; Based on the assimilation algorithm, the target analysis field is obtained by assimilating multiple observation sequences in the constructed initial background field.

2. The method according to claim 1, characterized in that, The process of transforming each type of initial data to obtain multiple observation sequences in a uniform format includes: Determine the mapping function between each type of initial data and the uniform target format; Based on the mapping function, the initial data is transformed to obtain the observation sequence in the target format; The observation sequence includes: observation point, observation value, three-dimensional location information, observation time, observation type identifier, and observation error variance.

3. The method according to claim 1, characterized in that, The assimilation algorithm, which involves assimilating multiple observation sequences in a constructed initial background field to obtain the target analysis field, includes: The initial background field for constructing a full atmospheric community climate model; Based on the assimilation algorithm, the initial analysis field is obtained by assimilation processing of multiple observation sequences in the initial background field; wherein, the assimilation algorithm includes: ensemble Kalman filter algorithm or localized particle filter algorithm; The initial analysis field is optimized to obtain the target analysis field.

4. The method according to claim 3, characterized in that, The initial background field for constructing the full atmospheric community climate model includes: Determine the current status data; wherein, the status data includes: time information and location information; Based on the state data, extract the corresponding historical data from the full atmospheric community climate model; Based on the historical data, the initial background field is constructed.

5. The method according to claim 3, characterized in that, The assimilation algorithm is used to assimilate multiple observation sequences in the initial background field to obtain the initial analysis field, including: The multiple observation sequences are input into the data assimilation research test platform; Based on the assimilation algorithm set in the data assimilation research and testing platform, multiple observation sequences are fused with the initial background field to obtain an analysis field set; The initial analysis field is obtained by processing the set of analysis fields based on the spatial localization mechanism and the covariance expansion mechanism.

6. The method according to claim 3, characterized in that, The optimization process of the initial analysis field to obtain the target analysis field includes: Determine the influencing factors that affect the initial analysis field; wherein the influencing factors include at least one of the following: weight information determined based on the type and accuracy requirements of the observation data, assimilation influence radius adjusted based on atmospheric variability characteristics at different levels in the near space, time decay factor, and adaptive assimilation residual error model; Based on the aforementioned influencing factors, the initial analysis field is adjusted to obtain the target analysis field.

7. The method according to any one of claims 1-6, characterized in that, The method further includes: The atmospheric analysis field is obtained by averaging the target analysis field. Based on the atmospheric analysis field, atmospheric analysis and processing are performed to obtain weather forecast data for the observation area corresponding to the near space.

8. The method according to any one of claims 1-6, characterized in that, in, The various types of observation data include at least one of the following: meteor radar data from different observation points, mesosphere-stratospheric-tropospheric radar data, lidar data, and satellite data; The preprocessing of various observation data in the near space yields various initial data, including: The meteor radar data is subjected to least squares inversion, effective event screening, error estimation, and spatiotemporal consistency verification to obtain the first initial data. And / or, perform sliding window smoothing, outlier removal, and beam consistency verification on the mesosphere-stratospheric-tropospheric radar data to obtain second initial data; And / or, perform data filtering and constraint processing on the lidar data to obtain third initial data; And / or, perform spatial matching and error processing on the satellite data to obtain fourth initial data.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing program instructions, and the processor executing the steps of the method according to any one of claims 1-8 when running the program instructions.

10. A computer program product, characterized in that, The computer program product includes a computer program / instruction that, when executed by a processor, implements the steps of the method according to any one of claims 1-8.

Images