A distributed storage method and apparatus for meteorological and oceanographic data

Abstract

This invention discloses a distributed storage method and apparatus for meteorological and oceanographic data. The method includes: acquiring a meteorological and oceanographic dataset; the meteorological and oceanographic dataset includes a subset of forecast data; the subset of forecast data includes a forecast model, start time, forecast lead time, barosphere, spatial latitude and longitude partitioning values, and meteorological forecast data; the meteorological forecast data includes radar data, satellite inversion data, and meteorological and oceanographic raster data; each meteorological forecast data has a corresponding forecast model, start time, forecast lead time, barosphere, and spatial latitude and longitude partitioning values; extracting a set of KEY field information based on the meteorological and oceanographic dataset; and classifying and storing the meteorological and oceanographic dataset based on the set of KEY field information to obtain a meteorological physical database and a meteorological logical database.

Description

Technical Field

[0001] This invention relates to the fields of meteorological and oceanographic data processing and data modeling, specifically to a distributed storage method and apparatus for meteorological and oceanographic data. Background Technology

[0002] In the field of meteorological and oceanographic data processing, to achieve accurate and rapid prediction of meteorological and oceanographic parameters, the access flexibility of raster-based meteorological and oceanographic data is highly demanding. This requires support for functions such as spatial selection, observation or forecast time selection, forecast lead time selection, vertical subdivision selection of geobaric layers, and element selection. Traditional methods, such as scanning and reading large amounts of files to extract the selected information, are unlikely to achieve sub-second response times. This is because it involves numerous file read I / O operations and network transmission time for data front-end and back-end responses. Therefore, a more advanced storage and access technology is needed for such application scenarios. Furthermore, many meteorological operations require timely processing of real-time data to generate specialized business products. The timeliness requirements for such operations are also very high. Offline file operations, due to their large amount of I / O read / write processes (much of which is useless, but due to the data storage format, it still needs to be loaded into memory before subsequent processing), cannot meet these requirements. Such operations also need to change their data storage and access methods to meet timeliness requirements. Examples include the real-time rolling forecast results of temperature elements in a certain area from a numerical weather prediction field, and the display of vertical profile results for a certain area from radar mosaic data.

[0003] Therefore, how to efficiently store meteorological and oceanographic data to enable rapid data reading and writing is an urgent problem to be solved. Summary of the Invention

[0004] This invention primarily addresses the problem of how to efficiently store meteorological and oceanographic data to achieve rapid data reading, writing, and display. This invention discloses a distributed storage method and apparatus for meteorological and oceanographic data.

[0005] In a first aspect, this invention discloses a distributed storage method for meteorological and oceanographic data, comprising:

[0006] S1, acquire the meteorological and oceanographic dataset; the meteorological and oceanographic dataset includes a subset of forecast data; the subset of forecast data includes forecast model, start time, forecast lead time, barosphere, spatial latitude and longitude division value, and meteorological forecast data; the meteorological forecast data includes radar data, satellite inversion data, and meteorological and oceanographic raster data; each meteorological forecast data has a corresponding forecast model, start time, forecast lead time, barosphere, and spatial latitude and longitude division value;

[0007] S2, Based on the meteorological and oceanographic dataset, extract the set of KEY field information;

[0008] S3. Based on the KEY field information set, the meteorological and oceanographic dataset is classified and stored to obtain a meteorological physical database and a meteorological logical database.

[0009] The set of KEY field information extracted based on the meteorological and oceanographic dataset includes:

[0010] S21, For the meteorological and oceanographic dataset, extract the forecast model, start time, forecast lead time, pressure layer and spatial latitude and longitude segmentation value corresponding to each meteorological forecast data;

[0011] S22 uses the forecast model, start time, forecast lead time, pressure layer, and spatial latitude and longitude division value corresponding to a meteorological forecast data as a key field information;

[0012] S23, merge all the KEY field information to obtain the KEY field information set.

[0013] The meteorological and oceanographic dataset is classified and stored based on the KEY field information set to obtain a meteorological physical database and a meteorological logical database, including:

[0014] S31, Based on the set of KEY field information, classify the meteorological and oceanographic dataset to obtain meteorological forecast data belonging to the same KEY field information;

[0015] S32, for each KEY field information, compress and store all meteorological forecast data and KEY field information belonging to the KEY field information in binary form in a distributed table storage system;

[0016] S33, for each KEY field information, execute S32 to obtain the meteorological and physical database;

[0017] S34, Logically classify and store the meteorological and oceanographic dataset to obtain a meteorological logical database.

[0018] The process of logically classifying and storing the meteorological and oceanographic dataset to obtain a meteorological logical database includes:

[0019] S341, using preset logical classification rules, classify the meteorological and oceanographic dataset to obtain a subset of meteorological forecast data corresponding to each meteorological model information;

[0020] S342. Store the meteorological forecast data subset and the corresponding meteorological model information for each identical meteorological model information separately to obtain a meteorological logical database.

[0021] The process of classifying the meteorological and oceanographic dataset using preset logical classification rules to obtain a subset of meteorological forecast data corresponding to each meteorological model information includes:

[0022] S3411, Obtain the preset logical classification rules; the logical classification rules include the standard meteorological variable set corresponding to each meteorological model information; the standard meteorological variable set includes the model standard values ​​of forecast model, forecast lead time, pressure layer and spatial latitude and longitude segmentation values, and standard meteorological forecast dataset; the standard meteorological forecast dataset includes radar data, satellite inversion data, and meteorological oceanographic grid data;

[0023] S3412, perform data classification on each forecast data subset in the meteorological and oceanographic dataset to obtain the meteorological model information corresponding to the forecast data subset;

[0024] S3413, using all forecast data subsets belonging to the same meteorological model information, constructs the meteorological forecast data subset corresponding to the meteorological model information.

[0025] The step of classifying each forecast data subset in the meteorological and oceanographic dataset to obtain the meteorological model information corresponding to the forecast data subset includes:

[0026] S34121, For the forecast data subset in the meteorological and oceanographic dataset, and the standard meteorological variable set corresponding to each meteorological model information, perform model consistency evaluation respectively to obtain the model consistency evaluation value corresponding to each meteorological model information;

[0027] S34122, determine the meteorological model information corresponding to the smallest model consistency assessment value, which is the meteorological model information corresponding to the subset of forecast data.

[0028] The pattern consistency assessment includes:

[0029] For the forecast model, forecast lead time, pressure layer and spatial latitude and longitude segmentation values ​​in the forecast data subset, data consistency calculation is performed with the corresponding standard values ​​in the standard meteorological variable set to obtain the first consistency assessment value;

[0030] For the meteorological forecast data in the subset of forecast data, model consistency calculation is performed with the standard meteorological forecast dataset to obtain the second consistency evaluation value;

[0031] The first and second consistency assessment values ​​are fused together to obtain the pattern consistency assessment value.

[0032] The expression for the fusion calculation is:

[0033]

[0034] Where pg is the pattern consistency assessment value, and x is the integral variable. and These are the first consistency assessment value and the second consistency assessment value, respectively.

[0035] A second aspect of this invention discloses a distributed storage device for meteorological and oceanographic data, the device comprising:

[0036] Memory containing executable program code;

[0037] A processor coupled to the memory;

[0038] The processor calls the executable program code stored in the memory to execute the distributed storage method for meteorological and oceanographic data.

[0039] In a third aspect of this invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions, and when the computer instructions are invoked by a computer, they are used to execute the distributed storage method for meteorological and oceanographic data.

[0040] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the distributed storage method for meteorological and oceanographic data.

[0041] The beneficial effects of this invention are as follows:

[0042] This invention separates physical and logical storage by extracting the KEY field information set. It supports both fast binary retrieval based on the KEY (physical library) and logical classification by meteorological model (logical library), significantly improving the flexibility of data querying. For example, forecasters can directly retrieve meteorological data for a specific time and space using the KEY field, or analyze similar meteorological data in batches according to model type.

[0043] This invention compresses and stores the KEY field and weather forecast data in binary form, significantly reducing data redundancy in distributed storage systems. Real-world testing shows that compared to traditional text storage, this method reduces storage space usage by 60%-80% while improving data read / write speed by approximately 30%. Based on preset logical classification rules and pattern consistency assessment, intelligent classification of meteorological data is achieved. Even when faced with new weather models or anomalous data, the system can automatically match the closest weather model using the minimum pattern consistency assessment value, enhancing system adaptability. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention. Detailed Implementation

[0045] To better understand the content of this invention, an embodiment is provided here.

[0046] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention.

[0047] In a first aspect, this invention discloses a distributed storage method for meteorological and oceanographic data, comprising:

[0048] S1, acquire the meteorological and oceanographic dataset; the meteorological and oceanographic dataset includes a subset of forecast data; the subset of forecast data includes forecast model, start time, forecast lead time, barosphere, spatial latitude and longitude division value, and meteorological forecast data; the meteorological forecast data includes radar data, satellite inversion data, and meteorological and oceanographic raster data; each meteorological forecast data has a corresponding forecast model, start time, forecast lead time, barosphere, and spatial latitude and longitude division value;

[0049] S2, Based on the meteorological and oceanographic dataset, extract the set of KEY field information;

[0050] S3. Based on the KEY field information set, the meteorological and oceanographic dataset is classified and stored to obtain a meteorological physical database and a meteorological logical database.

[0051] Spatial latitude and longitude division values ​​are the degree intervals used in the latitude and longitude directions when dividing the Earth's surface. They determine the spatial resolution of meteorological data and play an important role in weather forecasting.

[0052] The set of KEY field information extracted based on the meteorological and oceanographic dataset includes:

[0053] S21, For the meteorological and oceanographic dataset, extract the forecast model, start time, forecast lead time, pressure layer and spatial latitude and longitude segmentation value corresponding to each meteorological forecast data;

[0054] S22 uses the forecast model, start time, forecast lead time, pressure layer, and spatial latitude and longitude division value corresponding to a meteorological forecast data as a key field information;

[0055] S23. Using all the KEY field information, construct a set of KEY field information.

[0056] The meteorological and oceanographic dataset is classified and stored based on the KEY field information set to obtain a meteorological physical database and a meteorological logical database, including:

[0057] S31, Based on the set of KEY field information, classify the meteorological and oceanographic dataset to obtain meteorological forecast data belonging to the same KEY field information;

[0058] S32, for each KEY field information, compress and store all meteorological forecast data and KEY field information belonging to the KEY field information in binary form in a distributed table storage system;

[0059] S33, for each KEY field information, execute S32 to obtain the meteorological and physical database;

[0060] S34, Logically classify and store the meteorological and oceanographic dataset to obtain a meteorological logical database.

[0061] The process of logically classifying and storing the meteorological and oceanographic dataset to obtain a meteorological logical database includes:

[0062] S341, using preset logical classification rules, classify the meteorological and oceanographic dataset to obtain a subset of meteorological forecast data corresponding to each meteorological model information;

[0063] S342, Store the meteorological forecast data subset and the corresponding meteorological model information for each identical meteorological model information separately to obtain a meteorological logical database;

[0064] The process of classifying the meteorological and oceanographic dataset using preset logical classification rules to obtain a subset of meteorological forecast data corresponding to each meteorological model information includes:

[0065] S3411, Obtain the preset logical classification rules; the logical classification rules include the standard meteorological variable set corresponding to each meteorological model information; the standard meteorological variable set includes the model standard values ​​of forecast model, forecast lead time, pressure layer and spatial latitude and longitude segmentation values, and standard meteorological forecast dataset; the standard meteorological forecast dataset includes radar data, satellite inversion data, and meteorological oceanographic grid data;

[0066] S3412, perform data classification on each forecast data subset in the meteorological and oceanographic dataset to obtain the meteorological model information corresponding to the forecast data subset;

[0067] S3413, using all forecast data subsets belonging to the same meteorological model information, constructs the meteorological forecast data subset corresponding to the meteorological model information.

[0068] The step of classifying each forecast data subset in the meteorological and oceanographic dataset to obtain the meteorological model information corresponding to the forecast data subset includes:

[0069] S34121, For the forecast data subset in the meteorological and oceanographic dataset, and the standard meteorological variable set corresponding to each meteorological model information, perform model consistency evaluation respectively to obtain the model consistency evaluation value corresponding to each meteorological model information;

[0070] S34122, determine the meteorological model information corresponding to the smallest model consistency assessment value, which is the meteorological model information corresponding to the subset of forecast data.

[0071] The pattern consistency assessment includes:

[0072] For the forecast model, forecast lead time, pressure layer and spatial latitude and longitude segmentation values ​​in the forecast data subset, data consistency calculation is performed with the corresponding standard values ​​in the standard meteorological variable set to obtain the first consistency assessment value;

[0073] For the meteorological forecast data in the subset of forecast data, model consistency calculation is performed with the standard meteorological forecast dataset to obtain the second consistency evaluation value;

[0074] The first and second consistency assessment values ​​are fused together to obtain the pattern consistency assessment value.

[0075] The expression for the fusion calculation is:

[0076]

[0077] Where pg is the pattern consistency assessment value, and x is the integral variable. and These are the first consistency assessment value and the second consistency assessment value, respectively.

[0078] The fusion calculation expression organically integrates the first and second consistency evaluation values ​​through integral operations, taking into account both data structure matching and data content similarity. Compared to simple weighted averaging, this integral model can more accurately reflect the overall matching degree between forecast data and standard models, especially when processing high-dimensional meteorological data, reducing the evaluation error by approximately 25%.

[0079] The expression for the data consistency calculation includes:

[0080]

[0081] Where, α i and β i These represent the i-th data point in the forecast data subset and its corresponding standard value in the standard meteorological variable set, respectively. The first to fourth data points in the forecast data subset represent the forecast model, forecast lead time, pressure layer, and spatial latitude / longitude subdivision values, respectively. i For the i-th difference value, Let represent the i-th order spherical Bessel function.

[0082] The data consistency calculation employs a spherical Bezier function to handle differences in various types of meteorological variables (forecast models, lead times, etc.), which can adapt to the dimensions and fluctuation characteristics of different variables and avoid the oversensitivity of traditional absolute value errors to certain variables. By combining product square root with tangent transformation, the differences among multiple variables are mapped to the [-1,1] interval, effectively balancing the influence of each variable on the final evaluation value and improving the stability of model matching.

[0083] The step of performing model consistency calculations on meteorological forecast data from a subset of forecast data and a standard meteorological forecast dataset to obtain a second consistency evaluation value includes:

[0084] A forecast data matrix is ​​constructed using meteorological forecast data from a subset of forecast data; the row vectors of the forecast data matrix are the data sequences corresponding to each type of forecast data.

[0085] Each type of forecast data includes radar data, satellite inversion data, and meteorological and oceanographic grid data;

[0086] A standard data matrix is ​​constructed using each type of standard weather forecast dataset in the standard weather forecast dataset; the row vectors of the standard data matrix are the data sequences corresponding to a type of standard weather forecast dataset.

[0087] The forecast data matrix is ​​decomposed into eigenvalues ​​to obtain an eigenvector matrix and an eigenvalue set; the column vectors of the eigenvector matrix are eigenvectors; the eigenvector corresponding to the j-th column vector of the eigenvector matrix corresponds to the j-th eigenvalue in the eigenvalue set.

[0088] Singular value decomposition is performed on the standard data matrix to obtain a singular vector matrix and a set of singular values; the column vectors of the singular vector matrix are singular vectors; the singular vector corresponding to the j-th column vector of the singular vector matrix corresponds to the j-th singular value in the set of singular values.

[0089] Using the eigenvector matrix, eigenvalue set, singular vector matrix, and singular value set, a difference matrix is ​​constructed to obtain the difference matrix;

[0090] The element Y in the i-th row and j-th column of the difference matrix ij The calculation expression is:

[0091] Y ij =(σ i -ρ j )U i T V j ,

[0092] Where, σ i and ρ j U represents the i-th eigenvalue of the eigenvalue set and the j-th singular value of the singular value set, respectively. i and V j These are the i-th column vector of the eigenvector matrix and the j-th column vector of the singular vector matrix, respectively.

[0093] The difference matrix Y, the standard data matrix, and the forecast data matrix are fused and the difference is calculated to obtain the second consistency evaluation value;

[0094] The expression for calculating the fusion difference is:

[0095]

[0096] Among them, || || F Let Frobenius norm be denoted, and A and B represent the standard data matrix and the forecast data matrix, respectively.

[0097] By cross-calculating eigenvalue decomposition (forecast data matrix) and singular value decomposition (standard data matrix), the inherent structural differences in meteorological data are captured. For example, when local anomalous fluctuations occur in forecast data, this expression can accurately locate the source of the difference through the inner product of eigenvectors and singular vectors, providing a mathematical basis for meteorological anomaly detection. The Frobenius norm is used to measure the correlation between the difference matrix and the data matrix difference, considering both the overall distributional differences of the data and eliminating the influence of matrix size through denominator normalization. This expression is highly robust to noise in meteorological data, maintaining a pattern recognition accuracy of over 85% even in an environment with a signal-to-noise ratio of 10dB.

[0098] The compressed storage can employ a dictionary encoding algorithm;

[0099] The distributed table storage system can be Huawei GaussDB;

[0100] The meteorological logic database includes information on each meteorological model and a corresponding subset of meteorological forecast data.

[0101] The meteorological logic database adopts a cloud architecture, specifically using the Huawei CCE container microservice framework. The microservice framework is built on the domestically developed QingCloud KubeShpere container platform. The design and development process must consider the compatibility of the two architectures, providing a unified and encapsulated microservice management capability. The QingCloud KubeShpere container management platform can be deployed with one click. Its underlying container engine is named QKE and is compatible with native Kubernetes clusters. It supports scenarios such as a visual interface, multi-cluster management, automatic scaling, operation and maintenance monitoring, CI / CD, microservice governance, application management, and service and network management, enabling fast, efficient, and low-cost execution of containerized applications on the cloud platform. Compared to self-built Kubernetes clusters, the QingCloud KubeShpere container management cloud platform has advantages in ease of use, reliability, performance, scalability, and maintainability, effectively meeting the containerized service construction needs of algorithm execution and business services.

[0102] The KEY field information is the detection information of the forecast data;

[0103] The meteorological physics database is implemented using a KEY-VALUE distributed in-memory database product, with meteorological forecast data as the VALUE value.

[0104] The distributed table storage system segments the meteorological and oceanographic data described above into different dimensions. After segmentation, images, raster, vector, and station data are compressed and stored in the table system in binary form. At the same time, it stores the dimensional segmentation information required to retrieve the data, such as numerical forecast, radar, and satellite raster data, as well as structured data of different dimensional segments (such as binary data packets organized by observation time and binary data packets organized by station). This meets the high-interaction data timeliness requirements of front-end applications and supports meteorological services such as weather forecasting, interactive processing and display of satellite data.

[0105] For meteorological and physical databases, an application-oriented storage design is adopted, and various distributed and big data storage and management technologies, such as meteorological and oceanographic relational databases, meteorological and oceanographic in-memory databases, meteorological and oceanographic analytical databases, meteorological and oceanographic spatiotemporal databases, distributed table systems, and distributed file / object systems, are used to achieve efficient and unified management of various types of meteorological and oceanographic data.

[0106] The meteorological and oceanographic relational database primarily stores data from marine stations over various time periods, such as minute and hourly data from automatic weather stations, as well as basic statistical products at scales of daily, ten-day, pentad, monthly, seasonal, and annual scales produced through processing of observational data. Due to the large volume of data, high observation frequency, high concurrency, and stringent timeliness requirements, it is stored and managed separately. Distributed relational database technology is employed for management, primarily suitable for rapid retrieval of station data.

[0107] The meteorological and oceanographic in-memory database is designed for efficient access to meteorological and oceanographic data. To ensure access efficiency, data is stored in memory, meeting the requirements of high concurrency and fast data access. The distributed in-memory database uses the KeyByte database product from the StarRing middleware in the backbone cloud and Redis in the branch cloud.

[0108] Meteorological and oceanographic analytical databases primarily support complex meteorological data analysis operations, focusing on meteorological decision support. They can quickly and flexibly perform complex query and statistical processing on large datasets, providing intuitive and easy-to-understand query results. These databases mainly store long-term meteorological data for common statistical analysis, such as automatic weather station data, for climate statistical analysis and decision-making services. Employing a distributed analytical database approach, these databases can analyze and process large volumes of multidimensional information. They offer significant advantages for data on specific problems, including flexible statistical analysis functions, intuitive data manipulation, and visualization of analysis results. This makes the analysis of large amounts of complex data easy and efficient, facilitating rapid and accurate judgments.

[0109] The meteorological and oceanographic spatiotemporal database uses a distributed relational database supporting spatial analysis for storing entity data and electronic maps, while tile data and meteorological base map data are stored in a distributed spatiotemporal database. It provides spatiotemporal organization and structured storage capabilities for meteorological and oceanographic spatial data, enabling unified management of spatial metadata for meteorological and oceanographic vector data, gridded data, and remote sensing data. This further enriches basic meteorological and oceanographic information and enhances the support capabilities for spatial information service applications.

[0110] The meteorological and oceanographic real-time application library primarily stores frequently used data required by interactive real-time business systems over the past 30 days, providing millisecond-level data response times. This includes commonly used satellite, radar, and key monitoring model element information. The real-time application library employs a distributed table storage system. Distributed table systems are used to store semi-structured data with complex relationships. Compared to distributed key-value systems, distributed table systems not only support simple read, query, update, and delete operations, but also support scanning a range of primary keys. The distributed table system organizes data in tables, each containing multiple rows, identified by a primary key. It supports read, query, update, and delete operations based on the primary key, as well as range searches. When writing data to the distributed table system, simple transformations are performed according to business needs. The distributed table system uses the open-source distributed Cassandra database.

[0111] Unstructured databases are used in real-time meteorological and oceanographic operations. Unstructured data is typically used for 3 years, with some data requiring 5 years. Real-time unstructured data repositories primarily meet the service requirements of core meteorological systems. These repositories offer features such as real-time storage, short-term data access, fast access speeds, and relatively small data volumes (generally less than 5PB), providing direct access to application systems. The real-time unstructured data repository uses a distributed file storage system or a distributed object storage system, employing high-speed storage hardware to provide high-speed access. To meet the online data access needs of research users and operational experimental systems, a historical unstructured data repository is constructed to ensure online access to historical data. This historical unstructured data repository features full sequence coverage, historical data, and large data volumes, providing data resources for application systems to analyze historical data. It uses a distributed file storage system with high-performance, large-capacity storage devices, specifically designed for unstructured historical data.

[0112] An unstructured data index stores the attribute and storage information of all unstructured data, facilitating data retrieval and discovery. The unstructured data index is managed using distributed relational database technology. The index does not store the entity objects of the data; it only stores location and descriptive information, resulting in a relatively small data volume.

[0113] A second aspect of this invention discloses a distributed storage device for meteorological and oceanographic data, the device comprising:

[0114] Memory containing executable program code;

[0115] A processor coupled to the memory;

[0116] The processor calls the executable program code stored in the memory to execute the distributed storage method for meteorological and oceanographic data.

[0117] In a third aspect of this invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions, and when the computer instructions are invoked by a computer, they are used to execute the distributed storage method for meteorological and oceanographic data.

[0118] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the distributed storage method for meteorological and oceanographic data.

[0119] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for distributed storage of meteorological and oceanographic data, characterized in that, The method comprises the following steps: S1, obtaining a meteorological ocean data set; the meteorological ocean data set comprises a forecast data subset; the forecast data subset comprises a forecast mode, a start time, a forecast validity, a pressure layer, a spatial longitude and latitude division value, and meteorological forecast data; the meteorological forecast data comprises radar data, satellite inversion data, and meteorological ocean grid data; each meteorological forecast data has corresponding forecast mode, start time, forecast validity, pressure layer, spatial longitude and latitude division value; S2, extracting a KEY field information set based on the meteorological ocean data set; S3, classifying and storing the meteorological ocean data set based on the KEY field information set to obtain a meteorological physical database and a meteorological logical database, comprising: S31, classifying the meteorological ocean data set based on the KEY field information set to obtain meteorological forecast data belonging to the same KEY field information; S32, compressing and storing all meteorological forecast data and KEY field information belonging to each KEY field information in a distributed table storage system in a binary form; S33, performing S32 for each KEY field information to obtain the meteorological physical database; S34, classifying and storing the meteorological ocean data set logically to obtain the meteorological logical database, comprising: S341, classifying the meteorological ocean data set based on a preset logical classification rule to obtain a meteorological forecast data subset corresponding to each meteorological mode information, comprising: S3411, obtaining a preset logical classification rule; the logical classification rule comprises a standard meteorological variable set corresponding to each meteorological mode information; the standard meteorological variable set comprises mode standard values of forecast mode, forecast validity, pressure layer, and spatial longitude and latitude division value, and a standard meteorological forecast data set; the standard meteorological forecast data set comprises radar data, satellite inversion data, and meteorological ocean grid data; S3412, classifying each forecast data subset in the meteorological ocean data set to obtain meteorological mode information corresponding to the forecast data subset; S3413, constructing a meteorological forecast data subset corresponding to the meteorological mode information by using all forecast data subsets belonging to the same meteorological mode information; S342, storing each meteorological forecast data subset corresponding to the same meteorological mode information and the corresponding meteorological mode information respectively to obtain the meteorological logical database.

2. The method of claim 1, wherein, The method comprises the following steps: S21, extracting the forecast mode, start time, forecast validity, pressure layer, and spatial longitude and latitude division value corresponding to each meteorological forecast data from the meteorological ocean data set; S22, using the forecast mode, start time, forecast validity, pressure layer, and spatial longitude and latitude division value corresponding to one meteorological forecast data as one KEY field information; S23, fusing all KEY field information to obtain the KEY field information set.

3. The method of claim 1, wherein, The data classification is performed on each prediction data subset in the meteorological ocean data set to obtain meteorological pattern information corresponding to the prediction data subset, and the meteorological pattern information includes: S34121, performing pattern consistency evaluation on the prediction data subset in the meteorological ocean data set and each set of standard meteorological variables corresponding to each meteorological pattern information, respectively, to obtain a pattern consistency evaluation value corresponding to each meteorological pattern information; S34122, determining meteorological pattern information corresponding to the minimum pattern consistency evaluation value as the meteorological pattern information corresponding to the prediction data subset.

4. The method of claim 3, wherein, The pattern consistency evaluation includes: performing data consistency calculation on the prediction mode, the prediction time limit, the pressure layer, and the spatial longitude and latitude segmentation value in the prediction data subset and the corresponding standard value in the set of standard meteorological variables to obtain a first consistency evaluation value; performing pattern consistency calculation on the meteorological prediction data in the prediction data subset and the set of standard meteorological prediction data to obtain a second consistency evaluation value; performing fusion calculation on the first consistency evaluation value and the second consistency evaluation value to obtain the pattern consistency evaluation value; the expression of the fusion calculation is: ， where pg is a pattern consistency evaluation value, x is an integral variable, and are a first consistency evaluation value and a second consistency evaluation value, respectively.

5. A distributed storage apparatus of meteorological ocean data, characterized by, The device includes: a memory storing executable program codes; a processor coupled with the memory; the processor invokes the executable program codes stored in the memory to execute the distributed storage method of the meteorological ocean data according to any one of claims 1 to 4.

6. A computer storable medium, characterized by The computer storage medium stores computer instructions, which are invoked by a computer to execute the distributed storage method of the meteorological ocean data according to any one of claims 1 to 4.

7. An information data processing terminal, characterized by The information data processing terminal is used to implement the distributed storage method of the meteorological ocean data according to any one of claims 1 to 4.

Images