Meteorological forecast detection method, device, electronic equipment and program product

Abstract

The application provides a test method and device for a prediction mode, an electronic device and a program product. The method comprises: obtaining a target parameter input by a user, the target parameter comprising a target area, a target period and at least one meteorological element; determining a target element set according to the at least one meteorological element, the target element set comprising a plurality of target elements; determining original meteorological data corresponding to the plurality of target elements in the target period of the target area; performing anomaly detection processing on the original meteorological data corresponding to any one target element to determine target meteorological data corresponding to the target element; and generating a visual test result corresponding to a meteorological prediction according to the target meteorological data corresponding to the plurality of target elements, thereby improving the accuracy of detecting the meteorological prediction.

Description

Technical Field

[0001] This application relates to the field of weather forecasting, and more particularly to a weather forecasting detection method, apparatus, electronic equipment, and program product. Background Technology

[0002] With the continuous advancement of science and technology, numerical weather prediction models (NWP) have become a key tool in modern weather forecasting. NWP models use complex mathematical models and a large amount of observational data to predict future weather changes, covering various meteorological elements such as temperature, precipitation, wind direction, and wind speed.

[0003] Currently, weather forecast verification can be achieved in the following ways: multiple verification methods are used in ensemble forecast evaluation, supporting various verification tools for ensemble forecasts, and ensemble forecast results are evaluated through different verification methods; frequency matching and probability matching methods are used to perform real-time correction of precipitation forecasts, mainly focusing on the forecast verification of precipitation as a single meteorological element.

[0004] In the above methods, the meteorological data to be tested usually comes from different meteorological models and different data sources, resulting in low accuracy of the detected weather forecasts. Summary of the Invention

[0005] This application provides a method, apparatus, electronic device, and program product for detecting weather forecasts, in order to solve the technical problem of low accuracy in detecting weather forecasts in the prior art.

[0006] In a first aspect, this application provides a method for detecting weather forecasts, the method comprising:

[0007] Obtain target parameters input by the user, the target parameters including target area, target time period and at least one meteorological element;

[0008] Based on the at least one meteorological element, a target element set is determined, wherein the target element set includes multiple target elements;

[0009] Within the target time period in the target area, determine the original meteorological data corresponding to each of the multiple target elements;

[0010] For any target element, perform anomaly detection processing on the original meteorological data corresponding to the target element to determine the target meteorological data corresponding to the target element;

[0011] Based on the target meteorological data corresponding to the multiple target elements, a visual verification result corresponding to the meteorological forecast is generated.

[0012] In this way, by using the target parameters input by the user, a set of matching target elements is determined, and anomaly detection processing is performed based on the target elements to generate visualized inspection results, thereby improving the accuracy of weather forecast detection.

[0013] Optionally, the method described above involves performing anomaly detection processing on the original meteorological data corresponding to the target element to determine the target meteorological data corresponding to the target element, including:

[0014] Determine the element type corresponding to the target element;

[0015] Based on the element type, the original meteorological data is subjected to weighted difference processing to obtain weighted difference values;

[0016] Based on the weighted difference value and the original meteorological data, abnormal data in the original meteorological data are determined by a target algorithm.

[0017] Based on the abnormal data, the original meteorological data is filtered to obtain the target meteorological data.

[0018] In this way, different anomaly detection and processing are carried out according to different target elements, which improves the accuracy of weather forecast detection.

[0019] Optionally, the method described above, based on the weighted difference value and the original meteorological data, determines the abnormal data in the original meteorological data using a target algorithm, including:

[0020] A subsampling process is performed on the weighted difference value and the original meteorological data to obtain multiple sample sets. Each sample set includes multiple sample data, which are either the weighted difference value or metadata. The metadata is any one of the data in the original meteorological data.

[0021] For any given sample set, based on the multiple sample data corresponding to the sample set, an isolated tree for the sample set is established, and based on the isolated tree, the path lengths corresponding to the multiple sample data in the sample set are calculated respectively;

[0022] Based on multiple path lengths of multiple sampling sets, abnormal data in the original meteorological data are identified.

[0023] In this way, abnormal data can be identified based on different specific circumstances, thus improving the accuracy of weather forecasts.

[0024] Optionally, the method described above, determining anomalous data in the original meteorological data based on multiple path lengths of multiple sampling sets, includes:

[0025] For any metadata, determine the first path length corresponding to the metadata in multiple sampling sets respectively;

[0026] Determine the first average of multiple first path lengths;

[0027] Determine the dynamic threshold corresponding to the metadata;

[0028] If the first average value is greater than the dynamic threshold, then the metadata is determined to be in an abnormal state;

[0029] If the first average value is less than or equal to the dynamic threshold, then the metadata is determined to be in a normal state;

[0030] The metadata in the abnormal state is identified as abnormal data in the original meteorological data.

[0031] In this way, abnormal data can be identified based on different specific circumstances, thus improving the accuracy of weather forecasts.

[0032] Optionally, in the method described above, the raw meteorological data includes any multiple of observational data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data. Based on the element type, the raw meteorological data undergoes weighted differencing to obtain weighted difference values, including:

[0033] Determine whether the element type is a precipitation type;

[0034] If so, determine the first factor corresponding to the observed data, the second factor corresponding to the reanalysis data, and the third factor corresponding to the GPM data, and determine the sum of the first factor, the second factor, and the third factor as the weighted difference value, wherein the first factor is the product of the first weight value and the first difference corresponding to the observed data, and the first difference is the difference between any data in the observed data and the first weighted average value corresponding to the observed data;

[0035] If not, determine the first factor corresponding to the observed data and the second factor corresponding to the reanalysis data, and determine the sum of the first factor and the second factor as the weighted difference value.

[0036] In this way, different anomaly detection and processing are carried out according to different target elements, which improves the accuracy of weather forecast detection.

[0037] Optionally, the method described above, determining the target element set based on the at least one meteorological element, includes:

[0038] Obtain a set of multiple candidate elements, wherein the set of candidate elements includes multiple elements;

[0039] Determine multiple similarities between the at least one meteorological element and a set of multiple candidate elements;

[0040] The set of candidate elements with the highest similarity among the multiple similarity values ​​is determined as the target element set.

[0041] In this way, the most matching set of target elements is determined, which improves the accuracy of weather forecasting.

[0042] Optionally, the method described above generates a visual verification result corresponding to the weather forecast based on the target meteorological data corresponding to the multiple target elements, including:

[0043] For any target element, the target meteorological data corresponding to the target element is classified and processed to obtain multiple types of data, including any and multiple types of observation data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data.

[0044] Standardize each type of data to obtain the standard data corresponding to each type of data;

[0045] The standard data corresponding to multiple types of data are input into the target model to determine multiple test scores corresponding to the target elements;

[0046] Based on the multiple test scores corresponding to the multiple target elements, a visual test result corresponding to the weather forecast is generated.

[0047] This generates visual verification results corresponding to weather forecasts, improving the readability of the verification.

[0048] Secondly, this application provides a weather forecast detection device, the device comprising:

[0049] The acquisition module is used to acquire target parameters input by the user, the target parameters including target area, target time period and at least one meteorological element;

[0050] The first determining module is used to determine a set of target elements based on the at least one meteorological element, wherein the set of target elements includes multiple target elements;

[0051] The second determining module is used to determine the original meteorological data corresponding to the plurality of target elements respectively within the target time period in the target area;

[0052] The inspection module is used to perform anomaly detection processing on the original meteorological data corresponding to any target element, and to determine the target meteorological data corresponding to the target element.

[0053] The display module is used to generate visual verification results of weather forecasts based on the target meteorological data corresponding to the multiple target elements.

[0054] Optionally, in the apparatus described above, the inspection module is specifically used for:

[0055] Determine the element type corresponding to the target element;

[0056] Based on the element type, the original meteorological data is subjected to weighted difference processing to obtain weighted difference values;

[0057] Based on the weighted difference value and the original meteorological data, abnormal data in the original meteorological data are determined by a target algorithm.

[0058] Based on the abnormal data, the original meteorological data is filtered to obtain the target meteorological data.

[0059] Optionally, in the apparatus described above, the inspection module is specifically used for:

[0060] A subsampling process is performed on the weighted difference value and the original meteorological data to obtain multiple sample sets. Each sample set includes multiple sample data, which are either the weighted difference value or metadata. The metadata is any one of the data in the original meteorological data.

[0061] For any given sample set, based on the multiple sample data corresponding to the sample set, an isolated tree for the sample set is established, and based on the isolated tree, the path lengths corresponding to the multiple sample data in the sample set are calculated respectively;

[0062] Based on multiple path lengths of multiple sampling sets, abnormal data in the original meteorological data are identified.

[0063] Optionally, in the apparatus described above, the inspection module is specifically used for:

[0064] For any metadata, determine the first path length corresponding to the metadata in multiple sampling sets respectively;

[0065] Determine the first average of multiple first path lengths;

[0066] Determine the dynamic threshold corresponding to the metadata;

[0067] If the first average value is greater than the dynamic threshold, then the metadata is determined to be in an abnormal state;

[0068] If the first average value is less than or equal to the dynamic threshold, then the metadata is determined to be in a normal state;

[0069] The metadata in the abnormal state is identified as abnormal data in the original meteorological data.

[0070] Optionally, in the apparatus described above, the raw meteorological data includes any and multiple of observational data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data, and the verification module is specifically used for:

[0071] Determine whether the element type is a precipitation type;

[0072] If so, determine the first factor corresponding to the observed data, the second factor corresponding to the reanalysis data, and the third factor corresponding to the GPM data, and determine the sum of the first factor, the second factor, and the third factor as the weighted difference value, wherein the first factor is the product of the first weight value and the first difference corresponding to the observed data, and the first difference is the difference between any data in the observed data and the first weighted average value corresponding to the observed data;

[0073] If not, determine the first factor corresponding to the observed data and the second factor corresponding to the reanalysis data, and determine the sum of the first factor and the second factor as the weighted difference value.

[0074] Optionally, in the apparatus described above, the first determining module is specifically used for:

[0075] Obtain a set of multiple candidate elements, wherein the set of candidate elements includes multiple elements;

[0076] Determine multiple similarities between the at least one meteorological element and a set of multiple candidate elements;

[0077] The set of candidate elements with the highest similarity among the multiple similarity values ​​is determined as the target element set.

[0078] Optionally, in the apparatus described above, the display module is specifically used for:

[0079] For any target element, the target meteorological data corresponding to the target element is classified and processed to obtain multiple types of data, including any and multiple types of observation data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data.

[0080] Standardize each type of data to obtain the standard data corresponding to each type of data;

[0081] The standard data corresponding to multiple types of data are input into the target model to determine multiple test scores corresponding to the target elements;

[0082] Based on the multiple test scores corresponding to the multiple target elements, a visual test result corresponding to the weather forecast is generated.

[0083] Thirdly, this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0084] The memory stores computer-executed instructions;

[0085] The processor executes computer execution instructions stored in the memory to implement the method described in any of the first aspects.

[0086] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any of the first aspects.

[0087] Fifthly, this application provides a computer program product, including a computer program that, when executed by a computer, implements the method as described in any one of the first aspects.

[0088] The weather forecast detection method, apparatus, electronic equipment, and program product provided in this application improve the accuracy of weather forecast detection by acquiring user-input target parameters, including a target area, a target time period, and at least one meteorological element; determining a set of target elements, which includes multiple target elements, based on the at least one meteorological element; identifying the original meteorological data corresponding to each of the multiple target elements within the target area and target time period; performing anomaly detection processing on the original meteorological data corresponding to any one target element to determine the target meteorological data corresponding to the target element; and generating a visualized verification result for the weather forecast based on the target meteorological data corresponding to the multiple target elements. In this way, by determining the matching set of target elements through user-input target parameters and performing anomaly detection processing based on the target elements to generate a visualized verification result, the accuracy of weather forecast detection is improved. Attached Figure Description

[0089] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0090] Figure 1 This is a schematic diagram of the structure of an inspection system provided in an embodiment of this application;

[0091] Figure 2 A flowchart illustrating a method for verifying a forecast pattern provided in an embodiment of this application;

[0092] Figure 3 A flowchart illustrating another method for verifying a forecast pattern provided in this application embodiment;

[0093] Figure 4A schematic diagram of the structure of a forecast pattern verification device provided in an embodiment of this application;

[0094] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0095] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0096] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0097] It should be noted that although the terms "first," "second," etc., are used to describe various types of information in the embodiments of this application, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. Optionally, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information.

[0098] It should be understood that the terms "comprising" or "including" indicate the presence of the previously mentioned features, steps, or operations, but do not preclude the presence, occurrence, or addition of one or more other features, steps, or operations. The terms "and / or," etc., used in this application can be interpreted as inclusive, or mean any one or any combination thereof. Optionally, "A and / or B" means "any one of the following: A; B; A and B." Additionally, the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0099] With the continuous advancement of science and technology, numerical weather prediction models (NWP) have become a key tool in modern weather forecasting. NWP models use complex mathematical models and a large amount of observational data to predict future weather changes, covering various meteorological elements such as temperature, precipitation, wind direction, and wind speed.

[0100] Currently, weather forecast verification can be achieved in the following ways: multiple verification methods are used in ensemble forecast evaluation, supporting various verification tools for ensemble forecasts, and ensemble forecast results are evaluated through different verification methods; frequency matching and probability matching methods are used to perform real-time correction of precipitation forecasts, mainly focusing on the forecast verification of precipitation as a single meteorological element.

[0101] In the above methods, the meteorological data to be tested usually comes from different meteorological models and different data sources, resulting in low accuracy of the detected weather forecasts.

[0102] To address the aforementioned technical problems, this application provides a weather forecast detection method. The method involves acquiring user-input target parameters, determining a target element set based on at least one meteorological element from the target parameters, identifying the original meteorological data corresponding to each of the multiple target elements within a target area and time period, performing anomaly detection processing on the original meteorological data corresponding to any given target element to determine the target meteorological data corresponding to that element, and generating a visualized verification result for the weather forecast based on the target meteorological data corresponding to the multiple target elements. This method improves the accuracy of weather forecast detection by determining a matching target element set based on user-input target parameters, performing anomaly detection processing based on the target elements, and generating a visualized verification result.

[0103] Figure 1 This is a schematic diagram of a detection system provided in an embodiment of this application. Please refer to [link / reference]. Figure 1 , Figure 1 It can include a data layer, a quality processing layer, an analysis layer, and a visualization layer.

[0104] The data layer can be used to automatically download multiple meteorological data from various forecast models. These data can include various forecast data, observational data, GPM data from the Global Precipitation Measurement Plan (GPP), and reanalysis data. It supports multiple data formats, different pressure levels, and multidimensional meteorological elements. For example, multiple data formats can include Grib, NetCDF, and CSV formats. Different pressure levels can include 200 hPa, 500 hPa, 700 hPa, 850 hPa, and 925 hPa. Multidimensional meteorological elements can include precipitation, temperature, zonal wind, radial wind, and humidity.

[0105] The data downloaded by the data layer can have different spatiotemporal resolutions.

[0106] The data layer can automate data downloads through a timed task scheduling mechanism, ensuring the acquisition of real-time and accurate forecast and observation data. This significantly reduces the complexity of manual operations and greatly improves the real-time performance of data acquisition and system processing efficiency.

[0107] The quality processing layer can be used for data quality control through target algorithms.

[0108] Data quality control can include deleting abnormal files, checking for data anomalies and deleting files containing abnormal data, and repairing or correcting abnormal data that can be repaired.

[0109] For example, abnormal files are files that cannot be opened normally, or files with less data than normal.

[0110] After deleting a file, the data layer can resubmit the data download task for download. If the resubmission fails three times, the download task will not be submitted again.

[0111] Data quality control can ensure the accuracy and reliability of input data and avoid deviations in test results due to data defects.

[0112] The quality processing layer can also be used to convert downloaded data into a uniform file format, data format, and spatiotemporal resolution, and to align the data with a spatial grid structure.

[0113] The quality processing layer overcomes the compatibility issues caused by inconsistent data formats in existing technologies by standardizing data from various forecast models from different sources, significantly improving the efficiency and accuracy of data processing. The data layer successfully solves the technical challenges of diverse data sources and inconsistent formats, enabling unified management and analysis of multiple data sources.

[0114] The analysis layer can include multiple test indicators.

[0115] For example, multiple test indicators can include root mean square error, standard deviation, correlation coefficient, threat score, and reliability.

[0116] The analysis layer can be used to diagnose and attribute the errors of multiple forecast models based on multiple test index values.

[0117] The visualization layer can be designed based on a lightweight B / S architecture, supporting high-performance interactive operation on the web page, and has flexible layer switching, legend zooming and parameter setting capabilities, adapting to multi-terminal browsing environments.

[0118] The visualization layer can have a "one-click skill report generation" function, which automatically summarizes the core assessment results and visualization graphics to form standardized or customized output reports, greatly improving the automation and intelligence level of the business assessment process.

[0119] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0120] The technical solutions shown in this application will now be described in detail through specific embodiments. It should be noted that the following embodiments may exist independently or in combination with each other; for the same or similar content, the description will not be repeated in different embodiments.

[0121] Figure 2 This is a flowchart illustrating a weather forecast detection method provided in an embodiment of this application. The execution entity in this embodiment can be a processor. The processor can be implemented in software or a combination of software and hardware. Please refer to... Figure 2 The method includes:

[0122] S201. Obtain the target parameters input by the user.

[0123] The target parameters include the target area, the target time period, and at least one meteorological element.

[0124] The target area can be the location area to be detected.

[0125] Meteorological elements can include precipitation, temperature, zonal wind, radial wind, and humidity.

[0126] The target time period can be a historical period that needs to be detected.

[0127] Target parameters can be used to describe the parameters that users want to perform data analysis on.

[0128] The target parameters may include detection methods, which may include site detection and grid detection.

[0129] Site detection can be performed on a target area corresponding to a specific site.

[0130] Mesh detection can be performed on the target area corresponding to the target mesh.

[0131] An interactive interface can be displayed, which may include controls such as text boxes, selection boxes, and sliders. In response to user interaction with the target area selection control, meteorological element selection control, sliding operation of the slider corresponding to the target time period, and clicking operation of the submit button, the control values ​​corresponding to multiple controls are obtained, and the target parameters input by the user are determined based on the multiple control values.

[0132] The target parameters input by the user can be obtained through the command-line interface.

[0133] It should be noted that the target parameters input by the user can be obtained according to any feasible implementation method, and the embodiments of this application do not limit this.

[0134] S202. Determine the target element set based on at least one meteorological element.

[0135] The target element set includes multiple target elements.

[0136] Optionally, at least one meteorological element can be input into a preset model to output a set of target elements.

[0137] The preset model can be used to determine the set of target elements that best matches at least one meteorological element.

[0138] Optionally, multiple sets of candidate elements can be obtained, which include multiple elements; multiple similarities between at least one meteorological element and multiple sets of candidate elements can be determined; and the set of candidate elements with the highest similarity among the multiple similarities can be determined as the target element set.

[0139] It should be noted that the target element set can be determined according to any feasible implementation method, and the embodiments of this application do not limit this.

[0140] S203. Within the target time period of the target area, determine the original meteorological data corresponding to multiple target elements respectively.

[0141] Raw meteorological data includes any combination of observational data, forecast data, Global Precipitation Measurement Program (GPM) data, and reanalysis data.

[0142] Observational data can be physical quantities that cover the Earth's surface and near-surface atmosphere, collected by equipment such as ground weather stations and automatic weather stations.

[0143] Forecast data can be generated by predicting atmospheric conditions, weather phenomena, and related environmental parameters using meteorological models, statistical algorithms, or target algorithms corresponding to forecast models.

[0144] Global Precipitation Measurement Program (GPM) data can provide multi-source fusion global precipitation observation products by combining observation data from core satellites and their onboard radar and microwave radiometers with observation data from constellation satellites.

[0145] Reanalysis data can be a high spatiotemporal resolution meteorological dataset generated by combining observational data (such as ground observation stations, satellite observations, radiosonde data, etc.) with model simulation results.

[0146] Based on the target parameters, data can be downloaded through a preset program script to obtain the raw meteorological data corresponding to multiple target elements in the target area and target time period.

[0147] Based on the target parameters, a model can be generated through a script to determine the target script. Data can then be downloaded based on the target script to obtain the raw meteorological data corresponding to multiple target elements within the target area and target time period.

[0148] It should be noted that the original meteorological data corresponding to multiple target elements can be determined according to any feasible implementation method, and the embodiments of this application do not limit this.

[0149] S204. For any target element, perform anomaly detection processing on the original meteorological data corresponding to the target element to determine the target meteorological data corresponding to the target element.

[0150] Different anomaly detection and handling methods are used for different target elements.

[0151] The target meteorological data is the meteorological data after removing outliers.

[0152] Optionally, a first mapping relationship can be obtained, a target algorithm can be determined based on the target element and the first mapping relationship, and anomaly detection processing can be performed on the original meteorological data corresponding to the target element based on the target algorithm to determine the target meteorological data corresponding to the target element.

[0153] The first mapping relationship may include multiple meteorological elements and the algorithm corresponding to each meteorological element.

[0154] Optionally, the target element can be input into the algorithm selection model to obtain the target algorithm. Based on the target algorithm, anomaly detection processing is performed on the original meteorological data corresponding to the target element to determine the target meteorological data corresponding to the target element.

[0155] Optionally, the element type corresponding to the target element can be determined; based on the element type, the original meteorological data can be subjected to weighted difference processing to obtain a weighted difference value; based on the weighted difference value and the original meteorological data, abnormal data in the original meteorological data can be determined through a target algorithm; based on the abnormal data, the original meteorological data can be filtered to obtain the target meteorological data.

[0156] It should be noted that the target meteorological data corresponding to the target element can be determined according to any feasible implementation method, and the embodiments of this application do not limit this.

[0157] S205. Based on the target meteorological data corresponding to the multiple target elements, generate a visual verification result corresponding to the meteorological forecast.

[0158] Visualized test results can be used to visualize spatial, temporal, and corresponding multiple test scores.

[0159] Visualization results can include line charts, bar charts, scatter plots, contour plots, and color-coded plots to meet the analytical needs from different perspectives.

[0160] Based on the target meteorological data corresponding to the multiple target elements and the display method selected by the user, a visual verification result corresponding to the weather forecast can be generated.

[0161] For example, a line graph can be used to show the changing trend of the test scores (such as root mean square error, deviation, Pearson correlation coefficient, etc.) of the site test indicators with the forecast duration; a scatter plot can be used to visually present the spatial distribution deviation of the test scores of the site test indicators; and contour lines or color-coded maps can be used to represent the spatial field characteristics of the visualized test results.

[0162] Based on the target meteorological data corresponding to the multiple target elements, a visual verification result corresponding to the weather forecast can be generated through a big data model or an artificial intelligence model.

[0163] Optionally, for any target element, the target meteorological data corresponding to the target element can be classified to obtain multiple types of data, including any of the following: observation data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data; each type of data is standardized to obtain standard data corresponding to each type of data; the standard data corresponding to the multiple types of data are input into the target model to determine multiple test scores corresponding to the target element; and based on the multiple test scores corresponding to the multiple target elements, a visual test result corresponding to the meteorological forecast is generated.

[0164] Among them, classification processing can group data with the same data label into a dataset based on the data label corresponding to each data in the target meteorological data, thus obtaining a type of data.

[0165] Optionally, global and variable-level labels can be read from the target meteorological data. These labels include information such as "variable name," "latitude and longitude coordinates," "time dimension format," and "attribute description," thus forming an initial label set L. Simultaneously, a standard field set S is predefined, such as {precipitation, temperature, latitude, longitude, time}. By calculating string similarity (e.g., using Levenshtein distance) or regular expression matching, each label in the initial label set L is mapped to a corresponding standard field in the standard field set S, generating a mapping table. Classification is then performed based on this mapping table.

[0166] Optionally, if the target meteorological data is grid data of different resolutions (such as 0.1°×0.1° and 0.25°×0.25°), conservation interpolation can be used to interpolate the original field to the target grid. Subsequently, in order to ensure that the total information is not lost during the interpolation process, a two-way restoration check is performed: the target grid is mapped back to the original grid to obtain the original field interpolation, and the interpolation kernel parameters are iteratively optimized to minimize the following errors, thereby balancing conservation and error minimization.

[0167] Optionally, time interpolation can be performed on target meteorological data with different time resolutions (such as 30 minutes, 1 hour, and 3 hours): if hourly resolution is required, linear or cubic spline interpolation can be used on the original sequence.

[0168] For target meteorological data in areas with missing data, the neighborhood averaging method, regression prediction method, and spatial regression method can be used to determine the data.

[0169] The neighborhood averaging method can be filled with the average value of adjacent grid points or similar time periods.

[0170] Regression prediction methods can be used to train a regression model based on historical data from the same period and then use it for prediction and imputation.

[0171] Spatial regression can be used to estimate missing values ​​by combining multi-source data with weighted regression, thus ensuring spatial consistency.

[0172] Standardization processing can save individual meteorological elements of a pressure layer in each forecast model as a data file. The file data format is NetCDF, with a time resolution of hourly and a spatial resolution of 0.1°x0.1°.

[0173] The target model can be a learning model optimized from historical data, and there are no restrictions here.

[0174] Optionally, different test indicators may be used for different target elements.

[0175] For precipitation, the following metrics can be used for evaluation: root mean square error (RMSE), bias (BIAS), threat score (TS score), equivalent skill score (ETS score), false alarm rate (FAR), false negative rate (MAR), and probability density function (PDF).

[0176] For temperature, in the site inspection method, the root mean square error (RMSE), bias (BIAS), and Pearson correlation coefficient (PCC) can be used.

[0177] In the grid inspection method, RMSE and BIAS indices can be used for inspection, and the RMSE and BIAS of the daily maximum and minimum temperatures can also be inspected.

[0178] For wind speed, RMSE, BIAS, and PCC indicators are used for verification.

[0179] Optionally, for each test indicator (such as TS, RMSE, BIAS, PCC, etc.), the test score can be determined by linear mapping, which can be determined by the following formula:

[0180]

[0181] Where Iraw can represent the original test score, Imin can represent the minimum test score, and Imax can represent the maximum test score.

[0182] Optionally, for the rating distribution of each mode, confidence intervals are calculated using Bootstrap sampling. If the difference between the normalized scores of two models exceeds the sum of the half widths of their confidence intervals, it is considered a significant difference, providing users with a statistically supported ranking of their merits.

[0183] For example, a new sample set can be formed by sampling with replacement from the original N test samples (such as N observation-forecast data pairs). Repeat the above steps C times (C can be between 500 and 1000) to obtain the C group of RMSE values. Each time the normalized score I is calculated norm Arrange in ascending order {I norm,(1) ,...,I norm,(C)}, take the first and the The values ​​are used as the lower limit l and the upper limit u; if the confidence intervals of the two modes do not overlap, then a significant difference can be considered to exist, that is, if l A >u B This indicates that Model A is significantly better than Model B on this metric;

[0184] Where α can be a preset value, l A This can be the lower bound of pattern A, u B It can be the upper limit of mode B.

[0185] It should be noted that any feasible implementation method can be used to generate the visual verification results corresponding to the weather forecast, and the embodiments of this application do not limit this.

[0186] Optionally, after generating the visual verification results corresponding to the weather forecast, the method further includes: converting the visual verification results and target meteorological data into analytical text using natural language generation technology, and generating a graphic report by combining the visual verification results and analytical text.

[0187] The generated graphic reports can be edited and charts rearranged through an online fine-tuning interface. The system updates the graphic reports in real time, balancing automation and customization needs.

[0188] The weather forecast detection method provided in this embodiment obtains user-input target parameters, including a target area, a target time period, and at least one meteorological element; based on the at least one meteorological element, a target element set is determined, which includes multiple target elements; within the target area and target time period, the original meteorological data corresponding to each of the multiple target elements is determined; for any target element, anomaly detection processing is performed on the original meteorological data corresponding to the target element to determine the target meteorological data corresponding to the target element; based on the target meteorological data corresponding to the multiple target elements, a visualized verification result for the weather forecast is generated. In this way, by using user-input target parameters to determine a matching set of target elements, and performing anomaly detection processing based on the target elements to generate a visualized verification result, the accuracy of weather forecast detection is improved.

[0189] Below, in conjunction with Figure 3 The process of performing anomaly detection processing on the raw meteorological data corresponding to the target element for any forecast model and determining the target meteorological data corresponding to the target element (S204) is explained.

[0190] Figure 3 This is a flowchart illustrating another weather forecast detection method provided in this application embodiment. Based on the above embodiments, see also... Figure 3 The method includes:

[0191] S301. Determine the element type corresponding to the target element.

[0192] The element type can be precipitation, temperature, wind speed, humidity, etc.

[0193] The corresponding element type can be determined based on the target element.

[0194] S302. Based on the element type, perform weighted difference processing on the raw meteorological data to obtain weighted difference values.

[0195] Weighted differential processing can obtain more accurate differential values ​​by performing differential calculations on the raw meteorological data and assigning different weights according to the importance or reliability of the data points.

[0196] Optionally, the raw meteorological data can be weighted and differencing processed according to the element type to obtain the weighted difference value as follows: determine whether the element type is precipitation type; if so, determine the first factor corresponding to the observed data, the second factor corresponding to the reanalysis data, and the third factor corresponding to the GPM data, and determine the sum of the first factor, the second factor, and the third factor as the weighted difference value, where the first factor is the product of the first weight value and the first difference corresponding to the observed data, and the first difference is the difference between any data in the observed data and the first weighted average value corresponding to the observed data; if not, determine the first factor corresponding to the observed data and the second factor corresponding to the reanalysis data, and determine the sum of the first factor and the second factor as the weighted difference value.

[0197] S303. Subsampling is performed on the weighted difference values ​​and the original meteorological data to obtain multiple sample sets.

[0198] The sample set includes multiple sample data, which are either weighted difference values ​​or metadata.

[0199] Metadata can be any data point from the original meteorological data.

[0200] A certain number of sample data can be randomly selected from the weighted difference values ​​and the original meteorological data, and the set of multiple sample data can be determined as the sample set.

[0201] S304. For any given sample set, based on the multiple sample data corresponding to the sample set, establish an isolated tree for the sample set, and based on the isolated tree, calculate the path lengths corresponding to the multiple sample data in the sample set.

[0202] An isolated tree can be a binary tree structure.

[0203] During the construction process, a sample data point (such as temperature data or precipitation data) is randomly selected from the sampling set. Within the range of values ​​for this sample data, a split point is randomly determined. Based on this split point, multiple sample data points in the sampling set are divided into two parts: one part has values ​​less than the split point and enters the left subtree; the other part has values ​​greater than or equal to the split point and enters the right subtree. This process is recursively repeated until a preset stopping condition is met, such as when only one data point remains in the subset, or when the preset maximum tree depth is reached.

[0204] In the constructed isolated tree, for each sampled data in the sampling set, the number of edges traversed from the root node to the leaf node where the data is located is the path length corresponding to that sampled data.

[0205] S305. Based on the multiple path lengths of multiple sampling sets, identify the abnormal data in the original meteorological data.

[0206] Considering the characteristics of meteorological data and practical application scenarios, abnormal data in the original meteorological data are identified based on multiple path lengths of multiple sampling sets, thereby improving the accuracy of abnormal data identification.

[0207] Optionally, abnormal data in the original meteorological data can be determined based on multiple path lengths of multiple sampling sets as follows: for any metadata, determine the first path length corresponding to the metadata in multiple sampling sets respectively; determine the first average value of multiple first path lengths; determine the dynamic threshold corresponding to the metadata; if the first average value is greater than the dynamic threshold, determine that the metadata is in an abnormal state; if the first average value is less than or equal to the dynamic threshold, determine that the metadata is in a normal state; and determine the metadata in an abnormal state as abnormal data in the original meteorological data.

[0208] The dynamic threshold can be determined based on different target elements, different target areas, or different target time periods.

[0209] S306. Based on the abnormal data, the raw meteorological data is filtered and processed to obtain the target meteorological data.

[0210] For example, if there are outliers in the original temperature data series, these outliers can be deleted to make the data series more consistent with the actual temperature change trend.

[0211] The implementation details of each step in this application embodiment can be found in the description of the corresponding steps or operations in the above method embodiments; repeated content will not be repeated.

[0212] The weather forecast detection method provided in this embodiment determines the element type corresponding to the target element, performs weighted differencing on the original meteorological data according to the element type to obtain a weighted difference value, performs a subsampling process on the weighted difference value and the original meteorological data to obtain multiple sample sets, and establishes an isolation tree for any sample set based on the multiple sample data corresponding to the sample set. Based on the isolation tree, the path lengths corresponding to the multiple sample data in the sample set are calculated. Based on the multiple path lengths of the multiple sample sets, abnormal data in the original meteorological data are identified. Based on the abnormal data, the original meteorological data is filtered to obtain the target meteorological data. In this way, different anomaly detection processes are performed according to different target elements, improving the accuracy of weather forecast detection.

[0213] Figure 4 This is a schematic diagram of a weather forecast detection device provided in an embodiment of this application. Please refer to... Figure 4 The weather forecast detection device 400 includes an acquisition module 401, a first determination module 402, a second determination module 403, a verification module 404, and a display module 405, wherein...

[0214] The acquisition module 401 is used to acquire target parameters input by the user, the target parameters including target area, target time period and at least one meteorological element;

[0215] The first determining module 402 is used to determine a target element set based on the at least one meteorological element, wherein the target element set includes multiple target elements;

[0216] The second determining module 403 is used to determine the original meteorological data corresponding to the plurality of target elements respectively within the target time period in the target area;

[0217] The inspection module 404 is used to perform anomaly detection processing on the original meteorological data corresponding to any target element, and to determine the target meteorological data corresponding to the target element.

[0218] Display module 405 is used to generate visual verification results corresponding to weather forecasts based on the target meteorological data corresponding to the multiple target elements.

[0219] Optionally, in the apparatus described above, the inspection module 404 is specifically used for:

[0220] Determine the element type corresponding to the target element;

[0221] Based on the element type, the original meteorological data is subjected to weighted difference processing to obtain weighted difference values;

[0222] Based on the weighted difference value and the original meteorological data, abnormal data in the original meteorological data are determined by a target algorithm.

[0223] Based on the abnormal data, the original meteorological data is filtered to obtain the target meteorological data.

[0224] Optionally, in the apparatus described above, the inspection module 404 is specifically used for:

[0225] A subsampling process is performed on the weighted difference value and the original meteorological data to obtain multiple sample sets. Each sample set includes multiple sample data, which are either the weighted difference value or metadata. The metadata is any one of the data in the original meteorological data.

[0226] For any given sample set, based on the multiple sample data corresponding to the sample set, an isolated tree for the sample set is established, and based on the isolated tree, the path lengths corresponding to the multiple sample data in the sample set are calculated respectively;

[0227] Based on multiple path lengths of multiple sampling sets, abnormal data in the original meteorological data are identified.

[0228] Optionally, in the apparatus described above, the inspection module 404 is specifically used for:

[0229] For any metadata, determine the first path length corresponding to the metadata in multiple sampling sets respectively;

[0230] Determine the first average of multiple first path lengths;

[0231] Determine the dynamic threshold corresponding to the metadata;

[0232] If the first average value is greater than the dynamic threshold, then the metadata is determined to be in an abnormal state;

[0233] If the first average value is less than or equal to the dynamic threshold, then the metadata is determined to be in a normal state;

[0234] The metadata in the abnormal state is identified as abnormal data in the original meteorological data.

[0235] Optionally, in the apparatus described above, the raw meteorological data includes any and multiple of observational data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data, and the verification module 404 is specifically used for:

[0236] Determine whether the element type is a precipitation type;

[0237] If so, determine the first factor corresponding to the observed data, the second factor corresponding to the reanalysis data, and the third factor corresponding to the GPM data, and determine the sum of the first factor, the second factor, and the third factor as the weighted difference value, wherein the first factor is the product of the first weight value and the first difference corresponding to the observed data, and the first difference is the difference between any data in the observed data and the first weighted average value corresponding to the observed data;

[0238] If not, determine the first factor corresponding to the observed data and the second factor corresponding to the reanalysis data, and determine the sum of the first factor and the second factor as the weighted difference value.

[0239] Optionally, in the apparatus described above, the first determining module 402 is specifically used for:

[0240] Obtain a set of multiple candidate elements, wherein the set of candidate elements includes multiple elements;

[0241] Determine multiple similarities between the at least one meteorological element and a set of multiple candidate elements;

[0242] The set of candidate elements with the highest similarity among the multiple similarity values ​​is determined as the target element set.

[0243] Optionally, in the apparatus described above, the display module 405 is specifically used for:

[0244] For any target element, the target meteorological data corresponding to the target element is classified and processed to obtain multiple types of data, including any and multiple types of observation data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data.

[0245] Standardize each type of data to obtain the standard data corresponding to each type of data;

[0246] The standard data corresponding to multiple types of data are input into the target model to determine multiple test scores corresponding to the target elements;

[0247] Based on the multiple test scores corresponding to the multiple target elements, a visual test result corresponding to the weather forecast is generated.

[0248] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Please refer to... Figure 5 Electronic device 500 may include: memory 501, processor 502, and transceiver 503.

[0249] Memory 501 is used to store program instructions;

[0250] The processor 502 is used to execute the program instructions stored in the memory so that the electronic device 500 performs the above-described method.

[0251] Transceiver 503 may include a transmitter and / or a receiver. The transmitter may also be referred to as a transmitter, transmitter port, or transmitter interface, and the receiver may also be referred to as a receiver port, receiver interface, or similar descriptions. Exemplarily, memory 501, processor 502, and transceiver 503 are interconnected via bus 504.

[0252] This application also provides a computer program product that can be executed by a processor, and when the computer program product is executed, the above-described method can be implemented.

[0253] The weather forecast detection device, electronic device, computer-readable storage medium, and computer program product of this application embodiment can execute the technical solution shown in the above-described weather forecast detection method embodiment. Their implementation principles and beneficial effects are similar and will not be repeated here.

[0254] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0255] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0256] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0257] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0258] When integrated units / modules are implemented in hardware, the hardware can be digital circuits, analog circuits, etc. The physical implementation of the hardware structure includes, but is not limited to, transistors, memristors, etc. Unless otherwise specified, the processor can be any suitable hardware processor, such as a CPU, GPU, FPGA, DSP, and ASIC, etc. Unless otherwise specified, the storage unit can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc.

[0259] If the integrated unit / module is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory 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 of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0260] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as these combinations of technical features do not contradict each other, they should be considered within the scope of this specification.

[0261] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0262] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for detecting weather forecasts, characterized in that, The method includes: Obtain target parameters input by the user, the target parameters including target area, target time period and at least one meteorological element; Based on the at least one meteorological element, a target element set is determined, wherein the target element set includes multiple target elements; Within the target time period in the target area, the original meteorological data corresponding to the multiple target elements are determined, and the original meteorological data includes any multiple of observation data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data. For any target element, perform anomaly detection processing on the original meteorological data corresponding to the target element to determine the target meteorological data corresponding to the target element; Based on the target meteorological data corresponding to the multiple target elements, generate visual verification results for the meteorological forecast; Anomaly detection processing is performed on the raw meteorological data corresponding to the target element to determine the target meteorological data corresponding to the target element, including: Determine the element type corresponding to the target element; When the element type is precipitation type, the first factor corresponding to the observed data, the second factor corresponding to the reanalysis data, and the third factor corresponding to the GPM data are determined. The sum of the first factor, the second factor, and the third factor is determined as the weighted difference value. The first factor is the product of the first weight value and the first difference corresponding to the observed data. The first difference is the difference between any data in the observed data and the first weighted average value corresponding to the observed data. Based on the weighted difference value and the original meteorological data, abnormal data in the original meteorological data are determined by a target algorithm. Based on the abnormal data, the original meteorological data is filtered to obtain the target meteorological data.

2. The method according to claim 1, characterized in that, Based on the weighted difference value and the original meteorological data, anomalies in the original meteorological data are identified using a target algorithm, including: A subsampling process is performed on the weighted difference value and the original meteorological data to obtain multiple sample sets. Each sample set includes multiple sample data, which are either the weighted difference value or metadata. The metadata is any one of the data in the original meteorological data. For any given sample set, based on the multiple sample data corresponding to the sample set, an isolated tree for the sample set is established, and based on the isolated tree, the path lengths corresponding to the multiple sample data in the sample set are calculated respectively; Based on multiple path lengths of multiple sampling sets, abnormal data in the original meteorological data are identified.

3. The method according to claim 2, characterized in that, Based on multiple path lengths across multiple sampling sets, abnormal data in the original meteorological data are identified, including: For any metadata, determine the first path length corresponding to the metadata in multiple sampling sets respectively; Determine the first average of multiple first path lengths; Determine the dynamic threshold corresponding to the metadata; If the first average value is greater than the dynamic threshold, then the metadata is determined to be in an abnormal state; If the first average value is less than or equal to the dynamic threshold, then the metadata is determined to be in a normal state; Metadata in an abnormal state is identified as abnormal data in the original meteorological data.

4. The method according to any one of claims 1-3, characterized in that, The method further includes: If the element type is not the precipitation type, determine the first factor corresponding to the observation data and the second factor corresponding to the reanalysis data, and determine the sum of the first factor and the second factor as the weighted difference value.

5. The method according to any one of claims 1-3, characterized in that, Based on the at least one meteorological element, a set of target elements is determined, including: Obtain a set of multiple candidate elements, wherein the set of candidate elements includes multiple elements; Determine multiple similarities between the at least one meteorological element and a set of multiple candidate elements; The set of candidate elements with the highest similarity among the multiple similarity values ​​is determined as the target element set.

6. The method according to any one of claims 1-3, characterized in that, Based on the target meteorological data corresponding to the multiple target elements, a visual verification result corresponding to the weather forecast is generated, including: For any target element, the target meteorological data corresponding to the target element is classified and processed to obtain multiple types of data, including any and multiple types of observation data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data. Standardize each type of data to obtain the standard data corresponding to each type of data; The standard data corresponding to multiple types of data are input into the target model to determine multiple test scores corresponding to the target elements; Based on the multiple test scores corresponding to the multiple target elements, a visual test result corresponding to the weather forecast is generated.

7. A weather forecast detection device, characterized in that, The device includes: The acquisition module is used to acquire target parameters input by the user, the target parameters including target area, target time period and at least one meteorological element; The first determining module is used to determine a set of target elements based on the at least one meteorological element, wherein the set of target elements includes multiple target elements; The second determining module is used to determine the original meteorological data corresponding to the plurality of target elements in the target area during the target time period. The original meteorological data includes any multiple of observation data, forecast data, Global Precipitation Measurement Plan (GPM) data, and reanalysis data. The inspection module is used to perform anomaly detection processing on the original meteorological data corresponding to any target element, and to determine the target meteorological data corresponding to the target element. The display module is used to generate visual verification results of the weather forecast based on the target meteorological data corresponding to the multiple target elements. The second determining module is specifically used to determine the element type corresponding to the target element; When the element type is precipitation type, the first factor corresponding to the observed data, the second factor corresponding to the reanalysis data, and the third factor corresponding to the GPM data are determined. The sum of the first factor, the second factor, and the third factor is determined as the weighted difference value. The first factor is the product of the first weight value and the first difference corresponding to the observed data. The first difference is the difference between any data in the observed data and the first weighted average value corresponding to the observed data. Based on the weighted difference value and the original meteorological data, abnormal data in the original meteorological data are determined by a target algorithm. Based on the abnormal data, the original meteorological data is filtered to obtain the target meteorological data.

8. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1 to 6.

9. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1 to 6.

Images