Meteorological mode evaluation method and system based on cloud rainfall observation data

Through the meteorological mode evaluation method based on cloud precipitation observation data, the problem of unscientific meteorological mode evaluation is solved, the accuracy of precipitation prediction and the adaptability of meteorological mode are improved, and scientific evaluation results are provided to help users optimize the prediction effect.

CN119024465AInactive Publication Date: 2024-11-26HEFEI MAISUI NETWORK TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410924043.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-11
Publication Date
2024-11-26
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing technology lacks scientific and objective meteorological model evaluation methods, which leads to great uncertainty in the spatiotemporal distribution of precipitation prediction and makes it difficult to effectively prevent natural disasters.

Method used

The meteorological mode evaluation method based on cloud precipitation observation data, through the steps of storing prediction results, data collection, area division, meteorological factor prediction accuracy analysis, comprehensive performance evaluation, etc., calculate the performance comprehensive evaluation coefficient of the meteorological mode, identify and optimize the prediction model, Improve the accuracy of precipitation grade and time prediction.

Benefits of technology

Quantitative evaluation of meteorological mode performance is achieved, the accuracy of precipitation prediction and the adaptability of meteorological mode to different surface environments is improved, and scientific and objective evaluation results are provided to help users understand and improve the prediction effect.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119024465A_ABST
    Figure CN119024465A_ABST
Patent Text Reader

Abstract

The invention discloses a meteorological mode evaluation method and system based on cloud rainfall observation data, and particularly relates to the technical field of meteorological mode evaluation, and the method comprises the steps: generating a meteorological mode evaluation instruction based on the relation between a meteorological factor deviation index and a preset value; and executing a meteorological mode evaluation instruction, quantitatively evaluating the performance of the meteorological mode in the aspect of cloud rainfall prediction, and helping a user to know the prediction effect of the meteorological mode from the rainfall prediction probability, prediction time and abnormal region distribution condition. And eliminating the influence of the meteorological factor prediction accuracy index on the prediction probability spatio-temporal reliability index and the prediction time reliability index by using the meteorological factor deviation index, and obtaining a performance comprehensive evaluation coefficient of the meteorological mode based on the corrected prediction probability spatio-temporal reliability index, the prediction time reliability index and the abnormal distribution uniformity index. And the performance comprehensive evaluation coefficient is used for representing the performance of the meteorological mode, so that the problem of lack of scientific and objective meteorological mode evaluation methods in the prior art is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological model evaluation, and more specifically, to a meteorological model evaluation method and system based on cloud precipitation observation data. Background Art

[0002] Meteorological models are widely used in weather forecasting, climate prediction, environmental assessment, agricultural production and other fields. They can help people better understand the laws of weather changes, predict future weather conditions, and provide a scientific basis for various production and life activities; meteorological models are usually studied based on meteorological observation data, numerical simulation and theoretical analysis. Through statistical analysis of a large amount of meteorological data, the laws and trends of changes in various weather phenomena are revealed. In terms of numerical simulation, meteorological models simulate and predict weather changes by solving a series of atmospheric dynamic / thermodynamic equations and equations including water vapor and hydrometeors. Meteorological models provide users with forecast information on the probability of precipitation levels and the time of precipitation in a certain time and space scale in the future by analyzing and predicting the values ​​of meteorological factors related to precipitation. This information is of great significance for formulating disaster prevention and mitigation measures, agricultural production plans, and transportation planning.

[0003] However, it still has many shortcomings in actual use. For example, due to the complexity of the meteorological system, the temporal and spatial distribution of precipitation has a high degree of uncertainty, which poses a great challenge to the occurrence and prevention of natural disasters. The existing technology lacks a scientific and objective meteorological model evaluation method. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a meteorological model evaluation method and system based on cloud precipitation observation data to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a meteorological model evaluation method based on cloud precipitation observation data, comprising:

[0006] Prediction result storage: The meteorological model outputs the predicted probability of occurrence of precipitation level (d, fp) and the predicted precipitation occurrence time within the predicted time and space scale based on the predicted values ​​of precipitation-related meteorological factors; d represents the precipitation level, and fp represents the predicted probability of occurrence of precipitation level; the prediction results of the meteorological model are stored in the database;

[0007] Data collection: Collect cloud and precipitation observation data in the forecast time and space scale through ground meteorological stations, radar observations and satellite remote sensing; extract the actual precipitation occurrence time, precipitation level and precipitation related meteorological factors from the cloud and precipitation observation data;

[0008] Regional division: cloud precipitation observation data are divided into several regions according to the surface area and time; suppose it is divided into n surface regions and m time regions, i represents the number of the surface region, and j represents the number of the time region;

[0009] Analysis of meteorological factor prediction accuracy: Select atmospheric parameters that affect precipitation and mark them as precipitation-related meteorological factors. Based on the observed and predicted values ​​of precipitation-related meteorological factors, calculate the meteorological factor deviation index; if the meteorological factor deviation index Qy is within the preset range, generate a meteorological model evaluation instruction;

[0010] Meteorological model prediction performance analysis: execute meteorological model evaluation instructions, quantitatively evaluate the performance of meteorological models in cloud precipitation prediction, output the prediction probability spatiotemporal reliability index through the prediction probability accuracy analysis model; output the prediction time reliability index through the prediction time accuracy analysis model; output the anomaly distribution uniformity index through the anomaly area distribution model;

[0011] Comprehensive evaluation of meteorological model performance: The forecast probability spatiotemporal reliability index YP, the forecast time reliability index YT and the anomaly distribution uniformity index FJ are jointly analyzed to obtain the comprehensive performance evaluation coefficient of the meteorological model.

[0012] Preferably, the meteorological factor deviation index Qy is obtained as follows:

[0013] Assume that the number of precipitation-related meteorological factors is k, s represents the number of precipitation-related meteorological factors, and the sth precipitation-related meteorological factor is recorded as Q s ;

[0014] The linear coefficient between the sth precipitation-related meteorological factor and precipitation is obtained by using the Pearson correlation coefficient, and the nonlinear coefficient between the sth precipitation-related meteorological factor and precipitation is obtained by using the random forest. The linear coefficient and nonlinear coefficient of the sth precipitation-related meteorological factor are denoted as p s ,q s ; The influence coefficient of the sth precipitation-related meteorological factor on precipitation is recorded as p s Q s qs ;

[0015] The observed value of the sth precipitation-related meteorological factor is recorded as Q 观 , the predicted value of the sth precipitation-related meteorological factor is recorded as Q 预 ;

[0016] The meteorological factor deviation index Qy is calculated by the following formula:

[0017] Where Form represents the linear normalization function.

[0018] Preferably, it includes a meteorological model assessment environment judgment: if the meteorological factor offset index Qy is within a preset range, a meteorological model assessment instruction is generated; if the meteorological factor offset index Qy is not within the preset range, it indicates that the meteorological model is operating abnormally and no meteorological model assessment instruction is generated.

[0019] Preferably, the operation process of the prediction probability accuracy analysis model includes the following steps:

[0020] The actual precipitation level of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, and the predicted occurrence probability yp_ij of the precipitation level of the i-th surface area in the j-th time zone is obtained from the database;

[0021] Get the predicted precipitation of the i-th surface area in the j-th time zone, denoted as YL_ij, through the formula Calculate the first weight coefficient qaij of the i-th surface area in the j-th time zone;

[0022] By formula The spatial prediction probability accuracy index is calculated, and Ybi represents the spatial prediction accuracy index of the i-th surface area;

[0023] The actual precipitation level of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, and the predicted probability of occurrence of the precipitation level in the i-th surface area in the j-th time zone yp_ij is obtained from the database;

[0024] By formula The time prediction probability accuracy index is calculated; Ytj represents the time prediction probability accuracy index of the i-th time zone;

[0025] By formula The prediction probability spatiotemporal reliability index YP is calculated.

[0026] Preferably, the operation process of the prediction time accuracy analysis model includes the following steps:

[0027] The actual precipitation occurrence time of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, recorded as st_ij, and the predicted precipitation occurrence time of the i-th surface area in the j-th time zone is obtained from the database, recorded as yt_ij;

[0028] Get the actual precipitation of the i-th surface area in the j-th time zone, denoted as SL_ij, through the formula The second weight coefficient qbij of the i-th surface area in the j-th time zone is calculated;

[0029] By formula The prediction time reliability index YT is calculated.

[0030] Preferably, the operation process of the abnormal area distribution model includes the following steps:

[0031] Set the error range of the predicted probability of precipitation level and the predicted precipitation occurrence time, and filter out the abnormal area;

[0032] Get the abnormal probability of the jth time zone in the i-th surface area, denoted as yc_ij; through the formula Calculate the average abnormal probability yc_avg;

[0033] The distribution coefficient is expressed as the ratio of the abnormal probability of each area to the average abnormal probability, and the formula is Calculate the distribution coefficient fb_ij of each region; by formula The abnormal distribution uniformity index FJ is calculated. The smaller the abnormal distribution uniformity index is, the more uniform the distribution coefficient is.

[0034] Preferably, the prediction probability spatiotemporal reliability index is corrected based on the meteorological factor offset index Qy, by the formula YP" = YP*log2(1+Qy ) The revised forecast probability spatiotemporal reliability index YP" is obtained; the forecast time reliability index is corrected based on the meteorological factor offset index Qy, and the formula YT" = YT-e Qy The revised prediction time reliability index YT" is obtained by the formula The comprehensive performance evaluation coefficient Ynx is calculated.

[0035] Preferably, the meteorological model evaluation method based on cloud and precipitation observation data includes monitoring and early warning: taking measures based on the relationship between the prediction probability spatiotemporal reliability index YP, the prediction time reliability index YT and the anomaly distribution uniformity index FJ and the corresponding preset values; when the prediction probability spatiotemporal reliability index YP is lower than the preset value, it indicates that the prediction probability model needs to be optimized to improve the accuracy of precipitation level prediction; if YT is lower than the preset value, it indicates that the prediction time model needs to be improved to improve the accuracy of precipitation occurrence time prediction; if FJ is lower than the preset value, it indicates that it is necessary to check the impact of the surface environment in the abnormal area on the probability of occurrence of precipitation level prediction and the predicted precipitation occurrence time, so as to improve the adaptability of the meteorological model to different surface environments.

[0036] To achieve the above object, the present invention provides the following technical solution: a meteorological model evaluation system based on cloud precipitation observation data, comprising:

[0037] The prediction result storage module, based on the predicted values ​​of precipitation-related meteorological factors, outputs the predicted probability of occurrence of precipitation level (d, fp) and the predicted precipitation occurrence time within the predicted time and space scale, and stores the prediction results in the database;

[0038] The data acquisition module collects cloud precipitation observation data in the forecast time and space scale through ground meteorological stations, radar observations and satellite remote sensing; extracts the actual precipitation occurrence time, precipitation level and precipitation related meteorological factors real values ​​from the cloud precipitation observation data;

[0039] The regional division module divides the cloud precipitation observation data into several regions according to the surface area and time; suppose it is divided into n surface regions and m time regions, i represents the number of the surface region, and j represents the number of the time region;

[0040] The meteorological factor prediction accuracy analysis module selects atmospheric parameters that affect precipitation and marks them as precipitation-related meteorological factors. Based on the observed and predicted values ​​of precipitation-related meteorological factors, the meteorological factor deviation index is calculated.

[0041] The meteorological mode evaluation environment judgment module generates a meteorological mode evaluation instruction when the meteorological factor deviation index Qy is within the preset range; if the meteorological factor deviation index Qy is not within the preset range, it indicates that the meteorological mode is operating abnormally and no meteorological mode evaluation instruction is generated;

[0042] The meteorological model prediction performance analysis module quantitatively evaluates the performance of the meteorological model in cloud precipitation prediction. Through the prediction probability accuracy analysis model, it outputs the prediction probability spatiotemporal reliability index; through the prediction time accuracy analysis model, it outputs the prediction time reliability index; through the abnormal area distribution model, it outputs the abnormal distribution uniformity index;

[0043] The meteorological model performance comprehensive evaluation module jointly analyzes the prediction probability spatiotemporal reliability index YP, the prediction time reliability index YT and the anomaly distribution uniformity index FJ to obtain the comprehensive performance evaluation coefficient of the meteorological model.

[0044] Preferably, the meteorological model assessment system based on cloud precipitation observation data also includes: a data quality management module, used to ensure that the collected cloud precipitation observation data has a wide coverage area and a sufficient time range to obtain comprehensive and accurate assessment results; a data preprocessing module, used to clean the collected cloud precipitation observation data to remove duplicates, missing or outliers; and to standardize the data format to ensure the consistency and comparability of the cloud precipitation observation data.

[0045] Technical effects and advantages of the present invention:

[0046] (1) The meteorological model evaluation method based on cloud and precipitation observation data provided by the present invention, the accuracy of meteorological factor prediction is a prerequisite for the normal operation of the meteorological model, when the observed value and predicted value of the precipitation-related meteorological factor indicate that the meteorological model cannot operate normally; an early warning is given to the user, indicating that the predicted value of the meteorological factor is inaccurate, saving computing resources; executing the meteorological model evaluation instruction, quantitatively evaluating the performance of the meteorological model in cloud and precipitation prediction, helping the user understand the prediction effect of the meteorological model from the perspective of precipitation prediction probability, prediction time and abnormal area distribution, correcting the prediction probability spatiotemporal reliability index and prediction time reliability index based on the meteorological factor offset index, eliminating the influence of the meteorological factor prediction accuracy index on the prediction probability spatiotemporal reliability index and prediction time reliability index, and obtaining the performance comprehensive evaluation coefficient of the meteorological model based on the corrected prediction probability spatiotemporal reliability index, prediction time reliability index and abnormal distribution uniformity index, using the performance comprehensive evaluation coefficient to represent the performance of the meteorological model, when the performance comprehensive evaluation coefficient is within the preset range, it indicates that the meteorological model operates normally; otherwise, corresponding measures are taken to solve the problem of lack of scientific and objective meteorological model evaluation methods in the prior art.

[0047] (2) The meteorological model evaluation method based on cloud and precipitation observation data provided by the present invention can timely discover problems in the prediction probability model and optimize it when the prediction probability spatiotemporal reliability index YP is lower than the preset value, thereby improving the accuracy of precipitation level prediction; the introduction of the abnormal distribution uniformity index FJ enables the system to identify and process the impact of different surface environments on precipitation prediction; by checking the impact of the surface environment in abnormal areas on precipitation prediction, the adaptability of the meteorological model to different surface environments can be improved, making the prediction results more in line with the actual situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of the overall structure of the present invention. DETAILED DESCRIPTION

[0049] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0050] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0051] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present application, its application, or uses.

[0052] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered as part of the specification.

[0053] Example 1

[0054] The present invention provides Figure 1 The meteorological model evaluation method based on cloud and precipitation observation data shown in the figure includes:

[0055] Prediction result storage: The meteorological model outputs the predicted probability of occurrence of precipitation level and predicted precipitation occurrence time within the predicted time and space scale based on the predicted values ​​of precipitation-related meteorological factors; the prediction results of the meteorological model are stored in the database;

[0056] In the embodiment of the present invention, precipitation is divided into several precipitation levels; d represents the precipitation level, and fp represents the predicted probability of occurrence of the precipitation level; for example, the precipitation levels are divided into d1, d2, d3, d4, and d5 in descending order of precipitation; the predicted probability of occurrence of the precipitation levels includes (d1, fp1), (d2, fp2), (d3, fp3), (d4, fp4), and (d5, fp5);

[0057] Data collection: Collect cloud and precipitation observation data in the forecast time and space scale through ground meteorological stations, radar observations and satellite remote sensing; extract the actual precipitation occurrence time, precipitation level and precipitation related meteorological factors from the cloud and precipitation observation data;

[0058] Regional division: cloud precipitation observation data are divided into several regions according to the surface area and time; suppose it is divided into n surface regions and m time regions, i represents the number of the surface region, and j represents the number of the time region;

[0059] For the accuracy analysis of meteorological factor prediction, the atmospheric parameters that affect precipitation are selected and marked as precipitation-related meteorological factors. Based on the observed and predicted values ​​of precipitation-related meteorological factors, the meteorological factor deviation index is calculated.

[0060] In an embodiment of the present invention, the atmospheric parameters include atmospheric temperature, humidity, wind speed, wind direction, air pressure, water vapor content, vertical wind shear, atmospheric stability, radar reflectivity factor, radar speed and spectrum width;

[0061] Meteorological model assessment environment judgment: if the meteorological factor deviation index Qy is within the preset range, a meteorological model assessment instruction is generated; if the meteorological factor deviation index Qy is not within the preset range, it indicates that the meteorological model is operating abnormally and no meteorological model assessment instruction is generated;

[0062] Meteorological model prediction performance analysis: execute meteorological model evaluation instructions, quantitatively evaluate the performance of meteorological models in cloud precipitation prediction, output the prediction probability spatiotemporal reliability index through the prediction probability accuracy analysis model; output the prediction time reliability index through the prediction time accuracy analysis model; output the anomaly distribution uniformity index through the anomaly area distribution model;

[0063] Comprehensive evaluation of meteorological model performance: The forecast probability spatiotemporal reliability index YP, the forecast time reliability index YT and the anomaly distribution uniformity index FJ are jointly analyzed to obtain the comprehensive performance evaluation coefficient of the meteorological model.

[0064] It needs to be further explained in the embodiment of the present invention that the meteorological factor offset index Qy is obtained in the following manner:

[0065] Assume that the number of precipitation-related meteorological factors is k, s represents the number of precipitation-related meteorological factors, and the sth precipitation-related meteorological factor is recorded as Q s ;

[0066] The linear coefficient between the sth precipitation-related meteorological factor and precipitation is obtained by using the Pearson correlation coefficient, and the nonlinear coefficient between the sth precipitation-related meteorological factor and precipitation is obtained by using the random forest. The linear coefficient and nonlinear coefficient of the sth precipitation-related meteorological factor are denoted as p s ,q s ; The influence coefficient of the sth precipitation-related meteorological factor on precipitation is recorded as p s Q s qs ;

[0067] The observed value of the sth precipitation-related meteorological factor is recorded as Q 观 , the predicted value of the sth precipitation-related meteorological factor is recorded as Q 预 ;

[0068] The meteorological factor deviation index Qy is calculated by the following formula:

[0069] Where Form represents the linear normalization function.

[0070] It needs to be further explained in the embodiment of the present invention that the operation process of the prediction probability accuracy analysis model includes the following steps:

[0071] The actual precipitation level of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, and the predicted occurrence probability yp_ij of the precipitation level of the i-th surface area in the j-th time zone is obtained from the database;

[0072] Get the predicted precipitation of the i-th surface area in the j-th time zone, denoted as YL_ij, through the formula Calculate the first weight coefficient qaij of the i-th surface area in the j-th time zone;

[0073] By formula The spatial prediction probability accuracy index is calculated, and Ybi represents the spatial prediction accuracy index of the i-th surface area;

[0074] The actual precipitation level of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, and the predicted probability of occurrence of the precipitation level in the i-th surface area in the j-th time zone yp_ij is obtained from the database;

[0075] By formula The time prediction probability accuracy index is calculated; Ytj represents the time prediction probability accuracy index of the jth time zone;

[0076] By formula The prediction probability spatiotemporal reliability index YP is calculated.

[0077] It needs to be further explained in the embodiment of the present invention that the operation process of the prediction time accuracy analysis model includes the following steps:

[0078] The actual precipitation occurrence time of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, recorded as st_ij, and the predicted precipitation occurrence time of the i-th surface area in the j-th time zone is obtained from the database, recorded as yt_ij;

[0079] Get the actual precipitation of the i-th surface area in the j-th time zone, denoted as SL_ij, through the formula The second weight coefficient qbij of the i-th surface area in the j-th time zone is calculated;

[0080] By formula The prediction time reliability index YT is calculated.

[0081] It needs to be further explained in the embodiment of the present invention that the operation process of the abnormal area distribution model includes the following steps:

[0082] Set the error range of the predicted probability of precipitation level and the predicted precipitation occurrence time, and filter out the abnormal area;

[0083] Get the abnormal probability of the jth time zone in the i-th surface area, denoted as yc_ij; through the formula Calculate the average abnormal probability yc_avg;

[0084] The distribution coefficient is expressed as the ratio of the abnormal probability of each area to the average abnormal probability, and the formula is Calculate the distribution coefficient fb_ij of each region; by formula The abnormal distribution uniformity index FJ is calculated. The smaller the abnormal distribution uniformity index is, the more uniform the distribution coefficient is.

[0085] It is necessary to further explain in the embodiment of the present invention that the prediction probability spatiotemporal reliability index is corrected based on the meteorological factor offset index Qy, and the corrected prediction probability spatiotemporal reliability index YP" is obtained by the formula YP"=YP*log2(1+Qy); the prediction time reliability index is corrected based on the meteorological factor offset index Qy, and the corrected prediction probability spatiotemporal reliability index YP" is obtained by the formula YT"=YT*e Qy The revised prediction time reliability index YT" is obtained by the formula The comprehensive performance evaluation coefficient Ynx is calculated.

[0086] What needs to be further explained in the embodiments of the present invention is that the meteorological model evaluation method based on cloud precipitation observation data includes monitoring and early warning: taking measures based on the relationship between the prediction probability spatiotemporal reliability index YP, the prediction time reliability index YT and the abnormal distribution uniformity index FJ and the corresponding preset values; when the prediction probability spatiotemporal reliability index YP is lower than the preset value, it indicates that the prediction probability model needs to be optimized to improve the accuracy of precipitation level prediction; if YT is lower than the preset value, it indicates that the prediction time model needs to be improved to improve the accuracy of precipitation occurrence time prediction; if FJ is lower than the preset value, it indicates that it is necessary to check the influence of the surface environment in the abnormal area on the probability of occurrence of precipitation level prediction and the predicted precipitation occurrence time, so as to improve the adaptability of the meteorological model to different surface environments.

[0087] Example 2

[0088] The embodiment of the present invention provides a meteorological model evaluation system based on cloud precipitation observation data, including:

[0089] The prediction result storage module, based on the predicted values ​​of precipitation-related meteorological factors, outputs the predicted probability of occurrence of precipitation level (d, fp) and the predicted precipitation occurrence time within the predicted time and space scale, and stores the prediction results in the database;

[0090] The data acquisition module collects cloud precipitation observation data in the forecast time and space scale through ground meteorological stations, radar observations and satellite remote sensing; extracts the actual precipitation occurrence time, precipitation level and precipitation related meteorological factors real values ​​from the cloud precipitation observation data;

[0091] The regional division module divides the cloud precipitation observation data into several regions according to the surface area and time; suppose it is divided into n surface regions and m time regions, i represents the number of the surface region, and j represents the number of the time region;

[0092] The meteorological factor prediction accuracy analysis module selects atmospheric parameters that affect precipitation and marks them as precipitation-related meteorological factors. Based on the observed and predicted values ​​of precipitation-related meteorological factors, the meteorological factor deviation index is calculated.

[0093] The meteorological mode evaluation environment judgment module generates a meteorological mode evaluation instruction when the meteorological factor deviation index Qy is within the preset range; if the meteorological factor deviation index Qy is not within the preset range, it indicates that the meteorological mode is operating abnormally and no meteorological mode evaluation instruction is generated;

[0094] The meteorological model prediction performance analysis module quantitatively evaluates the performance of the meteorological model in cloud precipitation prediction. Through the prediction probability accuracy analysis model, it outputs the prediction probability spatiotemporal reliability index; through the prediction time accuracy analysis model, it outputs the prediction time reliability index; through the abnormal area distribution model, it outputs the abnormal distribution uniformity index;

[0095] The meteorological model performance comprehensive evaluation module jointly analyzes the prediction probability spatiotemporal reliability index YP, the prediction time reliability index YT and the anomaly distribution uniformity index FJ to obtain the comprehensive performance evaluation coefficient of the meteorological model.

[0096] What needs to be further explained in the embodiment of the present invention is that the meteorological model evaluation system based on cloud precipitation observation data also includes: a data quality management module, which is used to ensure that the collected cloud precipitation observation data has a wide coverage area and a sufficient time range to obtain comprehensive and accurate evaluation results; a data preprocessing module, which is used to clean the collected cloud precipitation observation data to remove duplicates, missing or outliers; and to standardize the data format to ensure the consistency and comparability of the cloud precipitation observation data.

[0097] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A meteorological model evaluation method based on cloud and precipitation observation data, characterized in that: include: Prediction result storage: The meteorological model outputs the predicted probability of occurrence of precipitation level and predicted precipitation occurrence time within the predicted time and space scale based on the predicted values ​​of precipitation-related meteorological factors; storing the forecast results of the weather model in a database; Data collection: Collect cloud and precipitation observation data in the forecast time and space scale through ground meteorological stations, radar observations and satellite remote sensing; extract the actual precipitation occurrence time, precipitation level and precipitation related meteorological factors from the cloud and precipitation observation data; Regional division: cloud precipitation observation data are divided into several regions according to the surface area and time; suppose it is divided into n surface regions and m time regions, i represents the number of the surface region, and j represents the number of the time region; Analysis of meteorological factor prediction accuracy: Select atmospheric parameters that affect precipitation and mark them as precipitation-related meteorological factors. Based on the observed and predicted values ​​of precipitation-related meteorological factors, calculate the meteorological factor deviation index; if the meteorological factor deviation index Qy is within the preset range, generate a meteorological model evaluation instruction; Meteorological model prediction performance analysis: execute meteorological model evaluation instructions, quantitatively evaluate the performance of meteorological models in cloud precipitation prediction, output the prediction probability spatiotemporal reliability index through the prediction probability accuracy analysis model; output the prediction time reliability index through the prediction time accuracy analysis model; output the anomaly distribution uniformity index through the anomaly area distribution model; Comprehensive evaluation of meteorological model performance: The forecast probability spatiotemporal reliability index YP, the forecast time reliability index YT and the anomaly distribution uniformity index FJ are jointly analyzed to obtain the comprehensive performance evaluation coefficient of the meteorological model.

2. The meteorological model evaluation method based on cloud precipitation observation data according to claim 1 is characterized in that: The meteorological factor deviation index Qy is obtained as follows: Assume that the number of precipitation-related meteorological factors is k, s represents the number of precipitation-related meteorological factors, and the sth precipitation-related meteorological factor is recorded as Q s ; The linear coefficient between the sth precipitation-related meteorological factor and precipitation is obtained by using the Pearson correlation coefficient, and the nonlinear coefficient between the sth precipitation-related meteorological factor and precipitation is obtained by using the random forest. The linear coefficient and nonlinear coefficient of the sth precipitation-related meteorological factor are denoted as p s ,q s ; The influence coefficient of the sth precipitation-related meteorological factor on precipitation is recorded as p s Q s qs ; The observed value of the sth precipitation-related meteorological factor is recorded as Q 观 , the predicted value of the sth precipitation-related meteorological factor is recorded as Q 预 ; The meteorological factor deviation index Qy is calculated by the following formula: Where Form represents the linear normalization function.

3. The meteorological model evaluation method based on cloud precipitation observation data according to claim 2 is characterized in that: Including meteorological model assessment environment judgment: if the meteorological factor offset index Qy is within the preset range, a meteorological model assessment instruction is generated; if the meteorological factor offset index Qy is not within the preset range, it indicates that the meteorological model is operating abnormally and no meteorological model assessment instruction is generated.

4. The meteorological model evaluation method based on cloud precipitation observation data according to claim 3 is characterized in that: The operation process of the prediction probability accuracy analysis model includes the following steps: The actual precipitation level of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, and the predicted occurrence probability yp_ij of the precipitation level of the i-th surface area in the j-th time zone is obtained from the database; By formula The spatial prediction probability accuracy index is calculated, Ybi represents the spatial prediction accuracy index of the i-th surface area; qaij represents the first weight coefficient qaij; The actual precipitation level of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, and the predicted probability of occurrence of the precipitation level in the i-th surface area in the j-th time zone yp_ij is obtained from the database; By formula The time prediction probability accuracy index is calculated; Ytj represents the time prediction probability accuracy index of the i-th time zone; By formula The prediction probability spatiotemporal reliability index YP is calculated.

5. The meteorological model evaluation method based on cloud precipitation observation data according to claim 4 is characterized in that: The operation process of the prediction time accuracy analysis model includes the following steps: The actual precipitation occurrence time of the i-th surface area in the j-th time zone is obtained from the cloud precipitation observation data, recorded as st_ij, and the predicted precipitation occurrence time of the i-th surface area in the j-th time zone is obtained from the database, recorded as yt_ij; By formula The prediction time reliability index YT is calculated, and qbij represents the second weight coefficient.

6. The meteorological model evaluation method based on cloud precipitation observation data according to claim 5 is characterized in that: The first weight coefficient is obtained by obtaining the predicted precipitation of the i-th surface area in the j-th time zone, recorded as YL_ij, through the formula Calculate the first weight coefficient qaij of the i-th surface area in the j-th time zone; The second weight coefficient is obtained by obtaining the actual precipitation of the i-th surface area in the j-th time zone, recorded as SL_ij, through the formula The second weight coefficient qbij of the i-th surface area in the j-th time zone is calculated.

7. The meteorological model evaluation method based on cloud precipitation observation data according to claim 1 is characterized in that: The operation process of the abnormal area distribution model includes the following steps: Set the error range of the predicted probability of precipitation level and the predicted precipitation occurrence time, and filter out the abnormal area; Get the abnormal probability of the jth time zone in the i-th surface area, denoted as yc_ij; through the formula Calculate the average abnormal probability yc_avg; The distribution coefficient is expressed as the ratio of the abnormal probability of each area to the average abnormal probability, and the formula is Calculate the distribution coefficient fb_ij of each region; by formula The abnormal distribution uniformity index FJ is calculated. The smaller the abnormal distribution uniformity index is, the more uniform the distribution coefficient is.

8. The meteorological model evaluation method based on cloud precipitation observation data according to claim 1 is characterized in that: Based on the meteorological factor offset index Qy, the prediction probability spatiotemporal reliability index is corrected, and the corrected prediction probability spatiotemporal reliability index YP" is obtained through the formula YP"=YP*log2(1+Qy); based on the meteorological factor offset index Qy, the prediction time reliability index is corrected through the formula YT"=YT*e Qy The revised prediction time reliability index YT" is obtained by the formula The comprehensive performance evaluation coefficient Ynx is calculated.

9. A meteorological model evaluation system based on cloud and precipitation observation data, characterized in that: include: The prediction result storage module, based on the predicted values ​​of precipitation-related meteorological factors, outputs the predicted probability of occurrence of precipitation level (d, fp) and the predicted precipitation occurrence time within the predicted time and space scale, and stores the prediction results in the database; The data acquisition module collects cloud precipitation observation data in the forecast time and space scale through ground meteorological stations, radar observations and satellite remote sensing; extracts the actual precipitation occurrence time, precipitation level and precipitation related meteorological factors real values ​​from the cloud precipitation observation data; The regional division module divides the cloud precipitation observation data into several regions according to the surface area and time; suppose it is divided into n surface regions and m time regions, i represents the number of the surface region, and j represents the number of the time region; The meteorological factor prediction accuracy analysis module selects atmospheric parameters that affect precipitation and marks them as precipitation-related meteorological factors. Based on the observed and predicted values ​​of precipitation-related meteorological factors, the meteorological factor deviation index is calculated. The meteorological mode evaluation environment judgment module generates a meteorological mode evaluation instruction when the meteorological factor deviation index Qy is within the preset range; The meteorological factor deviation index Qy is not within the preset range, indicating that the meteorological model is operating abnormally and no meteorological model evaluation instruction is generated; The meteorological model prediction performance analysis module quantitatively evaluates the performance of the meteorological model in cloud precipitation prediction. Through the prediction probability accuracy analysis model, it outputs the prediction probability spatiotemporal reliability index; through the prediction time accuracy analysis model, it outputs the prediction time reliability index; through the abnormal area distribution model, it outputs the abnormal distribution uniformity index; The meteorological model performance comprehensive evaluation module jointly analyzes the prediction probability spatiotemporal reliability index YP, the prediction time reliability index YT and the anomaly distribution uniformity index FJ to obtain the comprehensive performance evaluation coefficient of the meteorological model.

10. The meteorological model evaluation system based on cloud precipitation observation data according to claim 9, characterized in that: Also includes: A data quality management module to ensure that the cloud and precipitation observations collected have a wide coverage area and sufficient time frame to obtain comprehensive and accurate assessment results; The data preprocessing module is used to clean the collected cloud precipitation observation data, remove duplicate, missing or outlier values, and standardize the data format to ensure the consistency and comparability of the cloud precipitation observation data.