Winter rainfall phase state diagnosis method and system based on geostationary satellite temperature and humidity profile

Abstract

The invention discloses a winter rainfall phase state diagnosis method and system based on a geostationary satellite temperature and humidity profile, and the method comprises the steps: obtaining the atmosphere temperature and humidity profile data of a geostationary meteorological satellite, carrying out the preprocessing, and generating a basic data set; obtaining winter rainfall phase state data of the ground station, and performing classification preprocessing to generate a winter rainfall phase state data set; performing matching and partitioning processing on the basis of the basic data set and the winter rainfall phase state data to generate a matched and partitioned data set; constructing a diagnosis model based on a preset key meteorological characteristic quantity and the data set after matching and partitioning; and acquiring real-time atmosphere temperature and humidity profile data of the stationary meteorological satellite, inputting the real-time atmosphere temperature and humidity profile data into the diagnosis model, performing precipitation phase state diagnosis, and generating a precipitation phase state diagnosis result. Therefore, according to the winter rainfall phase state diagnosis method based on the stationary satellite temperature and humidity profile, the timeliness and accuracy of winter rainfall phase state monitoring are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of meteorological forecasting, and in particular to a winter precipitation phase diagnosis method and system based on geostationary satellite temperature and humidity profiles. Background Art

[0002] In winter, it is very important to correctly monitor the precipitation phase (rain, snow, freezing rain, etc.) in winter because it is directly related to various fields such as traffic safety, power supply, urban management and agricultural generation. Certain types of precipitation, such as freezing rain, will seriously affect transportation, power supply lines and crop frost damage in winter. Snowfall is more harmful to road safety and agricultural production than rainfall. Therefore, improving the accuracy of winter precipitation phase monitoring can not only reduce the economic losses caused by freezing rain or snow disasters, but also ensure the normal operation of society and the safety of people's lives and property.

[0003] At present, the forecast of precipitation phase mainly relies on numerical models and sounding data. However, there are still major problems in the forecast of complex precipitation phase in winter. Especially when the cold air invades in winter, the temperature and humidity environmental conditions will change rapidly. However, the model cannot reproduce the rapid changes of temperature and humidity due to the limitations of initial conditions and physical processes, which makes the forecast of precipitation phase obvious. At the same time, the sounding data is limited by the layout of its stations (there are only 82 sounding stations in my country, and stations cannot be deployed in mountainous areas and oceans) and the observation time (twice a day, at 8 am and 8 pm), resulting in the forecast of precipitation phase cannot be corrected in time through sounding data. Summary of the invention

[0004] The embodiment of the present invention provides a winter precipitation phase state diagnosis method and system based on geostationary satellite temperature and humidity profiles, which improves the timeliness and accuracy of winter precipitation phase state monitoring.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a winter precipitation phase diagnosis method based on geostationary satellite temperature and humidity profiles, comprising:

[0006] Obtain atmospheric temperature and humidity profile data from geostationary meteorological satellites, perform preprocessing, and generate basic data sets;

[0007] Obtain winter precipitation phase data from ground stations, perform classification preprocessing, and generate a winter precipitation phase data set;

[0008] Perform matching and partitioning processing based on the basic data set and the winter precipitation phase data to generate a matched and partitioned data set;

[0009] Constructing a diagnostic model based on preset key meteorological characteristics and the matched partitioned data set;

[0010] The real-time atmospheric temperature and humidity profile data and the ground station observation data of the geostationary meteorological satellite are obtained, and the real-time atmospheric temperature and humidity profile data and the ground station observation data are input into the diagnostic model to diagnose the precipitation phase and generate the precipitation phase diagnosis result.

[0011] In one embodiment of the present invention, the acquisition of atmospheric temperature and humidity profile data from a geostationary meteorological satellite and preprocessing to generate a basic data set includes:

[0012] Acquiring the atmospheric temperature and humidity profile data of the geostationary meteorological satellite through an infrared hyperspectral instrument, wherein the atmospheric temperature and humidity profile data carries a quality identification code;

[0013] Classifying the atmospheric temperature and humidity profile data based on the quality identification code to generate classified data;

[0014] Selection is performed based on the classified data to generate the basic data set.

[0015] In one embodiment of the present invention, the step of obtaining winter precipitation phase data from ground stations and performing classification preprocessing to generate a winter precipitation phase data set includes:

[0016] Acquiring the winter precipitation phase data through the ground station, wherein the winter precipitation phase data includes coding information;

[0017] The winter precipitation phase state data is classified according to the preset winter precipitation phase state parameters to generate the winter precipitation phase state data set.

[0018] In one embodiment of the present invention, the matching and partitioning processing is performed based on the basic data set and the winter precipitation phase data to generate a matched partitioned data set, which includes:

[0019] Taking the longitude and latitude of the winter precipitation phase data as the standard point, the basic data set and the winter precipitation phase data are matched one by one within a preset search radius to generate a matching data set, wherein the preset search radius is 12 km;

[0020] The matching data set is partitioned according to preset regions to generate the matching partitioned data set, wherein the preset regions include Northeast China, North China, Northwest China, Southwest China, the middle and lower reaches of the Yangtze River Plain, South China and the Qinghai-Tibet Plateau.

[0021] In one embodiment of the present invention, the construction of the diagnostic model based on the preset key meteorological feature and the matched partitioned data set includes:

[0022] Obtaining the preset key meteorological characteristic quantities, wherein the preset key meteorological characteristic quantities include temperature, temperature gradient, humidity, humidity gradient, pseudo-equivalent potential temperature, entropy map area, longitude and latitude information, 0°C temperature critical layer thickness, melting layer thickness, freezing layer thickness, 2-meter air temperature and surface temperature at different atmospheric altitude layers;

[0023] Integrate the preset key meteorological feature and the matched partitioned data set to generate a complete data set;

[0024] Based on the complete data set, the diagnostic model is constructed.

[0025] In one embodiment of the present invention, the step of acquiring real-time atmospheric temperature and humidity profile data and ground station observation data from a geostationary meteorological satellite, and inputting the real-time atmospheric temperature and humidity profile data and the ground station observation data into the diagnostic model to perform a diagnosis of precipitation phase, and generating a precipitation phase diagnosis result comprises:

[0026] Acquire the real-time atmospheric temperature and humidity profile data of the geostationary meteorological satellite and the ground station observation data, and extract key meteorological characteristic quantities;

[0027] The key meteorological characteristic quantities are input into the diagnostic model to generate precipitation phase diagnosis results.

[0028] In a second aspect, the present invention provides a winter precipitation phase diagnosis system based on geostationary satellite temperature and humidity profiles, comprising: a first generation module, a second generation module, a third generation module, a construction module and a fourth generation module. The first generation module is used to obtain the atmospheric temperature and humidity profile data of the geostationary meteorological satellite, and perform preprocessing to generate a basic data set. The second generation module is used to obtain the winter precipitation phase data of the ground station, and perform classification preprocessing to generate a winter precipitation phase data set. The third generation module is used to match and partition the basic data set and the winter precipitation phase data to generate a matching partitioned data set. The construction module is used to construct a diagnosis model based on preset key meteorological feature quantities and the matching partitioned data set. The fourth generation module is used to obtain the real-time atmospheric temperature and humidity profile data of the geostationary meteorological satellite and the ground station observation data, and input the real-time atmospheric temperature and humidity profile data and the ground station observation data into the diagnosis model to diagnose the precipitation phase and generate a precipitation phase diagnosis result.

[0029] In one embodiment of the present invention, the first generation module includes: a first acquisition unit, a first generation unit and a second generation unit. The first acquisition unit is used to acquire the atmospheric temperature and humidity profile data of the geostationary meteorological satellite through an infrared hyperspectral instrument, wherein the atmospheric temperature and humidity profile data carries a quality identification code. The first generation unit is used to classify the atmospheric temperature and humidity profile data based on the quality identification code to generate classified data. And the second generation unit is used to select based on the classified data to generate the basic data set.

[0030] In a third aspect, the present invention provides an electronic device, comprising:

[0031] at least one processor; and

[0032] a memory communicatively coupled to the at least one processor;

[0033] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the winter precipitation phase diagnosis method based on the geostationary satellite temperature and humidity profile as described above.

[0034] In a fourth aspect, the present invention provides a computer-readable storage medium comprising a computer program and instructions. When the computer program or the instructions are executed on a computer, the computer executes the winter precipitation phase diagnosis method based on the geostationary satellite temperature and humidity profile as described above.

[0035] Compared with the prior art, the winter precipitation phase diagnosis method and system based on the geostationary satellite temperature and humidity profile according to the present invention has the following beneficial effects:

[0036] 1. The present invention utilizes high temporal and spatial resolution temperature and humidity profile data from geostationary meteorological satellites. These data have a temporal resolution of once every 2 hours and a spatial resolution of 12 km, and can more accurately capture subtle changes in the atmosphere. Such high-resolution data significantly improves the accuracy of monitoring precipitation phases (such as rain, snow, freezing rain, etc.), which is particularly critical for coping with complex and rapidly changing winter precipitation phases.

[0037] 2. In data processing and model training, the present invention comprehensively considers the climate characteristics and altitudes of different geographical regions, and divides my country into seven regions for regional processing. This regional processing method can better capture and reflect the meteorological characteristics of different regions, thereby further improving the accuracy and pertinence of precipitation phase monitoring;

[0038] 3. The present invention uses XGBoost, an advanced machine learning algorithm, to construct a multi-level thermodynamic model of the atmosphere. Compared with traditional statistical methods, XGBoost can process more complex data features and provide more accurate labeled monitoring, which significantly improves the effect of winter precipitation phase monitoring and makes the monitoring results more reliable and accurate.

[0039] 4. Compared with sounding data, the data provided by geostationary meteorological satellites have a wide coverage and a large sample size, and can provide a richer data set for model training; this helps the model learn more meteorological characteristics and laws and improve the accuracy of monitoring; at the same time, geostationary meteorological satellite data can be obtained in real time, which further improves the timeliness and practicality of monitoring, and is of great significance for responding to winter rain, snow and freezing weather events;

[0040] 5. The present invention has broad application prospects and significant practical value. Accurate precipitation phase monitoring is crucial to various fields such as traffic safety, power supply, urban management and agricultural production. By improving the accuracy of monitoring, it can effectively reduce the economic losses caused by freezing rain, snowstorms and other weather conditions, and ensure the normal operation of society and the safety of people's lives and property. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flow chart of a winter precipitation phase state diagnosis method based on geostationary satellite temperature and humidity profiles in Embodiment 1 of the present invention;

[0042] Figure 2 1 is a schematic diagram of the structure of a winter precipitation phase diagnosis system based on geostationary satellite temperature and humidity profiles in Embodiment 2 of the present invention;

[0043] Figure 3 It is a structural schematic diagram of an electronic device in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0044] The embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention, rather than to limit the embodiments of the present invention. It is also necessary to explain that, for ease of description, only parts related to the embodiments of the present invention are shown in the accompanying drawings, rather than all structures.

[0045] For ease of understanding, the main implementation concepts of the embodiments of the present invention are first briefly described.

[0046] Geostationary meteorological satellites are able to provide full-range, high-frequency observations, and their high-temporal and spatial resolution temperature and humidity profile products are essential for accurate monitoring of precipitation phases. High-temporal and spatial resolution temperature and humidity profile products can provide detailed information on the vertical structure of the atmosphere, helping to identify key temperature and humidity layers that may lead to changes in precipitation types. For example, the formation of freezing rain requires a specific temperature structure, which usually occurs when the near-ground temperature is below zero but there is a melt layer in the upper atmosphere. By analyzing the temperature and humidity profiles, these conditions can be identified in advance, thereby improving the forecast accuracy of dangerous weather phenomena such as freezing rain.

[0047] The winter precipitation phase state diagnosis method and system based on the geostationary satellite temperature and humidity profile provided by the present invention are mainly used in the field of meteorological forecasting and services, especially in the environment where rain, snow and freezing weather occur frequently in winter. By acquiring the high temporal and spatial resolution temperature and humidity profile data of the geostationary meteorological satellite, combining the precipitation phase state data of the ground station, and using the advanced XGBoost algorithm to build a diagnostic model, accurate monitoring of the winter precipitation phase state (such as rain, snow, freezing rain, etc.) can be achieved.

[0048] In practical applications, this method can be widely used in meteorological departments at all levels, traffic management departments, power supply companies, agricultural management departments and other fields. The meteorological department can use this diagnostic system to issue accurate precipitation phase forecasts in advance, guide the public to arrange travel reasonably, and reduce the inconvenience and losses caused by bad weather. The traffic management department can take measures such as road snow removal and ice melting in advance according to the diagnosis results to ensure traffic safety and smoothness. Power supply companies can strengthen the protection of power facilities based on precipitation phase monitoring to prevent power outages caused by freezing rain or snow accumulation. The agricultural management department can guide farmers to take corresponding agricultural management measures based on the diagnosis results to reduce the risk of crop frost damage.

[0049] In addition, the system can also provide a scientific basis for urban emergency response, helping city managers to make quick decisions under winter rain, snow and freezing weather conditions to ensure the normal operation of the city and the safety of residents' lives and property. By acquiring geostationary meteorological satellite data in real time, the system can dynamically update diagnostic results, ensure the timeliness and accuracy of monitoring, and provide strong support for responding to winter disaster weather events.

[0050] Embodiment 1, Figure 1 is a flow chart of a winter precipitation phase state diagnosis method based on geostationary satellite temperature and humidity profiles in Embodiment 1 of the present invention, such as Figure 1 As shown, embodiment 1 provides a winter precipitation phase diagnosis method based on geostationary satellite temperature and humidity profiles, including:

[0051] Step S100, obtaining atmospheric temperature and humidity profile data from a geostationary meteorological satellite, and performing preprocessing to generate a basic data set;

[0052] Specifically, high temporal and spatial resolution atmospheric temperature and humidity profile data provided by geostationary meteorological satellites are obtained through infrared hyperspectral instruments, and classified and screened according to data quality identification codes to generate a basic data set for model training.

[0053] Step S200, obtaining winter precipitation phase data from ground stations, and performing classification preprocessing to generate a winter precipitation phase data set;

[0054] Specifically, winter precipitation phase data containing coded information are obtained through ground stations, and these data are classified and processed according to preset precipitation phase parameters to generate a winter precipitation phase data set for model training.

[0055] Step S300, performing matching and partitioning processing based on the basic data set and the winter precipitation phase data to generate a matched and partitioned data set;

[0056] Specifically, the preprocessed basic data set and the winter precipitation phase data after classification preprocessing were matched one by one within a search radius of 12 km, with the longitude and latitude of the ground station data as standard points. The matched data sets were divided into seven regions according to my country's natural geographical divisions: Northeast China, North China, Northwest China, Southwest China, the middle and lower reaches of the Yangtze River Plain, South China, and the Qinghai-Tibet Plateau, to generate a matched partitioned data set.

[0057] Step S400, constructing a diagnosis model based on preset key meteorological features and the matched partitioned data set;

[0058] Specifically, based on the pre-set key meteorological characteristics (such as temperature, humidity, temperature gradient, humidity gradient, etc.) and the data set after matching partitioning processing, the advanced XGBoost algorithm is used to construct an atmospheric multi-level thermodynamic diagnostic model for diagnosing the phase state of winter precipitation.

[0059] Step S500, obtaining real-time atmospheric temperature and humidity profile data and ground station observation data from the geostationary meteorological satellite, and inputting the real-time atmospheric temperature and humidity profile data and the ground station observation data into the diagnosis model to diagnose the precipitation phase state and generate a precipitation phase state diagnosis result;

[0060] Specifically, the latest real-time atmospheric temperature and humidity profile data are obtained from geostationary meteorological satellites, and real-time observation data are obtained from ground stations. After extracting key meteorological characteristics, they are input into the constructed diagnostic model, and the diagnostic results of the current winter precipitation phase (such as rain, snow, freezing rain, etc.) are generated through model calculation.

[0061] In this embodiment, the step S100 includes:

[0062] Step S101, acquiring the atmospheric temperature and humidity profile data of the geostationary meteorological satellite through an infrared hyperspectral instrument, wherein the atmospheric temperature and humidity profile data carries a quality identification code;

[0063] Specifically, infrared hyperspectral instruments are used to obtain atmospheric temperature and humidity profile data with high temporal and spatial resolution from geostationary meteorological satellites. These data not only contain detailed information on the vertical structure of the atmosphere, but also come with quality identification codes indicating the accuracy of the data inversion, facilitating subsequent data classification and screening.

[0064] In a specific example, assuming that the current time is a certain day in winter, the latest atmospheric temperature and humidity profile data is needed to diagnose the upcoming precipitation phase. First, the atmospheric temperature and humidity profile data is obtained through the infrared hyperspectral instrument (GIIRS) on the Fengyun geostationary meteorological satellite. These data are updated at intervals of 2 hours, with a spatial resolution of up to 12 kilometers, and the vertical level is subdivided into 101 layers, from 1000hPa to 0.005hPa, covering the main temperature and humidity change areas of the atmosphere. In the acquired data, each layer of data is accompanied by a quality identification code, which is used to indicate the inversion accuracy of the data. For example, data with a quality identification code of 0 indicates perfect data with the highest accuracy; data with a quality identification code of 1 indicates high-quality data with high accuracy; and data with quality identification codes of 2 and 3 indicate poor quality and may have certain errors. After obtaining these data, they will first be classified according to the quality identification code, and the data with quality identification codes 0 and 1 will be selected as the basic data set, because these data have higher accuracy and reliability, and can more accurately reflect the changes in temperature and humidity in the atmosphere, thus providing more reliable data support for subsequent precipitation phase diagnosis. Through this step, the quality of the data used can be ensured, and the accuracy and reliability of precipitation phase monitoring can be improved.

[0065] Step S102, classifying the atmospheric temperature and humidity profile data based on the quality identification code to generate classified data;

[0066] Specifically, the acquired atmospheric temperature and humidity profile data are divided into different accuracy levels (such as perfect data, high-quality data, and poor-quality data) based on the quality identification code, so that high-precision data can be selected to generate the basic data set later. For example, the quality identification code of each layer of data of each profile is first checked (a vertical profile contains 101 layers, each layer has a quality identification code, so there are 101 quality identification codes on a profile), and the data are divided into three categories according to these identification codes: Perfect data (quality identification code is 0): This type of data is considered to be the most accurate and reliable and suitable for high-precision analysis. High-quality data (quality identification code is 1): Although the accuracy is slightly lower than that of perfect data, it is still good enough to be used for most analysis tasks. Poor quality data (quality identification codes are 2 and 3): This type of data may have large errors, so it will be excluded in subsequent analysis.

[0067] Step S103, selecting based on the classified data to generate the basic data set;

[0068] Specifically, from the classified data, select the atmospheric temperature and humidity profile data with quality identification codes of 0 and 1 (i.e., perfect data and high-quality data) as the basic data set for building the diagnostic model to ensure the high accuracy and reliability of the data. This step requires selecting data suitable for building the diagnostic model from these classified data. Considering that the precipitation phase diagnosis has very high requirements for data accuracy, only the data with quality identification codes of 0 and 1 are selected because these data represent the highest and second highest accuracy levels. Specifically: Perfect data (quality identification code is 0): This type of data has the highest accuracy and reliability and is the first choice when building the model. High-quality data (quality identification code is 1): Although the accuracy is slightly lower than that of perfect data, it is still reliable enough to be used for model training. Therefore, in step S103, data entries with quality identification codes of 0 and 1 will be screened out from all classified data, and these data will be combined to form a basic data set. This basic data set will be used for the matching process in the subsequent step S200, the data partitioning process in the step S300, and the diagnostic model construction in the step S400. In this way, the data used to build the diagnostic model are ensured to be of high precision and reliability, thereby improving the accuracy of precipitation phase diagnosis.

[0069] In this embodiment, step S200 includes:

[0070] Step S201, acquiring the winter precipitation phase data through the ground station, wherein the winter precipitation phase data includes coding information;

[0071] Specifically, winter precipitation phase data are obtained through ground meteorological stations, which contain detailed precipitation type coding information for subsequent classification and model training. Winter precipitation phase data are obtained through ground meteorological stations distributed across the country. These stations regularly record and report local precipitation conditions, including information such as precipitation type, time, and intensity. In winter, special attention is paid to several common precipitation phases such as rain, snow, freezing rain, sleet, and ice pellets. In order to facilitate unified data processing and analysis, each station will be required to use a standardized coding system when recording precipitation phases. For example, it can be stipulated that rain is represented by codes "50-55", "58-59", "60-65", "80-82", and "91-92", which are further subdivided according to the density and continuity of rain. Freezing rain is represented by codes "56-57" and "66-67", corresponding to the freezing of drizzle and ordinary rain, respectively. Sleet is represented by the codes "68-69", "83-84", "87-88", and "93-94", covering a variety of mixed situations of rain, drizzle, and snow. Snow is represented by the codes "70-75", "76-78", and "85-86", which are subdivided according to the density and continuity of the snow. Ice pellets are represented by a separate code "79". In this way, when these data are obtained from the ground station, each record will contain a coded information that clearly indicates the precipitation phase at that time. For example, if a record shows the code "70", it means that intermittent or continuous snow with a low density was observed at that time. In step S201, not only these coded information are obtained, but also the accuracy and completeness of the data are ensured. These data will then be used for classification preprocessing in step S202 to generate a winter precipitation phase data set, providing a solid foundation for subsequent model training and diagnosis.

[0072] Step S202, classifying the winter precipitation phase state data according to the preset winter precipitation phase state parameters to generate the winter precipitation phase state data set;

[0073] Specifically, according to the preset winter precipitation phase classification standard (such as weather code), the coded precipitation phase data obtained from the ground station are classified and processed to form a winter precipitation phase data set for model training and verification. For example, it can be stipulated that weather codes "50-55", "58-59", "60-65", "80-82", and "91-92" are specifically subdivided according to the density and continuity of rain and classified as "rain". Weather codes "56-57" and "66-67" represent drizzle icing and ordinary rain icing, respectively, and are classified as "freezing rain". Weather codes "68-69", "83-84", "87-88", and "93-94" cover a variety of mixed situations of rain, drizzle, and snow, and are classified as "sleet". Weather codes "70-75", "76-78", and "85-86" represent snow of different densities and continuity, and are classified as "snow". Weather code "79" represents ice pellets alone. In step S202, each precipitation phase data record is traversed and classified into the corresponding precipitation phase category according to the weather code attached thereto. For example, if the weather code of a record is "75", it is known that it means "snow, intermittent or continuous, with high density during observation", so it is classified into the category of "snow". After such classification processing, a data set containing multiple winter precipitation phase categories is finally generated, and each category contains a large amount of sample data. This data set will be used in the subsequent model training and verification process, so that a more accurate and reliable winter precipitation phase diagnosis model can be constructed.

[0074] In this embodiment, step S300 includes:

[0075] Step S301, taking the longitude and latitude of the winter precipitation phase data as the standard point, matching the basic data set with the winter precipitation phase data one by one within a preset search radius to generate a matching data set, wherein the preset search radius is 12 km;

[0076] Specifically, step S301 is a key step for matching the basic data set with the winter precipitation phase data. It takes the longitude and latitude of the winter precipitation phase data as the standard point, and matches the atmospheric temperature and humidity profile data in the basic data set with the winter precipitation phase data one by one within the preset search radius (12km). This matching process aims to find the atmospheric temperature and humidity profile data corresponding to the precipitation phase data observed by the ground station, so as to provide accurate training data for subsequent model construction. For example, there is a ground station located at a certain place with a longitude and latitude of (116.40°E, 39.90°N), and the station observes a winter precipitation phase of snow. At the same time, there are atmospheric temperature and humidity profile data provided by geostationary meteorological satellites, which cover the entire area, and the spatial resolution of each data point is 12km. In step S301, a circular search area with a radius of 12km is set with the longitude and latitude of the station (116.40°E, 39.90°N) as the center. Then, in this search area, search for atmospheric temperature and humidity profile data that match the longitude and latitude of the station one by one. Since the spatial resolution of satellite data is 12km, there may be multiple atmospheric temperature and humidity profile data points within this search radius (if there are multiple profiles, take the nearest one). These data points need to be matched with the precipitation phase data (i.e. snow) of the station to generate a matching data set containing matching atmospheric temperature and humidity profiles and corresponding precipitation phases. In this way, a diagnostic model can be constructed based on these matching data to improve the accuracy of winter precipitation phase diagnosis.

[0077] Step S302, partitioning the matching data set according to preset regions to generate the matching partitioned data set, wherein the preset regions include Northeast China, North China, Northwest China, Southwest China, the middle and lower reaches of the Yangtze River Plain, South China and the Qinghai-Tibet Plateau;

[0078] Specifically, step S302 partitions the matching data set according to the preset seven regions (Northeast my country, North China, Northwest China, Southwest China, the middle and lower reaches of the Yangtze River Plain, South China and the Qinghai-Tibet Plateau) based on China's natural geography and climate characteristics, and generates a matching partitioned data set. This partitioning process takes into account the climate characteristics and altitude differences of different regions, which helps to capture and reflect the unique meteorological characteristics of each region, thereby improving the accuracy and pertinence of precipitation phase diagnosis. For example, a matching data set has been generated through step S301, which contains winter precipitation phase data from multiple ground stations across the country and their corresponding atmospheric temperature and humidity profile data. Now, it is necessary to partition this matching data set. First, according to the longitude and latitude information of each matching data point, determine the natural geographical region to which it belongs. For example, if the longitude and latitude of a data point are located at (120°E, 45°N), according to my country's geographical division, this point will belong to the Northeast region. Then, all matching data points belonging to the Northeast region are classified into one category to form a new subset. The same method is also applied to the other six regions: North China, Northwest China, Southwest China, the middle and lower reaches of the Yangtze River Plain, South China and the Qinghai-Tibet Plateau. Each region will form a subset containing all matching data points in the region. Finally, seven partitioned data sets were obtained, each of which represents the climate characteristics and precipitation phase patterns of its region. These partitioned data sets will be used in the subsequent model building steps to ensure that the diagnostic model can more accurately reflect the meteorological characteristics of each region, thereby improving the accuracy of precipitation phase diagnosis. For example, in the Northeast, due to the low winter temperature and common snowfall, the model in this region may pay more attention to the meteorological characteristics related to snowfall; while in South China, due to the relatively high winter temperature and more frequent rainfall, the model in this region may focus more on meteorological factors related to rainfall. Through this partitioning process, the diagnostic model can be made more refined and localized, and the practicality and reliability of the diagnostic results can be improved.

[0079] In this embodiment, step S400 includes:

[0080] Step S401, obtaining the preset key meteorological characteristic quantities, wherein the preset key meteorological characteristic quantities include temperature, temperature gradient, humidity, humidity gradient, pseudo-equivalent potential temperature, entropy map area, longitude and latitude information, 0°C temperature critical layer thickness, melting layer thickness, freezing layer thickness, 2-meter air temperature and surface temperature at different atmospheric altitude layers;

[0081] Specifically, step S401 involves obtaining preset key meteorological characteristic quantities. These characteristic quantities include temperature, temperature gradient, humidity, humidity gradient, pseudo-equivalent potential temperature, entropy map area, longitude and latitude information, 0°C critical layer thickness, melting layer thickness, freezing layer thickness, 2-meter air temperature and surface temperature (which can be directly read through ground station observation data) at different atmospheric altitude layers. These characteristic quantities are carefully selected because they have an important influence on the formation and change of precipitation phase (such as rain, snow, freezing rain, etc.). By extracting these characteristic quantities, rich and key information can be provided for subsequent model construction, thereby improving the accuracy of precipitation phase diagnosis. Assume that a matching partitioned data set from the Northeast region is being processed, which contains winter precipitation phase data and corresponding atmospheric temperature and humidity profile data of multiple ground stations in the region. In step S401, it is necessary to extract preset key meteorological characteristic quantities from this data set. First, pay attention to the temperature and humidity of different atmospheric altitude layers. These data can be directly read from the temperature and humidity profile products of the Fengyun geostationary meteorological satellite and ground station observations. For example, the surface temperature, the air temperature at a height of 2 meters, and the heights of each pressure layer from 1000hPa to 300hPa and the corresponding temperature and humidity values ​​can be extracted. These temperature and humidity data provide key information on the vertical structure of the atmosphere. Next, the temperature gradient and humidity gradient are calculated. The temperature gradient refers to the temperature difference between different altitude layers, and the humidity gradient refers to the humidity difference between different altitude layers. These gradient values ​​reflect the vertical changes in temperature and humidity in the atmosphere, which are crucial for understanding the formation mechanism of precipitation phases. In addition, it is necessary to calculate the pseudo-equivalent potential temperature, which is a thermodynamic parameter that comprehensively considers temperature and humidity and is of great significance for judging the stable state of the atmosphere. At the same time, the entropy map area is calculated, including the positive area (PA) and the negative area (NA), which reflect the energy changes in the temperature layers above and below the freezing point, respectively, which are very helpful for diagnosing the transition of precipitation phases. In addition to the above thermodynamic parameters, longitude and latitude information is also extracted to consider the influence of geographical location on precipitation phases. At the same time, the thickness of the critical layer at 0°C, the thickness of the melting layer, and the thickness of the freezing layer are calculated. These parameters reflect the thickness of the layers close to, above, and below the freezing point in the atmosphere, respectively, and play an important role in determining the precipitation phase (such as rain, snow, and freezing rain). Through the above steps, the preset key meteorological features are successfully extracted from the matching partitioned data set. These features will be used in the subsequent model building steps to improve the accuracy of precipitation phase diagnosis.

[0082] Step S402, integrating the preset key meteorological feature and the matched partitioned data set to generate a complete data set;

[0083] Specifically, step S402 is an important step before building the diagnostic model, which involves integrating the preset key meteorological feature quantity with the matching partitioned data set to generate a complete data set. This integration process is to combine the key meteorological feature quantity (such as temperature, humidity, pseudo-equivalent potential temperature, etc. of different atmospheric height layers) extracted from the matching partitioned data set with the corresponding precipitation phase data to form a comprehensive data set containing all necessary information. This complete data set will be used as the input of the subsequent construction of the diagnostic model to ensure that the model can fully learn the complex relationship between different meteorological features and precipitation phases. Assume that step S302 has been completed and the matching partitioned data set in Northeast China has been generated. This data set contains the winter precipitation phase data and its corresponding atmospheric temperature and humidity profile data of multiple ground stations in the region. In step S402, it is necessary to further integrate these data to build a complete data set. First, extract the preset key meteorological feature quantity from the matching partitioned data set. These feature quantities include but are not limited to: surface temperature, air temperature at a height of 2 meters, temperature and humidity values ​​of each pressure layer from 1000hPa to 300hPa, temperature gradient, humidity gradient, pseudo-equivalent potential temperature, entropy map area, 0℃ critical layer thickness, melting layer thickness and freezing layer thickness. These feature quantities reflect the key information of the vertical structure of the atmosphere and are crucial for diagnosing precipitation phases. Next, these extracted feature quantities are combined with the corresponding precipitation phase data (such as rain, snow, freezing rain, etc.). Specifically, the precipitation phase label (such as "rain", "snow", etc.) of each ground station is associated with the key meteorological feature quantities in its corresponding atmospheric temperature and humidity profile data. In this way, each sample contains complete input features (key meteorological feature quantities) and output labels (precipitation phases). After completing the above integration, a complete data set containing all necessary information is obtained. This data set will be used as input for the subsequent construction of the diagnostic model. In the model training stage, these data will be used to train the XGBoost model so that it can learn the complex relationship between different meteorological features and precipitation phases. Ultimately, the trained model will be able to diagnose the accurate precipitation phase based on real-time atmospheric temperature and humidity profile data.

[0084] Based on the above analysis, it can be seen that through the integration process of step S402, it is ensured that the diagnosis model can make full use of all available meteorological characteristic information, thereby improving the accuracy of precipitation phase diagnosis. This is of great significance for improving the accuracy of weather forecasts, ensuring traffic safety, power supply and agricultural production.

[0085] Step S403, constructing the diagnostic model based on the complete data set;

[0086] Specifically, step S403 builds a diagnostic model based on the complete data set generated in step S402. This data set contains key meteorological features (such as temperature, humidity, pseudo equivalent potential temperature, etc.) extracted from the matching partitioned data set and corresponding precipitation phase labels (such as rain, snow, freezing rain, etc.). In step S403, these data will be used to train a machine learning model, especially using the XGBoost algorithm to learn the complex relationship between meteorological features and precipitation phases. For example, step S402 has been completed, and a complete data set containing winter precipitation phases in Northeast China has been generated. This data set contains multiple samples, each of which has 45 features (such as temperature, humidity, pseudo equivalent potential temperature, etc. at different atmospheric height layers) and a corresponding precipitation phase label (such as "rain", "snow", etc.). In step S403, this data set will be used to train the XGBoost model. First, the data set needs to be divided into a training set and a test set, usually with a ratio of 80% as a training set and 20% as a test set. The training set is used to train the model, while the test set is used to evaluate the performance of the model. Next, set the initial parameters of the XGBoost model, such as learning rate, number of trees, maximum depth of trees, minimum weight of child nodes, etc. The selection of these parameters has an important impact on the performance of the model, so it is usually necessary to optimize them through methods such as grid search and cross-validation. Then, the training set data is used to train the XGBoost model. During the training process, the model will learn the complex relationship between meteorological characteristics and precipitation phase state, and try to minimize the diagnostic error. Through multiple iterations, the model will gradually converge to a better state. After the training is completed, the test set is used to evaluate the performance of the model. By calculating indicators such as accuracy, recall, and F1 score, you can understand how the model performs on unseen data. If the model performance is not ideal, you can adjust the model parameters or reselect the characteristics and retrain the model. Finally, a trained XGBoost model is obtained, which can accurately diagnose the corresponding precipitation phase state based on the input meteorological characteristics. This model can be applied to real-time precipitation phase diagnosis. By obtaining real-time atmospheric temperature and humidity profile data from geostationary meteorological satellites, extracting key meteorological characteristics, and inputting them into the model, the diagnosis results of precipitation phase state can be quickly obtained.

[0087] In this embodiment, step S500 includes:

[0088] Step S501, acquiring the real-time atmospheric temperature and humidity profile data of the geostationary meteorological satellite and the ground station observation data, and extracting key meteorological characteristic quantities;

[0089] Specifically, these key meteorological characteristics include temperature, temperature gradient, humidity, humidity gradient, pseudo-equivalent potential temperature, entropy map area, longitude and latitude information, 0°C critical layer thickness, melting layer thickness and freezing layer thickness, and 2-meter air temperature and surface temperature observed from ground stations. By extracting these characteristics, the latest atmospheric state information can be provided for the subsequent diagnostic model, thereby realizing real-time and accurate diagnosis of winter precipitation phase. For example, it is necessary to diagnose the upcoming precipitation phase in real time. In step S501, the latest atmospheric temperature and humidity profile data are first obtained through the infrared hyperspectral instrument of the geostationary meteorological satellite. These data cover different atmospheric height levels, and each level contains specific values ​​of temperature and humidity. After obtaining these data, start extracting key meteorological characteristics according to the preset method. For example, extract the temperature and humidity values ​​of each pressure layer from 1000hPa to 300hPa from the data. These values ​​reflect the temperature and humidity distribution in the vertical structure of the atmosphere, which is crucial for judging the precipitation phase. Next, the temperature gradient and humidity gradient are calculated. These gradient values ​​reflect the vertical rate of change of temperature and humidity in the atmosphere, which helps to understand the stability of the atmospheric state. In addition, the pseudo-equivalent potential temperature is calculated, which is a thermodynamic parameter that comprehensively considers temperature and humidity. It is very useful for judging the potential instability of the atmosphere. At the same time, information such as the thickness of the 0℃ critical temperature layer, the thickness of the melting layer, and the thickness of the freezing layer in the atmosphere are extracted. This information can be used to know which layers in the atmosphere may melt or freeze, which has a direct indicative effect on diagnosing precipitation phases (such as rain, snow, and freezing rain). In addition, the longitude and latitude information of the data is also recorded to associate the diagnostic results with specific geographical locations. Finally, the surface temperature and the air temperature at a height of 2 meters in the observation data are directly obtained through ground stations. After the extraction of these feature quantities is completed, they are input into the previously trained XGBoost diagnostic model. The model will quickly calculate the diagnostic results of the precipitation phase based on these real-time data, such as "rain", "snow" or "freezing rain". That is, the data entering the diagnostic model include three parts, which are: direct variables from the atmospheric temperature and humidity profile (such as temperature and humidity), indirect variables calculated by temperature and humidity (entropy map area, melting layer thickness, etc.), and finally auxiliary variables such as 2-meter air temperature and surface temperature observed from ground stations.

[0090] Based on the above analysis, it can be seen that through the real-time data acquisition and feature extraction of step S501, it can ensure that the diagnosis model always works based on the latest and most accurate atmospheric state information, thereby greatly improving the real-time and accuracy of precipitation phase diagnosis. This is of great significance to fields such as weather forecasting, traffic management, power supply and agricultural production.

[0091] Step S502, inputting the key meteorological characteristic quantity into the diagnosis model to generate a precipitation phase diagnosis result;

[0092] Specifically, in step S502, these key meteorological feature quantities are input into the diagnostic model constructed based on the XGBoost algorithm. XGBoost is an efficient machine learning algorithm that is particularly good at processing large-scale data sets and complex data features, and can automatically learn the nonlinear relationship between feature quantities and target variables (i.e., precipitation phases). Through multiple iterations of training, the model can gradually optimize its internal parameters to improve the accuracy of diagnosis. After the key meteorological feature quantities are input into the model, the model can quickly calculate the corresponding precipitation phase diagnosis results according to the rules and patterns learned internally. These results are usually expressed in the form of numerical labels, such as "rain", "snow", "freezing rain", etc., and can be converted back to the original precipitation phase category through a mapping relationship for ease of understanding and application. In step S501, the latest atmospheric temperature and humidity profile data has been obtained from the geostationary meteorological satellite, and key meteorological feature quantities including temperature and humidity values ​​of each pressure layer, temperature gradient, humidity gradient, pseudo-equivalent potential temperature, entropy map area, 0°C critical layer thickness, melting layer thickness and freezing layer thickness have been extracted, and key meteorological feature vectors (auxiliary variables) including surface temperature and temperature at 2 meters high have been extracted from ground station observation data. Next, in step S502, these key meteorological feature quantities are input into the previously trained XGBoost diagnostic model. After receiving these feature quantities, the model will immediately start its internal calculation mechanism to process and analyze the input data according to the learned rules and patterns. After a short calculation process, the model will output a numerical label as a diagnostic result, such as "1" for "rain", "2" for "snow", "3" for "freezing rain", etc. According to the preset mapping relationship, this numerical label can be converted back to the corresponding precipitation phase category, such as "1" for "rain", "2" for "snow", "3" for "freezing rain", etc. In this way, the precipitation phase diagnosis results based on the latest atmospheric temperature and humidity profile data are obtained. This result can be provided in real time to relevant units such as meteorological departments, traffic management departments, power supply companies and agricultural management departments, so that they can take corresponding measures in time to deal with the upcoming precipitation weather, thereby ensuring the normal operation of society and the safety of people's lives and property.

[0093] In a specific embodiment of the present invention, the winter precipitation phase diagnosis method based on the geostationary satellite temperature and humidity profile of the present invention includes:

[0094] Step 1: Quality control of temperature and humidity profile products of Fengyun geostationary meteorological satellite data. Specifically, the atmospheric temperature and humidity profiles provided by the Fengyun geostationary meteorological satellite infrared hyperspectral instrument (GIIRS) have a time resolution of once every 2 hours, a spatial resolution of 12km, and are divided into 101 vertical layers (1000hPa-0.005hPa). Each satellite data contains a quality identification code, which represents the inversion accuracy of the satellite data. Perfect data (quality identification code is 0), high-quality data (quality identification code is 1), and poor quality data (quality identification codes are 2 and 3). Since the diagnosis of precipitation phase is very sensitive to temperature changes, satellite temperature and humidity data with quality identification codes of 0 and 1 are selected as the basic data set.

[0095] Step 2: Classification and preprocessing of precipitation phase data from ground stations. Specifically, the precipitation data from my country's ground stations provide detailed phase information, and digital codes are used to characterize different precipitation phases. They are divided into five categories according to the common precipitation phases in winter: rain, freezing rain, sleet, snow, and ice pellets. The historical winter precipitation phase data from ground stations are classified according to Table 1.

[0096] Table 1 Classification rules for precipitation phase at ground stations

[0097]

[0098]

[0099] Step 3: Fengyun geostationary meteorological satellite temperature and humidity profile products are matched with ground station precipitation phase data, and data processing is divided according to my country's natural geographical divisions. Specifically: According to the current resolution of Fengyun geostationary meteorological satellite temperature and humidity profile products of 12km, when matching with ground station precipitation phase data, the station data longitude and latitude are used as standard points, and the search radius is set to 12km, and the Fengyun geostationary satellite temperature and humidity profile is matched with the station data one by one. Because the temperature and humidity data of the vertical layer of the atmosphere have obvious regional differences, the present invention comprehensively considers the altitude and climate characteristics, and divides my country into 7 regions: Northeast, North China, Northwest, Southwest, the middle and lower reaches of the Yangtze River Plain, South China, and the Qinghai-Tibet Plateau, and the matched data sets are partitioned according to the above 7 regions.

[0100] Step 4: Based on the XGBoost model, an atmospheric multilayer thermodynamic model of temperature, humidity and precipitation phases in different geographical regions is constructed. Specifically, through theoretical analysis of my country's winter precipitation phases and empirical statistics of physical research, based on the FY-4GIIRS satellite atmospheric temperature and humidity profiles and ground station observation data, the following key meteorological characteristic quantities are selected. The characteristic quantities and their data sources are as follows:

[0101] a: The temperature of different atmospheric altitude layers (such as 1000hPa, 925hPa, 850hPa, 700hPa, 500hPa, 400hPa, 300hPa) is derived from the temperature profile product of the Fengyun geostationary meteorological satellite, that is, the corresponding temperature data is directly read from the temperature profile product of the Fengyun geostationary meteorological satellite.

[0102] b: Temperature gradient (temperature difference between different altitude layers) at different altitude layers (1000hPa, 925hPa, 850hPa, 700hPa, 500hPa, 400hPa, 300hPa), which is derived from the temperature profile product of the Fengyun geostationary meteorological satellite; Where Δz is the height difference between pressure layer a and pressure layer b.

[0103] c: The humidity of different atmospheric altitude layers (such as 1000hPa, 925hPa, 850hPa, 700hPa, 500hPa, 400hPa, 300hPa) is derived from the humidity profile product of the Fengyun geostationary meteorological satellite, that is, the corresponding humidity data is directly read from the humidity profile product of the Fengyun geostationary meteorological satellite.

[0104] d: humidity gradient at different altitudes (1000hPa, 925hPa, 850hPa, 700hPa, 500hPa, 400hPa, 300hPa), which is derived from the temperature and humidity profile products of the Fengyun geostationary meteorological satellite;

[0105]

[0106] Where Δz is the height difference between pressure layer a and pressure layer b.

[0107] e: Pseudo-equivalent potential temperature at different altitude layers (1000hPa, 925hPa, 850hPa, 700hPa, 500hPa, 400hPa, 300hPa), which is derived from the temperature profile product of the Fengyun geostationary meteorological satellite;

[0108]

[0109] Among them, p 0 is the standard atmospheric pressure of 1000 hPa, R is the gas constant of dry air (287 J (kg·k)), C p The constant pressure specific heat capacity of dry air (1004J (kg·k)), L v is the latent heat of vaporization of water (2.5*10 6 ).

[0110] f: Entropy map area (positive area PA and negative area NA); Positive area (PA): Appears in the layer above the freezing point, which is the area between the 0℃ isotherm and the ambient temperature; when hydrometeors descend in this layer, they absorb energy, causing warming, which may cause them to transform from solid to liquid. Negative area (NA): Appears in the layer below the freezing point, which is the area between the 0℃ isotherm and the ambient temperature; when hydrometeors descend in this layer, they lose energy, causing cooling, which may cause liquid water droplets to transform into ice particles. It comes from the temperature and humidity profile product of the Fengyun geostationary meteorological satellite;

[0111]

[0112] Among them, c p is the specific heat capacity of air at constant pressure, which is a constant and is usually taken as 1005 J·kg -1 ·K -1 , is the average absolute temperature of the layer, θ top is the potential temperature at the top of the layer, θ bottom is the potential temperature at the bottom of the layer.

[0113] g: longitude and latitude information, which comes from the grid center point information of the Fengyun geostationary meteorological satellite temperature and humidity profile product, that is, the longitude and latitude information of the observation grid center is directly read from the Fengyun geostationary meteorological satellite temperature and humidity profile product.

[0114] h: 0℃ critical temperature layer thickness (thickness of the layer between -1℃ and 1℃), which is derived from the temperature profile product of the Fengyun geostationary meteorological satellite;

[0115] H=z 2 -z 1

[0116] Among them, z 1 is the height of the lower boundary of the critical layer (height of -1°C), z 2 is the height of the upper boundary of the critical layer (the height of 1°C).

[0117] i: thickness of the melting layer in the vertical layer of the atmosphere (the total thickness of the layer with a temperature above 0°C, which usually indicates the area where melting may occur; it is derived from the temperature profile product of the Fengyun geostationary meteorological satellite;

[0118] H 融化层 =∑z 2 -z 1

[0119] Among them, z 1 and z 2 They are the lower and upper boundary heights of each interval above 0℃ respectively; when calculating, the thickness of all layers above 0℃ are accumulated.

[0120] j: thickness of the freezing layer (the total thickness of the layer with a temperature below 0°C). This layer usually indicates the area where freezing may occur. It is derived from the temperature profile product of the Fengyun geostationary meteorological satellite.

[0121] H 冻结层 =∑z 2 -z 1

[0122] Among them, z 1 and z 2 They are the lower and upper boundary heights of each interval below 0℃ respectively; when calculating, the thickness of all layers below 0℃ are accumulated.

[0123] The surface temperature and 2-meter air temperature can be directly read from ground station observation data.

[0124] There are a total of 45 feature quantities in 10 categories, and 45 feature quantities are obtained. The labels of precipitation phases include the five types of precipitation phases divided in step 2: rain, sleet, freezing rain and ice pellets. The above satellite temperature and humidity profiles and ground precipitation phase data are matched with the partitioned data set, and the data set is divided into a training set and a test set. The usual division ratio is 80% training set and 20% test set. Select the XGBoost model and set the initial parameters, including the learning rate, the number of trees, the maximum depth of the tree, the minimum weight of the child node, and other hyperparameters. Use the training set data to train the XGBoost model, and evaluate the model performance through the validation set data, and adjust the model parameters to prevent overfitting or underfitting.

[0125] Specifically, the initial parameters were set as follows: learning rate 0.1, number of trees 100, maximum tree depth 6, minimum child node weight 1, gamma value 0, subsampling rate and feature sampling rate 0.8, L1 regularization term coefficient 0, L2 regularization term coefficient 1. Then, these parameters were further optimized through grid search and cross-validation to obtain the best model performance, and the best parameter models for different geographical regions were recorded. The model was retrained on the entire training set using the best parameters, and its performance was evaluated on the test set.

[0126] Step 5: Diagnose precipitation phase based on the atmospheric multi-layer thermodynamic model; extract the 45 key feature quantities defined above through the atmospheric temperature and humidity profile data of the Fengyun geostationary meteorological satellite that can be obtained in real time on the business platform, input them into the trained XGBoost model, and map the numerical labels in the results of the diagnosed precipitation phase back to the original precipitation phase categories, which are: rain, snow, sleet, freezing rain and ice pellets. In this way, the latest generation of geostationary meteorological satellites can be used to accurately and efficiently diagnose the precipitation phase in the atmosphere, providing scientific basis and technical support for weather forecasting and meteorological analysis.

[0127] Embodiment 2: Figure 2 FIG. 1 is a schematic diagram of a winter precipitation phase state diagnosis system based on geostationary satellite temperature and humidity profiles in Embodiment 2 of the present invention. Figure 2 As shown, embodiment 2 provides a winter precipitation phase diagnosis system based on geostationary satellite temperature and humidity profile, including: a first generation module 201, a second generation module 202, a third generation module 203, a construction module 204 and a fourth generation module 205. The first generation module 201 is used to obtain the atmospheric temperature and humidity profile data of the geostationary meteorological satellite, and perform preprocessing to generate a basic data set. The second generation module 202 is used to obtain the winter precipitation phase data of the ground station, and perform classification preprocessing to generate a winter precipitation phase data set. The third generation module 203 is used to match and partition the basic data set and the winter precipitation phase data to generate a matching partitioned data set. The construction module 204 is used to construct a diagnosis model based on preset key meteorological feature quantities and the matching partitioned data set. The fourth generation module 205 is used to obtain the real-time atmospheric temperature and humidity profile data of the geostationary meteorological satellite, and input the real-time atmospheric temperature and humidity profile data into the diagnosis model to diagnose the precipitation phase and generate a precipitation phase diagnosis result.

[0128] In this embodiment, the first generation module 201 includes: a first acquisition unit, a first generation unit, and a second generation unit. The first acquisition unit is used to acquire the atmospheric temperature and humidity profile data of the geostationary meteorological satellite through an infrared hyperspectral instrument, wherein the atmospheric temperature and humidity profile data carries a quality identification code. The first generation unit is used to classify the atmospheric temperature and humidity profile data based on the quality identification code to generate classified data. And the second generation unit is used to select based on the classified data to generate the basic data set.

[0129] In this embodiment, the second generation module 202 includes: a second acquisition unit and a third generation unit. The second acquisition unit is used to acquire the winter precipitation phase data through the ground station, wherein the winter precipitation phase data includes coding information. The third generation unit is used to classify the winter precipitation phase data according to the preset winter precipitation phase parameters to generate the winter precipitation phase data set.

[0130] In this embodiment, the third generation module 203 includes: a fourth generation unit and a fifth generation unit. The fourth generation unit is used to match the basic data set and the winter precipitation phase data one by one within a preset search radius with the latitude and longitude of the winter precipitation phase data as the standard point to generate a matching data set, wherein the preset search radius is 12km. The fifth generation unit is used to partition the matching data set according to preset regions to generate the matching partitioned data set, wherein the preset regions include Northeast China, North China, Northwest China, Southwest China, the middle and lower reaches of the Yangtze River Plain, South China and the Qinghai-Tibet Plateau.

[0131] In this embodiment, the construction module 204 includes: a third acquisition unit, a sixth generation unit and a construction unit. The third acquisition unit is used to acquire the preset key meteorological feature quantity, wherein the preset key meteorological feature quantity includes the temperature, temperature gradient, humidity, humidity gradient, pseudo equivalent potential temperature, entropy map area, longitude and latitude information, 0°C temperature critical layer thickness, melting layer thickness, freezing layer thickness, 2-meter air temperature and surface temperature of different atmospheric altitude layers. The sixth generation unit is used to integrate the preset key meteorological feature quantity and the matched partitioned data set to generate a complete data set. The construction unit is used to construct the diagnostic model based on the complete data set.

[0132] In this embodiment, the fourth generation module 205 includes: an extraction unit and a seventh generation unit. The extraction unit is used to obtain the real-time atmospheric temperature and humidity profile data of the geostationary meteorological satellite and extract key meteorological feature quantities. The seventh generation unit is used to input the key meteorological feature quantities into the diagnosis model to generate a precipitation phase diagnosis result.

[0133] The various variations and specific examples of the winter precipitation phase diagnosis method based on the geostationary satellite temperature and humidity profile provided in Example 1 are also applicable to the winter precipitation phase diagnosis system based on the geostationary satellite temperature and humidity profile provided in this example. Through the above detailed description of a winter precipitation phase diagnosis method based on the geostationary satellite temperature and humidity profile, those skilled in the art can clearly know the implementation method of a winter precipitation phase diagnosis system based on the geostationary satellite temperature and humidity profile in this example. Therefore, for the sake of brevity of the specification, it will not be described in detail here.

[0134] Embodiment 3, Figure 3 is a schematic diagram of the structure of an electronic device in Embodiment 3 of the present invention, such as Figure 3 As shown, the third embodiment further provides an electronic device 300 , which may include: a processor 301 and a memory 302 .

[0135] The memory 302 is used to store programs; the memory 302 may include volatile memory (English: volatile memory), such as random-access memory (English: random-access memory, abbreviated: RAM), such as static random-access memory (English: static random-access memory, abbreviated: SRAM), double data rate synchronous dynamic random access memory (English: Double Data Rate Synchronous Dynamic Random Access Memory, abbreviated: DDR SDRAM), etc.; the memory may also include non-volatile memory (English: non-volatile memory), such as flash memory (English: flash memory). The memory 302 is used to store computer programs (such as applications, functional modules, etc. that implement the above method), computer instructions, etc., and the above computer programs, computer instructions, etc. can be partitioned and stored in one or more memories 302. And the above computer programs, computer instructions, data, etc. can be called by the processor 301.

[0136] The above-mentioned computer programs, computer instructions, etc. may be stored in partitions in one or more memories 302 . And the above-mentioned computer programs, computer instructions, etc. may be called by the processor 301 .

[0137] The processor 301 is used to execute the computer program stored in the memory 302 to implement each step of the method involved in the above embodiment.

[0138] For details, please refer to the relevant description in the previous method embodiment.

[0139] The processor 301 and the memory 302 may be independent structures or integrated structures. When the processor 301 and the memory 302 are independent structures, the memory 302 and the processor 301 may be coupled and connected via a bus 303 .

[0140] The electronic device of this embodiment can execute the technical solution in the above method, and its specific implementation process and technical principle are the same, which will not be repeated here.

[0141] Embodiment 4, Embodiment 4 also provides a computer-readable storage medium, including a computer program and instructions. When the computer program or instructions are executed on a computer, the computer executes the winter precipitation phase diagnosis method based on the geostationary satellite temperature and humidity profile of any embodiment of the present invention.

[0142] Computer-readable storage media include: USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks, and other media that can store program codes.

[0143] This embodiment also provides a computer program product, which includes: a computer program, which is stored in a readable storage medium. At least one processor of an electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device executes the solution provided by any of the above embodiments.

[0144] It should be understood that the various forms of the above-described processes may be used to reorder, add or delete steps. For example, the steps described in the present disclosure may be performed in parallel, sequentially or in a different order.

[0145] As long as the expected results of the technical solution disclosed in the present invention can be achieved, this document does not impose any limitations here.

[0146] In summary, the winter precipitation phase diagnosis method and system based on geostationary satellite temperature and humidity profiles of the present invention have the following beneficial effects:

[0147] 1. The present invention utilizes high temporal and spatial resolution temperature and humidity profile data from geostationary meteorological satellites. These data have a temporal resolution of once every 2 hours and a spatial resolution of 12 km, which can more accurately capture subtle changes in the atmosphere. Such high-resolution data significantly improves the accuracy of diagnosing precipitation phases (such as rain, snow, freezing rain, etc.), which is particularly critical for coping with complex and rapidly changing winter precipitation phases.

[0148] 2. In data processing and model training, the present invention comprehensively considers the climate characteristics and altitudes of different geographical regions and divides my country into seven regions for regional processing. This regional processing method can better capture and reflect the meteorological characteristics of different regions, thereby further improving the accuracy and pertinence of precipitation phase monitoring;

[0149] 3. The present invention uses XGBoost, an advanced machine learning algorithm, to construct a multi-level thermodynamic model of the atmosphere. Compared with traditional statistical methods, XGBoost can process more complex data features and provide more accurate labeled monitoring, which significantly improves the effect of winter precipitation phase monitoring and makes the diagnostic results more reliable and accurate.

[0150] 4. Compared with sounding data, the data provided by geostationary meteorological satellites have a wide coverage and a large sample size, and can provide a richer data set for model training; this helps the model learn more meteorological characteristics and laws and improve the accuracy of monitoring; at the same time, geostationary meteorological satellite data can be obtained in real time, which further improves the timeliness and practicality of monitoring, and is of great significance for responding to winter rain, snow and freezing weather events;

[0151] 5. The present invention has broad application prospects and significant practical value. Accurate diagnosis of precipitation phase is crucial to various fields such as traffic safety, power supply, urban management and agricultural production. By improving the accuracy of diagnosis, the economic losses caused by freezing rain, snowstorms and other weather conditions can be effectively reduced, and the normal operation of society and the safety of people's lives and property can be guaranteed.

[0152] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A winter precipitation phase diagnosis method based on geostationary satellite temperature and humidity profiles, characterized in that: include: Obtain atmospheric temperature and humidity profile data from geostationary meteorological satellites, perform preprocessing, and generate basic data sets; Obtain winter precipitation phase data from ground stations, perform classification preprocessing, and generate a winter precipitation phase data set; Perform matching and partitioning processing based on the basic data set and the winter precipitation phase data to generate a matched and partitioned data set; Constructing a diagnostic model based on preset key meteorological characteristics and the matched partitioned data set; The real-time atmospheric temperature and humidity profile data and the ground station observation data of the geostationary meteorological satellite are obtained, and the real-time atmospheric temperature and humidity profile data and the ground station observation data are input into the diagnostic model to diagnose the precipitation phase and generate the precipitation phase diagnosis result.

2. The winter precipitation phase diagnosis method based on geostationary satellite temperature and humidity profiles according to claim 1 is characterized in that: The acquisition of atmospheric temperature and humidity profile data from a geostationary meteorological satellite and preprocessing to generate a basic data set includes: Acquiring the atmospheric temperature and humidity profile data of the geostationary meteorological satellite through an infrared hyperspectral instrument, wherein the atmospheric temperature and humidity profile data carries a quality identification code; Classifying the atmospheric temperature and humidity profile data based on the quality identification code to generate classified data; Selection is performed based on the classified data to generate the basic data set.

3. The winter precipitation phase diagnosis method based on geostationary satellite temperature and humidity profiles according to claim 1 is characterized in that: The step of obtaining winter precipitation phase data from ground stations and performing classification preprocessing to generate a winter precipitation phase data set includes: Acquiring the winter precipitation phase data through the ground station, wherein the winter precipitation phase data includes coding information; The winter precipitation phase state data is classified according to the preset winter precipitation phase state parameters to generate the winter precipitation phase state data set.

4. The winter precipitation phase diagnosis method based on geostationary satellite temperature and humidity profiles according to claim 1 is characterized in that: The matching and partitioning processing is performed based on the basic data set and the winter precipitation phase data to generate a matched partitioned data set, including: Taking the longitude and latitude of the winter precipitation phase data as the standard point, the basic data set and the winter precipitation phase data are matched one by one within a preset search radius to generate a matching data set, wherein the preset search radius is 12 km; The matching data set is partitioned according to preset regions to generate the matching partitioned data set, wherein the preset regions include Northeast China, North China, Northwest China, Southwest China, the middle and lower reaches of the Yangtze River Plain, South China and the Qinghai-Tibet Plateau.

5. The winter precipitation phase diagnosis method based on geostationary satellite temperature and humidity profiles according to claim 1 is characterized in that: The step of constructing a diagnostic model based on the preset key meteorological feature and the matched partitioned data set includes: Obtaining the preset key meteorological characteristic quantities, wherein the preset key meteorological characteristic quantities include temperature, temperature gradient, humidity, humidity gradient, pseudo-equivalent potential temperature, entropy map area, longitude and latitude information, 0°C temperature critical layer thickness, melting layer thickness, freezing layer thickness, 2-meter air temperature and surface temperature at different atmospheric altitude layers; Integrate the preset key meteorological feature and the matched partitioned data set to generate a complete data set; Based on the complete data set, the diagnostic model is constructed.

6. The winter precipitation phase diagnosis method based on geostationary satellite temperature and humidity profiles according to claim 1 is characterized in that: The step of acquiring real-time atmospheric temperature and humidity profile data and ground station observation data from a geostationary meteorological satellite, and inputting the real-time atmospheric temperature and humidity profile data and the ground station observation data into the diagnosis model to diagnose the precipitation phase state, and generating the precipitation phase state diagnosis result includes: Acquire the real-time atmospheric temperature and humidity profile data of the geostationary meteorological satellite and the ground station observation data, and extract key meteorological characteristic quantities; The key meteorological characteristic quantities are input into the diagnostic model to generate precipitation phase diagnosis results.

7. A winter precipitation phase diagnosis system based on geostationary satellite temperature and humidity profiles, characterized in that: include: The first generation module is used to obtain atmospheric temperature and humidity profile data from geostationary meteorological satellites, and perform preprocessing to generate a basic data set; The second generation module is used to obtain the winter precipitation phase data of the ground station, and perform classification preprocessing to generate the winter precipitation phase data set; A third generation module is used to perform matching and partitioning processing based on the basic data set and the winter precipitation phase data to generate a matched and partitioned data set; A construction module, used to construct a diagnosis model based on preset key meteorological characteristics and the matched partitioned data set; as well as The fourth generation module is used to obtain the real-time atmospheric temperature and humidity profile data and the ground station observation data of the geostationary meteorological satellite, and input the real-time atmospheric temperature and humidity profile data and the ground station observation data into the diagnostic model to diagnose the precipitation phase and generate the precipitation phase diagnosis result.

8. The winter precipitation phase diagnosis system based on geostationary satellite temperature and humidity profiles according to claim 7 is characterized in that: The first generation module comprises: A first acquisition unit is used to acquire the atmospheric temperature and humidity profile data of the geostationary meteorological satellite through an infrared hyperspectral instrument, wherein the atmospheric temperature and humidity profile data carries a quality identification code; A first generating unit is used to classify the atmospheric temperature and humidity profile data based on the quality identification code to generate classified data; and The second generating unit is used to select and generate the basic data set based on the classified data.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the winter precipitation phase diagnosis method based on the geostationary satellite temperature and humidity profiles described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that: The method comprises a computer program and instructions. When the computer program or the instructions are executed on a computer, the computer is enabled to execute the winter precipitation phase diagnosis method based on the geostationary satellite temperature and humidity profiles as described in any one of claims 1 to 6.

Images