Multi-source meteorological data networking puzzle method supporting rapid meteorological early warning

Abstract

The invention relates to the technical field of meteorological early warning, in particular to a multi-source meteorological data networking puzzle method supporting rapid meteorological early warning. The method comprises the following steps: obtaining corresponding to-be-networked puzzle multi-source meteorological data in a target area, and carrying out meteorological source numbering to obtain meteorological element source numbering data corresponding to each single source; obtaining a single-time data size and a single-time data analysis time consumption corresponding to each single-source numbered meteorological data, and carrying out meteorological complexity evaluation and meteorological complex ascending order arrangement to generate a multi-source meteorological element data sequence; performing target coverage gridding on the target area to obtain a target square or volume mesh coverage area; and carrying out meteorological element laying marking and value domain proportion calculation on the target square or volume grid coverage area, and meanwhile, carrying out early-warning full-occupation networking puzzle so as to output a multi-source meteorological data networking puzzle result. According to the method, the utilization rate of the multi-source meteorological data can be greatly improved, and the networking puzzle speed is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of meteorological warning technology, and in particular to a multi-source meteorological data networking method that supports rapid meteorological warning. Background Art

[0002] Existing meteorological warning technologies often involve a variety of meteorological data. These data include not only real-time observations from various meteorological observation devices, but also future forecasts from numerical weather prediction models and artificial intelligence models. Due to the multitude of data sources and the significant differences in spatial and temporal resolution, much of the data must first be uploaded to specialized meteorological data platforms such as Tianqing before being downloaded and used. This consumes considerable time, limiting the timeliness of weather warnings. For example, a typical weather radar has a scan cycle of approximately six minutes, and data upload and download takes another three minutes. This means that from the moment the radar starts observing to the moment the observation data is acquired, the data transmission process, without any processing, takes at least nine minutes. Furthermore, considering the time required to produce and release weather warning products, a weather warning based solely on data from a single radar can take at least 15 minutes to be issued to its intended recipients. If data from multiple radars is involved, the time taken to issue a warning can be even longer due to factors such as incomplete synchronization of radar observations, the large volume of data, and inefficient transmission and processing. In the context of low-altitude economic development, the objects of meteorological warning services may have a higher speed of movement, and untimely meteorological warnings cannot meet the demand for meteorological warning services in related fields at all.

[0003] Currently, a relatively effective method for improving warning speed is to use artificial intelligence technology to infer the future development of relevant weather processes. Artificial intelligence models trained with massive samples can infer the occurrence and development of weather processes over a long period of time in just minutes, significantly improving their speed compared to manual and numerical weather forecasting models. However, as mentioned above, factors that seriously affect warning speed occur not only in the prediction phase but also in the observation phase. It is worth noting that many artificial intelligence models require real-time observation data input that matches the model training samples during the inference process. However, the observation phase often suffers from delayed data transmission and low data quality due to various practical issues. This means that while artificial intelligence algorithms improve forecast speed, they also pose the risk of unstable forecast efficiency. Summary of the Invention

[0004] Based on this, it is necessary for the present invention to provide a multi-source meteorological data networking method that supports rapid meteorological warnings to solve at least one of the above technical problems.

[0005] To achieve the above objectives, a multi-source meteorological data networking method supporting rapid meteorological warnings includes the following steps:

[0006] Step S1: obtaining the corresponding multi-source meteorological data to be networked in the target area, and numbering the meteorological sources according to the corresponding multi-source meteorological data to be networked in the target area to obtain meteorological element source number data corresponding to each single source;

[0007] Step S2: Obtain the amount of single-time data corresponding to each single-source number meteorological data and the time required to parse the single-time data through the meteorological element source number data corresponding to each single source number, and perform a meteorological complexity assessment on the meteorological element source number data corresponding to each single source based on the amount of single-time data corresponding to each single-source number meteorological data and the time required to parse the single-time data, so as to obtain the complexity corresponding to each single-source number meteorological data;

[0008] Step S3: Based on the complexity corresponding to each single source numbered meteorological data, the meteorological element source numbered data corresponding to each single source are arranged in ascending order of meteorological complexity to generate a multi-source meteorological element data sequence; the target area is gridded to obtain a target square or volume grid coverage area; based on the multi-source meteorological element data sequence, the target square or volume grid coverage area is marked with meteorological elements to obtain a target meteorological grid blank area and a target meteorological grid value area;

[0009] Step S4: Calculate the proportion of the target meteorological grid with values ​​based on the blank domain of the target meteorological grid and the domain with values ​​to obtain the proportion of the target meteorological domain with values; obtain the meteorological warning threshold corresponding to each single source, and perform warning occupancy network puzzle based on the meteorological warning threshold corresponding to each single source combined with the blank domain of the target meteorological grid and the domain with values ​​to output the multi-source meteorological data network puzzle result.

[0010] Furthermore, step S1 includes the following steps:

[0011] Step S11: Acquire the corresponding multi-source meteorological data to be networked in the target area;

[0012] Step S12: obtaining the meteorological data source corresponding to each single source in the target area through the corresponding multi-source meteorological data to be networked in the target area;

[0013] Step S13: extracting meteorological elements from the multi-source meteorological data to be networked in the target area, so as to obtain a representative meteorological element description corresponding to each single source in the target area;

[0014] Step S14: Based on the meteorological data source and representative meteorological element description corresponding to each single source in the target area, the meteorological source of the corresponding multi-source meteorological data to be networked in the target area is numbered to obtain meteorological element source number data corresponding to each single source.

[0015] Furthermore, step S2 includes the following steps:

[0016] Step S21: Obtaining a meteorological multidimensional feature vector corresponding to each single source including the geographical location, observation equipment type, and data collection frequency by parsing the meteorological element source number data corresponding to each single source;

[0017] Step S22: Obtain the fractal dimension of each data source in the spatiotemporal dimension using the meteorological element source number data corresponding to each single source, and construct a spatiotemporal representation of the meteorological multidimensional feature vector corresponding to each single source, including the geographic location, observation equipment type, and data collection frequency, based on the fractal dimension of each data source in the spatiotemporal dimension, to generate a spatiotemporal feature representation of the meteorological elements corresponding to each single source;

[0018] Step S23: performing meteorological data fluctuation analysis on the meteorological element source number data corresponding to each single source based on the spatiotemporal characteristic expression of the meteorological element corresponding to each single source, so as to obtain the meteorological element data fluctuation wavelength corresponding to each single source;

[0019] Step S24: Obtaining the amount of single-hour data corresponding to each single-source numbered meteorological data through the meteorological element source number data corresponding to each single source number, and evaluating the impact of the time consumption of parsing the meteorological element source number data corresponding to each single source number based on the amount of single-hour data corresponding to each single-source numbered meteorological data and the fluctuation wavelength of the meteorological element data corresponding to each single source number, so as to obtain the time consumption of parsing the single-hour data corresponding to each single-source numbered meteorological data;

[0020] Step S25: Based on the single-time data volume corresponding to each single-source numbered meteorological data and the time consumed in parsing the single-time data, the meteorological complexity of the meteorological element source numbered data corresponding to each single source is evaluated to obtain the complexity corresponding to each single-source numbered meteorological data.

[0021] Furthermore, step S23 includes the following steps:

[0022] Step S231: Calculating meteorological time series entropy values ​​for meteorological element source number data corresponding to each single source based on the spatiotemporal characteristic expression of meteorological elements corresponding to each single source, so as to obtain temporal entropy value changes of meteorological data of each single source at different time scales;

[0023] Step S232: Convert the time domain entropy value changes of each single source meteorological data at different time scales to generate corresponding entropy value change frequency domain representations, and perform peak and energy estimation based on the entropy value change frequency domain representations to obtain the peak frequency and energy distribution corresponding to each single source meteorological data;

[0024] Step S233: performing meteorological data fluctuation analysis on the meteorological element source number data corresponding to each single source based on the peak frequency and energy distribution corresponding to each single source meteorological data, so as to obtain the meteorological element data fluctuation wavelength corresponding to each single source.

[0025] Furthermore, the time-consuming impact assessment of parsing the meteorological element source number data corresponding to each single source based on the single-time data volume corresponding to each single source number meteorological data and the fluctuation wavelength of the meteorological element data corresponding to each single source in step S24 includes the following steps:

[0026] The network bandwidth and device performance factor received during the transmission process are obtained through the meteorological element source number data corresponding to each single source, and the transmission impedance analysis is performed on the meteorological element source number data corresponding to each single source based on the network bandwidth and device performance factor received during the transmission process to obtain the data transmission impedance corresponding to each single source meteorological data;

[0027] Based on the single-time data volume corresponding to each single-source numbered meteorological data, combined with the fluctuation wavelength of the meteorological element data corresponding to each single source and the data transmission impedance corresponding to each single source meteorological data, a meteorological parsing tensor analysis is performed to construct a meteorological element parsing tensor field corresponding to each single source meteorological data, where each component of each tensor represents the degree of influence of data volume, data fluctuation frequency, and data transmission impedance on the parsing time.

[0028] The Gaussian curvature at each position is obtained through the meteorological element analytical tensor field corresponding to each single source meteorological data, and the peak area of ​​the meteorological element analytical tensor field corresponding to each single source meteorological data is determined based on the Gaussian curvature at each position, so as to obtain the peak area of ​​the analytical time consumption of each single source meteorological data in the data analysis process;

[0029] The tensor corresponding to the peak area of ​​the parsing time of each single source meteorological data in the data parsing process is converted into the corresponding meteorological data feature vector, where the modulus of the meteorological data feature vector represents the data volume, and the direction represents the numerical combination of the influencing factors corresponding to the data parsing time;

[0030] Based on the direction of the meteorological data characteristic vector corresponding to each single source meteorological data during the data parsing process, the impact of the parsing time consumption of the meteorological element source number data corresponding to each single source is evaluated to obtain the single-time data parsing time consumption corresponding to each single source number meteorological data.

[0031] Furthermore, step S25 includes the following steps:

[0032] Step S251: constructing a data volume-time dimension tensor based on the single-time data volume and the time required to analyze the single-time data corresponding to each single-source numbered meteorological data. A three-dimensional data volume-time dimension tensor corresponding to the meteorological data is constructed with the single-time data volume as the X-axis, the time required to analyze the single-time data as the Y-axis, and the time series as the Z-axis.

[0033] Step S252: Calculate the information entropy corresponding to each tensor element based on the three-dimensional data volume-time dimension tensor corresponding to the meteorological data, and use graph theory knowledge to treat tensor elements with adjacent time relationships as nodes. The weight of the connection edge between each node is determined by the change in information entropy, and the direction of the edge represents the flow direction of information entropy. A meteorological tensor element information entropy flow network is constructed.

[0034] Step S253: Based on the meteorological tensor element information entropy flow network and combined with the size of the single-time data corresponding to each single source numbered meteorological data and the time consumed in parsing the single-time data, the meteorological complexity evaluation is performed on the meteorological element source numbered data corresponding to each single source to obtain the complexity corresponding to each single source numbered meteorological data.

[0035] Furthermore, step S253 includes the following steps:

[0036] The meteorological element information flow connection relationship between each single source number meteorological data is obtained through the meteorological tensor element information entropy flow network, and the network agglomeration evaluation is performed based on the meteorological element information flow connection relationship between each single source number meteorological data to obtain the meteorological tensor node aggregation degree and meteorological tensor edge density between each single source number meteorological data;

[0037] Based on the degree of meteorological tensor node aggregation and the density of meteorological tensor edges between each single source numbered meteorological data, the processing resistance coefficient of the meteorological element source numbered data corresponding to each single source is quantified to obtain the meteorological processing resistance coefficient corresponding to each single source numbered meteorological data;

[0038] Based on the size of the single-time data corresponding to each single-source numbered meteorological data and the time required to parse the single-time data, and combined with the meteorological processing resistance coefficient corresponding to each single-source numbered meteorological data, a meteorological complexity assessment is performed to obtain the complexity corresponding to each single-source numbered meteorological data.

[0039] Furthermore, step S3 includes the following steps:

[0040] Step S31: arranging the meteorological element source number data corresponding to each single source in ascending order of meteorological complexity based on the complexity corresponding to each single source number meteorological data, so as to generate a multi-source meteorological element data sequence;

[0041] Step S32: setting the target grid side length and performing target coverage meshing on the target area based on the target grid side length to obtain a target square or volume grid coverage area. If the target area is a two-dimensional space area, it is divided into a square grid coverage area; if the target area is a three-dimensional space area, it is divided into a cube grid coverage area.

[0042] Step S33: Based on the multi-source meteorological element data sequence, the target square or volume grid coverage area is marked with meteorological elements in order from low to high complexity. During the paving process, the location information of the meteorological element data is used to determine which unit square or volume area in the target square or volume grid coverage area it belongs to. If no single meteorological element is laid in a unit square or volume area, the unit square or volume area is marked as a blank domain to obtain a blank domain of the target meteorological grid; if multiple single meteorological elements are laid in a unit square or volume area, the maximum value of the multiple single meteorological elements is taken to represent the value range intensity of the unit square or volume area, and the unit square or volume area is marked as having a value range to obtain a target meteorological grid with a value range.

[0043] Furthermore, step S4 includes the following steps:

[0044] Step S41: Calculating the proportion of target meteorological grid with value ranges based on the blank domain of the target meteorological grid and the target meteorological grid with value ranges to obtain the proportion of target meteorological grid with value ranges;

[0045] Step S42: Obtaining a meteorological warning threshold corresponding to each single source, wherein the meteorological warning threshold is flexibly determined based on the industry requirements, data characteristics, and expert experience corresponding to each single source;

[0046] Step S43: Based on the meteorological warning threshold corresponding to each single source, a meteorological element exceeding warning judgment is performed on each target meteorological grid blank area and the target meteorological grid with a value range. If the target meteorological grid is a blank area, the blank mark corresponding to the target meteorological grid area is still maintained; if the target meteorological grid has a value range, the meteorological warning threshold corresponding to each single source is compared with the meteorological elements in the corresponding unit square or volume area. If the meteorological elements in the corresponding unit square or volume area exceed the meteorological warning threshold, the unit square or volume area is marked as a warning area; otherwise, the unit square or volume area is marked as a non-warning area.

[0047] Step S44: Calculating the warning occupancy rate based on the warning domain and the non-warning domain corresponding to the target area to obtain the meteorological element warning occupancy rate corresponding to the target area;

[0048] Step S45: Based on the meteorological element warning occupancy rate corresponding to the target area, all the warning domains and non-warning domains that have been judged are filled and repeatedly laid into the network puzzle. If the meteorological element warning occupancy rate corresponding to the target area is lower than the preset threshold, the meteorological element laying mark and re-warning mark of step S3 are repeated. Otherwise, it stops immediately and outputs the multi-source meteorological data network puzzle result.

[0049] Furthermore, step S44 includes the following steps:

[0050] Obtain the total area of ​​the target meteorological zone excluding the blank area of ​​the target meteorological grid in the target area;

[0051] Obtain the corresponding total number of target warning domains and the total number of target non-warning domains through the warning domains and non-warning domains corresponding to the target area;

[0052] The location information of a single meteorological element in each warning domain is obtained through the warning domain corresponding to the target area, and the location profile of the meteorological element in the corresponding warning domain is determined based on the location information of the single meteorological element in each warning domain, so as to obtain the size of the location profile corresponding to each single meteorological element in the warning domain;

[0053] Determine the size of the location coverage area between the single meteorological elements in the warning domain based on the size of the location contour range corresponding to each single meteorological element in the warning domain, and perform total meteorological element coverage statistics based on the size of the location coverage area between the single meteorological elements in the warning domain combined with the size of the location contour range corresponding to each single meteorological element in the warning domain and sum up each warning domain to obtain the total meteorological element warning area size corresponding to the target area;

[0054] Based on the total area of ​​the target meteorological zone, combined with the total number of target warning domains and the total number of target non-warning domains, the warning occupancy rate of the total area of ​​meteorological element warnings corresponding to the target area is calculated to obtain the meteorological element warning occupancy rate corresponding to the target area.

[0055] Beneficial effects of the present invention:

[0056] The multi-source meteorological data networking method for supporting rapid meteorological warnings proposed in the present invention has the beneficial effect of providing the necessary raw data for the meteorological analysis system by acquiring multi-source meteorological data for the target area. These meteorological data come from a variety of sources, including multiple meteorological stations, satellite observation data, meteorological model prediction results, etc., thereby ensuring the comprehensiveness and accuracy of the data. Subsequently, each meteorological source is identified using a meteorological source number, which helps to accurately distinguish and track the contribution and quality of each data source during subsequent processing. This number can help to more efficiently identify sources when processing multi-source data, ensure the accuracy and reliability of the data integration process, and lay the foundation for data analysis in subsequent steps. In addition, through source numbering, different weights can be set for source data of different meteorological elements, so that when evaluating and calculating data, the influence of certain sources can be targeted to be strengthened or weakened, thereby optimizing data utilization efficiency. Secondly, by monitoring and analyzing the amount of time data and the time required for data parsing for each meteorological data source, the complexity of meteorological data can be quantified, which not only provides a basis for evaluating the computing and storage resource requirements of meteorological data, but also provides an efficient strategy for subsequent processing steps. Through this assessment, the most challenging meteorological data sources can be identified, which is crucial for the system to optimize the configuration of computing resources when processing multi-source data. For example, some data sources require more computing time or storage resources due to excessive data volume or high parsing complexity. By assessing the complexity, it is possible to identify and take optimization measures in advance, such as parallel computing or data compression, thereby solving the problems of untimely data transmission and low data quality, and improving overall processing efficiency. In addition, complexity assessment can also help prioritize more important source data when faced with resource constraints, ensuring that meteorological data in key areas and time periods can be processed and parsed in a timely and accurate manner.Then, by sorting the meteorological element source number data corresponding to each single source in ascending order based on the complexity corresponding to each single source number meteorological data, the meteorological data with lower complexity can be effectively prioritized to generate a complex ascending sequence of meteorological data, which helps to optimize the order and efficiency of data processing. Further gridding processing can make the application of meteorological data more flexible and accurate by dividing the target area into regular grid cells. After the gridding processing divides the area into multiple small areas, meteorological data can be allocated and marked in each grid cell on demand, ensuring that data in different areas are fully covered and clearly identifying which areas have data gaps. This can solve the problem of large differences in spatial resolution of multi-source meteorological data. Meteorological data often comes from diverse sources and has great spatial differences. In order to achieve data networking, traditional networking puzzle methods need to make the spatial resolution of different data sources consistent through spatial interpolation and other means. By gridding the target space and then determining which grid in the space the meteorological data from any source falls, the spatial position alignment of multi-source meteorological data can be achieved, ensuring the accuracy and regional adaptability of subsequent warnings. Finally, by calculating the proportion of available domains, the coverage of effective meteorological data within a region can be quantified. This process helps assess which areas have sufficient meteorological data support and which areas still lack data and require additional data sources or optimized warning strategies. When conducting meteorological warnings, the setting of warning thresholds is closely related to meteorological data coverage. Each data source has a different warning threshold. Based on these thresholds, the meteorological data of the target area can be analyzed to generate a multi-source meteorological data mosaic. This mosaic accurately shows which areas meet warning conditions and which areas have missing data or do not meet the warning threshold. By using the warning domain occupancy rate assessment to optimize the time misalignment of multi-source data, each unit cell in the gridded spatial area can be identified as a warning domain or non-warning domain based on the value of one or more meteorological elements within it. Warning domains within a given spatial range do not need to wait until all member data are available. Real-time assessment of the warning domain occupancy rate can determine whether a warning product can be issued. When the warning domain occupancy rate reaches a certain threshold, data that has not arrived in time can be ignored, thereby improving both forecast speed and warning efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0058] Figure 1 This is a schematic flow chart of the steps of the multi-source meteorological data networking method for supporting rapid meteorological warnings according to the present invention;

[0059] Figure 2A flowchart of a computer execution of a multi-source meteorological data networking method for supporting rapid meteorological warnings according to the present invention;

[0060] Figure 3 This is a partial schematic diagram of the mosaic result of multi-source meteorological data networking corresponding to the two-dimensional target area of ​​the present invention. DETAILED DESCRIPTION

[0061] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0062] To achieve this, please refer to Figures 1 to 3 The present invention provides a multi-source meteorological data networking method for supporting rapid meteorological warning, the method comprising the following steps:

[0063] Step S1: obtaining the corresponding multi-source meteorological data to be networked in the target area, and numbering the meteorological sources according to the corresponding multi-source meteorological data to be networked in the target area to obtain meteorological element source number data corresponding to each single source;

[0064] Step S2: Obtain the amount of single-time data corresponding to each single-source number meteorological data and the time required to parse the single-time data through the meteorological element source number data corresponding to each single source number, and perform a meteorological complexity assessment on the meteorological element source number data corresponding to each single source based on the amount of single-time data corresponding to each single-source number meteorological data and the time required to parse the single-time data, so as to obtain the complexity corresponding to each single-source number meteorological data;

[0065] Step S3: Based on the complexity corresponding to each single source numbered meteorological data, the meteorological element source numbered data corresponding to each single source are arranged in ascending order of meteorological complexity to generate a multi-source meteorological element data sequence; the target area is gridded to obtain a target square or volume grid coverage area; based on the multi-source meteorological element data sequence, the target square or volume grid coverage area is marked with meteorological elements to obtain a target meteorological grid blank area and a target meteorological grid value area;

[0066] Step S4: Calculate the proportion of the target meteorological grid with values ​​based on the blank domain of the target meteorological grid and the domain with values ​​to obtain the proportion of the target meteorological domain with values; obtain the meteorological warning threshold corresponding to each single source, and perform warning occupancy network puzzle based on the meteorological warning threshold corresponding to each single source combined with the blank domain of the target meteorological grid and the domain with values ​​to output the multi-source meteorological data network puzzle result.

[0067] In the embodiment of the present invention, please refer to Figure 1 FIG. 1 is a flow chart of the steps of the multi-source meteorological data networking method for supporting rapid meteorological warnings according to the present invention. In this example, the multi-source meteorological data networking method for supporting rapid meteorological warnings includes the following steps:

[0068] Step S1: obtaining the corresponding multi-source meteorological data to be networked in the target area, and numbering the meteorological sources according to the corresponding multi-source meteorological data to be networked in the target area to obtain meteorological element source number data corresponding to each single source;

[0069] In an embodiment of the present invention, a city cluster in the middle and lower reaches of the Yangtze River is used as the target area. 120 ground meteorological observation stations, 3 meteorological satellites and 5 meteorological radars are deployed in the area. The ground meteorological observation stations collect temperature, humidity, wind speed and other data once a minute; the meteorological satellites transmit cloud maps and atmospheric temperature vertical distribution data every 30 minutes; and the meteorological radars send precipitation intensity and storm movement trajectory data every 6 minutes. Through dedicated data transmission lines, these meteorological data from different devices are aggregated to the meteorological data center to form multi-source meteorological data to be networked. In order to distinguish the data source, a "letter + number" numbering rule is formulated. For ground meteorological observation stations, "A" is used to represent it, and "01" represents temperature, such as the temperature of the observation station numbered "A01". At the same time, a source identifier is added to the data record header, such as the "AMS-" prefix is ​​marked on the ground meteorological observation station data header, the satellite name and orbit information are marked on the meteorological satellite data, and the radar station-specific code is marked on the meteorological radar data. The data header identifier is automatically scanned, and a unique meteorological element source number is assigned to each data according to the rules, forming meteorological element source number data corresponding to each single source, such as the temperature data of the "A01" observation station.

[0070] Step S2: Obtain the amount of single-time data corresponding to each single-source number meteorological data and the time required to parse the single-time data through the meteorological element source number data corresponding to each single source number, and perform a meteorological complexity assessment on the meteorological element source number data corresponding to each single source based on the amount of single-time data corresponding to each single-source number meteorological data and the time required to parse the single-time data, so as to obtain the complexity corresponding to each single-source number meteorological data;

[0071] In an embodiment of the present invention, by taking the ground meteorological observation station numbered "A01" as an example, the size of its single-hour data volume can be determined by counting the number of data bytes collected each time. Each time the observation station collects temperature information, the data volume is stable at about 150 bytes. The time consumption for parsing the single-hour data is obtained by actual testing on the data processing server. The server is equipped with a corresponding processor and 64AB memory. After multiple tests, the average parsing time consumption of the observation station data is 12 milliseconds. When evaluating the complexity of the meteorological element, the data volume and parsing time consumption are comprehensively considered. The larger the data volume and the longer the parsing time consumption, the higher the complexity. An evaluation standard is set, and the data volume and parsing time consumption are weighted and calculated according to a certain ratio. Assuming that the data volume weight accounts for 40% and the parsing time weight accounts for 60%, the complexity of "A01" is calculated as: data volume score (score corresponding to 150 bytes) × 40% + parsing time score (score corresponding to 12 milliseconds) × 60%; the remaining complexity calculations are similar. In this way, all single-source numbered meteorological data are calculated to obtain their corresponding complexity, providing a quantitative basis for subsequent data processing.

[0072] Step S3: Based on the complexity corresponding to each single source numbered meteorological data, the meteorological element source numbered data corresponding to each single source are arranged in ascending order of meteorological complexity to generate a multi-source meteorological element data sequence; the target area is gridded to obtain a target square or volume grid coverage area; based on the multi-source meteorological element data sequence, the target square or volume grid coverage area is marked with meteorological elements to obtain a target meteorological grid blank area and a target meteorological grid value area;

[0073] In an embodiment of the present invention, all single-source numbered meteorological data are arranged in ascending order of complexity, starting with the "A05" ground meteorological observation station with the lowest complexity, and arranged in sequence to form a multi-source meteorological element data sequence, and the target area is gridded, and the side length of the target grid is set to 5 kilometers. Since the area is a two-dimensional plane, it is divided into square grids with a side length of 5 kilometers. The entire area is divided into several unit square grids, and each grid has a clear coordinate identification, such as the grid coordinates of the upper left corner are (0, 0), which increase in sequence to the right and downward. According to the order of the multi-source meteorological element data sequence, starting from the data with the lowest complexity, the meteorological elements are laid into the grid. Taking the "A05" observation station as an example, according to its geographical location information, it is determined that it is located in the grid of coordinates (10, 15). The meteorological data of the observation station is spread into this grid, and the grid is marked as a valued domain. If a grid has not received any meteorological element data after all data are spread, it is marked as a blank domain. By spreading and marking the data one by one, the gridded result containing the blank domain of the target meteorological grid and the valued domain of the target meteorological grid is finally obtained.

[0074] Step S4: Calculate the proportion of the target meteorological grid with values ​​based on the blank domain of the target meteorological grid and the domain with values ​​to obtain the proportion of the target meteorological domain with values; obtain the meteorological warning threshold corresponding to each single source, and perform warning occupancy network puzzle based on the meteorological warning threshold corresponding to each single source combined with the blank domain of the target meteorological grid and the domain with values ​​to output the multi-source meteorological data network puzzle result.

[0075] In an embodiment of the present invention, by counting the number of blank domains and value domains of the target meteorological grid, assuming that there are 200 blank domain grids, 800 value domain grids, and the total number of grids in the target area is 1000, the value domain ratio is calculated, that is, the number of value domain grids divided by the total number of grids, and the value domain ratio is 80%. For different single sources, the meteorological warning threshold is determined according to the industry needs, data characteristics and expert experience of the industry it serves. For example, for the temperature data of the ground meteorological observation station serving agriculture, the high temperature warning threshold is set to 35°C and the low temperature warning threshold is set to 5°C; for the precipitation intensity data of the meteorological radar serving urban flood control, the 1-hour precipitation intensity exceeding 50 mm is set as the warning threshold. The meteorological elements in each grid are compared with the warning threshold of the corresponding single source. For the grid with value domain, if the meteorological element exceeds the warning threshold, the grid is marked as a warning domain; if it does not exceed, it is marked as a non-warning domain; the blank domain maintains the original mark, counts the number of warning domains and non-warning domains, and calculates the warning occupancy rate = 10×(meteorological radar precipitation intensity data). =Total number of target warning domains / (total number of target warning domains + total number of target non-warning domains))×(total area size of meteorological element warnings ÷ total area size of target meteorological area)×100%. Assuming that the calculated warning occupancy rate is 16%, by pre-setting the meteorological element warning occupancy rate threshold as 20%, the previously calculated target area meteorological element warning occupancy rate of 16% is compared with the threshold. Since 16% is lower than 20%, the meteorological element paving and marking operation of step S3 is re-executed. From the multi-source meteorological element data sequence, the meteorological elements that have not participated in paving or have not been completed are continued to be paved into the target meteorological grid in order of complexity from low to high. The grid is marked with blank domains and value domains again, and the warning domain and non-warning domain are re-determined according to the meteorological warning threshold. If the warning occupancy rate reaches 22% after a certain calculation, exceeding the pre-set 20% threshold, the paving and marking operations are immediately stopped, and the multi-source meteorological data network puzzle result is output (partial network puzzle result is shown in FIG). Figure 3 The results are presented in the form of a visual grid map, which clearly marks the distribution of warning areas, non-warning areas and blank areas in the target area, providing intuitive information for meteorological warnings.

[0076] Furthermore, step S1 includes the following steps:

[0077] Step S11: Acquire the corresponding multi-source meteorological data to be networked in the target area;

[0078] In an embodiment of the present invention, a provincial meteorological monitoring center conducts meteorological data collection for a target area within a city cluster in the middle and lower reaches of the Yangtze River. Through an established meteorological data collection network, the center acquires multi-source meteorological data corresponding to the target area, including the networked jigsaw puzzle, from multiple channels. Utilizing 100 ground-based meteorological observation stations deployed in the area, the center collects temperature, humidity, wind speed, wind direction, and other data in real time. These observation stations transmit data to the data center via dedicated communication lines at one-minute intervals. Simultaneously, the center accesses remote sensing data from two meteorological satellites, which transmit cloud maps and vertical distribution of atmospheric temperature covering the target area every 30 minutes. Furthermore, radar echo data is obtained from three surrounding meteorological radar stations, which transmit information such as precipitation intensity and storm movement trajectory every six minutes. These data from various channels, including ground-based observation stations, meteorological satellites, and meteorological radar stations, are integrated to form a multi-source meteorological data set containing multiple types and different time frequencies, including the networked jigsaw puzzle, providing raw data support for subsequent processing.

[0079] Step S12: obtaining the meteorological data source corresponding to each single source in the target area through the corresponding multi-source meteorological data to be networked in the target area;

[0080] In an embodiment of the present invention, the data source is identified based on previously acquired multi-source meteorological data to be networked. For each data record, the source is determined by the identification information carried in the data header. The data header of the ground meteorological observation station is prefixed with a fixed station number "AMS-", for example, "AMS-001" represents the ground meteorological observation station numbered 001; the header of meteorological satellite data is marked with the satellite name and orbit information, such as "FY-4B-AeostationaCy" for the geostationary orbit satellite data of Fengyun-4B; and the data of the meteorological radar station is marked with a radar station-specific code, such as "CD-03" for the meteorological radar station numbered 03. The header identifier of each data record is automatically scanned, and data with the same identifier is classified into the corresponding source category. After identification, all data are accurately divided into three categories: 100 ground meteorological observation stations, two meteorological satellites, and three meteorological radar stations, totaling 105 single data sources. A meteorological data source list corresponding to each single source in the target area is formed, which clearly presents the source of the data and facilitates subsequent differentiated processing of data from different sources.

[0081] Step S13: extracting meteorological elements from the multi-source meteorological data to be networked in the target area, so as to obtain a representative meteorological element description corresponding to each single source in the target area;

[0082] In an embodiment of the present invention, meteorological elements are extracted from the corresponding multi-source meteorological data to be networked in the target area. For the data from the ground meteorological observation station, key meteorological elements such as temperature, humidity, wind speed, and wind direction are extracted from each record. For example, from the observation station data record numbered "AMS-001" "2024-10-0110:00:00, 25℃, 60%, 3m / s, southeast wind", the temperature "25℃", humidity "60%", wind speed "3m / s", and wind direction "southeast wind" are accurately separated. For meteorological satellite remote sensing data, image processing technology is used to extract elements such as cloud top height and cloud coverage from cloud maps, and temperature values ​​of different altitude layers are extracted from the atmospheric temperature vertical distribution data; for meteorological radar station data, elements such as precipitation area range and precipitation intensity distribution are analyzed from radar echo images. For each single data source, the most representative meteorological elements are selected for description. For example, ground meteorological observation stations use temperature and wind speed as representative elements, meteorological satellites use cloud top height and cloud coverage as representatives, and meteorological radar stations use precipitation intensity and precipitation area range as representatives. In this way, a description of representative meteorological elements corresponding to each single source in the target area is obtained, simplifying the data structure while retaining key meteorological information.

[0083] Step S14: Based on the meteorological data source and representative meteorological element description corresponding to each single source in the target area, the meteorological source of the corresponding multi-source meteorological data to be networked in the target area is numbered to obtain meteorological element source number data corresponding to each single source.

[0084] In this embodiment of the present invention, based on the previously determined meteorological data source and representative meteorological element description corresponding to each single source, the target area's corresponding multi-source meteorological data to be networked is assigned a meteorological source number. A unified "letter + number" encoding rule is used, with the code consisting of an alphabetic identifier for the data source type and a numerical number representing the meteorological element. For ground-based meteorological observation stations, the letter "A" is used to indicate the data source type. For a ground-based meteorological observation station numbered "AMS-001," if temperature is the representative element, it is assigned the number "A01," where "A" represents the ground-based meteorological observation station and "01" represents the temperature element provided by the station. If wind speed is the representative element, it is assigned the number "A02." For meteorological satellites, the letter "B" is used to indicate the data source type. For example, if cloud top height is the representative element, the Fengyun-4B satellite is assigned the number "B01," where "B" represents the meteorological satellite and "01" represents the cloud top height element provided by the satellite. If cloud coverage is the representative element, the number is "B02." For meteorological radar stations, the letter "C" is used to indicate the data source type. For a meteorological radar station numbered "CD-03", if precipitation intensity is used as the representative element, it is numbered "C01", where "C" represents the meteorological radar station and "01" represents the precipitation intensity element provided by the station; if precipitation area range is used as the representative element, it is numbered "C03". According to this rule, all data from 105 single data sources are numbered one by one, and finally meteorological element source number data corresponding to each single source is generated. These numbered data provide a unique, concise and standardized identifier for the subsequent networking puzzle of meteorological data, facilitating rapid retrieval, matching and integration of data, and ensuring the efficient use of multi-source meteorological data in the rapid meteorological warning process.

[0085] Furthermore, step S2 includes the following steps:

[0086] Step S21: Obtaining a meteorological multidimensional feature vector corresponding to each single source including the geographical location, observation equipment type, and data collection frequency by parsing the meteorological element source number data corresponding to each single source;

[0087] In an embodiment of the present invention, after completing the meteorological element source numbering, the meteorological element source number data corresponding to each single source is parsed to obtain a meteorological multidimensional feature vector. Taking the number "A01" (the temperature element of the ground meteorological observation station 001) as an example, the "A" in the number is used to identify the data source type as a ground meteorological observation station. Combined with the pre-established data source metadata index, it is found that the geographical location of the observation station is 30.5 degrees north latitude and 114.3 degrees east longitude, the observation equipment type is an automatic weather station, and the data collection frequency is once per minute. This information is converted into a feature vector format, in which the geographical location is represented by longitude and latitude coordinates, the observation equipment type is digitally encoded (the automatic weather station is encoded as 1), and the data collection frequency is represented by the number of collections per minute (1 time / minute), and finally the corresponding "A01" is formed. The meteorological multidimensional feature vector is [30.5, 114.3, 1, 1]. For the number "B01" (the cloud top height element of the meteorological satellite), the data source type is identified as a meteorological satellite through "B". The satellite orbital position is 105 degrees east longitude in the geostationary orbit, the observation equipment type is a multi-channel imager, and the data collection frequency is once every 30 minutes. This information is converted into a feature vector, with the orbital position expressed in longitude (105 degrees), the observation equipment type is coded as 2, and the data collection frequency is expressed as 0.033 times / minute (1 / 30), forming a feature vector [105, 0, 2, 0.033]. Using the same method, the meteorological element source number data corresponding to all 105 single sources are parsed to obtain the meteorological multidimensional feature vectors containing the geographical location, observation equipment type, and data collection frequency.

[0088] Step S22: Obtain the fractal dimension of each data source in the spatiotemporal dimension using the meteorological element source number data corresponding to each single source, and construct a spatiotemporal representation of the meteorological multidimensional feature vector corresponding to each single source, including the geographic location, observation equipment type, and data collection frequency, based on the fractal dimension of each data source in the spatiotemporal dimension, to generate a spatiotemporal feature representation of the meteorological elements corresponding to each single source;

[0089] In this embodiment of the present invention, the fractal dimension of each data source in the spatiotemporal dimension is calculated based on the meteorological element source number data corresponding to each single source. Taking "C01" (the precipitation intensity element of meteorological radar station 03) as an example, this radar station collects precipitation data every 6 minutes, forming time series data. The fractal dimension is calculated using the box counting method, which divides the time axis into intervals of different scales and counts the fluctuations in precipitation intensity data within each interval. For example, within a 1-hour time window, it is divided into 10 small intervals. The range and frequency of precipitation intensity changes within each small interval are counted. As the interval size decreases, the number of boxes required to cover the data point is calculated. The slope is obtained by fitting a curve of the relationship between the number of boxes and the size in a double logarithmic coordinate system. This slope is the fractal dimension in the time dimension. For the spatial dimension, the spatial distribution of precipitation intensity within the radar station's monitoring area was analyzed. The box counting method was also used to divide the monitoring area into grids of different sizes. The changes in precipitation intensity within each grid were counted, and the spatial fractal dimension was calculated. Assuming that the calculated temporal fractal dimension is 1.8 and the spatial fractal dimension is 1.5, the fractal dimension of "C01" in the spatiotemporal dimension is (1.8, 1.5). Based on this fractal dimension, the spatiotemporal expression of the data source was constructed for the meteorological multidimensional feature vector [30.2, 115.4, 3, 0.167] corresponding to "C01". The fractal dimension was used as a weight factor to adjust the geographic location and data acquisition frequency components in the feature vector. For example, multiply the geographic location component by the spatial fractal dimension, and multiply the data collection frequency component by the temporal fractal dimension to obtain the adjusted eigenvector [30.2×1.5, 115.4×1.5, 3, 0.167×1.8]=[45.3, 173.1, 3, 0.3]. In this way, the spatiotemporal characteristic expression of the meteorological element corresponding to "C01" is generated, reflecting the characteristics of the data source in the spatiotemporal dimension. The same processing is performed on all single-source meteorological data to finally generate the corresponding spatiotemporal characteristic expression of the meteorological element.

[0090] Step S23: performing meteorological data fluctuation analysis on the meteorological element source number data corresponding to each single source based on the spatiotemporal characteristic expression of the meteorological element corresponding to each single source, so as to obtain the meteorological element data fluctuation wavelength corresponding to each single source;

[0091] In an embodiment of the present invention, a meteorological data fluctuation analysis is performed on the meteorological element source number data based on the spatiotemporal characteristic expression of meteorological elements corresponding to each single source previously generated. Taking the number "B02" (cloud coverage element of the meteorological satellite) as an example, its spatiotemporal characteristic expression of meteorological elements is [105, 0, 2, 0.033, 1.7, 1.4], where the last two components are the time and space fractal dimensions, respectively. The cloud coverage data of the satellite is analyzed in the time domain, and the time series data is converted into the frequency domain using the Fourier transform method. In the frequency domain, the main frequency components of the data fluctuation are identified. For example, through analysis, it is found that there is an obvious peak at a frequency of 0.02 times / minute (corresponding to a 50-minute period), and the energy of this frequency component accounts for 40% of the total energy, indicating that the fluctuation of the cloud coverage data in the 50-minute period is most significant. At the same time, combined with the spatial fractal dimension, we analyzed the spatial variation of cloud cover and found that it exhibits self-similarity at specific spatial scales. This self-similarity is associated with the 50-minute periodic fluctuation in time. The peak frequency of 0.02 times / minute is used as the fluctuation wavelength of the meteorological element data corresponding to "B02". In the case of multiple significant peaks, the frequency that best reflects the essential characteristics of data fluctuation is selected as the main fluctuation frequency, taking into account the influence of their energy distribution and spatiotemporal fractal dimension. By analyzing the meteorological element source number data corresponding to all 105 single sources, the fluctuation wavelength of the meteorological element data corresponding to each single source is finally obtained, providing key parameters for subsequent data analysis and network puzzle construction.

[0092] Step S24: Obtaining the amount of single-hour data corresponding to each single-source numbered meteorological data through the meteorological element source number data corresponding to each single source number, and evaluating the impact of the time consumption of parsing the meteorological element source number data corresponding to each single source number based on the amount of single-hour data corresponding to each single-source numbered meteorological data and the fluctuation wavelength of the meteorological element data corresponding to each single source number, so as to obtain the time consumption of parsing the single-hour data corresponding to each single-source numbered meteorological data;

[0093] In an embodiment of the present invention, the size of the single-time data corresponding to each single-source numbered meteorological data is obtained through the source number data of the meteorological element corresponding to each single source. Taking the number "A02" (the wind speed element of the ground meteorological observation station 002) as an example, the observation station collects wind speed data once a minute. The data collected each time contains information such as wind speed value, wind direction angle, and measurement accuracy. The data size is 128 bytes. A statistical analysis is performed on the historical data of the data source, and it is found that the size of the single-time data size is stable at about 128 bytes. Combined with the previously obtained meteorological element data fluctuation wavelength, the time consumption for parsing the single-time data of "A02" is evaluated. It is known that the fluctuation frequency of the data source is 0.05 times / minute (corresponding to a 20-minute period), which means that the data changes frequently. In the process, data that changes frequently requires more computing resources to process, so the parsing time will increase accordingly. Through experimental testing and statistical analysis, a relationship model between data fluctuation frequency and parsing time is established. For data with a fluctuation frequency of 0.05 times / minute, the parsing time is 30% longer than that of stable data. Assuming that the single-time parsing time of stable data is 10 milliseconds, the single-time data parsing time of "A02" is 10×(1+30%)=13 milliseconds. According to the same method, all single-source numbered meteorological data are analyzed to obtain the corresponding single-time data size, and the parsing time is evaluated in combination with the fluctuation wavelength of the meteorological element data. Finally, the single-time data parsing time corresponding to each single-source numbered meteorological data is obtained, providing data support for the subsequent meteorological complexity assessment.

[0094] Step S25: Based on the single-time data volume corresponding to each single-source numbered meteorological data and the time consumed in parsing the single-time data, the meteorological complexity of the meteorological element source numbered data corresponding to each single source is evaluated to obtain the complexity corresponding to each single-source numbered meteorological data.

[0095] In an embodiment of the present invention, a meteorological complexity evaluation is performed on meteorological element source number data based on the size of single-time data corresponding to each single-source numbered meteorological data obtained previously and the length of time it takes to parse the single-time data. Taking number "C03" (precipitation area range element of meteorological radar station 03) as an example, the size of single-time data of the data source is 512 bytes, and the time it takes to parse the single-time data is 25 milliseconds. A meteorological complexity evaluation index system is defined, including factors such as data size, parsing time, and data fluctuation frequency. The size of single-time data and the time it takes to parse the single-time data are used as the main evaluation indicators, and the influence of data fluctuation frequency is considered. For "C03", its data fluctuation frequency is 0.03 times / minute (corresponding to a 33.3-minute period), indicating that the data changes are at a medium frequency. A complexity calculation formula is established: meteorological complexity = data size × parsing time Coefficient + parsing time × data volume coefficient + fluctuation frequency impact value, where the parsing time coefficient is determined according to the length of the parsing time, the data volume coefficient is determined according to the size of the data, and the fluctuation frequency impact value is determined according to the high or low fluctuation frequency. Assuming that the parsing time coefficient is 0.6, the data volume coefficient is 0.4, and the fluctuation frequency impact value is 5 (the corresponding value for medium frequency), the meteorological complexity of "C03" = 512×0.6+25×0.4+5=307.2+10+5=322.2. According to this method, the meteorological complexity of all 105 single-source numbered meteorological data is evaluated, and finally the complexity corresponding to each single-source numbered meteorological data is obtained. These complexity values ​​reflect the difficulty of each data source in the data processing and analysis process, providing an important reference for subsequent meteorological data networking puzzles and rapid meteorological warnings, and helping to reasonably allocate computing resources and optimize processing processes.

[0096] Furthermore, step S23 includes the following steps:

[0097] Step S231: Calculating meteorological time series entropy values ​​for meteorological element source number data corresponding to each single source based on the spatiotemporal characteristic expression of meteorological elements corresponding to each single source, so as to obtain temporal entropy value changes of meteorological data of each single source at different time scales;

[0098] In the embodiment of the present invention, the meteorological time series entropy value of the meteorological element source number data is calculated based on the spatiotemporal feature expression of the meteorological elements corresponding to each single source. Taking the temperature element of the ground meteorological observation station 001 as an example, its spatiotemporal feature expression includes the geographical location (30.5°, 114.3°), the type of observation equipment (automatic weather station), the data collection frequency (1 time / minute) and the fractal dimension (1.8, 1.5). The data source provides temperature data for the past 7 days, with a total of 10080 observations. The time scale is divided into three levels: minute level (Δ t=1 minute), hourly level (Δt=60 minutes), and daily level (Δt=1440 minutes). At each time scale, the sliding window method is used to calculate the time domain entropy value. Taking the hourly level as an example, the window size is 24 hours, the step size is 1 hour, and a total of 143 data segments are generated. For each data segment, the probability distribution P(Ti) of the temperature value is calculated, where Ti is the temperature interval. The StuCAes formula is used to determine the number of intervals k=1+loA2(n)≈14, where n is the length of the data segment. The time domain entropy value calculation formula is: H=-∑P(Ti)×loA2P(Ti). For example, in the 50th hour window, the temperature range is 20-25°C, which is divided into 14 intervals. The frequency of occurrence of each interval is counted. If the frequency of the interval [22.0-22.5)°C is 0.15, then the contribution of this interval to the entropy value is -0.15×loA2(0.15)≈0.41. The contributions of all intervals are accumulated to obtain the time domain entropy value H=3.24 bits of this window. This process is repeated for all time windows to generate the hourly time domain entropy value sequence {H1, H2, ..., H143}. Similarly, the minute-level and daily time domain entropy value sequences are calculated, and finally the time domain entropy value changes of "A01" on the three time scales are obtained. For example, the minute-level entropy value range is 2.8-3.5 bits, reflecting the uncertainty of temperature fluctuations in a short period of time.

[0099] Step S232: Convert the time domain entropy value changes of each single source meteorological data at different time scales to generate corresponding entropy value change frequency domain representations, and perform peak and energy estimation based on the entropy value change frequency domain representations to obtain the peak frequency and energy distribution corresponding to each single source meteorological data;

[0100] In an embodiment of the present invention, frequency domain conversion and feature extraction are performed based on the previously obtained time domain entropy value change. Taking the hourly time domain entropy value sequence {H1, H2, ..., H143} of "A01" as an example, discrete Fourier transform (DFT) is used to convert it into frequency domain representation. In order to improve the calculation efficiency, the fast Fourier transform (FFT) algorithm is used to fill the sequence length with zeros to 256 points. The frequency domain representation is a complex sequence {X0, X1, ..., X255}, where the amplitude |Xk| represents the intensity of the frequency component k, the phase ∠Xk represents the phase information, and the energy Ek = |Xk|2 of each frequency component is calculated. The total energy Etotal = ∑Ek, the frequency resolution is Δf = 1 / (N×Δt) = 1 / (256×60)≈0.000065Hz, and the corresponding period is approximately 15385 minutes (approximately 10.7 days). ), by analyzing the frequency domain energy distribution, significant peaks are identified. For example, an obvious peak is found at k=8, corresponding to a frequency f=8×0.000065≈0.00052 Hz (with a period of approximately 32.1 hours). The energy share of this frequency component is 35%, and the peak position is accurately estimated by using the parabolic interpolation method to obtain a more accurate frequency value fpeak=0.00053 Hz. The energy distribution is further calculated, and the frequency range is divided into 10 frequency bands. The energy share of each frequency band is calculated. For example, the 0-0.0001 Hz frequency band accounts for 5%, the 0.0001-0.0002 Hz frequency band accounts for 10%, and so on. Finally, the peak frequency of "A01" is 0.00053 Hz, and the energy distribution presents a single peak feature centered on this frequency, indicating that the temperature entropy value changes in a periodic fluctuation of about 32 hours.

[0101] Step S233: performing meteorological data fluctuation analysis on the meteorological element source number data corresponding to each single source based on the peak frequency and energy distribution corresponding to each single source meteorological data, so as to obtain the meteorological element data fluctuation wavelength corresponding to each single source.

[0102] In an embodiment of the present invention, based on the previously obtained peak frequency and energy distribution, a fluctuation analysis is performed on the meteorological element source number data. Taking the number "B01" (the cloud top height element of the meteorological satellite) as an example, its peak frequency is 0.00034 Hz ​​(the period is about 29.4 hours), and the energy distribution shows that the first three frequency components account for 75% of the total energy. A bandpass filter is used to separate the signal components near the peak frequency, and the filter bandwidth is set to ±10% of the peak frequency. The filtered signal is subjected to a Hilbert transform to extract the instantaneous amplitude and phase information. The intensity variation pattern of the fluctuation is determined through instantaneous amplitude analysis. For example, it is found that the cloud top height fluctuation reaches a peak at around 10 am every day, and the amplitude variation range is ±500 meters. By combining the spatiotemporal fractal dimension (1.7, 1.4), The spatiotemporal correlation of the fluctuation frequency was calculated. The spatial fractal dimension affects the propagation range of the fluctuation, and the temporal fractal dimension affects the persistence of the fluctuation. The fluctuation wavelength was calculated using the formula λ=v / f, where v is the fluctuation propagation speed, which is estimated to be approximately 120 km / h through the phase difference between adjacent observation points. The calculated fluctuation wavelength of "B01" is approximately 10286 km, which matches the coverage range of satellite observations. The peak frequency of "B01" was finally determined to be 0.00034 Hz, corresponding to a fluctuation period of approximately 29.4 hours. The fluctuation intensity shows a daily variation pattern, and the fluctuation wavelength is approximately 10286 km. This result provides key parameters for subsequent meteorological data fusion and early warning. For example, when matching multi-source data, other data sources with a period close to 29.4 hours can be given priority to improve the accuracy of the early warning model.

[0103] Furthermore, the time-consuming impact assessment of parsing the meteorological element source number data corresponding to each single source based on the single-time data volume corresponding to each single source number meteorological data and the fluctuation wavelength of the meteorological element data corresponding to each single source in step S24 includes the following steps:

[0104] The network bandwidth and device performance factor received during the transmission process are obtained through the meteorological element source number data corresponding to each single source, and the transmission impedance analysis is performed on the meteorological element source number data corresponding to each single source based on the network bandwidth and device performance factor received during the transmission process to obtain the data transmission impedance corresponding to each single source meteorological data;

[0105] In the embodiment of the present invention, by taking the meteorological element source number data numbered "A01" (temperature element of ground meteorological observation station 001) as an example, the data is transmitted to the meteorological data center via a dedicated communication line, and the network bandwidth during the transmission process is obtained by monitoring the network equipment. Assuming that the real-time available bandwidth of the line is 10 Mbps (megabits per second), it is converted into bytes, that is, 10×1024×1024÷8=1310720 bytes / second. The device performance factor is determined by evaluating the hardware configuration and operating status of the data receiving end server. The baseline performance value of the server CPU is set to 100; the current CPU utilization rate is 30%. According to the formula device performance factor = baseline performance value × (1-CPU utilization rate), the device performance factor of the server is 100 × (1-0.3) = 70. When performing transmission impedance analysis, the formula data transmission impedance = single-time data volume / (network bandwidth × device performance factor) is used. It is known that the single-time data volume of "A01" is 128 bytes, so its data transmission impedance is 128 / (1310720×70)≈1.37×10 -7 ,Through the same method, the source number data of meteorological ,elements corresponding to all single sources are used to obtain their network bandwidth and ,equipment performance factors, and calculate their corresponding data ,transmission impedance, providing quantitative indicators at the transmission level for ,subsequent analysis.

[0106] Preferably, a meteorological parsing tensor analysis is performed based on the amount of single-time data corresponding to each single-source numbered meteorological data, in combination with the fluctuation wavelength of the meteorological element data corresponding to each single source and the data transmission impedance corresponding to each single-source meteorological data, so as to construct a meteorological element parsing tensor field corresponding to each single-source meteorological data, wherein each component of each tensor represents the degree of influence of the data amount, the data fluctuation frequency, and the data transmission impedance on the parsing time.

[0107] In the embodiment of the present invention, by continuing to take "A01" as an example, it is known that the data size of a single time is 128 bytes. In the previous step, it has been determined that the fluctuation wavelength of its meteorological element data is λ = 10286 × 1000 meters (converting kilometers to meters). The calculated data transmission impedance is 1.37 × 10 -7 When constructing the meteorological element analytical tensor field, the three components of the tensor T are T1 (indicating the influence of data size), T2 (indicating the influence of data fluctuation frequency), and T3 (indicating the influence of data transmission impedance). The formula is used. Where n is the number of all single sources (assuming it is 105). First calculate T1. Assuming the total amount of data from all single sources at a single time is 100,000 bytes, then T1 = 128 / 100,000 = 0.00128; T2 ≈ 9.72 × 10 -8 ; T3=1.37×10 -7, and thus the meteorological element analytical tensor corresponding to "A01" is obtained as T = [0.00128, 9.72×10 -8 , 1.37×10 -7 ], for all 105 single-source meteorological data, according to the above formula and method, the three components of their respective tensors are calculated respectively, and the meteorological element analytical tensor field corresponding to each single-source meteorological data is constructed, which clearly presents the influence of each factor on the analysis time.

[0108] Preferably, the Gaussian curvature at each position is obtained through the meteorological element parsing tensor field corresponding to each single source meteorological data, and the meteorological element parsing tensor field corresponding to each single source meteorological data is parsed based on the Gaussian curvature at each position to determine the parsing peak area, so as to obtain the parsing time peak area of ​​each single source meteorological data in the data parsing process;

[0109] In the embodiment of the present invention, by taking the meteorological element analytical tensor field corresponding to "A01" as an example, the Gaussian curvature calculation method in differential geometry is used to determine the Gaussian curvature at each position. For the tensor T = [T1, T2, T3], a three-dimensional space surface is constructed, where the x-axis corresponds to T1, the y-axis corresponds to T2, and the z-axis corresponds to T3. A series of discrete points are selected on the surface, and the first-order and second-order partial derivatives of each point on the surface are approximately calculated by the finite difference method. Assume that at a certain point, according to the Gaussian curvature formula (in L, M, and N are calculated by the second-order partial derivatives), the Gaussian curvature of the point is calculated, all discrete points in the entire tensor field are traversed, and after calculating the Gaussian curvature of each point, the threshold is set to 0.0001. The area where the Gaussian curvature is greater than the threshold is determined as the analysis peak area. After calculation and screening, the peak area of ​​analysis time for the "A01" meteorological data in the data analysis process is determined. The above operation is repeated for all single-source meteorological data, and finally the peak area of ​​analysis time is obtained for each.

[0110] Preferably, the tensor corresponding to the peak area of ​​the parsing time of each single source meteorological data in the data parsing process is converted into a corresponding meteorological data feature vector, wherein the modulus of the meteorological data feature vector represents the size of the data volume, and the direction represents the numerical combination of the influencing factors corresponding to the data parsing time;

[0111] In the embodiment of the present invention, the tensor T of “A01” in the peak time-consuming region is [0.00128, 9.72×10 -8 , 1.37×10 -7 ] as an example, it is converted into a meteorological data feature vector V. According to the rule, the modulus length of the meteorological data feature vector |V| represents the amount of data, and the formula is used. Then |V|≈0.00128, the direction of the eigenvector is determined by the numerical combination of each component, using the formula The direction angles with respect to the T1 axis on the two-dimensional plane and in three-dimensional space are calculated respectively, and the calculated values ​​are θ1≈0 and θ2≈0.00616 radians. Thus, the meteorological data feature vector V=[0.00128,0,0.00616] corresponding to "A01" is obtained, where the modulus reflects the amount of data, and the direction reflects the numerical combination of influencing factors corresponding to the data analysis time. The tensors of all single-source meteorological data in the peak area of ​​analysis time are converted into meteorological data feature vectors according to the same method.

[0112] Preferably, the time consumption impact of parsing the meteorological element source number data corresponding to each single source is evaluated based on the direction of the meteorological data feature vector corresponding to each single source meteorological data during the data parsing process, so as to obtain the time consumption of parsing the single-time data corresponding to each single source number meteorological data.

[0113] In an embodiment of the present invention, by taking the meteorological data characteristic vector V=[0.00128, 0, 0.00616] corresponding to "A01" as an example, the impact of its direction on the analysis time is evaluated, and a relationship model between the direction angle and the analysis time is set. It is assumed that the direction angle θ2 is linearly related to the analysis time t, and the formula is t=a×θ2+b, where a and b are coefficients obtained by fitting historical data. Assuming a=1000 and b=5, substituting θ2=0.00616 into the formula, it can be obtained that the analysis time of the single-time data of "A01" is t=1000×0.00616+5=11.16 milliseconds. For other single-source numbered meteorological data, the same relationship model and calculation method are used according to the direction of their meteorological data characteristic vectors to evaluate the length of their respective single-time data analysis times. In this way, the single-time data analysis time corresponding to each single-source numbered meteorological data is comprehensively and accurately obtained, providing an important basis for reasonably arranging the data processing sequence and resource allocation in rapid meteorological warning.

[0114] Furthermore, step S25 includes the following steps:

[0115] Step S251: constructing a data volume-time dimension tensor based on the single-time data volume and the time required to analyze the single-time data corresponding to each single-source numbered meteorological data. A three-dimensional data volume-time dimension tensor corresponding to the meteorological data is constructed with the single-time data volume as the X-axis, the time required to analyze the single-time data as the Y-axis, and the time series as the Z-axis.

[0116] In an embodiment of the present invention, by taking the meteorological element source number data numbered "A01" (temperature element of ground meteorological observation station 001) as an example, it is known that its single-time data size is 128 bytes, and the time taken to parse the single-time data is 11.16 milliseconds. In 100 consecutive time series (assuming the time interval is 1 minute), the single-time data size and the time taken to parse the single-time data at each time point are recorded. According to the construction rules, the single-time data size is used as the X-axis, and its value range fluctuates between 120-130 bytes; the time taken to parse the single-time data is used as the Y-axis, and its value range fluctuates between 10-12 milliseconds; the time series is used as the Z-axis, from 1 to 100, in three-dimensional space, the single-time data size, the time taken to parse the single-time data and the time series corresponding to each time point are used as a tensor element to determine its coordinates in the three-dimensional space. For example, at the 10th time point, the single-time data volume is 125 bytes, and the single-time data parsing time is 10.8 milliseconds. Then the coordinates of the tensor element are (125, 10.8, 10). The data of 100 time points are processed in turn, and the tensor elements corresponding to each time point are filled into the three-dimensional space to form a three-dimensional data volume-time dimension tensor corresponding to the meteorological data. This tensor fully presents the changes in the "A01" meteorological data in the three dimensions of data volume, parsing time and time, providing a structured data foundation for subsequent analysis. For the other 104 single-source numbered meteorological data, the same method is used to construct the corresponding three-dimensional data volume-time dimension tensor based on their respective single-time data volume, single-time data parsing time and time series.

[0117] Step S252: Calculate the information entropy corresponding to each tensor element based on the three-dimensional data volume-time dimension tensor corresponding to the meteorological data, and use graph theory knowledge to treat tensor elements with adjacent time relationships as nodes. The weight of the connection edge between each node is determined by the change in information entropy, and the direction of the edge represents the flow direction of information entropy. A meteorological tensor element information entropy flow network is constructed.

[0118] In the embodiment of the present invention, the information entropy corresponding to each tensor element is calculated by constructing the three-dimensional data volume-time dimension tensor for "A01". The information entropy calculation formula is: where p i It is the probability of data appearing in a certain state. Taking the X-axis (data size) as an example, 120-130 bytes are divided into 5 intervals. The number of times the data volume falls into each interval at each time point is counted, and the probability of each interval appearing is calculated. i, and then calculate the information entropy of the tensor element in the X-axis direction. Similarly, calculate the information entropy in the Y-axis (analysis time) and Z-axis (time series) directions, and finally obtain the comprehensive information entropy of each tensor element. Use graph theory knowledge to construct the information entropy flow network of meteorological tensor elements, and use tensor elements with adjacent time relationships as nodes. For example, the tensor elements corresponding to the first time point and the second time point are used as adjacent nodes. The weight of the connecting edge between each node is determined by the change in information entropy. The calculation formula is W = |H t+1 -H t |, where H t+1 is the information entropy of the tensor elements at time t+1, H t is the information entropy of the tensor element at time t, and the direction of the edge indicates the flow direction of the information entropy. If H t+1 >H t , the direction of the edge is from the node at time t to the node at time t+1; conversely, it is from the node at time t+1 to the node at time t. In this way, all tensor elements of "A01" are connected to form a meteorological tensor element information entropy flow network. In this network, the nodes represent the meteorological data status at different time points, and the weight and direction of the edge reflect the change and flow trend of information entropy in the time series. The three-dimensional data volume-time dimension tensor constructed by the other 104 single-source numbered meteorological data is also subjected to information entropy calculation and network construction operations.

[0119] Step S253: Based on the meteorological tensor element information entropy flow network and combined with the size of the single-time data corresponding to each single source numbered meteorological data and the time consumed in parsing the single-time data, the meteorological complexity evaluation is performed on the meteorological element source numbered data corresponding to each single source to obtain the complexity corresponding to each single source numbered meteorological data.

[0120] In the embodiment of the present invention, the meteorological complexity evaluation is performed by combining the meteorological tensor element information entropy flow network based on "A01" with the amount of single-time data and the time taken to analyze the single-time data, and setting evaluation indicators, including the average information entropy H of the nodes in the network. avg , the average weight of the edge W avg , Coefficient of variation CV of data size size and the coefficient of variation CV of the time taken to analyze time , average information entropy Where N is the number of nodes, H i is the information entropy of each node; the average weight of the edge Where M is the number of edges, W j is the weight of each edge; the coefficient of variation of the data size where σ size is the standard deviation of the data size, is the average value of the data size; the coefficient of variation of the time it takes to analyze where σ time is the standard deviation of the parsing time, It is the average of the time taken to analyze. The meteorological complexity calculation formula is set as complexity = α × H avg +β×W avg +γ×CV size +δ×CV time , where α=0.3、β=0.3、γ=0.2、δ=0.2, for example, if the H of “A01” is calculated, avg =2.5,W avg =0.3, CV size =0.03, CV time =0.05, and substituting it into the formula, we get complexity = 0.856. For the other 104 single-source numbered meteorological data, according to their respective meteorological tensor element information entropy flow network and related data characteristics, we calculated according to the same indicators and formulas to obtain the complexity corresponding to each single-source numbered meteorological data. These complexity values ​​reflect the comprehensive complexity of each single-source meteorological data in terms of data volume changes, analysis time changes, and information entropy flow, providing an important reference for the formulation of data processing strategies and optimal resource allocation in rapid meteorological warnings.

[0121] Furthermore, step S253 includes the following steps:

[0122] The meteorological element information flow connection relationship between each single source number meteorological data is obtained through the meteorological tensor element information entropy flow network, and the network agglomeration evaluation is performed based on the meteorological element information flow connection relationship between each single source number meteorological data to obtain the meteorological tensor node aggregation degree and meteorological tensor edge density between each single source number meteorological data;

[0123] In an embodiment of the present invention, by taking the meteorological tensor element information entropy flow network numbered "A01" (temperature element of ground meteorological observation station 001) and "C01" (precipitation intensity element of meteorological radar station 03) as an example, the meteorological element information flow connection relationship between the two data sources is obtained. In the network, the tensor element at each time point is used as a node, and the edge between adjacent time nodes represents the information entropy flow. The number of directly connected edges between the nodes of "A01" and "C01" in the same time window (such as 100 time points) is counted. Assuming that there are 35 edges in total, accounting for 28% of the total number of edges of "A01" and 25% of the total number of edges of "C01", it is indicated that there is a moderate information flow association between the two data sources. When performing network clustering evaluation, the degree of clustering of the meteorological tensor nodes is calculated, and the local clustering coefficient formula is used. Among them E i is the number of edges that actually exist between neighbor nodes of node i, ki is the degree of node i. The clustering coefficient is calculated for all nodes in the "A01" network and averaged. Assuming that the average clustering coefficient is 0.65, it indicates that the nodes are highly clustered and information flows frequently in the local area. The edge density of the meteorological tensor is calculated using the formula Where M is the actual number of edges in the network, and N is the number of nodes. Assuming that the "A01" network has 100 nodes and 450 edges, the edge density D ≈ 0.091, indicating that the network connections are sparse and the information flow paths are limited. Using the same method, the connection relationship between all 105 single-source numbered meteorological data is evaluated, and finally the node clustering degree and edge density of each are obtained, providing a structural feature basis for the subsequent quantitative processing resistance.

[0124] Preferably, based on the degree of clustering of meteorological tensor nodes and the degree of density of meteorological tensor edges between each single source numbered meteorological data, the meteorological element source numbered data corresponding to each single source is processed with a quantified resistance coefficient to obtain the meteorological processing resistance coefficient corresponding to each single source numbered meteorological data;

[0125] In the embodiment of the present invention, the resistance coefficient is quantified based on the meteorological tensor node aggregation degree (0.65) and the meteorological tensor edge density (0.091) of "A01". The calculation formula of the resistance coefficient is defined as R = α × (1-C avg )+β×(1-D), where C avg is the average clustering coefficient, D is the edge density, and the corresponding weights are set to α = 0.7 and β = 0.3. Substituting them into the calculation, we can get the meteorological processing resistance coefficient R = 0.518 for "A01". Analyzing the physical meaning of the coefficient, a high degree of node clustering (close to 1) indicates that information flows efficiently in the local area and the processing resistance is low; a high edge density (close to 1) indicates that the network is closely connected, with many information transmission paths and low processing resistance. Therefore, (1-C avg ) and (1-D) quantify the contribution of node structure and edge structure to processing resistance, respectively. For "C01", assuming its average clustering coefficient is 0.42 and edge density is 0.12, the processing resistance coefficient R = 0.67, which is higher than that of "A01", indicating that the network structure of radar data leads to higher processing resistance. Through this method, the processing resistance coefficients of all single-source meteorological data are finally obtained.

[0126] Preferably, a meteorological complexity assessment is performed based on the amount of single-time data corresponding to each single source numbered meteorological data and the time consumed in parsing the single-time data, combined with the meteorological processing resistance coefficient corresponding to each single source numbered meteorological data, to obtain the complexity corresponding to each single source numbered meteorological data.

[0127] In the embodiment of the present invention, the meteorological complexity evaluation is performed based on the single-time data size of "A01" (128 bytes), the single-time data parsing time (11.16 milliseconds) and the meteorological processing resistance coefficient (0.518), and the complexity calculation formula is defined as follows: Where size and time are observations from a single source, and is the average value of all data sources, assuming byte, milliseconds, substituting it into the calculation, the complexity C of "A01" is ≈ 1.44. For "C01", assuming that its single-time data size is 512 bytes, the single-time data parsing time is 25 milliseconds, and the processing resistance coefficient is 0.67, the complexity C is ≈ 14.28, which is much higher than "A01", indicating that the high data volume, long parsing time and high processing resistance of radar data lead to higher complexity. Calculations are performed on all 105 single-source numbered meteorological data to obtain their respective complexity values. These values ​​reflect the comprehensive difficulty of data processing. Data sources with high complexity need to prioritize optimization of processing strategies, such as allocating more computing resources or adopting more efficient algorithms, to provide a decision-making basis for data processing in rapid meteorological warnings.

[0128] Furthermore, step S3 includes the following steps:

[0129] Step S31: arranging the meteorological element source number data corresponding to each single source in ascending order of meteorological complexity based on the complexity corresponding to each single source number meteorological data, so as to generate a multi-source meteorological element data sequence;

[0130] In an embodiment of the present invention, the meteorological element source number data corresponding to all 105 single sources are arranged in ascending order based on the complexity corresponding to each single source numbered meteorological data previously obtained. The complexity value range is 0.85 (such as ground observation station "A01") to 14.28 (such as meteorological radar station "C01"). The data sources are arranged in order from small to large in complexity to form a multi-source meteorological element data sequence. This sorting ensures that in the subsequent gridding process, data with lower complexity (usually basic data with simple processing and uniform spatial coverage) is given priority to participate in the laying, providing initial basic data for the grid, while data with higher complexity (usually fine data with complex processing and high spatial resolution) is supplemented in subsequent stages to optimize and refine the grid information, thereby realizing meteorological data fusion from coarse to fine.

[0131] Step S32: setting the target grid side length and performing target coverage meshing on the target area based on the target grid side length to obtain a target square or volume grid coverage area. If the target area is a two-dimensional space area, it is divided into a square grid coverage area; if the target area is a three-dimensional space area, it is divided into a cube grid coverage area.

[0132] In the embodiment of the present invention, by setting the side length of the target grid to 1 kilometer, a two-dimensional target area with a coverage area of ​​100 kilometers × 100 kilometers is gridded. The entire area is divided into 100 × 100 unit square grids using a square grid division method. The side length of each grid is 1 kilometer and the area is 1 square kilometer. The coordinates of each grid are determined by its lower left corner vertex. For example, the grid coordinates of the lower left corner are (0, 0), and the coordinates to the right are (1, 0), (2, 0), and so on. The coordinates upward are (0, 1). , (0, 2), etc. The grid coordinates of the upper right corner are (90, 90). If the target area is a three-dimensional space area (such as a three-dimensional space area of ​​100 km × 100 km × 100 km), the side length of the target grid is set to 1 km × 1 km × 1 km, and the area is divided into 100 × 100 × 100 unit cube grids. The volume of each grid is 1 cubic kilometer. Through this gridding method, the continuous meteorological observation area is discretized into regular grid units, providing a spatial basis for the subsequent laying of meteorological elements.

[0133] Step S33: Based on the multi-source meteorological element data sequence, the target square or volume grid coverage area is marked with meteorological elements in order from low to high complexity. During the paving process, the location information of the meteorological element data is used to determine which unit square or volume area in the target square or volume grid coverage area it belongs to. If no single meteorological element is laid in a unit square or volume area, the unit square or volume area is marked as a blank domain to obtain a blank domain of the target meteorological grid; if multiple single meteorological elements are laid in a unit square or volume area, the maximum value of the multiple single meteorological elements is taken to represent the value range intensity of the unit square or volume area, and the unit square or volume area is marked as having a value range to obtain a target meteorological grid with a value range.

[0134] In an embodiment of the present invention, based on the previously generated multi-source meteorological element data sequence, the obtained target square grid coverage area is marked with meteorological elements in order from low to high complexity. Taking the temperature element as an example, the ground meteorological observation station "A01" with the lowest complexity (position coordinates are 30.5 degrees north latitude and 114.3 degrees east longitude, and the corresponding grid coordinates are (50, 40)) is processed first. The temperature value of this observation station is 25°C. This value is spread into the corresponding grid (50, 40), and the grid is marked as having a value range with a value range intensity of 25. Then the automatic weather station "A02" with the second lowest complexity (position coordinates are 30.6 degrees north latitude and 114.5 degrees east longitude, and the corresponding grid coordinates are (50, 40)) is processed. )), its temperature value is 25.5℃. Since the grid has been paved with temperature values, the two values ​​are compared, and the maximum value 25.5 is taken as the new value range strength of the grid. The marking information of the grid is updated, and other data sources are processed continuously. If a grid (such as the grid with coordinates (20, 30)) has not been paved with any meteorological elements after all data sources are processed, the grid is marked as a blank domain. After processing all 105 data sources, the target meteorological grid is finally obtained, of which some grids are marked as having value domains and some are marked as blank domains. In the grids with value domains, the value range strength is determined by the maximum value of multiple single meteorological elements paved, forming a complete spatial distribution grid map of meteorological elements, providing a structured data basis for subsequent meteorological warning analysis.

[0135] Furthermore, step S4 includes the following steps:

[0136] Step S41: Calculating the proportion of target meteorological grid with value ranges based on the blank domain of the target meteorological grid and the target meteorological grid with value ranges to obtain the proportion of target meteorological grid with value ranges;

[0137] In an embodiment of the present invention, after completing the target meteorological grid division and meteorological element laying marking, the target meteorological grid blank domain and the target meteorological grid with value domain are calculated. Taking the target area being divided into 100×100 unit square grids as an example, according to statistics, the number of grids with value domains is 7500, and the number of grids with blank domains is 2500. The calculation method of the value domain ratio is: value domain ratio = (number of grids with value domains ÷ total number of grids) × 100%. Substituting the specific value into the formula, that is, (7500÷10000)×100%=75%. In this way, the degree of effective coverage of meteorological elements in the target area is clearly quantified, and a clear numerical basis is provided for the subsequent evaluation of the integrity and availability of meteorological data. The higher the ratio, the more sufficient the coverage of meteorological elements in the target area.

[0138] Step S42: Obtaining a meteorological warning threshold corresponding to each single source, wherein the meteorological warning threshold is flexibly determined based on the industry requirements, data characteristics, and expert experience corresponding to each single source;

[0139] In an embodiment of the present invention, a meteorological warning threshold is determined for the meteorological data corresponding to each single source. Taking the temperature element of the ground meteorological observation station numbered "A01" as an example, considering that the area where the observation station is located is an agricultural planting area, based on the temperature tolerance range of agricultural production, combined with the fluctuation characteristics of the historical temperature data of the observation station, and referring to the experience of agricultural meteorological experts, its temperature warning threshold is determined to be: a high temperature warning threshold of 35°C, and a low temperature warning threshold of 5°C. For the precipitation intensity element of the meteorological radar station numbered "C01", since it mainly serves the urban flood control department, combined with the carrying capacity of the urban drainage system, the distribution of the historical precipitation intensity data of the radar station, and the advice of hydrological experts and urban planning experts, its precipitation intensity warning threshold is set to: triggering a warning when the precipitation intensity exceeds 50 mm within 1 hour. In this way, for the different meteorological elements corresponding to 105 single sources, the corresponding meteorological warning thresholds are determined one by one according to the needs of the industry to which they belong, the characteristics of the data themselves, and the experience of experts, providing an accurate standard for the warning judgment of subsequent meteorological elements.

[0140] Step S43: Based on the meteorological warning threshold corresponding to each single source, a meteorological element exceeding warning judgment is performed on each target meteorological grid blank area and the target meteorological grid with a value range. If the target meteorological grid is a blank area, the blank mark corresponding to the target meteorological grid area is still maintained; if the target meteorological grid has a value range, the meteorological warning threshold corresponding to each single source is compared with the meteorological elements in the corresponding unit square or volume area. If the meteorological elements in the corresponding unit square or volume area exceed the meteorological warning threshold, the unit square or volume area is marked as a warning area; otherwise, the unit square or volume area is marked as a non-warning area.

[0141] In an embodiment of the present invention, a warning judgment is made on the target meteorological grid based on the meteorological warning threshold corresponding to each single source. For the blank area of ​​the target meteorological grid, such as the blank grid with coordinates (20, 30), no element comparison is performed, and its blank mark is directly maintained. For the target meteorological grid with a value range, taking the grid with coordinates (50, 40) as an example, the temperature value of 25°C of the ground meteorological observation station numbered "A01" and the temperature value of 25.5°C of the automatic meteorological observation station numbered "A02" are laid in the grid, and the maximum value of 25.5°C is taken as the meteorological element value of the grid. Compare this value with the temperature warning thresholds corresponding to "A01" and "A02" (high temperature 35℃, low temperature 5℃). Since 25.5℃ is within the range of 5℃-35℃, the grid is marked as a non-warning domain. For example, the grid with coordinates (60, 70) is paved with the precipitation intensity value of 60 mm / hour of the weather radar station numbered "C01", while the precipitation intensity warning threshold of "C01" is 50 mm / hour. 60 mm / hour exceeds the warning threshold, so the grid is marked as a warning domain. By comparing and judging all grids with value domains one by one, the warning marking work of the entire target meteorological grid is completed.

[0142] Step S44: Calculating the warning occupancy rate based on the warning domain and the non-warning domain corresponding to the target area to obtain the meteorological element warning occupancy rate corresponding to the target area;

[0143] In an embodiment of the present invention, it is assumed that the target area is a square two-dimensional area with a side length of 100 kilometers, which is divided into 100×100 unit square grids, each grid has a side length of 1 kilometer and an area of ​​1 square kilometer. It is known that the number of blank domains in the target meteorological grid is 2500. Then, to calculate the total area of ​​the target meteorological zone, it is necessary to first calculate the total grid area and then subtract the blank domain area. The total grid area is 100×100×1=10,000 square kilometers, and the blank domain area is 2500×1=2500 square kilometers. Therefore, the total area of ​​the target meteorological zone is 10,000-2500=7,500 square kilometers. In the target meteorological grid that has completed the warning marking, the warning domain and non-warning domain are counted, and by traversing all 10,000 grids, their marking status is identified one by one. After statistics, it is found that the number of grids marked as warning domains is 1,500, and the number of grids marked as non-warning domains is 6,000.Taking the warning domain grid of the meteorological radar station numbered "C01" at coordinates (60, 70) as an example, the grid contains the precipitation intensity element. It is known that the location information of a single meteorological element (precipitation intensity (assuming it is a square)) in the grid is the grid center coordinates (60.5, 70.5). To determine the size of its location outline range, the method of expanding a certain distance around the location is adopted with this location as the center. Assuming the expansion distance is 0.2 kilometers, the length of the location outline range in the horizontal and vertical directions is 0.2×2=0.4 kilometers, and its area is 0.4×0.4=0.16 square kilometers. For single meteorological elements in other warning domain grids, the size of the location outline range is determined according to the same expansion rule based on their location information. Taking the two adjacent warning domain grids with coordinates (60, 70) and (61, 70) as examples, the location outlines of the precipitation intensity elements of the meteorological radar station numbered "C01" in these two grids partially overlap. It is known that the location outline ranges of the two elements are both 0.16 square kilometers. The location coverage area is determined by the geometric calculation method. The size is 0.04 square kilometers. Perform the above calculations on all 1500 warning domains and sum the results. Assume that the total area of ​​meteorological element warnings corresponding to the target area is 600 square kilometers. Then, based on the data obtained in the previous steps, calculate the meteorological element warning occupancy rate corresponding to the target area. It is known that the total area of ​​the target meteorological area is 7500 square kilometers, the total number of target warning domains is 1500, the total number of target non-warning domains is 6000, and the total area of ​​meteorological element warnings is 600 square kilometers. The formula for calculating the warning coverage rate is: Warning coverage rate = 10 × (total number of target warning areas / (total number of target warning areas + total number of target non-warning areas)) × (total area of ​​meteorological element warnings ÷ total area of ​​target meteorological zones) × 100%. Substituting the specific value into the formula, we get 10 × (1500 / (1500 + 6000)) × (600 ÷ 7500) × 100% = 16%. This result directly reflects the actual coverage percentage of meteorological elements within the target area that meet warning conditions, providing a quantitative basis for meteorological warning decision-making. By comparing it with a pre-set threshold, it can determine whether further meteorological data supplementation or corresponding warning information issuance is needed, helping meteorological departments to conduct meteorological warning work more scientifically and accurately.

[0144] Step S45: Based on the meteorological element warning occupancy rate corresponding to the target area, all the warning domains and non-warning domains that have been judged are filled and repeatedly laid into the network puzzle. If the meteorological element warning occupancy rate corresponding to the target area is lower than the preset threshold, the meteorological element laying mark and re-warning mark of step S3 are repeated. Otherwise, it stops immediately and outputs the multi-source meteorological data network puzzle result.

[0145] In an embodiment of the present invention, by pre-setting the meteorological element warning occupancy rate threshold as 20%, the previously calculated target area meteorological element warning occupancy rate of 16% is compared with the threshold. Since 16% is lower than 20%, the meteorological element paving and marking operation of step S3 is re-executed. From the multi-source meteorological element data sequence, the meteorological elements that have not participated in the paving or have not been completed are continued to be paved into the target meteorological grid in order of complexity from low to high. The grid is marked again as blank domain and value domain, and the warning domain and non-warning domain are re-judged based on the meteorological warning threshold. If the warning occupancy rate reaches 22% after a certain calculation, exceeding the pre-set 20% threshold, the paving and marking operations are immediately stopped, and the multi-source meteorological data networking puzzle result is output. The result is presented in the form of a visual grid map, clearly marking the distribution of warning domains, non-warning domains and blank domains in the target area, providing intuitive and accurate meteorological information to the meteorological warning department so that corresponding warning and response measures can be taken in a timely manner.

[0146] Furthermore, step S44 includes the following steps:

[0147] Obtain the total area of ​​the target meteorological zone excluding the blank area of ​​the target meteorological grid in the target area;

[0148] In this embodiment of the present invention, the target area is assumed to be a two-dimensional square area with a side length of 100 kilometers, divided into 100×100 unit square grids, each with a side length of 1 kilometer and an area of ​​1 square kilometer. It is known that the number of blank areas in the target meteorological grid is 2500. Then, to calculate the total area of ​​the target meteorological zone, it is necessary to first calculate the total grid area and then subtract the blank area area. The total grid area is 100×100×1=10,000 square kilometers, and the blank area area is 2500×1=2500 square kilometers. Therefore, the total area of ​​the target meteorological zone is 10,000-2500=7500 square kilometers. In this way, the total area of ​​the area that actually contains meteorological element data after removing the blank areas is accurately obtained, providing basic data for the subsequent calculation of the warning occupancy rate.

[0149] Preferably, the total number of corresponding target warning areas and the total number of target non-warning areas are obtained through the warning areas and non-warning areas corresponding to the target area;

[0150] In an embodiment of the present invention, by counting the warning domains and non-warning domains in the target meteorological grids that have completed the warning marking, by traversing all 10,000 grids and identifying their marking status one by one, it is found that the number of grids marked as warning domains is 1,500, and the number of grids marked as non-warning domains is 6,000. This statistical process adopts the method of scanning the grids row by row and column by column, just like checking the marking status of each square on a map one by one to ensure that no grid is missed. Through this comprehensive and detailed statistics, the total number of target warning domains and the total number of target non-warning domains are clearly obtained. These data are the key basis for further analyzing the warning situation and calculating the warning occupancy rate.

[0151] Preferably, the position information of a single meteorological element in each warning domain is obtained through the warning domain corresponding to the target area, and the meteorological element position profile of the corresponding warning domain is determined based on the position information of the single meteorological element in each warning domain, so as to obtain the range size of the position profile corresponding to each single meteorological element in the warning domain;

[0152] In the embodiment of the present invention, the warning domain grid of the weather radar station numbered "C01" at the coordinates (60, 70) is taken as an example. The grid contains the precipitation intensity element. It is known that the position information of a single meteorological element (precipitation intensity (assuming it is a square)) in the grid is the grid center coordinate (60.5, 70.5). In order to determine the size of its position contour range, a method is adopted in which the position is taken as the center and a certain distance is extended in all directions. Assuming that the extension distance is 0.2 kilometers, the length of the side of the position contour range in the horizontal and vertical directions is 0.2×2=0.4 kilometers, and its area is 0.4 ×0.4=0.16 square kilometers. For other single meteorological elements within the warning domain grid, the location contour range size is determined according to its location information and the same expansion rule. For example, the temperature element of the ground meteorological observation station numbered "A01" in the warning domain grid at coordinates (55, 55) is determined by expanding 0.2 kilometers with the observation point position in the grid as the center, and the area is calculated to be 0.16 square kilometers. Through this method, the location contour range size corresponding to each single meteorological element in the warning domain is finally obtained, laying the foundation for the subsequent calculation of the element coverage area.

[0153] Preferably, the size of the position coverage area between the single meteorological elements in the warning domain is determined according to the size of the position contour range corresponding to each single meteorological element in the warning domain, and the total coverage statistics of the meteorological elements are performed based on the size of the position coverage area between the single meteorological elements in the warning domain and the size of the position contour range corresponding to each single meteorological element in the warning domain, and the sum of each warning domain is calculated to obtain the total warning area size of the meteorological elements corresponding to the target area;

[0154] In the embodiment of the present invention, by taking the coordinates (60, 70) and (61, 70) of two adjacent warning domain grids as examples, the position contours of the precipitation intensity elements of the meteorological radar station numbered "C01" in the two grids partially overlap. It is known that the range of the two element position contours is 0.16 square kilometers. Through the geometric calculation method, it is determined that their position coverage area is 0.04 square kilometers. For example, it has been determined that the range of the position contour of a single meteorological element is constructed by taking the element position as the center and extending a certain distance to the surrounding areas. Here, the position contours of the precipitation intensity elements in the two grids are both squares with a side length of 0.4 kilometers. The shape of the overlapping area is analyzed: since the two grids are adjacent (laterally adjacent), the overlapping part of their square position contours is also a regular square. The two square position contours overlap horizontally and their respective extension distances are the same, so the side length of the overlapping square is |0.4-(1-0.4)|=0.2 kilometers (1 kilometer is the grid side length, and the length of the position contour side is subtracted from the length of the non-overlapping part to obtain the side length of the overlapping part). According to the square The area formula is "area = side length × side length". Substituting the side length of the overlapping square, 0.2 kilometers, into the formula, we get a location coverage area of ​​0.2 × 0.2 = 0.04 square kilometers. When calculating the total coverage area of ​​meteorological elements, for each warning domain, add the location contour range areas of the single meteorological elements within it, and then subtract the location coverage areas between elements (to avoid double counting). For example, there are three single meteorological elements in a warning domain (1 and 2 overlap, 2 and 3 overlap), with location contour range areas of 0.16 square kilometers, 0.16 square kilometers, and 0.16 square kilometers, respectively. The sum of their location coverage areas is 0.08 square kilometers. Then the total coverage area of ​​meteorological elements in the warning domain is 0.16 + 0.16 + 0.16 - 0.08 = 0.4 square kilometers. Perform the above calculation for all 1,500 warning domains and sum the results. Assuming that the total warning area of ​​meteorological elements corresponding to the target area is 600 square kilometers, this precise calculation can accurately calculate the total area actually covered by meteorological elements in the warning area.

[0155] Preferably, the warning occupancy rate of the total area size of the target meteorological zone is calculated based on the total number of target warning domains and the total number of target non-warning domains to obtain the meteorological element warning occupancy rate corresponding to the target area.

[0156] In an embodiment of the present invention, the meteorological element warning occupancy rate corresponding to the target area is calculated based on the data obtained in the previous steps. It is known that the total area of ​​the target meteorological area is 7,500 square kilometers, the total number of target warning domains is 1,500, the total number of target non-warning domains is 6,000, and the total area of ​​meteorological element warnings is 600 square kilometers. The warning occupancy rate calculation formula is: warning occupancy rate = 10 × (total number of target warning domains / (total number of target warning domains + total number of target non-warning domains)) × (total area size of meteorological element warnings ÷ total area size of target meteorological area) × 100%. Substituting the specific values ​​into the formula, that is, 10 × (1,500 / (1,500 + 6,000)) × (600 ÷ 7,500) × 100% = 16%. The calculation result intuitively reflects the actual coverage ratio of meteorological elements in the target area that meet the warning conditions, providing a quantitative basis for meteorological warning decision-making. By comparing with the pre-set threshold, it can be determined whether it is necessary to further supplement meteorological data or issue corresponding warning information, helping the meteorological department to carry out meteorological warning work more scientifically and accurately.

[0157] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A multi-source meteorological data networking method supporting rapid meteorological warning, characterized in that: The following steps are involved: Step S1: obtaining the corresponding multi-source meteorological data to be networked in the target area, and numbering the meteorological sources according to the corresponding multi-source meteorological data to be networked in the target area to obtain meteorological element source number data corresponding to each single source; Step S2: Obtain the amount of single-time data corresponding to each single-source number meteorological data and the time required to parse the single-time data through the meteorological element source number data corresponding to each single source number, and perform a meteorological complexity assessment on the meteorological element source number data corresponding to each single source based on the amount of single-time data corresponding to each single-source number meteorological data and the time required to parse the single-time data, so as to obtain the complexity corresponding to each single-source number meteorological data; Step S3: arranging the meteorological element source number data corresponding to each single source in ascending order of meteorological complexity based on the complexity corresponding to each single source numbered meteorological data, so as to generate a multi-source meteorological element data sequence; Performing target coverage gridding on the target area to obtain a target square or volume grid coverage area; Based on the multi-source meteorological element data sequence, the target square or volume grid coverage area is marked with meteorological elements to obtain the target meteorological grid blank area and the target meteorological grid value area; Step S4: Calculate the proportion of the target meteorological grid with values ​​based on the blank domain of the target meteorological grid and the domain with values ​​to obtain the proportion of the target meteorological domain with values; obtain the meteorological warning threshold corresponding to each single source, and perform warning occupancy network puzzle based on the meteorological warning threshold corresponding to each single source combined with the blank domain of the target meteorological grid and the domain with values ​​to output the multi-source meteorological data network puzzle result.

2. The multi-source meteorological data networking method supporting rapid meteorological warning according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Acquire the corresponding multi-source meteorological data to be networked in the target area; Step S12: obtaining the meteorological data source corresponding to each single source in the target area through the corresponding multi-source meteorological data to be networked in the target area; Step S13: extracting meteorological elements from the multi-source meteorological data to be networked in the target area, so as to obtain a representative meteorological element description corresponding to each single source in the target area; Step S14: Based on the meteorological data source and representative meteorological element description corresponding to each single source in the target area, the meteorological source of the corresponding multi-source meteorological data to be networked in the target area is numbered to obtain meteorological element source number data corresponding to each single source.

3. The multi-source meteorological data networking method supporting rapid meteorological warning according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: Obtaining a meteorological multidimensional feature vector corresponding to each single source including the geographical location, observation equipment type, and data collection frequency by parsing the meteorological element source number data corresponding to each single source; Step S22: Obtain the fractal dimension of each data source in the spatiotemporal dimension using the meteorological element source number data corresponding to each single source, and construct a spatiotemporal representation of the meteorological multidimensional feature vector corresponding to each single source, including the geographic location, observation equipment type, and data collection frequency, based on the fractal dimension of each data source in the spatiotemporal dimension, to generate a spatiotemporal feature representation of the meteorological elements corresponding to each single source; Step S23: performing meteorological data fluctuation analysis on the meteorological element source number data corresponding to each single source based on the spatiotemporal characteristic expression of the meteorological element corresponding to each single source, so as to obtain the meteorological element data fluctuation wavelength corresponding to each single source; Step S24: Obtaining the amount of single-hour data corresponding to each single-source numbered meteorological data through the meteorological element source number data corresponding to each single source number, and evaluating the impact of the time consumption of parsing the meteorological element source number data corresponding to each single source number based on the amount of single-hour data corresponding to each single-source numbered meteorological data and the fluctuation wavelength of the meteorological element data corresponding to each single source number, so as to obtain the time consumption of parsing the single-hour data corresponding to each single-source numbered meteorological data; Step S25: Based on the single-time data volume corresponding to each single-source numbered meteorological data and the time consumed in parsing the single-time data, the meteorological complexity of the meteorological element source numbered data corresponding to each single source is evaluated to obtain the complexity corresponding to each single-source numbered meteorological data.

4. The multi-source meteorological data networking method for supporting rapid meteorological warning according to claim 3 is characterized in that: Step S23 includes the following steps: Step S231: Calculating meteorological time series entropy values ​​for meteorological element source number data corresponding to each single source based on the spatiotemporal characteristic expression of meteorological elements corresponding to each single source, so as to obtain temporal entropy value changes of meteorological data of each single source at different time scales; Step S232: Convert the time domain entropy value changes of each single source meteorological data at different time scales to generate corresponding entropy value change frequency domain representations, and perform peak and energy estimation based on the entropy value change frequency domain representations to obtain the peak frequency and energy distribution corresponding to each single source meteorological data; Step S233: performing meteorological data fluctuation analysis on the meteorological element source number data corresponding to each single source based on the peak frequency and energy distribution corresponding to each single source meteorological data, so as to obtain the meteorological element data fluctuation wavelength corresponding to each single source.

5. The multi-source meteorological data networking method for supporting rapid meteorological warning according to claim 3 is characterized in that: The time-consuming impact assessment of parsing the meteorological element source number data corresponding to each single source based on the single-time data volume corresponding to each single source number meteorological data and the fluctuation wavelength of the meteorological element data corresponding to each single source in step S24 includes the following steps: The network bandwidth and device performance factor received during the transmission process are obtained through the meteorological element source number data corresponding to each single source, and the transmission impedance analysis is performed on the meteorological element source number data corresponding to each single source based on the network bandwidth and device performance factor received during the transmission process to obtain the data transmission impedance corresponding to each single source meteorological data; Based on the single-time data volume corresponding to each single-source numbered meteorological data, combined with the fluctuation wavelength of the meteorological element data corresponding to each single source and the data transmission impedance corresponding to each single source meteorological data, a meteorological parsing tensor analysis is performed to construct a meteorological element parsing tensor field corresponding to each single source meteorological data, where each component of each tensor represents the degree of influence of data volume, data fluctuation frequency, and data transmission impedance on the parsing time. The Gaussian curvature at each position is obtained through the meteorological element analytical tensor field corresponding to each single source meteorological data, and the peak area of ​​the meteorological element analytical tensor field corresponding to each single source meteorological data is determined based on the Gaussian curvature at each position, so as to obtain the peak area of ​​the analytical time consumption of each single source meteorological data in the data analysis process; The tensor corresponding to the peak area of ​​the parsing time of each single source meteorological data in the data parsing process is converted into the corresponding meteorological data feature vector, where the modulus of the meteorological data feature vector represents the data volume, and the direction represents the numerical combination of the influencing factors corresponding to the data parsing time; Based on the direction of the meteorological data characteristic vector corresponding to each single source meteorological data during the data parsing process, the impact of the parsing time consumption of the meteorological element source number data corresponding to each single source is evaluated to obtain the single-time data parsing time consumption corresponding to each single source number meteorological data.

6. The multi-source meteorological data networking method supporting rapid meteorological warning according to claim 3 is characterized in that: Step S25 includes the following steps: Step S251: constructing a data volume-time dimension tensor based on the single-time data volume and the time required to analyze the single-time data corresponding to each single-source numbered meteorological data. A three-dimensional data volume-time dimension tensor corresponding to the meteorological data is constructed with the single-time data volume as the X-axis, the time required to analyze the single-time data as the Y-axis, and the time series as the Z-axis. Step S252: Calculate the information entropy corresponding to each tensor element based on the three-dimensional data volume-time dimension tensor corresponding to the meteorological data, and use graph theory knowledge to treat tensor elements with adjacent time relationships as nodes. The weight of the connection edge between each node is determined by the change in information entropy, and the direction of the edge represents the flow direction of information entropy. A meteorological tensor element information entropy flow network is constructed. Step S253: Based on the meteorological tensor element information entropy flow network and combined with the size of the single-time data corresponding to each single source numbered meteorological data and the time consumed in parsing the single-time data, the meteorological complexity evaluation is performed on the meteorological element source numbered data corresponding to each single source to obtain the complexity corresponding to each single source numbered meteorological data.

7. The multi-source meteorological data networking method for supporting rapid meteorological warning according to claim 6 is characterized in that: Step S253 includes the following steps: The meteorological element information flow connection relationship between each single source number meteorological data is obtained through the meteorological tensor element information entropy flow network, and the network agglomeration evaluation is performed based on the meteorological element information flow connection relationship between each single source number meteorological data to obtain the meteorological tensor node aggregation degree and meteorological tensor edge density between each single source number meteorological data; Based on the degree of meteorological tensor node aggregation and the density of meteorological tensor edges between each single source numbered meteorological data, the processing resistance coefficient of the meteorological element source numbered data corresponding to each single source is quantified to obtain the meteorological processing resistance coefficient corresponding to each single source numbered meteorological data; Based on the size of the single-time data corresponding to each single-source numbered meteorological data and the time required to parse the single-time data, and combined with the meteorological processing resistance coefficient corresponding to each single-source numbered meteorological data, a meteorological complexity assessment is performed to obtain the complexity corresponding to each single-source numbered meteorological data.

8. The multi-source meteorological data networking method supporting rapid meteorological warning according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: arranging the meteorological element source number data corresponding to each single source in ascending order of meteorological complexity based on the complexity corresponding to each single source number meteorological data, so as to generate a multi-source meteorological element data sequence; Step S32: setting the target grid side length and performing target coverage meshing on the target area based on the target grid side length to obtain a target square or volume grid coverage area. If the target area is a two-dimensional space area, it is divided into a square grid coverage area; if the target area is a three-dimensional space area, it is divided into a cube grid coverage area. Step S33: Based on the multi-source meteorological element data sequence, the target square or volume grid coverage area is marked with meteorological elements in order from low to high complexity. During the paving process, the location information of the meteorological element data is used to determine which unit square or volume area in the target square or volume grid coverage area it belongs to. If no single meteorological element is laid in a unit square or volume area, the unit square or volume area is marked as a blank domain to obtain a blank domain of the target meteorological grid; if multiple single meteorological elements are laid in a unit square or volume area, the maximum value of the multiple single meteorological elements is taken to represent the value range intensity of the unit square or volume area, and the unit square or volume area is marked as having a value range to obtain a target meteorological grid with a value range.

9. The multi-source meteorological data networking method supporting rapid meteorological warning according to claim 1 is characterized in that: Step S4 includes the following steps: Step S41: Calculating the proportion of target meteorological grid with value ranges based on the blank domain of the target meteorological grid and the target meteorological grid with value ranges to obtain the proportion of target meteorological grid with value ranges; Step S42: Obtaining a meteorological warning threshold corresponding to each single source, wherein the meteorological warning threshold is flexibly determined based on the industry requirements, data characteristics, and expert experience corresponding to each single source; Step S43: Based on the meteorological warning threshold corresponding to each single source, a meteorological element exceeding warning judgment is performed on each target meteorological grid blank area and the target meteorological grid with a value range. If the target meteorological grid is a blank area, the blank mark corresponding to the target meteorological grid area is still maintained; if the target meteorological grid has a value range, the meteorological warning threshold corresponding to each single source is compared with the meteorological elements in the corresponding unit square or volume area. If the meteorological elements in the corresponding unit square or volume area exceed the meteorological warning threshold, the unit square or volume area is marked as a warning area; otherwise, the unit square or volume area is marked as a non-warning area. Step S44: Calculating the warning occupancy rate based on the warning domain and the non-warning domain corresponding to the target area to obtain the meteorological element warning occupancy rate corresponding to the target area; Step S45: Based on the meteorological element warning occupancy rate corresponding to the target area, all the warning domains and non-warning domains that have been judged are filled and repeatedly laid into the network puzzle. If the meteorological element warning occupancy rate corresponding to the target area is lower than the preset threshold, the meteorological element laying mark and re-warning mark of step S3 are repeated. Otherwise, it stops immediately and outputs the multi-source meteorological data network puzzle result.

10. The multi-source meteorological data networking method supporting rapid meteorological warning according to claim 9 is characterized in that: Step S44 includes the following steps: Obtain the total area of ​​the target meteorological zone excluding the blank area of ​​the target meteorological grid in the target area; Obtain the corresponding total number of target warning domains and the total number of target non-warning domains through the warning domains and non-warning domains corresponding to the target area; The location information of a single meteorological element in each warning domain is obtained through the warning domain corresponding to the target area, and the location profile of the meteorological element in the corresponding warning domain is determined based on the location information of the single meteorological element in each warning domain, so as to obtain the size of the location profile corresponding to each single meteorological element in the warning domain; Determine the size of the location coverage area between the single meteorological elements in the warning domain based on the size of the location contour range corresponding to each single meteorological element in the warning domain, and perform total meteorological element coverage statistics based on the size of the location coverage area between the single meteorological elements in the warning domain combined with the size of the location contour range corresponding to each single meteorological element in the warning domain and sum up each warning domain to obtain the total meteorological element warning area size corresponding to the target area; Based on the total area of ​​the target meteorological zone, combined with the total number of target warning domains and the total number of target non-warning domains, the warning occupancy rate of the total area of ​​meteorological element warnings corresponding to the target area is calculated to obtain the meteorological element warning occupancy rate corresponding to the target area.

Images