High-resolution low-altitude meteorological data and short-term nowcasting methods, systems and equipment

Abstract

This application discloses a method, system, and device for high-resolution low-air meteorological data and short-term nowcasting, relating to the technical field of low-air meteorological assessment. The method includes: dividing the target area based on the similarity of meteorological physical processes and according to the underlying surface characteristics and the long-term climate background field used to characterize the long-term state of climate elements, obtaining multiple sub-regions and corresponding meteorological physical characteristics; for each sub-region, matching and selecting a suitable microscale meteorological model configuration scheme based on the meteorological physical characteristics of that sub-region; loading multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements into the invoked microscale meteorological model configuration scheme, and generating a high-resolution meteorological data field for each sub-region through rapid update cyclic assimilation and parallel computation. This application has the effect of improving forecast accuracy.

Description

Technical Field

[0001] This application relates to the technical field of low-air meteorological assessment, and in particular to high-resolution low-air meteorological data and short-term nowcasting methods, systems and equipment. Background Technology

[0002] Meteorological data analysis and short-term nowcasting play a crucial supporting role in many fields such as aviation, transportation, agriculture, energy, and disaster prevention and mitigation. With the increasing sophistication of socio-economic activities, related industries have an increasingly urgent need for high spatiotemporal resolution meteorological information covering the near-surface to low-altitude range, especially for forecast products that can accurately reflect the influence of complex local underlying surfaces and whose uncertainties are controllable.

[0003] Currently, obtaining high-resolution meteorological fields mainly relies on numerical weather prediction models. Existing common techniques typically involve increasing the grid resolution of global or regional models within a single, fixed model framework. However, this approach faces significant challenges when dealing with vast target areas with diverse geographical features: First, complex and diverse terrain and land cover types significantly alter local energy and material exchange processes, and traditional homogenization or coarse parameterization schemes struggle to accurately characterize these "microclimate" effects, leading to systematic biases in local weather forecast simulations. Second, globally covering the entire region with a high-resolution computational grid incurs enormous computational costs, limiting the update frequency of data assimilation and forecast cycles, making it difficult to meet the stringent timeliness requirements of short-term nowcasting.

[0004] In terms of data processing and initial field construction, existing methods typically integrate observational data from multiple sources, such as radar, satellites, ground stations, and radiosondes, into a single assimilation system for global fusion analysis. However, because the spatial representativeness and error characteristics of different observational data vary across different geographical environments, this "one-size-fits-all" assimilation strategy may not optimally integrate all information at the microscale, affecting the accuracy and rationality of the actual analysis field and introducing initial errors into subsequent forecasts.

[0005] Furthermore, the uncertainty in short-term nowcasting mainly stems from initial condition errors and imperfections in the models themselves. Traditional single deterministic forecasts or ensemble forecasts with a limited number of members sometimes struggle to adequately quantify this uncertainty. Consequently, the forecast products provided are insufficient in characterizing the probability of extreme weather events or the evolutionary possibilities of refined weather elements, failing to fully meet the needs of risk warning and decision-making.

[0006] Therefore, current technologies in high-resolution meteorological data construction and short-term nowcasting generally suffer from problems such as insufficient adaptability to local ground surface characteristics, contradiction between computational efficiency and update frequency, insufficient consideration of spatial heterogeneity in the utilization of multi-source observation data, and insufficient precision in quantifying forecast uncertainty, which restrict further improvement in forecast accuracy and product practicality. Summary of the Invention

[0007] To improve forecast accuracy, this application provides a high-resolution low-air meteorological data and short-term nowcasting method, system, and equipment.

[0008] Firstly, this application provides a high-resolution low-air meteorological data and short-term nowcasting method, employing the following technical solution: High-resolution low-air meteorological data and short-term nowcasting methods include: Based on the similarity of meteorological and physical processes, and according to the underlying surface characteristics of the target area and the long-term climate background field used to characterize the long-term state of climate elements, the target area is divided into multiple sub-regions and corresponding meteorological and physical characteristics. For each sub-region, a suitable microscale meteorological model configuration scheme is matched and selected based on the meteorological and physical characteristics of that sub-region. The multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements are respectively loaded into the microscale meteorological model configuration scheme called. Through rapid update cyclic assimilation and parallel computing, a high-resolution meteorological field of each sub-region is generated. By merging the high-resolution meteorological fields of all the aforementioned sub-regions, a unified high-resolution meteorological field for the target region is formed. Using the unified high-resolution meteorological field as the initial field, the microscale ensemble forecasting system is driven by the multi-source real-time meteorological observation data to generate multiple high-resolution meteorological forecast scenarios within a short-term near-term timeframe. Based on the aforementioned high-resolution weather forecast scenario, probabilistic high-resolution forecast products are generated.

[0009] By adopting the above technical solution, based on the similarity of meteorological and physical processes, and according to the underlying surface characteristics of the target area and the long-term climate field used to characterize the long-term state of climate elements, the target area is divided into multiple sub-regions and corresponding meteorological and physical characteristics. Then, for each sub-region, a suitable microscale meteorological model configuration scheme is matched and selected according to the meteorological and physical characteristics of that sub-region. Then, the multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements are loaded into the called microscale meteorological model configuration scheme respectively. Through rapid update cyclic assimilation and parallel computing, a high-resolution meteorological field of each sub-region is generated. Then, the high-resolution meteorological fields of all sub-regions are merged to form a unified high-resolution meteorological field of the target area. Then, using the unified high-resolution meteorological field as the initial field, the microscale ensemble forecasting system is driven by multi-source real-time meteorological observation data to generate multiple high-resolution meteorological forecast scenarios within the short-term near-term time. Finally, based on the high-resolution meteorological forecast scenarios, probabilistic high-resolution forecast products are generated. This invention constructs a complete chain from refined real-time data to probabilistic short-term and nowcasting, overcoming the problems of insufficient adaptability to local ground surface, low computational efficiency, coarse fusion of multi-source observations, and insufficient quantification of uncertainties in traditional high-resolution numerical forecasts. As a result, it improves the accuracy and geographical fit of near-surface to low-air meteorological field simulation, increases the update frequency of data assimilation and forecasts, and can provide UAVs with high-resolution forecast products with uncertainty information, achieving a synergistic enhancement of forecast accuracy, timeliness, and decision support capabilities.

[0010] Optionally, the step of dividing the target region into multiple sub-regions and corresponding meteorological and physical characteristics based on the similarity of meteorological and physical processes and according to the underlying surface characteristics of the target region and the long-term climate background field used to characterize the long-term state of climate elements includes: Based on a preset resolution, the target area is discretized into multiple geographic grid points; Based on the long-term climate background field, the corresponding climate background parameters are obtained for each geographic grid point through spatial interpolation. The long-term climate background field is a spatial distribution data field used to characterize the long-term average state of climate elements. The climate background parameters include at least the average wind field parameters, average temperature field parameters, average humidity field parameters, and precipitation climatology parameters. For each geographic grid point, based on the underlying surface characteristics and the climate background parameters of that geographic grid point, a set of key influencing factors are determined through a physical diagnostic algorithm. The key influencing factors include at least dynamic forcing factors, thermal non-uniformity factors, and local circulation climate factors. Based on the aforementioned set of key influencing factors, spatial clustering is performed on all the aforementioned geographic grid points to form multiple sub-regions; For each sub-region, the meteorological and physical characteristics of the sub-region are determined based on the central value and distribution characteristics of the sub-region.

[0011] By adopting the above technical solution, in order to achieve the division of the target area, the target area is discretized into multiple geographic grid points based on a preset resolution. Then, based on the long-term climate background field, the corresponding climate background parameters are obtained for each geographic grid point through spatial interpolation. The long-term climate background field is a spatial distribution data field used to characterize the long-term average state of climate elements. The climate background parameters include at least the mean wind field parameters, mean temperature field parameters, mean humidity field parameters, and precipitation climatology parameters. Then, for each geographic grid point, based on the underlying surface characteristics and climate background parameters of that geographic grid point, a set of key influencing factors is determined through a physical diagnostic algorithm. The key influencing factors include at least the dynamic forcing factor, the thermal non-uniformity factor, and the local circulation climate factor. Then, based on the set of key influencing factors, all geographic grid points are spatially clustered to form multiple sub-regions. Then, for each sub-region, the meteorological and physical characteristics of that sub-region are determined based on the central value and distribution characteristics of that sub-region.

[0012] Optionally, the step of matching and selecting a suitable microscale meteorological model configuration scheme based on the meteorological and physical characteristics of each sub-region includes: A microscale meteorological model configuration scheme library is constructed, wherein the microscale meteorological model configuration scheme library defines corresponding core configuration parameters for each typical meteorological scenario. The core configuration parameters include physical parameterization scheme, grid resolution, boundary layer scheme and initial perturbation scheme. For each sub-region, the similarity between the meteorological and physical characteristics of the sub-region and each typical meteorological scenario in the microscale meteorological model configuration scheme library is calculated to determine multiple candidate schemes; For each sub-region, based on the historical high-resolution observation data or reanalysis data of the sub-region, a rapid historical back-calculation verification is performed on each of the candidate schemes, and the optimal adaptation scheme for the sub-region is determined from the candidate schemes. For each sub-region, based on the real-time meteorological conditions of that sub-region, the core configuration parameters of the optimal adaptation scheme corresponding to that sub-region are dynamically adjusted to generate the corresponding microscale meteorological model configuration scheme.

[0013] By adopting the above technical solution, in order to select microscale meteorological model configuration schemes for each sub-region, a microscale meteorological model configuration scheme library is constructed. This library defines corresponding core configuration parameters for each typical meteorological scenario, including physical parameterization scheme, grid resolution, boundary layer scheme, and initial perturbation scheme. Then, for each sub-region, the similarity between the meteorological physical characteristics of that sub-region and each typical meteorological scenario in the microscale meteorological model configuration scheme library is calculated to determine multiple candidate schemes. Next, for each sub-region, based on the historical high-resolution observation data or reanalysis data of that sub-region, rapid historical back-calculation verification is performed on each candidate scheme to determine the optimal fit scheme for that sub-region. Finally, for each sub-region, based on the real-time meteorological conditions of that sub-region, the core configuration parameters of the corresponding optimal fit scheme are dynamically adjusted to generate the corresponding microscale meteorological model configuration scheme.

[0014] Optionally, the step of loading multi-source real-time meteorological observation data and real-time climate background fields of each of the sub-regions into the called microscale meteorological model configuration scheme, and generating high-resolution meteorological fields of each of the sub-regions through rapid update cyclic assimilation and parallel computing, includes: Multi-source real-time meteorological observation data is received and assimilated in a preset high-frequency update cycle to obtain a three-dimensional incremental field. The multi-source real-time meteorological observation data includes ground meteorological station data, wind profiler radar data, GPS water vapor detection data, and meteorological satellite data. For each sub-region, the three-dimensional incremental field and the real-time climate background field are jointly loaded into the microscale meteorological model configuration scheme called by that sub-region, wherein the real-time climate background field is a spatial distribution data field used to characterize the recent average state of climate elements. Using independent computing nodes allocated to each sub-region, the rapid integral calculation of the microscale meteorological model configuration scheme is performed in parallel to generate high-resolution meteorological fields for each sub-region, wherein the meteorological fields include wind field, temperature field, humidity field and pressure field.

[0015] By adopting the above technical solution, in order to generate high-resolution meteorological fields for each sub-region, multi-source real-time meteorological observation data is received and assimilated cyclically at a preset high-frequency update cycle to obtain a three-dimensional incremental field. The multi-source real-time meteorological observation data includes ground meteorological station data, wind profiler radar data, GPS water vapor detection data, and meteorological satellite data. Then, for each sub-region, the three-dimensional incremental field and the real-time climate background field are jointly loaded into the microscale meteorological model configuration scheme called by that sub-region. The real-time climate background field is a spatial distribution data field used to characterize the recent average state of climate elements. Then, using the independent computing nodes allocated to each sub-region, the rapid integration calculation of the microscale meteorological model configuration scheme is executed in parallel to generate the high-resolution meteorological fields for each sub-region. The meteorological fields include wind field, temperature field, humidity field, and pressure field.

[0016] Optionally, the step of fusing the high-resolution meteorological fields of all the sub-regions to form a unified high-resolution meteorological field of the target region includes: A weighting function based on physical constraints is used to fuse the high-resolution meteorological field in the overlapping area of ​​adjacent sub-regions to obtain a preliminary fused field. The weighting function is dynamically adjusted according to the forecast uncertainty or topographic representativeness of each sub-region at the boundary. By applying conservation constraints on key physical quantities to the preliminary fusion field, a coordinated field is generated, wherein the key physical quantities include at least one of mass, momentum and water vapor flux. The coordinated fields of all the sub-regions are stitched together to form a unified high-resolution meteorological field for the target region.

[0017] By adopting the above technical solution, in order to generate a unified high-resolution meteorological field, a weight function based on physical constraints is used to fuse the high-resolution meteorological fields in the overlapping areas of adjacent sub-regions to obtain a preliminary fused field. The weight function is dynamically adjusted according to the forecast uncertainty or topographic representativeness of each sub-region at the boundary. Then, conservation constraints of key physical quantities are applied to the preliminary fused field to generate a coordinated field. The key physical quantities include at least one of mass, momentum and water vapor flux. Finally, the coordinated fields of all sub-regions are spliced ​​together to form a unified high-resolution meteorological field for the target region.

[0018] Optionally, the step of generating multiple high-resolution meteorological forecast scenarios within a short-term, near-term timeframe by using the unified high-resolution meteorological field as the initial field and driving a microscale ensemble forecasting system based on the multi-source real-time meteorological observation data includes: A set of perturbation initial fields is generated by applying a multi-scale hybrid initial value perturbation and physical process perturbation to the unified high-resolution meteorological field. Based on the set of perturbation initial fields, multiple ensemble forecast members are constructed, wherein each of the ensemble forecast members corresponds to one of the perturbation initial fields; The microscale ensemble forecasting system, based on the microscale meteorological model configuration scheme of each sub-region, drives each ensemble forecasting member to perform short-term near-term integration in parallel to generate multiple high-resolution meteorological forecast scenarios. During the integration process, based on the real-time weather characteristics derived from the assimilation analysis of the multi-source real-time meteorological observation data, the disturbance physical parameterization schemes of each forecast member in the predefined set of disturbance physical parameterization schemes are selectively enabled or switched.

[0019] By adopting the above technical solution, in order to generate high-resolution weather forecast scenarios, a multi-scale hybrid initial value perturbation and physical process perturbation method is used to apply perturbation to a unified high-resolution weather field, generating a set of perturbation initial fields. Then, based on a set of perturbation initial fields, multiple ensemble forecast members are constructed, where each ensemble forecast member corresponds to a perturbation initial field. Then, through a microscale ensemble forecast system, based on the microscale meteorological model configuration scheme of each sub-region, each ensemble forecast member is driven in parallel to perform short-term near-term integration, generating multiple high-resolution weather forecast scenarios. During the integration process, based on the real-time weather characteristics obtained from the assimilation analysis of multi-source real-time meteorological observation data, the perturbation physical parameterization schemes of each ensemble forecast member are selectively enabled or switched from a predefined set of perturbation physical parameterization schemes.

[0020] Optionally, the step of generating probabilistic high-resolution forecast products based on the high-resolution weather forecast scenario includes: The high-resolution weather forecast scenarios are aggregated and statistically analyzed to calculate the probability distribution of key meteorological variables at each of the geographic grid points. Based on the probability distribution, spatially continuous probability forecast products are generated for at least one preset meteorological threshold. Based on historical fitting statistical correction methods and spatial smoothing algorithms, the probabilistic forecast products are processed to generate probabilistic high-resolution forecast products.

[0021] By adopting the above technical solution, in order to generate high-resolution forecast products, the high-resolution meteorological forecast scenarios are aggregated and statistically analyzed to calculate the probability distribution of key meteorological variables at each geographic grid point. Then, based on the probability distribution, spatially continuous probability forecast products are generated for at least one preset meteorological threshold. Finally, based on the statistical correction method of historical fitting and the spatial smoothing algorithm, the probability forecast products are processed to generate probabilistic high-resolution forecast products.

[0022] Optionally, the method further includes: Based on at least one preset meteorological disaster early warning threshold, a visualized risk level area map and early warning information are generated according to the high-resolution forecast product.

[0023] By adopting the above technical solution, in order to achieve visualized early warning, based on at least one preset meteorological disaster early warning threshold, a visualized risk level area map and early warning information are generated according to high-resolution forecast products.

[0024] Secondly, this application also provides a high-resolution low-air meteorological data and short-term nowcasting system, which adopts the following technical solution: High-resolution low-air weather data and short-term nowcasting system, including: The segmentation module is used to segment the target area based on the similarity of meteorological and physical processes and according to the underlying surface characteristics of the target area and the long-term climate background field used to characterize the long-term state of climate elements, to obtain multiple sub-regions and corresponding meteorological and physical characteristics. The configuration scheme generation module is used to match and select an appropriate microscale meteorological model configuration scheme for each sub-region based on the meteorological and physical characteristics of that sub-region. The meteorological field generation module is used to load the multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements into the called microscale meteorological model configuration scheme, and generate high-resolution meteorological fields of each sub-region through rapid update cyclic assimilation and parallel computing. The fusion module is used to fuse the high-resolution meteorological fields of all the sub-regions to form a unified high-resolution meteorological field of the target region; The weather forecast scenario generation module is used to generate multiple high-resolution weather forecast scenarios within a short-term, near-term timeframe by using the unified high-resolution meteorological field as the initial field and driving the microscale ensemble forecast system based on the multi-source real-time meteorological observation data. The forecast product generation module is used to generate probabilistic high-resolution forecast products based on the high-resolution meteorological forecast scenario.

[0025] Thirdly, this application also provides a computer device, which adopts the following technical solution: A computer device includes a memory and a processor, the memory storing a computer program executable on the processor, the processor executing the computer program to implement the method described in the first aspect.

[0026] In summary, this application includes at least the following beneficial technical effects: Based on the similarity of meteorological and physical processes, and according to the underlying surface characteristics of the target area and the long-term climate field used to characterize the long-term state of climate elements, the target area is divided into multiple sub-regions and corresponding meteorological and physical characteristics. Then, for each sub-region, a suitable microscale meteorological model configuration scheme is matched and selected according to the meteorological and physical characteristics of that sub-region. Then, the multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements are loaded into the called microscale meteorological model configuration scheme respectively. Through rapid update cyclic assimilation and parallel computing, a high-resolution meteorological field of each sub-region is generated. Then, the high-resolution meteorological fields of all sub-regions are merged to form a unified high-resolution meteorological field of the target area. Then, using the unified high-resolution meteorological field as the initial field, the microscale ensemble forecasting system is driven by multi-source real-time meteorological observation data to generate multiple high-resolution meteorological forecast scenarios within the short-term near-term time. Finally, based on the high-resolution meteorological forecast scenarios, a probabilistic high-resolution forecast product is generated. This invention constructs a complete chain from refined real-time data to probabilistic short-term and nowcasting, overcoming the problems of insufficient adaptability to local ground surface, low computational efficiency, coarse fusion of multi-source observations, and insufficient quantification of uncertainties in traditional high-resolution numerical forecasts. As a result, it improves the accuracy and geographical fit of near-surface to low-air meteorological field simulation, increases the update frequency of data assimilation and forecasts, and can provide UAVs with high-resolution forecast products with uncertainty information, achieving a synergistic enhancement of forecast accuracy, timeliness, and decision support capabilities. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the overall process of an embodiment of this application.

[0028] Figure 2 This is a schematic diagram of the system structure of this application.

[0029] Figure 3 This is a structural block diagram of the computer device described in this application. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0031] This application discloses a high-resolution low-air meteorological data and short-term nowcasting method.

[0032] Reference Figure 1 High-resolution low-air meteorological data and short-term nowcasting methods include: Step S11: Based on the similarity of meteorological and physical processes, and according to the underlying surface characteristics of the target area and the long-term climate background field used to characterize the long-term state of climate elements, the target area is divided into multiple sub-regions and corresponding meteorological and physical characteristics.

[0033] It should be noted that step S11 addresses the issue of spatial heterogeneity of meteorological conditions within a large target area. By analyzing key physical factors influencing weather change processes (such as underlying surface characteristics composed of topography, altitude, vegetation, and water distribution) and long-term climate conditions (i.e., the climate background field), and based on the principle of similarity in meteorological physical processes, the complex large area is scientifically divided into multiple relatively homogeneous sub-regions. This division ensures greater consistency in the meteorological driving mechanisms and change patterns within each sub-region, providing support for the subsequent selection of targeted meteorological models and avoiding errors caused by a "one-size-fits-all" simulation.

[0034] Step S12: For each sub-region, match and select a suitable microscale meteorological model configuration scheme based on the meteorological and physical characteristics of that sub-region.

[0035] It should be noted that after the regional division is completed, step S12 is used to adapt the corresponding simulation scheme to each sub-region. Since the dominant physical processes in different sub-regions are different (for example, mountainous areas may be dominated by topographic dynamics and thermal effects, while urban areas need to focus on anthropogenic heat sources and roughness effects), it is necessary to match and select the most suitable microscale meteorological model configuration from the model library based on its unique meteorological and physical characteristics. This includes selecting appropriate parameterization schemes, physical process modules, and spatial resolution settings to ensure that the model can best characterize the key meteorological processes of the sub-region, thereby improving the accuracy and physical rationality of the simulation.

[0036] Step S13: Load the multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements into the called microscale meteorological model configuration scheme respectively. Through rapid update cyclic assimilation and parallel computing, generate high-resolution meteorological field of each sub-region.

[0037] It should be noted that in step S13, for each sub-region, its corresponding multi-source real-time observation data (such as data from ground stations, radar, satellites, etc.) and recent climate conditions (i.e., real-time climate background field) are input into the customized microscale meteorological model called in step S12. Through rapid update cyclic assimilation technology, the model can continuously incorporate the latest observation information into its operation, constantly correct the simulation state, and make it closer to reality. Combined with parallel computing technology, multiple sub-regions are processed simultaneously, and finally, high spatiotemporal resolution meteorological fields (such as detailed temperature, wind field, humidity field, etc.) are generated efficiently and synchronously for each sub-region.

[0038] Step S14: Merge the high-resolution meteorological fields of all sub-regions to form a unified high-resolution meteorological field of the target region.

[0039] It should be noted that since step S13 generates meteorological fields for multiple independent sub-regions, these fields may be discontinuous or overlapping at the boundaries of the sub-regions. The task of step S14 is to perform seamless stitching and fusion. Specifically, using data fusion algorithms, the high-resolution meteorological fields of all sub-regions are processed through boundary coordination, weighted averaging of overlapping areas, etc., to eliminate stitching traces and integrate them into a unified high-resolution meteorological field that covers the entire target area and has continuous and consistent physical quantities. This unified field integrates the detailed features of each sub-region and is the accurate initial field for subsequent forecasting work.

[0040] Step S15: Using a unified high-resolution meteorological field as the initial field, the microscale ensemble forecasting system is driven by multi-source real-time meteorological observation data to generate multiple high-resolution meteorological forecast scenarios within the short-term near-term timeframe.

[0041] It should be noted that step S15 transitions from real-time analysis to short-term nowcasting. The unified high-resolution real-time field obtained in step S14 is used as the initial field, which contains the most accurate spatial detail information currently available. Meanwhile, continuously input multi-source real-time observation data provides dynamic constraints for the model. Utilizing a microscale ensemble forecasting system, by introducing small perturbations to the initial field, physical processes, or boundary conditions, multiple slightly different forecast scenarios are run in parallel, thereby generating a set of high-resolution forecast scenarios within a short-term, near-term (e.g., within 1 hour). This ensemble approach can characterize forecast uncertainties and provide a sample basis for probabilistic forecasts. Furthermore, in this invention, the microscale ensemble forecasting system includes multiple parallel-running forecast member instances. The operation of each forecast member instance is defined and controlled by its corresponding microscale meteorological model configuration scheme. Specifically, the microscale meteorological model configuration scheme provides the microscale ensemble forecasting system with core operating parameters for a specific sub-region. The microscale ensemble forecasting system initializes and executes the numerical integration calculation of each forecast member instance based on the corresponding core operating parameters. The relationship between the two is that the microscale meteorological model configuration scheme defines the behavioral pattern of the model for each local region or ensemble member, while the microscale ensemble forecasting system is the engine that executes these behavioral patterns and manages their integrated computation.

[0042] Step S16: Generate probabilistic high-resolution forecast products based on high-resolution weather forecast scenarios.

[0043] It should be noted that step S16 aims to transform deterministic ensemble forecast scenarios into probabilistic products that are more conducive to risk decision-making. By statistically analyzing multiple high-resolution forecast scenarios generated in step S15, the probability of various meteorological events (such as rainfall exceeding a certain threshold, strong winds, etc.) occurring can be calculated, ultimately generating probabilistic forecast products, such as "precipitation probability maps" and "severe convection probability maps." These products intuitively display the probability of different weather outcomes and can provide users with more valuable risk warning information that goes beyond a single deterministic forecast.

[0044] In the above implementation, based on the similarity of meteorological and physical processes, and according to the underlying surface characteristics of the target area and the long-term climate field used to characterize the long-term state of climate elements, the target area is divided into multiple sub-regions and corresponding meteorological and physical characteristics. Then, for each sub-region, a suitable microscale meteorological model configuration scheme is matched and selected according to the meteorological and physical characteristics of that sub-region. Then, the multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements are loaded into the called microscale meteorological model configuration scheme respectively. Through rapid update cyclic assimilation and parallel computing, a high-resolution meteorological field of each sub-region is generated. Then, the high-resolution meteorological fields of all sub-regions are merged to form a unified high-resolution meteorological field of the target area. Then, using the unified high-resolution meteorological field as the initial field, the microscale ensemble forecast system is driven by multi-source real-time meteorological observation data to generate multiple high-resolution meteorological forecast scenarios within the short-term near-term time. Then, based on the high-resolution meteorological forecast scenarios, a probabilistic high-resolution forecast product is generated. This invention constructs a complete chain from refined real-time data to probabilistic short-term and nowcasting, overcoming the problems of insufficient adaptability to local ground surface, low computational efficiency, coarse fusion of multi-source observations, and insufficient quantification of uncertainties in traditional high-resolution numerical forecasts. As a result, it improves the accuracy and geographical fit of near-surface to low-air meteorological field simulation, increases the update frequency of data assimilation and forecasts, and can provide UAVs with high-resolution forecast products with uncertainty information, achieving a synergistic enhancement of forecast accuracy, timeliness, and decision support capabilities.

[0045] As a further implementation of the method, the steps of dividing the target area into multiple sub-regions and corresponding meteorological and physical characteristics based on the similarity of meteorological and physical processes and according to the underlying surface characteristics of the target area and the long-term climate background field used to characterize the long-term state of climate elements include: Step S21: Based on the preset resolution, the target area is discretized into multiple geographic grid points.

[0046] Step S22: Based on the long-term climate background field, obtain the corresponding climate background parameters for each geographic grid point through spatial interpolation. The long-term climate background field is a spatial distribution data field used to characterize the long-term average state of climate elements. The climate background parameters include at least the mean wind field parameters, mean temperature field parameters, mean humidity field parameters, and precipitation climatology parameters.

[0047] Step S23: For each geographic grid point, based on the underlying surface characteristics and climate background parameters of that geographic grid point, a set of key influencing factors are determined through a physical diagnostic algorithm. The key influencing factors include at least dynamic forcing factors, thermal heterogeneity factors, and local circulation climate factors.

[0048] Step S24: Based on a set of key influencing factors, spatial clustering is performed on all geographic grid points to form multiple sub-regions.

[0049] Step S25: For each sub-region, determine the meteorological and physical characteristics of the sub-region based on the central value and distribution characteristics of the sub-region.

[0050] In the above implementation, in order to achieve the division of the target area, the target area is discretized into multiple geographic grid points based on a preset resolution. Then, based on the long-term climate background field, the corresponding climate background parameters are obtained for each geographic grid point through spatial interpolation. The long-term climate background field is a spatial distribution data field used to characterize the long-term average state of climate elements. The climate background parameters include at least the mean wind field parameters, mean temperature field parameters, mean humidity field parameters, and precipitation climatology parameters. Then, for each geographic grid point, based on the underlying surface characteristics and climate background parameters of the geographic grid point, a set of key influencing factors is determined through a physical diagnostic algorithm. The key influencing factors include at least the dynamic forcing factor, the thermal non-uniformity factor, and the local circulation climate factor. Then, based on the set of key influencing factors, all geographic grid points are spatially clustered to form multiple sub-regions. Then, for each sub-region, the meteorological and physical characteristics of the sub-region are determined based on the central value and distribution characteristics of the sub-region.

[0051] As a further implementation of the method, the step of matching and selecting a suitable microscale meteorological model configuration scheme based on the meteorological and physical characteristics of each sub-region includes: Step S31: Construct a microscale meteorological model configuration scheme library. The microscale meteorological model configuration scheme library defines corresponding core configuration parameters for each typical meteorological scenario. The core configuration parameters include physical parameterization scheme, grid resolution, boundary layer scheme, and initial perturbation scheme.

[0052] It should be noted that step S31 creates a pre-defined, standardized scheme library. This scheme library predefines various typical weather or climate scenarios (e.g., complex mountainous terrain, areas with significant urban heat island effects, coastal areas affected by sea and land breezes, etc.) and configures a set of validated or optimized core model parameters for each scenario. These parameters directly determine the physical processes and computational behavior of the model, specifically including: physical parameterization scheme, i.e., how to simulate processes that cannot be directly analyzed, such as cloud microphysics and radiation; grid resolution, i.e., the level of detail in the model simulation; boundary layer scheme, i.e., how to characterize near-surface turbulence and energy exchange; and initial perturbation scheme, i.e., how to generate meaningful initial condition differences for ensemble forecasts.

[0053] Step S32: For each sub-region, calculate the similarity between the meteorological and physical characteristics of the sub-region and each typical meteorological scenario in the microscale meteorological model configuration scheme library, and determine multiple candidate schemes.

[0054] It should be noted that in step S32, for each divided sub-region, the meteorological and physical characteristics of that region are quantitatively compared with the characteristics of various typical scenarios in the scheme library. By calculating the similarity, multiple model configuration schemes that are close to the characteristics of the current sub-region and are theoretically applicable are selected from the library.

[0055] Step S33: For each sub-region, based on the historical high-resolution observation data or reanalysis data of that sub-region, perform rapid historical back-calculation verification on each candidate scheme, and determine the optimal fit scheme for that sub-region from the candidate schemes.

[0056] It should be noted that, in order to select the best-performing option from multiple theoretically feasible candidate schemes, the system uses historical high-precision observation data or reanalysis data of the sub-region over a past period as a basis to drive each candidate scheme to perform short-term historical time-period simulations (i.e., rapid backcalculation). By comparing the degree of agreement between the simulation results of each scheme and the real historical data (such as the magnitude of error, the ability to characterize key processes, etc.), the system objectively evaluates its actual performance in the specific sub-region, thereby scientifically selecting an optimal, locally adapted model configuration scheme.

[0057] Step S34: For each sub-region, based on the real-time meteorological conditions of that sub-region, the core configuration parameters of the optimal adaptation scheme corresponding to that sub-region are dynamically adjusted to generate the corresponding microscale meteorological model configuration scheme.

[0058] It should be noted that even if the historically validated optimal scheme is selected, current weather conditions (such as extreme heat or heavy precipitation events) may differ from historical averages. Therefore, the system monitors the meteorological conditions of the sub-region in real time (such as temperature gradient, humidity, and stability) and fine-tunes some core parameters of the selected optimal scheme accordingly (e.g., adjusting the mixing length in boundary layer parameterization or fine-tuning microphysical scheme parameters based on cloud water content). This allows the model configuration to adapt not only to the regional climate but also to real-time weather, further improving the accuracy and flexibility of the simulation.

[0059] In the above implementation, to select microscale meteorological model configuration schemes for each sub-region, a microscale meteorological model configuration scheme library is constructed. This library defines corresponding core configuration parameters for each typical meteorological scenario, including physical parameterization scheme, grid resolution, boundary layer scheme, and initial perturbation scheme. Then, for each sub-region, the similarity between the meteorological physical characteristics of that sub-region and each typical meteorological scenario in the microscale meteorological model configuration scheme library is calculated to determine multiple candidate schemes. Next, for each sub-region, based on historical high-resolution observation data or reanalysis data, rapid historical back-calculation verification is performed on each candidate scheme to determine the optimal fit scheme for that sub-region. Finally, for each sub-region, based on the real-time meteorological conditions of that sub-region, the core configuration parameters of the corresponding optimal fit scheme are dynamically adjusted to generate the corresponding microscale meteorological model configuration scheme.

[0060] As a further implementation of the method, the steps of loading multi-source real-time meteorological observation data and real-time climate background fields of each sub-region into the called microscale meteorological model configuration scheme, and generating high-resolution meteorological fields of each sub-region through rapid update cyclic assimilation and parallel computing, include: Step S41: Receive and assimilate multi-source real-time meteorological observation data in a cyclical manner with a preset high-frequency update cycle to obtain a three-dimensional incremental field. The multi-source real-time meteorological observation data includes ground meteorological station data, wind profiler radar data, GPS water vapor detection data, and meteorological satellite data.

[0061] It should be noted that step S41, through a preset high-frequency update cycle (e.g., minutes), cyclically receives and assimilates real-time meteorological observation data from multiple sources, including ground-based meteorological stations, wind profiler radar, GPS water vapor detection, and meteorological satellites, to generate a three-dimensional incremental field. This process utilizes advanced data assimilation technology to fuse observational information at different spatiotemporal resolutions and correct the background field, continuously capturing the rapid evolution details of atmospheric conditions. This provides subsequent models with key information reflecting the differences between the latest observations and short-term forecasts, thereby ensuring the timeliness and accuracy of the actual field.

[0062] Step S42: For each sub-region, the three-dimensional incremental field and the real-time climate background field are jointly loaded into the microscale meteorological model configuration scheme called for that sub-region, wherein the real-time climate background field is a spatial distribution data field used to characterize the recent average state of climate elements.

[0063] It should be noted that in step S42, for each divided sub-region, the three-dimensional incremental field generated in step S41 and the real-time climate background field representing the recent average state of climate elements are jointly loaded into the microscale meteorological model scheme called and configured for that sub-region. This process essentially provides optimal input for model initialization. The incremental field injects real-time weather disturbance information, while the real-time climate background field provides stable recent climate state constraints. The combination of the two and their matching with the model configuration (such as the physical parameterization scheme) adapted to the sub-region generates a coordinated and refined initial field for the next step of rapid integration.

[0064] Step S43: Using the independent computing nodes allocated to each sub-region, perform fast integral calculations of the microscale meteorological model configuration scheme in parallel to generate high-resolution meteorological fields for each sub-region. The meteorological fields include wind field, temperature field, humidity field and pressure field.

[0065] It should be noted that step S43 utilizes independent computing nodes allocated to each sub-region to perform rapid integration calculations of each microscale meteorological model configuration scheme in parallel. This parallel architecture based on physically similar sub-regions greatly improves computational efficiency, enabling the system to complete high-resolution numerical integration of multiple sub-regions in a short time, thereby independently generating high-resolution meteorological fields for each sub-region, including key elements such as wind field, temperature field, humidity field, and pressure field.

[0066] In the above implementation, in order to generate high-resolution meteorological fields for each sub-region, multi-source real-time meteorological observation data are received and assimilated cyclically at a preset high-frequency update cycle to obtain a three-dimensional incremental field. The multi-source real-time meteorological observation data includes ground meteorological station data, wind profiler radar data, GPS water vapor detection data, and meteorological satellite data. Then, for each sub-region, the three-dimensional incremental field and the real-time climate background field are loaded together into the microscale meteorological model configuration scheme called by that sub-region. The real-time climate background field is a spatial distribution data field used to characterize the recent average state of climate elements. Then, using the independent computing nodes allocated to each sub-region, the fast integration calculation of the microscale meteorological model configuration scheme is executed in parallel to generate high-resolution meteorological fields for each sub-region. The meteorological fields include wind field, temperature field, humidity field, and pressure field.

[0067] As a further implementation of the method, the step of fusing high-resolution meteorological fields from all sub-regions to form a unified high-resolution meteorological field for the target region includes: Step S51: Based on the weighting function of physical constraints, the high-resolution meteorological field in the boundary overlapping area of ​​adjacent sub-regions is fused to obtain a preliminary fused field. The weighting function is dynamically adjusted according to the forecast uncertainty or topographic representativeness of each sub-region at the boundary.

[0068] Step S52: Apply conservation constraints on key physical quantities to the preliminary fusion field to generate a coordinated field, wherein the key physical quantities include at least one of mass, momentum and water vapor flux.

[0069] Step S53: The coordinated fields of all sub-regions are stitched together to form a unified high-resolution meteorological field for the target region.

[0070] In the above implementation, in order to generate a unified high-resolution meteorological field, a weight function based on physical constraints is used to fuse the high-resolution meteorological fields in the overlapping areas of adjacent sub-regions to obtain a preliminary fused field. The weight function is dynamically adjusted according to the forecast uncertainty or topographic representativeness of each sub-region at the boundary. Then, conservation constraints of key physical quantities are applied to the preliminary fused field to generate a coordinated field. The key physical quantities include at least one of mass, momentum and water vapor flux. Finally, the coordinated fields of all sub-regions are spliced ​​together to form a unified high-resolution meteorological field of the target region.

[0071] As a further implementation of the method, the steps of generating multiple high-resolution meteorological forecast scenarios within a short-term near-term timeframe, using a unified high-resolution meteorological field as the initial field and driving a microscale ensemble forecasting system based on multi-source real-time meteorological observation data, include: Step S61: A set of perturbation initial fields is generated by applying a multi-scale hybrid initial value perturbation and physical process perturbation to the unified high-resolution meteorological field.

[0072] Step S62: Based on a set of perturbation initial fields, construct multiple ensemble forecast members, where each ensemble forecast member corresponds to a perturbation initial field.

[0073] Step S63: Through the microscale ensemble forecasting system, based on the microscale meteorological model configuration scheme of each sub-region, the ensemble forecasting members are driven in parallel to perform short-term near-term integration, generating multiple high-resolution meteorological forecast scenarios.

[0074] Step S64: During the integration process, based on the real-time weather characteristics obtained from the assimilation analysis of multi-source real-time meteorological observation data, the disturbance physical parameterization schemes of each set forecast member are selectively enabled or switched from the predefined set of disturbance physical parameterization schemes.

[0075] In the above implementation, in order to generate high-resolution weather forecast scenarios, a multi-scale hybrid initial value perturbation and physical process perturbation method is adopted to apply perturbation to a unified high-resolution weather field, generating a set of perturbation initial fields. Then, based on a set of perturbation initial fields, multiple ensemble forecast members are constructed, where each ensemble forecast member corresponds to a perturbation initial field. Then, through a microscale ensemble forecast system, based on the microscale meteorological model configuration scheme of each sub-region, each ensemble forecast member is driven in parallel to perform short-term near-term integration, generating multiple high-resolution weather forecast scenarios. During the integration process, based on the real-time weather characteristics obtained from the assimilation analysis of multi-source real-time meteorological observation data, the perturbation physical parameterization schemes of each ensemble forecast member are selectively enabled or switched from a predefined set of perturbation physical parameterization schemes.

[0076] As a further implementation of the method, the step of generating probabilistic high-resolution forecast products based on high-resolution weather forecast scenarios includes: Step S71: Perform ensemble statistics on the high-resolution weather forecast scenarios and calculate the probability distribution of key weather variables at each geographic grid point.

[0077] Step S72: Based on the probability distribution, generate a spatially continuous probability forecast product for at least one preset meteorological threshold.

[0078] Step S73: Based on the statistical correction method and spatial smoothing algorithm of historical fitting, the probabilistic forecast products are processed to generate probabilistic high-resolution forecast products.

[0079] In the above implementation, in order to generate high-resolution forecast products, the high-resolution weather forecast scenarios are aggregated and statistically analyzed to calculate the probability distribution of key meteorological variables at each geographic grid point. Then, based on the probability distribution, spatially continuous probability forecast products are generated for at least one preset meteorological threshold. Finally, based on the statistical correction method of historical fitting and the spatial smoothing algorithm, the probability forecast products are processed to generate probabilistic high-resolution forecast products.

[0080] As a further implementation of the method, the method also includes: Based on at least one preset meteorological disaster early warning threshold, a visualized risk level area map and early warning information are generated according to high-resolution forecast products.

[0081] In the above implementation, in order to achieve visualized early warning, a visualized risk level area map and early warning information are generated based on at least one preset meteorological disaster early warning threshold and high-resolution forecast products.

[0082] This application also discloses a high-resolution low-air meteorological data and short-term nowcasting system.

[0083] refer to Figure 2 High-resolution low-air weather data and short-term nowcasting system, including: The segmentation module is used to segment the target area based on the similarity of meteorological and physical processes and according to the underlying surface characteristics of the target area and the long-term climate background field used to characterize the long-term state of climate elements, resulting in multiple sub-regions and corresponding meteorological and physical characteristics. The configuration scheme generation module is used to match and select a suitable microscale meteorological model configuration scheme for each sub-region based on the meteorological and physical characteristics of that sub-region. The meteorological field generation module is used to load multi-source real-time meteorological observation data of each sub-region and real-time climate background field used to characterize the recent state of climate elements into the called microscale meteorological model configuration scheme, and generate high-resolution meteorological fields of each sub-region through rapid update cyclic assimilation and parallel computing. The fusion module is used to fuse high-resolution meteorological fields from all sub-regions to form a unified high-resolution meteorological field for the target region. The weather forecast scenario generation module is used to generate multiple high-resolution weather forecast scenarios within a short-term, near-term timeframe by using a unified high-resolution real-time weather field as the initial field and driving a microscale ensemble forecast system based on multi-source real-time weather observation data. The forecast product generation module is used to generate probabilistic high-resolution forecast products based on high-resolution weather forecast scenarios.

[0084] The high-resolution low-air meteorological data and short-term nowcasting system of the present invention can implement any of the methods in the high-resolution low-air meteorological data and short-term nowcasting method, and the specific working process of the high-resolution low-air meteorological data and short-term nowcasting system of the present invention can refer to the corresponding process in the above-mentioned high-resolution low-air meteorological data and short-term nowcasting method.

[0085] This application also discloses a computer device.

[0086] refer to Figure 3 A computer device includes a memory and a processor, the memory storing a computer program that can run on the processor, the processor executing the computer program to implement any of the above-described methods for high-resolution low-air weather conditions and short-term nowcasting.

[0087] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A high-resolution method for low-altitude meteorological observation and short-term nowcasting, characterized in that, include: Based on the similarity of meteorological and physical processes, and according to the underlying surface characteristics of the target area and the long-term climate background field used to characterize the long-term state of climate elements, the target area is divided into multiple sub-regions and corresponding meteorological and physical characteristics. For each sub-region, a suitable microscale meteorological model configuration scheme is matched and selected based on the meteorological and physical characteristics of that sub-region. The multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements are respectively loaded into the microscale meteorological model configuration scheme called. Through rapid update cyclic assimilation and parallel computing, a high-resolution meteorological field of each sub-region is generated. By merging the high-resolution meteorological fields of all the aforementioned sub-regions, a unified high-resolution meteorological field for the target region is formed. Using the unified high-resolution meteorological field as the initial field, the microscale ensemble forecasting system is driven by the multi-source real-time meteorological observation data to generate multiple high-resolution meteorological forecast scenarios within a short-term near-term timeframe. Based on the aforementioned high-resolution weather forecast scenario, a probabilistic high-resolution forecast product is generated. The step of matching and selecting a suitable microscale meteorological model configuration scheme for each sub-region based on the meteorological and physical characteristics of that sub-region includes: A microscale meteorological model configuration scheme library is constructed, wherein the microscale meteorological model configuration scheme library defines corresponding core configuration parameters for each typical meteorological scenario. The core configuration parameters include physical parameterization scheme, grid resolution, boundary layer scheme and initial perturbation scheme. For each sub-region, the similarity between the meteorological and physical characteristics of the sub-region and each typical meteorological scenario in the microscale meteorological model configuration scheme library is calculated to determine multiple candidate schemes; For each sub-region, based on the historical high-resolution observation data or reanalysis data of the sub-region, a rapid historical back-calculation verification is performed on each of the candidate schemes, and the optimal adaptation scheme for the sub-region is determined from the candidate schemes. For each sub-region, based on the real-time meteorological conditions of that sub-region, the core configuration parameters of the optimal adaptation scheme corresponding to that sub-region are dynamically adjusted to generate the corresponding microscale meteorological model configuration scheme.

2. The high-resolution low-air meteorological data and short-term nowcasting method according to claim 1, characterized in that, The step of dividing the target region into multiple sub-regions and corresponding meteorological and physical characteristics based on the similarity of meteorological and physical processes and according to the underlying surface characteristics of the target region and the long-term climate background field used to characterize the long-term state of climate elements includes: Based on a preset resolution, the target area is discretized into multiple geographic grid points; Based on the long-term climate background field, the corresponding climate background parameters are obtained for each geographic grid point through spatial interpolation. The long-term climate background field is a spatial distribution data field used to characterize the long-term average state of climate elements. The climate background parameters include at least the average wind field parameters, average temperature field parameters, average humidity field parameters, and precipitation climatology parameters. For each geographic grid point, based on the underlying surface characteristics and the climate background parameters of that geographic grid point, a set of key influencing factors are determined through a physical diagnostic algorithm. The key influencing factors include at least dynamic forcing factors, thermal non-uniformity factors, and local circulation climate factors. Based on the aforementioned set of key influencing factors, spatial clustering is performed on all the aforementioned geographic grid points to form multiple sub-regions; For each sub-region, the meteorological and physical characteristics of the sub-region are determined based on the central value and distribution characteristics of the sub-region.

3. The high-resolution low-air meteorological data and short-term nowcasting method according to claim 1, characterized in that, The step of loading multi-source real-time meteorological observation data and real-time climate background fields of each of the sub-regions into the called microscale meteorological model configuration scheme, and generating high-resolution meteorological fields of each of the sub-regions through rapid update cyclic assimilation and parallel computing, includes: Multi-source real-time meteorological observation data is received and assimilated in a preset high-frequency update cycle to obtain a three-dimensional incremental field. The multi-source real-time meteorological observation data includes ground meteorological station data, wind profiler radar data, GPS water vapor detection data, and meteorological satellite data. For each sub-region, the three-dimensional incremental field and the real-time climate background field are jointly loaded into the microscale meteorological model configuration scheme called by that sub-region, wherein the real-time climate background field is a spatial distribution data field used to characterize the recent average state of climate elements. Using independent computing nodes allocated to each sub-region, the rapid integral calculation of the microscale meteorological model configuration scheme is performed in parallel to generate high-resolution meteorological fields for each sub-region, wherein the meteorological fields include wind fields, temperature fields, humidity fields and pressure fields.

4. The high-resolution low-air meteorological data and short-term nowcasting method according to claim 1, characterized in that, The step of fusing the high-resolution meteorological fields of all the sub-regions to form a unified high-resolution meteorological field of the target region includes: A weighting function based on physical constraints is used to fuse the high-resolution meteorological field in the overlapping area of ​​adjacent sub-regions to obtain a preliminary fused field. The weighting function is dynamically adjusted according to the forecast uncertainty or topographic representativeness of each sub-region at the boundary. By applying conservation constraints on key physical quantities to the preliminary fusion field, a coordinated field is generated, wherein the key physical quantities include at least one of mass, momentum and water vapor flux. The coordinated fields of all the sub-regions are stitched together to form a unified high-resolution meteorological field for the target region.

5. The high-resolution low-air meteorological data and short-term nowcasting method according to claim 1, characterized in that, The step of generating multiple high-resolution meteorological forecast scenarios within a short-term, near-term timeframe by using the unified high-resolution meteorological field as the initial field and driving a microscale ensemble forecasting system based on the multi-source real-time meteorological observation data includes: A set of perturbation initial fields is generated by applying a multi-scale hybrid initial value perturbation and physical process perturbation to the unified high-resolution meteorological field. Based on the set of perturbation initial fields, multiple ensemble forecast members are constructed, wherein each of the ensemble forecast members corresponds to one of the perturbation initial fields; The microscale ensemble forecasting system, based on the microscale meteorological model configuration scheme of each sub-region, drives each ensemble forecasting member to perform short-term near-term integration in parallel to generate multiple high-resolution meteorological forecast scenarios. During the integration process, based on the real-time weather characteristics derived from the assimilation analysis of the multi-source real-time meteorological observation data, the disturbance physical parameterization schemes of each forecast member in the predefined set of disturbance physical parameterization schemes are selectively enabled or switched.

6. The high-resolution low-air meteorological data and short-term nowcasting method according to claim 2, characterized in that, The step of generating probabilistic high-resolution forecast products based on the high-resolution weather forecast scenario includes: The high-resolution weather forecast scenarios are aggregated and statistically analyzed to calculate the probability distribution of key meteorological variables at each of the geographic grid points. Based on the probability distribution, spatially continuous probability forecast products are generated for at least one preset meteorological threshold. Based on historical fitting statistical correction methods and spatial smoothing algorithms, the probabilistic forecast products are processed to generate probabilistic high-resolution forecast products.

7. The high-resolution low-air meteorological data and short-term nowcasting method according to any one of claims 1 to 6, characterized in that, The method further includes: Based on at least one preset meteorological disaster early warning threshold, a visualized risk level area map and early warning information are generated according to the high-resolution forecast product.

8. A high-resolution low-air meteorological data and short-term nowcasting system, characterized in that, include: The segmentation module is used to segment the target area based on the similarity of meteorological and physical processes and according to the underlying surface characteristics of the target area and the long-term climate background field used to characterize the long-term state of climate elements, to obtain multiple sub-regions and corresponding meteorological and physical characteristics. The configuration scheme generation module is used to match and select an appropriate microscale meteorological model configuration scheme for each sub-region based on the meteorological and physical characteristics of that sub-region. The meteorological field generation module is used to load the multi-source real-time meteorological observation data of each sub-region and the real-time climate background field used to characterize the recent state of climate elements into the called microscale meteorological model configuration scheme, and generate high-resolution meteorological fields of each sub-region through rapid update cyclic assimilation and parallel computing. The fusion module is used to fuse the high-resolution meteorological fields of all the sub-regions to form a unified high-resolution meteorological field of the target region; The weather forecast scenario generation module is used to generate multiple high-resolution weather forecast scenarios within a short-term, near-term timeframe by using the unified high-resolution meteorological field as the initial field and driving the microscale ensemble forecast system based on the multi-source real-time meteorological observation data. The forecast product generation module is used to generate probabilistic high-resolution forecast products based on the high-resolution weather forecast scenario. The step of matching and selecting a suitable microscale meteorological model configuration scheme for each sub-region based on the meteorological and physical characteristics of that sub-region includes: A microscale meteorological model configuration scheme library is constructed, wherein the microscale meteorological model configuration scheme library defines corresponding core configuration parameters for each typical meteorological scenario. The core configuration parameters include physical parameterization scheme, grid resolution, boundary layer scheme and initial perturbation scheme. For each sub-region, the similarity between the meteorological and physical characteristics of the sub-region and each typical meteorological scenario in the microscale meteorological model configuration scheme library is calculated to determine multiple candidate schemes; For each sub-region, based on the historical high-resolution observation data or reanalysis data of the sub-region, a rapid historical back-calculation verification is performed on each of the candidate schemes, and the optimal adaptation scheme for the sub-region is determined from the candidate schemes. For each sub-region, based on the real-time meteorological conditions of that sub-region, the core configuration parameters of the optimal adaptation scheme corresponding to that sub-region are dynamically adjusted to generate the corresponding microscale meteorological model configuration scheme.

9. A computer device, characterized in that, The method includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the method of any one of claims 1 to 7.

Images