LES super-high-resolution tc atmospheric flow field numerical simulation method and system

Abstract

The application discloses a LES super-high-resolution TC atmospheric flow field numerical simulation method and system, and relates to the fields of atmospheric observation, numerical assimilation and high-resolution numerical simulation. The method is characterized in that: three-dimensional dynamic-thermal information is obtained by implementing airship platform radar observation and downward sounding observation in the core area of a tropical cyclone; the radar radial wind and profile observation operators are preprocessed and constructed; a data assimilation strategy is adopted to integrate multi-source observation in a multi-nested grid numerical model to generate a three-dimensional initial / reanalysis field suitable for large eddy simulation; then, a large eddy simulation solver is configured on the innermost grid covering the boundary layer of the tropical cyclone; fine horizontal grids and boundary layer encryption vertical stratification are adopted to explicitly analyze the boundary layer vortex, small-scale convection and turbulent structure; and a super-high-resolution three-dimensional flow field product for fine evaluation of strong wind and disaster risk is output.

Description

Technical Field

[0001] This invention belongs to the fields of atmospheric observation, numerical assimilation, and high-resolution numerical simulation. In particular, it addresses the method and system for generating high-precision initial / reanalysis fields by combining radar and drop-sonde observations on an airship platform, using data assimilation technology, and then initiating Large-Eddy Simulation (LES) based on these fields to conduct ultra-high-resolution numerical simulations of tropical cyclone (TC) atmospheric flow fields. This is used to accurately analyze the boundary layer turbulence structure, small-scale eddies, and multi-scale interaction processes of TC, thereby improving the prediction and mechanism research level of tropical cyclone intensity and structural evolution. Background Technology

[0002] Traditional meteorological numerical simulation has been successfully applied in simulating mesoscale and even large-scale weather processes, and its ability to predict the paths of high-impact weather events such as tropical cyclones (TCs) is constantly improving. However, for the fine flow field characteristics such as small-scale turbulent structures, strong convective cells, and strong wind bands near the eyewall within the boundary layer of tropical cyclones, existing operational and research models still struggle to achieve high-precision characterization and evolution simulation due to the scarcity of observational data, insufficient initial field resolution, and large model grid scale.

[0003] In terms of observation, conventional ground-based observation networks and weather radars are mainly distributed in land and near-shore areas, while real-time data at sea is relatively scarce, and observations of the three-dimensional wind field and thermal structure of the core area of ​​tropical cyclones in the open ocean are particularly weak. Airborne or high-altitude platform-mounted Doppler radars can acquire information on the wind field inside storms to some extent, but due to the influence of platform type, trajectory planning, and detection geometry constraints, their observation coverage and vertical distribution often cannot simultaneously cover multi-scale structures such as the boundary layer, eyewall, and peripheral rainbands. Drop-sondes can provide high-vertical-resolution temperature, humidity, and wind profiles, but their deployment is spatiotemporally and spatially dispersed and costly, making it difficult to form high-density three-dimensional observation constraints for the core area of ​​tropical cyclones.

[0004] In terms of data assimilation, existing operational and research systems are mostly geared towards mesoscale models with kilometer-level or coarser resolutions, focusing on improving elements such as large-scale environmental fields and cyclone center locations, but with limited ability to constrain fine boundary layer structures and small-scale vortices. Although assimilation methods for different types of data, such as radar volume scan observations and profile observations, have been developed, in practical applications they are often distributed across different systems or conducted as separate experiments, lacking an integrated analytical framework for small-scale flow fields in the core region of tropical cyclones.

[0005] With the improvement of computing resources, boundary layer turbulence (LES) is gradually being applied to the study of boundary layer turbulence, energy transport, and extreme gusts. LES, through explicit analysis of large eddies with dominant energy, has the potential to characterize fine phenomena such as banded vortex structures, tornado-scale vortices, and strong upward motion. However, existing LES studies mostly use idealized environmental fields or initial conditions obtained by interpolation from coarse-resolution reanalysis fields, resulting in low coupling with actual observations. Furthermore, a mature multi-scale integrated technical system has not yet been formed between mesoscale models and LES in terms of grid configuration, physical parameterization, and boundary condition transfer.

[0006] In summary, existing technologies still have significant shortcomings in acquiring and collaboratively utilizing observational data in the core region of the TC (turbulent flow field), constructing high-precision initial fields for fine-scale structures, and integrating multi-scale numerical models with the LES (Low-Earth Streaming System), making it difficult to support ultra-high-resolution numerical simulations of the TC flow field. Therefore, how to provide high-quality initial conditions for the LES and construct a multi-scale integrated numerical simulation scheme for the TC flow field based on the comprehensive utilization of multi-source observational data is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0007] (a) Purpose of the invention

[0008] To address the aforementioned deficiencies and shortcomings of existing technologies, this invention aims to provide a method and system for ultra-high resolution TC flow field numerical simulation driven by LES (Large Eddy Simulation) data assimilation of airship radar and dropsonde data. This method acquires Doppler radar and dropsonde profile observation data from an airship platform in parallel within the target meteorological environment. A high-precision three-dimensional initial field or reanalysis field is constructed using a numerical assimilation strategy suitable for multi-source observations. This field then drives Large Eddy Simulation (LES) to conduct ultra-high resolution TC flow field numerical simulations with grid scales ranging from tens to hundreds of meters. This enables detailed analysis of the TC flow field, particularly the turbulent structure within the boundary layer, small-scale vortices, convective-scale systems, and their multi-scale interactions. This significantly enhances the ultra-high resolution numerical simulation capability of the TC flow field, providing physically consistent and reliable flow field data support for the detailed assessment of strong winds, extreme precipitation, and disaster risks.

[0009] (II) Technical Solution

[0010] To achieve the objective of this invention and solve its technical problems, the present invention adopts the following technical solution:

[0011] The first objective of this invention is to provide a LES-based ultra-high resolution numerical simulation method for tropical cyclone (TC) flow fields driven by the assimilation of airship radar and drop-sonde data. This method is used to construct a high-precision initial / reanalysis field adapted to Large Eddy Simulation (LES) in a tropical cyclone (TC) environment and to conduct ultra-high resolution flow field numerical simulations. It enables refined simulation and analysis of small-scale flow field structures such as the TC boundary layer, turbulence, and convection. The method includes at least the following steps:

[0012] S100. Collaborative observation data acquisition by airship platform: An airship platform is deployed at a preset altitude above the target TC. The onboard Doppler radar is used to acquire radial wind observation data covering the core area of ​​TC. At the same time, a drop-down radiosonde is released to acquire vertical profile observation data of pressure, temperature, humidity and three-dimensional wind field.

[0013] S200. Observation data preprocessing and observation operator construction: Time synchronization, quality control and spatial registration of airship radar and dropsonde observation data are performed. Radar radial wind observation operator is constructed based on airship platform position and radar beam geometry. Profile observation operator is constructed based on sounding trajectory. The mapping relationship between observations and model variables is established.

[0014] S300. Multi-source data assimilation and 3D analysis field construction: In the TC numerical model framework with multiple nested grids, the data assimilation method is used to fuse the preprocessed airship radar and dropsonde observation data with the model background field to obtain a 3D analysis field that matches the resolution of the nested grids, and the boundary conditions of the outer and inner grids are generated accordingly.

[0015] S400. LES Nested Domain Configuration: Configure the LES solver on the innermost mesh covering the TC boundary layer and core region, set the horizontal mesh spacing to below 100 meters, and refine the boundary layer in the vertical direction to initialize the LES nested domain with the constructed three-dimensional analysis field and its boundary conditions.

[0016] S500. Ultra-high resolution LES numerical simulation: Integrating the non-static control equations within the nested domain of LES, the subgrid-scale turbulence parameterization scheme is used to explicitly analyze the dominant energy of large eddy motion and some small-scale turbulence processes, obtaining the TC ultra-high resolution three-dimensional gas field including boundary layer vortices, small-scale convection generation, turbulence structure, and large eddy evolution.

[0017] S600. Simulation Result Output: Post-process the LES simulation results to output the TC three-dimensional flow field, turbulent kinetic energy, vortex structure identification, boundary layer height change, and convection generation mechanism analysis results.

[0018] The second objective of this invention is to provide a LES ultra-high resolution TC atmospheric flow field numerical simulation system driven by airship radar and drop-sonde data assimilation. This system is used to construct high-precision initial / reanalysis fields adapted to Large Eddy Simulation (LES) and to conduct ultra-high resolution flow field numerical simulations under tropical cyclone (TC) environments. The system includes the following modules:

[0019] The airship platform collaborative observation module includes an airship platform deployed over the target TC, which controls the onboard Doppler radar to acquire radial wind observation data covering the core area of ​​TC, and controls the drop-down sounding release and receiving device to acquire vertical profile observation data of pressure, temperature, humidity and three-dimensional wind field.

[0020] The observation data preprocessing and observation operator construction module is used to perform time synchronization, quality control and spatial registration of airship radar and dropsonde observation data, and to construct radar radial wind observation operator and profile observation operator based on airship platform position, radar beam geometry and sounding trajectory to establish the mapping relationship between observations and numerical model state variables.

[0021] The multi-source data assimilation and three-dimensional analysis field construction module is configured in the TC numerical model framework with multiple nested grids. It is used to call various observation operators, fuse the preprocessed airship radar and dropsonde observation data with the model background field, and generate a three-dimensional analysis field and corresponding boundary conditions that match the resolution of each nested grid.

[0022] The LES nested domain configuration module is used to configure the LES solver in the innermost mesh covering the TC boundary layer and the core region. It sets the horizontal mesh spacing to below 100 meters and refines the boundary layer in the vertical direction. It also initializes the LES nested domain using the three-dimensional analysis field and boundary conditions.

[0023] The LES numerical simulation module is used to integrate non-hydrostatic control equations within the LES nested domain. It employs a subgrid-scale turbulence parameterization scheme to explicitly analyze the dominant energy of large eddy motion and some small-scale turbulence processes, obtaining the TC ultra-high resolution three-dimensional gas flow field.

[0024] The simulation results output module is used to post-process the LES numerical simulation results and output the TC three-dimensional flow field and turbulence statistics, vortex structure and extreme wind field diagnostic results.

[0025] (III) Technical Effects

[0026] Compared with existing technologies, the LES ultra-high resolution TC atmospheric flow field numerical simulation method and system based on airship radar and drop-down sounding data assimilation driven by the present invention has the following significant technical effects:

[0027] (1) This invention introduces high spatiotemporal resolution three-dimensional dynamic-thermal information into the same analysis framework by coordinating Doppler radar observation and drop-in radiosonde observation on the airship platform in the core area of ​​the TC. It constructs a unified observation mapping and assimilation process for radial wind, temperature, humidity and three-dimensional wind profile, which overcomes the problems of insufficient coverage and single constraint elements of traditional observation methods in the core area of ​​the TC. It also improves the accuracy of the three-dimensional initial / reanalysis field adapted to LES in terms of structural integrity and physical consistency.

[0028] (2) The present invention configures the LES solver in the innermost domain of the multi-nested grid numerical model. By turning off the traditional planetary boundary layer parameterization, using horizontal grids below 100 meters and vertically layering the boundary layer with densification, and using the high-precision three-dimensional analysis field obtained by assimilation for initialization, the large-scale environmental field and the small-scale turbulent process can achieve dynamic consistency multi-scale coupling in the same framework. It can explicitly analyze key structures such as TC boundary layer vortices, small-scale convective cells and large eddy evolution, and significantly improve the credibility of ultra-high resolution flow field simulation.

[0029] (3) This invention systematically processes the LES simulation results to form a multi-source diagnostic product that includes three-dimensional wind field, turbulent kinetic energy distribution, vortex structure identification, boundary layer height evolution and extreme wind field index. It can not only be used for fine analysis of TC intensity and structural evolution, but also provide quantitative flow field support for fine risk assessment such as strong wind, extreme precipitation and engineering load assessment. Attached Figure Description

[0030] Figure 1 This is a flowchart of the LES ultra-high resolution TC atmospheric flow field numerical simulation method based on airship radar and drop-down sounding data assimilation driven by the present invention.

[0031] Figure 2 This is a system architecture diagram of the LES ultra-high resolution TC atmospheric flow field numerical simulation system based on the assimilation of airship radar and drop-down sounding data provided by the present invention.

[0032] Figure 3 The diagram shows a comparison of the path and intensity of a typical tropical cyclone. In diagram a, the cyclone path is plotted with longitude (°E) on the horizontal axis and latitude (°N) on the vertical axis. The solid line represents the cyclone center path obtained from the natural experiment of the numerical model, and the dashed line represents the operational best path data. Diagram b shows the intensity evolution, with time (UTC) on the horizontal axis, maximum surface wind speed (m / s, red) on the left vertical axis, and minimum sea level pressure (hPa, blue) on the right vertical axis. The comparison of the two types of curves in the diagram is used to characterize the model's ability to simulate the evolution of cyclone trajectory and intensity over time.

[0033] Figure 4The figure shows the planar distribution of the 2-minute average 10m wind speed and gust coefficient in the core region of a tropical cyclone under different horizontal grid resolutions. A to D represent the distribution of the 2-minute average 10m wind speed (m / s) at t=9h when the grid spacing is 2km, 500m, 166m, and 55m, respectively. E to H correspond to A to D but represent the gust coefficient distribution. The dashed circles in the figure represent the radius range of 30km from the cyclone center, used to compare the differences in the spatial structure and intensity of persistent winds and gusts under different grid resolutions.

[0034] Figure 5 The diagram shows the vertical velocity field and scale decomposition results of the boundary layer of a tropical cyclone under different horizontal grid resolutions. A~C represent the horizontal distribution of instantaneous vertical velocity (m / s) at t=8.75h and 183m altitude in the northeast quadrant, with grid spacing of 500m, 166m, and 55m, respectively. D~F correspond to A~C but retain only the large-scale filtered fields with horizontal wavelengths greater than 1600m. G~I correspond to A~C but are small-scale perturbation fields with wavelengths less than 1600m. Different shades and textures in the diagram are used to characterize the intensity and spatial morphology of the rising and sinking motions, revealing the contribution of boundary layer vortices and multi-scale turbulent structures to the strong rising motion. Detailed Implementation

[0035] This invention aims to provide a method and system for numerical simulation of ultra-high resolution TC atmospheric flow fields using LES (Light Surface Energy) driven by the assimilation of airship radar and drop-sonde data. To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention will be described in more detail below with reference to the accompanying drawings. The described embodiments are some, but not all, of this invention, and are exemplary, intended to explain the invention, and should not be construed as limiting the invention.

[0036] Example 1: LES Ultra-High Resolution TC Gas Flow Field Numerical Simulation Method

[0037] like Figure 1 As shown in the embodiments of the present invention, the LES ultra-high resolution TC flow field numerical simulation method based on airship radar and drop-sonde data assimilation is used to construct a high-precision initial / reanalysis field adapted to Large Eddy Simulation (LES) and carry out ultra-high resolution flow field numerical simulation in tropical cyclone (TC) environments. This enables refined simulation and analysis of small-scale macro-flow field structures such as the TC boundary layer, turbulence, and convection. The main steps include:

[0038] S100. Acquisition of collaborative observation data from airship platforms:

[0039] An airship platform is deployed at a predetermined altitude above the target TC. The onboard Doppler radar acquires radial wind observation data covering the core area of ​​TC. At the same time, a drop-sonde is released to acquire vertical profile observation data of pressure, temperature, humidity, and three-dimensional wind field.

[0040] In this embodiment of the invention, the airship platform is preferably deployed in the lower stratosphere near the tropopause, and is in a near-hovering or slow-cruising state above the TC, with its altitude range controlled between 16 and 22 km. The horizontal distance DIS (Distance Index from Storm-center) between the airship and the center of the TC is determined based on the real-time position of the TC and the radius of maximum wind (RMW), so that the DIS is located within the range of 0.8 to 1.5 RMW, ensuring that the radar beam passes through or covers the maximum wind speed zone of the TC during the scanning process. At the same time, the radar is controlled to perform volume scanning within the pitch angle range of 10° to 60° according to the preset scanning strategy, taking into account the acquisition of information on the lower boundary layer structure and the middle and upper core wind field. In addition, the Doppler radar preferably adopts a multi-elevation angle and multi-azimuth repeated volume scanning mode within the same scanning cycle, and sets its scanning cycle to no more than 10 minutes to ensure the temporal resolution of wind field observation in the core area of ​​the TC. During radar scanning, the scanning starting azimuth and scanning sector width are dynamically adjusted according to the TC's moving speed and the relative position of the airship platform, so that the radar observation sector always covers the TC's eyewall and near-eye area. When releasing the drop-in radiosonde, multiple profile lines are laid out at different azimuth radii in the core of the TC and the surrounding rainband according to the preset radial and azimuth point layout strategy to ensure the sampling density of the radiosonde profile in the radial and azimuth directions.

[0041] It should be noted that the Doppler radar on the airship platform preferably adopts a dual-polarization or full-polarization configuration. By simultaneously acquiring horizontal and vertical polarized echo signals, it can achieve fine identification of precipitation particle phase, particle orientation, and liquid water content, providing microphysical parameter constraints for terminal velocity correction in subsequent radial wind inversion. Before release, the drop-in radiosonde needs to undergo attitude sensor calibration and GPS signal quality inspection to ensure that it can maintain measurement accuracy under strong wind shear and turbulent conditions. The sampling frequency of the radiosonde is set to no less than 2Hz to capture vertical fine structural changes on the order of 10 meters within the boundary layer.

[0042] S200. Observation data preprocessing and observation operator construction:

[0043] Time synchronization, quality control, and spatial registration of airship radar and dropsonde observation data are performed. Radar radial wind observation operator is constructed based on the airship platform position and radar beam geometry, and profile observation operator is constructed based on the sounding trajectory. The mapping relationship between observations and model variables is established.

[0044] In this embodiment of the invention, when constructing the radar radial wind observation operator, the radar beam azimuth, elevation angle, and geographical height of the detection point are explicitly considered. Based on the geometric projection relationship between the three-dimensional wind field components and the radar line of sight, the tangential wind, radial wind, and vertical velocity in the model are decomposed into the radial observation space. At the same time, the terminal falling velocity of precipitation particles is introduced as a parameter, and the terminal falling velocity component is subtracted from the radial velocity to improve the consistency between the radial wind and the actual wind field. When constructing the profile observation operator, based on the falling trajectory and timestamp of the dropsonde, the model grid field is mapped to the location of the radiosonde measurement point through time interpolation and spatial interpolation to ensure the consistent comparison of temperature, humidity, and wind field profiles.

[0045] More specifically, the above observation operator is preferably expressed in the following form. Let the radar radial wind observation operator be H. r Then, for any model grid point, the wind vector (u,v,w) and radar line-of-sight geometric parameters ( Radial wind observations are expressed as V(θ), where θ is the radial wind. r =H r (u,v,w)=u·cosθcos +v·cosθsin +w·sinθ-V t ,in θ is the azimuth angle, θ is the elevation angle, and V is the azimuth angle. t Let H be the terminal velocity of precipitation particles, which is given by the empirical relationship between particle size distribution and rainfall intensity; let H be the profile observation operator. p Then, for the drop-type probe at time t d In position (x) d ,y d ,z d The observation point can be obtained through the neighboring model grid field X(x) i ,y j ,z k Spatiotemporal interpolation of ,t) is obtained , where y d This includes profile elements such as temperature, humidity, and wind vectors. w ijk Spatial interpolation weights and satisfying Through the above formal construction, a one-to-one mapping between observations and mode state variables is achieved.

[0046] In addition, when preprocessing the observation data, echo quality control and ground clutter removal are required for the original radar volume scan data. Abnormal observations are removed using reflectivity thresholds, radial velocity de-aliasing discrimination, and neighborhood consistency checks. Before assimilation of radial wind observation data, high-density radar observations are spatially aggregated according to preset radial distance and azimuth intervals to form representative super-observations in the radial and azimuth directions. The equivalent observation error variance is estimated for each super-observation to reduce the spatial correlation of observation errors and the data volume. For drop-sonde observation data, a combination of vertical smoothing and outlier removal is used for preprocessing to ensure the quality of profile observations.

[0047] S300. Multi-source data assimilation and 3D analysis field construction:

[0048] In the TC numerical model framework with multiple nested grids, a data assimilation method is used to fuse preprocessed airship radar and dropsonde observation data with the model background field to obtain a three-dimensional analysis field that matches the resolution of the nested grids, and the boundary conditions of the outer and inner grids are generated accordingly.

[0049] In this embodiment of the invention, data assimilation preferably employs an ensemble Kalman filter method, maintaining no fewer than 30 ensemble members on both the outer and inner nested grids. Flow-dependent background error covariance is constructed by perturbing the initial field and boundary conditions of the model. During the assimilation period, a joint observation vector is formed using radar radial wind observations and downsinking radiosonde profile observations. The constructed radar radial wind observation operator and profile observation operator are used to project the observation space onto each ensemble member to obtain the simulated observation value corresponding to each observation. Based on the difference between the actual observation value and the simulated observation value, the model state of each ensemble member is weighted and corrected to obtain a three-dimensional analysis field that matches the resolution of the nested grid.

[0050] More specifically, in the embodiments of the present invention, the following unified mathematical form is preferably used to describe the multi-source observation assimilation process. Let the joint observation vector be denoted as... , where y radar For radar radial wind observation sets, y drop This is a set of profile observations of temperature, humidity, and wind vectors from a dropsonde radiosonde; the model background state vector is denoted as x. f Then, based on the radar radial wind observation operator H constructed in step S200, r and profile observation operator H p Define the joint observation operator Thus, the corresponding simulated observation vector y is obtained. f =H(x f For any set member k in the ensemble Kalman filter, calculate the observation residual of that member. The gain matrix K is constructed using the background error covariance matrix obtained from the set estimation, and the analysis and update are completed. Under a unified framework, information from both radar radial wind observation and drop-in radiosonde profile observation is simultaneously absorbed to obtain a three-dimensional analysis field set that satisfies the multi-source observation constraints.

[0051] Furthermore, step S300 also includes a multi-scale localization and weight allocation process. Different localization radii of covariance are set for different nested grids, so that the outer grid reflects the smooth adjustment of the environmental field at a scale of 100 to 500 km, and the inner grid highlights the fine correction of the TC core structure at a scale of no more than 100 km. At the same time, observation error covariance matrices are set according to the differences in spatial representativeness and error characteristics between radar radial wind and sounding profile. High weights are given to radar radial wind, which is sensitive to TC core circulation, and high weights are given to temperature and humidity profiles, which are sensitive to TC thermal stratification characteristics. This achieves the synergistic optimization of dynamic and thermal elements in the three-dimensional analysis field.

[0052] It should be noted that the ensemble multi-source joint assimilation framework constructed in this step unifies the complementary constraints of airship radar and dropsonde on the TC core circulation and thermo-stratification in a single assimilation update process. By setting multi-scale localization and differentiated weights, it achieves coordinated adjustment of the environmental field and the core vortex, thereby more fully recovering the detailed features of the TC boundary layer and warm core structure while maintaining large-scale equilibrium constraints. This provides initial and boundary conditions for the LES nested domain that are closer to the real three-dimensional structure, improving the physical reliability of subsequent ultra-high resolution numerical simulations.

[0053] S400. LES Nested Domain Configuration:

[0054] The LES solver is configured on the innermost mesh covering the TC boundary layer and the core region. The horizontal mesh spacing is set to less than 100 meters and the boundary layer is refined in the vertical direction. The constructed three-dimensional analysis field and its boundary conditions are used to initialize the LES nested domain.

[0055] In this embodiment of the invention, the horizontal grid spacing is preferably set to within the range of 50-100 m in the innermost grid, and the TC boundary layer region of 0-2 km in the vertical direction is densified to achieve a near-surface vertical resolution on the order of 10-40 m. Furthermore, the planetary boundary layer parameterization scheme is turned off in the LES nested domain, the LES solution mode is enabled, and a subgrid-scale turbulence parameterization scheme with anisotropic and nonlinear divergence characteristics is selected to close the unanalyzed small-scale turbulence flux. At the same time, the wind field, temperature field, and humidity field of the outer mesoscale model are transmitted to the LES nested domain in the form of time interpolation through unidirectional or bidirectional nested boundary conditions.

[0056] When selecting the horizontal grid spacing for the LES nested domain, two criteria must be met simultaneously: First, the grid scale should be located within the inertial subregion of the turbulent energy spectrum so that most of the turbulent kinetic energy can be explicitly analyzed; second, the grid scale should be smaller than the characteristic scale of the dominant vortex in the TC boundary layer. The typical horizontal scale of the boundary layer vortex is in the range of 200-500 meters. Therefore, setting the horizontal grid spacing to 50-200 meters can effectively capture the dominant vortex structure. The vertical grid is refined within the boundary layer using a hyperbolic tangent stretching function so that the center height of the first-layer grid near the ground is controlled at 5-10 meters, which meets the logarithmic law distribution assumption for surface layer flux calculation.

[0057] Furthermore, in step S400, when initializing the LES nested domain, the following refined processing is preferably adopted: the assimilated three-dimensional analysis field is projected onto the LES mesh using cubic spline interpolation to ensure the continuity and smoothness of the physical field; the initial field after interpolation is dynamically balanced, high-frequency gravity wave noise is suppressed by numerical filtering, and the non-static equilibrium equations are iteratively solved until the residuals converge; for the initial turbulent field within the boundary layer, based on the wind shear and thermal stratification characteristics of the assimilated three-dimensional analysis field, random perturbations conforming to the Kolmogorov spectrum characteristics are superimposed, with the perturbation amplitude controlled within 5-10% of the mean flow field, to accelerate the turbulence development process in the LES simulation.

[0058] It should be noted that the LES nested domain configuration and initialization process proposed in this step organically combines high-resolution grid design, boundary condition propagation, dynamic balance adjustment, and turbulent perturbation injection. Under the premise of ensuring numerical stability and energy conservation characteristics, the LES initial field inherits the large-scale structure of the assimilation analysis field and has turbulent level perturbations consistent with boundary layer wind shear and thermal stratification. This allows the simulation to enter a reasonable turbulent evolution stage in the early stages of integration, which helps to avoid the long-term self-adjustment and spurious gravity wave interference problems common in idealized initial fields.

[0059] S500. Ultra-high resolution LES numerical simulation:

[0060] By integrating the non-static control equations within the nested domain of LES, and employing a subgrid-scale turbulence parameterization scheme, the dominant energy of large eddy motion and some small-scale turbulent processes are explicitly analyzed, resulting in a TC ultra-high resolution three-dimensional gas field that includes boundary layer vortices, small-scale convection generation, turbulent structure, and large eddy evolution.

[0061] In this embodiment of the invention, the non-hydrostatic control equations are explicitly integrated within the LES nested domain. The integration time covers at least several hours (e.g., 6-12 hours) of stable development of the TC core structure to ensure the representativeness of the turbulence statistics. During the integration process, the analytical large eddy component and the subgrid-scale flux component are calculated to obtain the vertical momentum flux, turbulent kinetic energy, and their distribution with height. At the same time, the instantaneous field and time-averaged field of vertical velocity and vorticity are extracted at a predetermined height to identify and statistically analyze the spatial scale, intensity, and frequency of occurrence of the boundary layer strip vortex structure and tornado-like scale vortices, so as to characterize the multi-scale turbulence structure features of the TC boundary layer. Furthermore, the preferred methods for handling the physical processes of the TC boundary layer include: The sea surface flux parameterization scheme employs the COARE (Coupled Ocean–Atmosphere Response Experiment) algorithm, which considers the influence of ocean waves. Sensible and latent heat fluxes are calculated based on the wind speed at 10 meters and the sea-air temperature difference. The sea surface roughness length is dynamically adjusted according to wind speed and wave conditions. The boundary layer top entrainment process is characterized by explicit analytical turbulent mixing, capturing the erosion of the inversion layer at the boundary layer top and the downward transmission of dry air from the free atmosphere. The radiative transfer process adopts the RRTMG (Rapid Radiative Transfer Model for GCMs) scheme, with a time step set to 5–10 minutes, fully considering the feedback influence of cloud-radiative interaction on the warm core structure of the TC and the development of convection.

[0062] It should be noted that this step, through instantaneous-average joint diagnosis under long-term integration, transforms the high-frequency three-dimensional flow field obtained by LES analysis into statistically representative turbulent characteristic quantities, and forms a repeatable diagnostic process in boundary layer vortex identification, strip structure extraction, and extreme wind field index construction. This invention can not only reproduce the generation and evolution process of multi-scale turbulence and small-scale convection in the TC boundary layer, but also provide a quantitative flow field statistical basis for subsequent boundary layer parameterization scheme improvement and engineering load and disaster risk assessment. Furthermore, the selection of subgrid-scale turbulence parameterization schemes in LES simulations needs to be optimized based on grid resolution and TC boundary layer characteristics. When the horizontal grid spacing is less than 100 meters, the dynamic Smagorinsky model or scale-adaptive simulation method is preferred, with model coefficients dynamically adjusted based on the invariants of the local strain rate tensor and rotation rate tensor. In strong convection regions, subgrid parameterization needs to be coupled with cloud microphysical processes, considering the contribution of cloud condensation latent heat release to subgrid turbulent kinetic energy. By solving the turbulent kinetic energy equations, explicit closure of turbulent transport and buoyancy generation terms can be achieved, improving the simulation's response to TC eyewall convection bursts.

[0063] S600. Simulation result output:

[0064] Post-processing of the LES simulation results yields the following outputs: TC three-dimensional flow field, turbulent kinetic energy, vortex structure identification, boundary layer height variation, and convection generation mechanism analysis. Preferably, based on the three-dimensional wind and pressure fields obtained from the LES simulation, the minimum sea-level pressure, maximum near-surface wind speed, and maximum wind radius of the TC are calculated and compared with results from outer mesoscale models. Gust coefficients, turbulence intensity, and shear parameters at 10m wind speeds are calculated at multiple radii and heights to assess the risk of extreme wind fields at different azimuths and radii. Simultaneously, radial-height and azimuth-height profiles are constructed to diagnose the warm core structure, eyewall rising motion zone, and boundary layer inflow and upper-level outflow channels, providing multi-dimensional flow field diagnostic information for TC intensity evolution and disaster risk assessment.

[0065] Furthermore, in step S600, it is preferable to perform statistical analysis on the long-term series data of LES simulation, and use time averaging and perturbation decomposition methods to obtain the mean circulation field and turbulent perturbation field within the TC boundary layer. The boundary layer structure obtained by analysis is compared with the observational statistical results to verify the rationality of LES configuration and assimilation scheme. On this basis, a risk indicator field including the spatial distribution of maximum wind speed, the probability distribution of extreme gusts, and the high-frequency occurrence area of ​​vortex structure is constructed and output in gridded or vectorized form to provide quantitative flow field reference data for disaster prevention and mitigation engineering design, offshore operation safety assessment, and subsequent model physical process improvement.

[0066] Example 2: LES Ultra-High Resolution TC Gas Flow Field Numerical Simulation System

[0067] Based on the method flow described in Embodiment 1 above, Embodiment 2 further provides a specific implementation structure for a LES ultra-high resolution TC atmospheric flow field numerical simulation system driven by the assimilation of airship radar and drop-in sounding data. For example... Figure 2 As shown, the system covers the entire process of steps S100 to S600 in Example 1. It adopts a modular architecture and integrates observation, preprocessing and observation operator construction, multi-source data assimilation, LES nested domain configuration, LES numerical simulation and result output on the same software and hardware platform to achieve integrated operation of observation-assimilation-large eddy simulation.

[0068] In this embodiment 2, the system includes an airship platform collaborative observation module, an observation data preprocessing and observation operator construction module, a multi-source data assimilation and three-dimensional analysis field construction module, a LES nested domain configuration module, a LES numerical simulation module, and a simulation result output module. Each module is deployed on at least one numerical computing platform that includes a central processing unit (CPU) and / or a graphics processing unit (GPU), a large-capacity storage unit, and a high-speed network interface. The modules interact and schedule processes with each other through standardized data interfaces and a unified task control program.

[0069] The airship platform collaborative observation module is communicatively connected to the Doppler radar subsystem and the drop-down radiosonde release and receiving device on the airship platform. It is used to control the radar volume scan parameters, scanning sector and radiosonde release time according to the preset observation scheme and the real-time position of the TC. It receives and caches radial wind observation data and profile observation data such as temperature, humidity and wind field, and performs preliminary marking according to the observation timestamp and platform position information, providing system-level support for the execution of step S100 in Example 1.

[0070] The observation data preprocessing and observation operator construction module is used to call the original radar and radiosonde observation data, and perform preprocessing operations such as time synchronization, echo quality control, ground clutter removal, de-aliasing processing, vertical smoothing and outlier removal on it. Based on this, according to the airship position, radar beam geometry and drop radiosonde trajectory, the radar radial wind observation operator and profile observation operator described in step S200 of Example 1 are generated. This module outputs the observation operators and the preprocessed observation data in a unified format to the multi-source data assimilation and three-dimensional analysis field construction module.

[0071] The multi-source data assimilation and 3D analysis field construction module is configured in the TC numerical model framework with multiple nested grids. It is used to receive the joint observation vector and corresponding observation operator within the set Kalman filter assimilation period described in step S300 of Example 1, to perform observation space projection and state update on the background field set, and to set the localization radius and observation weight coefficient according to the different grid resolutions of the outer and inner layers. It outputs the 3D analysis field and boundary condition file that matches each nested grid, providing input for the subsequent LES nested domain configuration module.

[0072] The LES nested domain configuration module is used to complete the mesh generation, initial field interpolation, and dynamic balance adjustment of the LES nested domain described in step S400 of Example 1 on the innermost mesh covering the TC boundary layer and the core region, according to the preset mesh resolution range, vertical densification level, and boundary condition transfer strategy. At the same time, it superimposes the initial turbulent disturbance on the boundary layer according to the assimilation analysis field characteristics, and generates the configuration file and initial field data of the LES solver.

[0073] The LES numerical simulation module is used to call the non-static large eddy simulation solution program. In step S500 of Example 1, it controls the integration duration, time step, output frequency, and physical process schemes such as sea surface flux and radiative transfer. It performs LES integration and generates high-frequency output data including three-dimensional wind field, temperature field, turbulence statistics, etc.

[0074] The simulation results output module manages and processes the LES numerical simulation output in a unified manner. Based on the diagnostic requirements set by the user, it automatically calculates products such as minimum sea level pressure, maximum near-ground wind speed, maximum wind radius, gust coefficient, turbulence intensity, vortex structure identification index, and radial-height and azimuth-height profiles. It also supports output in various forms such as gridded data files, graphical profiles, and statistical reports, thus achieving the function corresponding to step S600 in Example 1.

[0075] It should be noted that in this embodiment 2, each module can be deployed on the same physical computing node as an independent software unit, or it can run in a multi-node cluster environment through a distributed deployment method; the interface protocols, data formats and physical implementation methods between each module can be equivalently replaced without changing the basic idea of ​​the present invention, and all should be considered to fall within the protection scope of the present invention.

[0076] Example 3: Typical TC Case Description

[0077] Based on Examples 1 and 2 above, this example selects a typical tropical cyclone in the Northwest Pacific to illustrate the specific application process and effects of the LES ultra-high resolution TC atmospheric flow field numerical simulation method and system driven by airship radar and drop-sonde data assimilation. First, based on the same mode and assimilation configuration as in Example 1, a multi-nested grid system covering the troposphere and lower stratosphere is constructed. Continuous integration is performed on the outer grid to obtain the natural experiment "true value" evolution of a typical TC, and this is compared with operational best path data, such as... Figure 3 As shown, the natural experiment can reproduce the rapid intensification process of the TC well in terms of both the path and the evolution of the lowest sea level pressure, providing a reliable reference for the observation simulation and assimilation assessment of this invention.

[0078] Building upon this, this embodiment follows the observation deployment and system architecture described in Embodiments 1 and 2. An airship platform is deployed at an altitude of 16–22 km above the TC core, Doppler radar volume scan parameters are configured, and multiple downward-dropped sounding profile lines are established near the eyewall and in the surrounding rainband. Utilizing the flight platform collaborative observation module and observation data preprocessing and observation operator construction module described in Embodiment 2, radar radial wind and profile observation operators are generated. The simulated airship radar radial wind and downward-dropped sounding profile observations are assimilated into an ensemble Kalman filter framework to obtain a three-dimensional analysis field matching the resolution of the inner nested grid. Compared to the unassimilated case, the analysis field in this embodiment more closely resembles natural experiments in terms of low-level convergence, eyewall tangential wind, and warm core structure, providing initial conditions containing a fine core structure for the subsequent LES nested domain.

[0079] Subsequently, on the innermost mesh covering the TC boundary layer and eyewall region, based on the LES nested domain configuration module in Examples 1 and 2, the horizontal mesh spacing was set to different resolution combinations such as 500m, 166m, and 55m, and the mesh was densified at an altitude of 0–2km. The planetary boundary layer parameterization scheme was turned off, and the LES solution mode and subgrid-scale turbulence parameterization scheme were enabled. Using the assimilated three-dimensional analysis field and boundary conditions as the initial driving field for LES, ultra-high resolution integration was performed on a typical TC for 7–10 hours, outputting diagnostic quantities such as 10m wind field, turbulent kinetic energy, and vertical velocity at multiple times.

[0080] like Figure 4 As shown, the distribution of 10m wind speed and gust coefficient obtained under different horizontal grid resolutions in this embodiment indicates that compared with the mesoscale model that only uses a kilometer-level grid, the hundred-meter-level LES nesting can resolve narrower strong wind bands, more concentrated extreme gust areas, and more complex near-ground wind field asymmetric structures, providing more refined flow field input for extreme wind field and gust risk assessment. Figure 5 The results show the horizontal distribution of vertical velocity and boundary layer vortex structure at a height of 183m. It can be seen that as the grid resolution is refined from 500m to 166m and 55m, the wavelengths of the horizontal cirrus bands and tornado-scale vortices in the boundary layer shorten and the structure becomes clearer. The boundary layer vortex bands (roll bands) near the eyewall exhibit detailed features of alternating strip-shaped rising and sinking regions, verifying the effectiveness of the LES nested domain of this invention in resolving the multi-scale turbulent structure of the TC boundary layer.

[0081] As can be seen from this embodiment, based on the integrated observation-assimilation-large eddy simulation chain constructed in Embodiments 1 and 2, this invention can realize the assimilation application of airship radar and drop-in sounding in actual (or idealized) TC cases, and obtain boundary layer vortices, small-scale convection and extreme near-ground wind field characteristics with high spatiotemporal resolution within the LES nested domain, providing an operable technical path for TC intensity structure analysis and disaster risk assessment.

[0082] The objectives of this invention have been fully and effectively achieved through the above embodiments. Those skilled in the art will understand that this invention includes, but is not limited to, the contents described in the accompanying drawings and the specific embodiments described above. Although the invention has been described with reference to what is currently considered the most practical and preferred embodiments, it should be understood that the invention is not limited to the disclosed embodiments, and any modifications that do not depart from the functional and structural principles of the invention will be included within the scope of the claims.

Claims

1. A LES super-high resolution TC atmospheric flow field numerical simulation method based on airship radar and downcast sounding data assimilation driving, characterized in that, The method comprises the following steps: S100. The airship platform is arranged at a preset height above the target TC, and a Doppler radar carried by the airship platform is used to obtain radial wind observation data covering the core area of the TC, and a downward-projection sounding instrument is released to obtain vertical profile observation data of pressure, temperature, humidity and three-dimensional wind field; S200. The airship radar and downward-projection sounding observation data are time-synchronized, quality-controlled and spatially registered, a radar radial wind observation operator is constructed based on the position of the airship platform and the geometric relationship of the radar beam, a profile observation operator is constructed based on the sounding trajectory, and a mapping relationship between the observation and the model variable is established; S300. In the TC numerical model framework with multiple nested grids, a data assimilation method is used to fuse the preprocessed airship radar and downward-projection sounding observation data with the model background field, so as to obtain a three-dimensional analysis field matched with the resolution of the nested grid, and the boundary conditions of the outer and inner grids are generated accordingly; S400. In the innermost grid covering the boundary layer and the core area of the TC, a LES solver is configured, the horizontal grid spacing is set to be less than 100 meters, and the boundary layer is encrypted in the vertical direction, so as to initialize the LES nested domain based on the three-dimensional analysis field and the boundary conditions; S500. The non-static control equation is integrated in the LES nested domain, a sub-grid scale turbulence parameterization scheme is used to explicitly analyze the dominant energy of the large eddy motion and part of the small-scale turbulence process, and the TC three-dimensional atmospheric flow field is obtained, which includes the boundary layer vortex, the small-scale convection generation, the turbulence structure and the large eddy evolution; S600. The LES simulation results are post-processed, and the TC three-dimensional flow field, the turbulence kinetic energy, the vortex structure identification, the boundary layer height variation and the convection generation mechanism analysis results are output.

2. The method of claim 1, wherein, In step S100, the airship platform is arranged in the lower stratosphere near the tropopause, and is in an approximate hovering or slow cruising state above the TC, and the height range is controlled in the range of 16-22 km; the horizontal distance DIS between the airship and the center of the TC is determined according to the real-time position of the TC and the maximum wind radius RMW, so that DIS is located in the range of 0.8-1.5 RMW, and the radar beam is ensured to pass through or cover the maximum wind speed band of the TC during scanning; meanwhile, the radar is controlled to perform volume scanning in the range of 10°-60° of the pitch angle according to the preset scanning strategy, so as to obtain the low-layer boundary layer structure and the middle-high layer core wind field information.

3. The method according to claim 1 or 2, characterized in that, In step S100, the Doppler radar adopts a multi-elevation and multi-direction repeated volume scanning mode in the same scanning period. During the radar scanning, the scanning starting direction and the scanning sector width are dynamically adjusted according to the TC moving speed and the relative position of the airship platform, so that the radar observation sector always covers the TC eye wall and the near-eye area; when the downward-projection sounding instrument is released, a plurality of profile lines are arranged on different azimuth radii of the TC core and the peripheral rain belt according to a preset radial and azimuth distribution strategy, so as to ensure the sampling density of the sounding profile in the radial and azimuth directions.

4. The method of claim 1, wherein, In step S200, when constructing the radar radial wind observation operator, the radar beam azimuth, elevation and geographic height of the detection point are explicitly considered, according to the geometric projection relationship between the three-dimensional wind field components and the radar line-of-sight direction, the tangential wind, radial wind and vertical velocity in the model are decomposed into the radial observation space, and the terminal falling velocity parameterization of the precipitation particles is introduced, and the terminal falling velocity component is deducted from the radial velocity to improve the consistency of the radial wind and the real wind field; when constructing the profile observation operator, based on the falling trajectory and time stamp of the dropsonde, the model grid field is mapped to the sounding point position through time interpolation and spatial interpolation.

5. The method of claim 1, wherein, In step S200, the original radar volume scan data is subjected to echo quality control and ground clutter removal, abnormal observations are removed by using reflectivity threshold, radial velocity aliasing discrimination and neighborhood consistency test; before the radial wind observation data assimilation, the high-density radar observation is spatially aggregated according to the preset radial distance interval and azimuth interval to form representative super-observations in the radial and azimuth directions, and the equivalent observation error variance is estimated for each super-observation to reduce the spatial correlation of observation errors and the data volume; the dropsonde observation data is preprocessed by combining vertical smoothing and outlier removal.

6. The method of claim 1, wherein, In step S300, the data assimilation adopts the ensemble Kalman filter method, and maintains not less than 30 ensemble members on the outer and inner nested grids respectively, and constructs the flow-dependent background error covariance by perturbing the model initial field and boundary conditions; during the assimilation period, the radar radial wind observation and the dropsonde profile observation form a joint observation vector, the radar radial wind observation operator and the profile observation operator constructed are used to project the observation space of each ensemble member respectively, to obtain the simulated observation values corresponding to each observation, and the model state of each ensemble member is weightedly corrected based on the difference between the actual observation value and the simulated observation value, to obtain a three-dimensional analysis field matched with the resolution of the nested grid.

7. The method of claim 1, wherein, In step S300, the multi-scale localization and weight distribution process is also included, different covariance localization radii are set for different nested grids, so that the outer grid reflects the environmental field smoothing adjustment on the scale of 100-500 km, and the inner grid highlights the fine correction of the TC core structure on the scale of not more than 100 km; at the same time, according to the differences in spatial representativeness and error characteristics of the radar radial wind and the sounding profile, the observation error covariance matrix is set, the radar radial wind sensitive to the TC core circulation is given a high weight, and the temperature and humidity profile sensitive to the thermal stratification characteristics of the TC core is given a high weight, to realize the collaborative optimization of dynamic and thermal elements in the three-dimensional analysis field.

8. The method of claim 1, wherein, In step S400, the horizontal grid spacing is set in the innermost layer grid in the range of 50-100 m, and the TC boundary layer region of 0-2 km is encrypted in the vertical direction to achieve a near-surface vertical resolution of 10-40 m; the planetary boundary layer parameterization scheme is closed in the LES nesting domain, the LES solving mode is enabled, and the sub-grid scale turbulence parameterization scheme with anisotropy and nonlinear backscatter characteristics is selected to close the unresolved small-scale turbulence flux; at the same time, the wind field, temperature field and humidity field of the outer mesoscale mode are transmitted to the LES nesting domain in the form of time interpolation through one-way or two-way nested boundary conditions.

9. The method according to claim 1 or 8, characterized in that, In step S400, the three-dimensional analysis field generated by assimilation is projected to the LES grid by a cubic spline interpolation method; the initial field after interpolation is dynamically balanced and adjusted, high-frequency gravity wave noise is suppressed by numerical filtering technology, and the non-hydrostatic balance equation set is iteratively solved until the residual converges; for the initial field of the boundary layer turbulence, based on the wind shear and thermal stratification characteristics of the three-dimensional analysis field generated by assimilation, random disturbance conforming to the Kolmogorov spectrum characteristics is superimposed, and the disturbance amplitude is controlled in the range of 5-10% of the average flow field.

10. The method of claim 1, wherein, In step S500, the non-hydrostatic control equation is explicitly time-integrated in the LES nesting domain, and the integration time covers several hours of stable development of the TC core structure; during the integration process, the analytical large eddy component and the sub-grid scale flux component are calculated respectively to obtain the vertical momentum flux, the turbulent kinetic energy and its distribution with height; at the same time, the instantaneous field and the time-averaged field of the vertical velocity and the vorticity at the predetermined height are extracted to identify and count the spatial scale, intensity and occurrence frequency of the boundary layer strip-shaped vortex structure and the tornado-like scale vortex.

11. The method according to claim 1 or 10, characterized in that, In step S500, the processing of the TC boundary layer physical process includes: the COARE algorithm considering the influence of sea waves is used for the sea surface flux parameterization scheme, the sensible heat flux and latent heat flux are calculated according to the 10-meter height wind speed and the sea-air temperature difference, and the sea surface roughness length is dynamically adjusted according to the wind speed and the sea wave state; the top boundary layer entrainment process is described by an explicitly analytical turbulent mixing to capture the erosion of the boundary layer top inversion layer and the downward transmission of the free atmospheric dry air; the RRTMG scheme is used for the radiation transmission process, and the time step is set to 5-10 min.

12. The method of claim 1, wherein, In step S600, the long-time sequence data simulated by LES is statistically analyzed, the time-averaged and disturbance decomposition method is used to obtain the average circulation field and the turbulent disturbance field in the TC boundary layer, the analytically obtained boundary layer structure is compared with the observation statistical results to verify the rationality of the LES configuration and the assimilation scheme; the risk indicator field including the maximum wind speed spatial distribution, the extreme gust probability distribution, and the high-frequency occurrence area of the vortex structure is constructed and output in the form of gridding or vectorization.

13. A LES super-high resolution TC atmospheric flow field numerical simulation system based on airship radar and lower-pouring sounding data assimilation driving, adopting the method of any one of claims 1-12, characterized in that, It comprises: A flying boat platform cooperative observation module, comprising a flying boat platform arranged above a target TC, controlling a mounted Doppler radar to obtain radial wind observation data covering the TC core area, and controlling a downward-projection sounding release and receiving device to obtain pressure, temperature, humidity and three-dimensional wind field vertical profile observation data; The data preprocessing and observation operator construction module is used for time synchronization, quality control and spatial registration of the airship radar and the sounding observation data, and is used for constructing a radar radial wind observation operator and a profile observation operator based on an airship platform position, a radar beam geometric relation and a sounding trajectory. The data assimilation and three-dimensional analysis field construction module is configured in a TC numerical model framework, is used for calling each observation operator, fuses the preprocessed airship radar and the sounding observation data with a model background field, and generates a three-dimensional analysis field and a corresponding boundary condition matched with each nested grid resolution of the model. The LES nested domain configuration module is used for configuring an LES solver in the innermost layer grid covering a TC boundary layer and a core region, setting a horizontal grid spacing to be below a hundred meters, encrypting the boundary layer in a vertical direction, and initializing the LES nested domain by using the three-dimensional analysis field and the boundary condition. The LES numerical simulation module is used for integrating a non-static control equation in the LES nested domain, explicitly analyzing a dominant energy large eddy motion and part of small scale turbulent process by using a sub-grid scale turbulence parameterization scheme, and obtaining a TC super-high resolution three-dimensional atmospheric flow field. The simulation result output module is used for post-processing the LES numerical simulation result, and outputting a TC three-dimensional flow field, a turbulent statistic, a vortex structure and an extreme wind field diagnostic result.

Images