Beidou-based multi-dimensional adaptive atmospheric river detection method and system

Abstract

The invention discloses a Beidou-based multi-dimensional adaptive atmospheric river detection method and system, and belongs to the field of atmospheric sciences, and the method comprises the steps: obtaining original meteorological data; based on the original meteorological data, comprehensive water vapor transportation is obtained through calculation, the comprehensive water vapor transportation is judged through a Gaussian kernel density smooth threshold value and a fixed threshold value, and a preliminary atmospheric river is obtained; geometric features of the preliminary atmospheric river are calculated through a three-way axis search mechanism; and based on the geometrical characteristics of the preliminary atmospheric river, removing a non-atmospheric river by combining relative vorticity with comprehensive water vapor transportation to obtain a final atmospheric river. The Beidou satellite is used for providing high-precision time service, real-time positioning and short message communication services, time synchronism and space accuracy of meteorological data are ensured, atmospheric river axis smoothness and length calculation accuracy are optimized through a three-way axis search mechanism, relative vortex and comprehensive water vapor transmission, and the atmospheric river axis smoothness and length calculation accuracy are improved. Therefore, the atmospheric river recognition reliability is remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of atmospheric science, specifically relating to a multi-dimensional adaptive atmospheric river detection method and system based on BeiDou. Background Technology

[0002] Atmospheric rivers are narrow, elongated jet streams in the lower troposphere that resemble ribbons and can transport large amounts of water vapor in a short time, triggering extreme precipitation. They play a crucial role in the global water cycle and hydrological influences. Addressing the shortcomings of current atmospheric river detection algorithms, this system utilizes the BeiDou satellite system to construct an integrated air-space-ground collaborative monitoring network, achieving high-precision, real-time, and intelligent atmospheric river detection and early warning services.

[0003] In recent years, with the popularization of high-resolution reanalysis data and satellite remote sensing technology, domestic and foreign scientific research and operational departments have widely adopted the integrated water vapor transport threshold method, geometric morphology and other means to automatically identify atmospheric rivers. These traditional methods perform well in the context of dry and cold mid-latitudes, but they show obvious shortcomings in the active season of tropical cyclones or subtropical sea areas, specifically: (1) Unstable axis tracking: Traditional methods mostly extract the main body of atmospheric rivers by setting an integrated water vapor transport threshold, and then use the centroid or the maximum integrated water vapor transport path to establish the axis. However, when the atmospheric river and the water vapor flux of the tropical cyclone are superimposed, the axis often shows a "sawtooth" or a return with unclear physical meaning, which in turn produces systematic errors. (2) Tropical cyclones are misjudged as atmospheric rivers: The strong vortex structure of tropical cyclones is easy to form a high integrated water vapor transport area similar to that of typical atmospheric rivers, especially in the 24 hours before landfall. It is difficult to distinguish between the two by a single threshold or simple morphological parameters, which leads to the core area of ​​the cyclone being mislabeled as an atmospheric river, interfering with the attribution analysis of extreme precipitation. (3) Limitations of existing solutions: Although existing studies have introduced techniques such as curvature constraints, K-means clustering and convolutional neural networks to try to solve the above problems, they still face many limitations such as insufficient training samples, poor model interpretability, and high real-time computing costs. Summary of the Invention

[0004] The purpose of this invention is to address the problems existing in the prior art by providing a multi-dimensional adaptive atmospheric river detection method based on BeiDou. This method utilizes the high-precision timing, real-time positioning, and short message communication services provided by BeiDou satellites to ensure the temporal synchronization and spatial accuracy of meteorological data. Furthermore, by combining a three-axis search mechanism with relative vorticity and comprehensive water vapor transport, it accurately eliminates tropical cyclone interference and optimizes the smoothness of atmospheric river axes and the accuracy of length calculation, thereby significantly improving the reliability of atmospheric river identification.

[0005] According to one aspect of this specification, a multi-dimensional adaptive atmospheric river detection method based on BeiDou is provided, comprising:

[0006] S1. Obtain raw meteorological data;

[0007] S2. The comprehensive water vapor transport is calculated based on the original meteorological data. The comprehensive water vapor transport is judged by Gaussian kernel density smoothing threshold and fixed threshold to obtain the preliminary atmospheric river.

[0008] S3. Calculate the geometric characteristics of the preliminary atmospheric river using a three-axis search mechanism;

[0009] S4. Based on the geometric characteristics of the preliminary atmospheric river, non-atmospheric rivers are excluded by relative vorticity and comprehensive water vapor transport to obtain the final atmospheric river.

[0010] Furthermore, by using a Gaussian kernel density smoothing threshold and a fixed threshold to determine the integrated water vapor transport, a preliminary atmospheric river is obtained, including:

[0011] The latitude and longitude corresponding to the comprehensive water vapor transport are organized into grid units;

[0012] By finding all grid cells that exceed a Gaussian kernel density smoothing threshold and a fixed threshold, one or more preliminary atmospheric rivers are formed from continuous grids that meet both thresholds; wherein, the Gaussian kernel density smoothing threshold is obtained by spatially smoothing the composite water vapor transport at a set percentage quantile; the fixed threshold is a set percentage quantile of the composite water vapor transport of all grid cells in the annual atmospheric river detection area.

[0013] Further, S3 includes:

[0014] The largest integrated water vapor transport grid unit in the preliminary atmospheric river is taken as the starting center;

[0015] The number of grids with integrated water vapor transport in three set directions is calculated based on the starting center, and the initial semicircle search direction is determined based on the number of grids.

[0016] An initial search is performed based on the initial semicircle search direction. The largest integrated water vapor transport grid cell in the same direction semicircle is found. The search continues with the largest integrated water vapor transport grid cell in the same direction semicircle as the center. The initial search is completed when the largest integrated water vapor transport of the last grid cell is equal to a fixed threshold.

[0017] After the initial search is completed, a reverse search is performed. The reverse search is the same as the initial search, but the direction of the semicircle is opposite to the direction of the initial search.

[0018] Connect all the largest integrated water vapor transport grid cells found to obtain the axis of a preliminary atmospheric river, and calculate the geometric characteristics of the axis.

[0019] Furthermore, the method also includes:

[0020] The geometric features of the axis include length and width, wherein the length is calculated using the following expression:

[0021]

[0022] lon i and lat i These are the longitude and latitude of the i-th largest integrated water vapor transport grid unit, respectively. and These are the longitude and latitude of the (i+1)th largest integrated water vapor transport grid unit, respectively;

[0023] The width is the Earth's total surface area of ​​the initial atmospheric river divided by the length of the initial atmospheric river.

[0024] Furthermore, tropical cyclone identification is performed based on the geometric characteristics of the preliminary atmospheric rivers, including:

[0025] Set the relative vorticity threshold;

[0026] Set a threshold for the number of clusters to reflect the spatial scale of the relative vorticity threshold corresponding to tropical cyclones;

[0027] When the sliding window detects a cluster that meets the quantity threshold, the average comprehensive water vapor transport value of the preliminary atmospheric river within the corresponding sliding window range is calculated.

[0028] When the average comprehensive water vapor transport value is greater than a set threshold, the roundness and eccentricity of the corresponding preliminary atmospheric river are calculated, and a first roundness threshold, a second roundness threshold, and an eccentricity threshold are set; among them, preliminary atmospheric rivers with roundness greater than the first roundness threshold, eccentricity less than the eccentricity threshold, and meeting the cluster number threshold are excluded; in addition, preliminary atmospheric rivers with roundness greater than the second roundness threshold are excluded.

[0029] Furthermore, based on the raw meteorological data, the comprehensive water vapor transport was calculated, including:

[0030]

[0031]

[0032]

[0033] Where g is the gravitational acceleration in meters per second squared, u and v are the eastward and northward components of the wind in meters per second, q is the specific humidity in kilograms per kilogram, dp is the pressure difference between two adjacent pressure levels, IVTu is the zonal component of the Integrated Water Transport (IVT), and IVTv is the polar component of the Integrated Water Transport.

[0034] Further, S1 includes:

[0035] Specific humidity, wind, and relative vorticity information are periodically observed using fixed-site equipment and uploaded to the BeiDou satellite system.

[0036] Using mobile observation equipment equipped with BeiDou positioning terminals, the position is dynamically adjusted in real time to capture the best observation area;

[0037] The system uses meter-level positioning services to track the trajectory of mobile observation equipment in real time and collect raw meteorological data.

[0038] According to one aspect of this specification, a multi-dimensional adaptive atmospheric river detection system based on BeiDou is provided, comprising:

[0039] The data acquisition module is used to acquire raw meteorological data;

[0040] The preliminary module is used to calculate the comprehensive water vapor transport based on the original meteorological data, and to judge the comprehensive water vapor transport by using a Gaussian kernel density smoothing threshold and a fixed threshold to obtain the preliminary atmospheric river;

[0041] The three-axis search module is used to calculate the geometric features of the preliminary atmospheric river through a three-axis search mechanism;

[0042] The atmospheric river acquisition module is used to determine tropical cyclones based on the geometric characteristics of the preliminary atmospheric river, exclude non-atmospheric rivers according to the determination results, and obtain the final atmospheric river.

[0043] According to one aspect of this specification, an electronic device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the BeiDou-based multidimensional adaptive atmospheric river detection method.

[0044] According to one aspect of this specification, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the BeiDou-based multidimensional adaptive atmospheric river detection method.

[0045] Compared with the prior art, the beneficial effects of the present invention are:

[0046] 1. This invention provides high-precision time synchronization, real-time positioning and short message communication services, ensuring the time synchronization and spatial accuracy of data within the system, and minimizing data transmission delay.

[0047] 2. This invention introduces cyclone dynamics feature discrimination to accurately identify atmospheric rivers under tropical cyclone interference scenarios, ensuring the continuity and smoothness of atmospheric river axes.

[0048] 3. This invention utilizes real-time BeiDou-transmitted observation data and real-time calculation of integrated water vapor transport to generate accurate atmospheric river early warning information, which is then rapidly broadcast to terminals in affected areas via the BeiDou short message communication link. This enables precise delivery and efficient response to early warning decisions, significantly improving the efficiency and reliability of disaster risk management. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart of a method according to an embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of the meteorological data acquisition structure according to an embodiment of the present invention;

[0052] Figure 3 This is a preliminary atmospheric river detection schematic diagram according to an embodiment of the present invention;

[0053] Figure 4 This is a flowchart illustrating the generation of a preliminary atmospheric river axis according to an embodiment of the present invention.

[0054] Figure 5 This is a flowchart illustrating the generation of the final realistic atmospheric river in an embodiment of the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] like Figure 1 As shown, this embodiment of the invention provides a multi-dimensional adaptive atmospheric river detection method based on BeiDou, including: S1, acquiring raw meteorological data; S2, calculating comprehensive water vapor transport based on the raw meteorological data, and judging the comprehensive water vapor transport by using a Gaussian kernel density smoothing threshold and a fixed threshold to obtain a preliminary atmospheric river; S3, calculating the geometric features of the preliminary atmospheric river through a three-axis search mechanism;

[0057] S4. Based on the geometric characteristics of the preliminary atmospheric rivers, tropical cyclones are determined, and non-atmospheric rivers are excluded according to the determination results to obtain the final atmospheric rivers.

[0058] Specifically, such as Figure 2 As shown, this embodiment of the invention also provides steps for acquiring raw meteorological data. Fixed stations such as coastal and high-altitude automatic weather stations are used to periodically observe specific humidity, wind, and relative vorticity information and upload it to the BeiDou satellite system. Mobile observation equipment, such as buoys and drones equipped with BeiDou positioning terminals, are used to dynamically adjust their positions in real time to capture the optimal observation area. The BeiDou satellite system provides high-precision time synchronization, ensuring consistent timestamps for all ground observation data; meter-level positioning services are used to track the mobile observation equipment's trajectory in real time and collect meteorological observation data.

[0059] Specifically, all meteorological observation equipment is equipped with a built-in BeiDou timing module to receive two-way time comparison signals from BeiDou-3 satellites. Ground observation equipment collects and transmits key meteorological data such as specific humidity, wind, and relative vorticity in real time, using BeiDou short message service for data transmission.

[0060] Specifically, such as Figure 3 As shown, this embodiment of the invention also provides preliminary atmospheric river detection. Based on raw meteorological data, comprehensive water vapor transport is calculated, and then a dual threshold consisting of a Gaussian kernel density smoothing threshold and a fixed threshold is set for preliminary atmospheric river detection. Specifically, this can be divided into the following steps:

[0061] 1. Calculation of Integrated Water Vapor Transport (IVT): IVT is used as the tracking variable to monitor atmospheric rivers. The calculation of IVT is based on the vertical integration of water vapor fluxes from 1000 hPa to 300 hPa within the Eulerian framework.

[0062]

[0063]

[0064]

[0065] Where g is the gravitational acceleration in meters per second squared, u and v are the eastward and northward components of the wind in meters per second, q is the specific humidity in kilograms per kilogram, dp is the pressure difference between two adjacent pressure levels, IVTu is the zonal component of the Integrated Water Transport (IVT), and IVTv is the polar component of the Integrated Water Transport.

[0066] 2. Dual-threshold detection: Gaussian kernel density smoothing threshold and fixed threshold are used to process composite water vapor transport. All grid cells exceeding the dual thresholds are identified. A grid cell refers to a grid composed of latitude and longitude coordinates, with each latitude and longitude corresponding to a value representing composite water vapor transport. One or more preliminary atmospheric rivers are formed by consecutive grids meeting the dual thresholds. First, the 85th percentile of composite water vapor transport for all time steps per year is calculated for each grid. This calculated 85th percentile composite water vapor transport is then spatially smoothed using Gaussian kernel density smoothing (GKS) to obtain the Gaussian kernel density smoothing threshold. The fixed threshold is defined as the 90th percentile of composite water vapor transport across all grids within the annual atmospheric river detection area of ​​the data.

[0067] Specifically, the composite water vapor transport at the 85th percentile for all time steps in each year is first calculated for each grid. This calculated 85th percentile composite water vapor transport is then spatially smoothed using Gaussian kernel density smoothing (GKS) to obtain a Gaussian kernel density smoothing threshold. A Gaussian kernel density smoothing threshold is obtained for each year. Next, a fixed threshold is calculated: the 90th percentile of the composite water vapor transport for all grids within the atmospheric river detection area each year. The composite water vapor transport values ​​are then filtered against the fixed threshold for their respective years, and then filtered again using the Gaussian kernel density smoothing threshold. Values ​​meeting both threshold requirements are retained; those not meeting the requirements are set to 0 (meaning none exist). All retained values ​​are considered preliminary atmospheric river data.

[0068] Specifically, such as Figure 4 As shown, this embodiment of the invention provides the generation of the axis of a preliminary atmospheric river and the calculation of its geometric features. Based on the detected preliminary atmospheric river, its corresponding axis is generated through a three-axis search mechanism, and then the geometric features (length, width) of the preliminary atmospheric river are calculated, as follows:

[0069] 1. Three-axis search: For the detection area at the same time step, traverse every connected region to find the largest integrated water vapor transport grid cell in the initial atmospheric river. Using this grid cell as the starting center "A", begin the three-axis search method. Specifically, at center "A", calculate the number of grids with integrated water vapor transport within the eastward, northeastward, and northward semicircles. Determine the initial semicircle search direction based on the number of grids. The standard width of the atmospheric river is 1000 km, and the semicircle radius is set to 500 km. Assuming the initial semicircle search direction is eastward, using "A" as the center, find the largest integrated water vapor transport grid cell "B" among all grids within the eastward semicircle. Then, using "B" as the center, continue searching for the next largest integrated water vapor transport grid cell in the eastward semicircle. The search terminates when the largest integrated water vapor transport of the last grid cell equals the previously calculated fixed threshold. The initial search is then complete, and the reverse search begins. The reverse search is the same as the initial search, but the semicircle direction is opposite to the initial semicircle search direction. Subsequent searches with the initial semicircle search direction determined to be northeast or north are the same as those with the initial semicircle search direction determined to be east. Then, all the maximum integrated water vapor transport grids found are connected to generate a preliminary axis of atmospheric river, solving the problems of jagged edges, deflections, and underestimated length of traditional axes.

[0070] 2. Calculate the geometric characteristics of the preliminary atmospheric river: The axis of the preliminary atmospheric river, the length of which can be defined by formula (4).

[0071]

[0072] lon i and lat i These are the longitude and latitude of the i-th largest integrated water vapor transport grid cell, respectively. Length is defined as the length of the initial atmospheric river. The width is calculated by dividing the total Earth surface area of ​​the initial atmospheric river by its length.

[0073] Specifically, such as Figure 5 As shown, this embodiment of the invention, based on the geometric characteristics of preliminary atmospheric rivers, utilizes relative eddy current combined with comprehensive water vapor transport to exclude non-atmospheric rivers, thus obtaining the final real atmospheric river. The specific steps are as follows:

[0074] 1. The initial atmospheric river length threshold is set to 2000 km, with an aspect ratio of 2. All initial atmospheric rivers with an axis not exceeding 2000 km and an aspect ratio less than 2 will be excluded.

[0075] 2. Relative vorticity combined with comprehensive water vapor transport: The relative vorticity at 850 hPa and the comprehensive water vapor transport value within the calculation sliding window are used to determine tropical cyclones. The determination process is as follows:

[0076] (1) The relative vorticity threshold is set to 10*10. -5 s -1 Clusters exceeding the relative vorticity threshold have geometric features similar to those of tropical cyclones, exhibiting certain circular and nonlinear characteristics.

[0077] (2) The number of clusters reflects the spatial scale of the relative vorticity of the tropical cyclone. The detection of tropical cyclones based on relative vorticity needs to consider the spatial scale and size of the relative vorticity. Therefore, in this embodiment of the invention, it is set that when the number of clusters is between 50 and 300, the roundness of the clusters should be greater than 0.3 and the eccentricity of the ellipse fitted to the clusters should be less than 0.9. When the number of clusters is between 300 and 500, the roundness should be greater than 0.1 and the eccentricity should be less than 0.9.

[0078] (3) When the sliding window detects a cluster that meets the conditions, the average comprehensive water vapor transport value of the preliminary atmospheric river within the corresponding sliding window range will be calculated. The average comprehensive water vapor transport value must be greater than 600 kg per meter per second.

[0079] (4) Calculate the roundness and eccentricity of the corresponding preliminary atmospheric river. For a roundness greater than 0.2 and an eccentricity less than 0.9, and the geographical area of ​​the preliminary atmospheric river includes the geographical area of ​​a cluster that meets the conditions, it will be identified as having potential tropical cyclones. If the roundness is greater than 0.8, the preliminary atmospheric river will be directly considered to have tropical cyclone-like characteristics and will be excluded. If the above four judgment conditions are met, it is considered that the preliminary atmospheric river is affected by a tropical cyclone, and the preliminary atmospheric river will be excluded to obtain the final real atmospheric river. Finally, atmospheric river early warning information is generated and accurately pushed to Beidou terminal devices in the affected area.

[0080] The implementation of the various embodiments of the present invention is based on programmed processing by a device with processor functionality. Therefore, in practical engineering, the technical solutions and functions of the various embodiments of the present invention are encapsulated into various modules. Based on this reality, and building upon the above embodiments, the embodiments of the present invention provide a BeiDou-based multi-dimensional adaptive atmospheric river detection system, which is used to execute a BeiDou-based multi-dimensional adaptive atmospheric river detection method from the above method embodiments.

[0081] The system includes: a data acquisition module for acquiring raw meteorological data; a preliminary module for calculating comprehensive water vapor transport based on the raw meteorological data, and judging the comprehensive water vapor transport through a Gaussian kernel density smoothing threshold and a fixed threshold to obtain preliminary atmospheric rivers; a triaxial search module for calculating the geometric characteristics of the preliminary atmospheric rivers through a triaxial search mechanism; and an atmospheric river acquisition module for determining tropical cyclones based on the geometric characteristics of the preliminary atmospheric rivers, excluding non-atmospheric rivers according to the determination results, and obtaining the final atmospheric rivers.

[0082] The multi-dimensional adaptive atmospheric river detection system based on BeiDou provided in this invention addresses the problems of unstable axis tracking, misjudgment of tropical cyclones, insufficient training samples, poor model interpretability, and excessively high real-time computation costs in existing technologies. It employs several modules, utilizing BeiDou satellites to provide high-precision time synchronization, real-time positioning, and short message communication services, ensuring the temporal synchronization and spatial accuracy of meteorological data. Furthermore, through a three-axis search mechanism and a combination of relative vorticity and comprehensive water vapor transport, it accurately eliminates tropical cyclone interference, optimizes the smoothness of atmospheric river axes and the accuracy of length calculation, thereby significantly improving the reliability of atmospheric river identification.

[0083] Based on the same inventive concept as the foregoing embodiments, this embodiment of the invention also provides an electronic device, including a memory and a processor. The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize a multi-dimensional adaptive atmospheric river detection method based on BeiDou as proposed in the above embodiments.

[0084] This invention also provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, this program overcomes problems such as unstable axis tracking, misjudgment of tropical cyclones, insufficient training samples, poor model interpretability, and excessively high real-time computation costs. It ensures the temporal synchronization and spatial accuracy of data within the system, minimizes data transmission latency, introduces cyclone dynamics feature discrimination, accurately identifies atmospheric rivers under tropical cyclone interference scenarios, and ensures the continuity and smoothness of atmospheric river axes.

[0085] The storage medium can be any non-volatile storage device such as a hard disk, solid-state drive, flash drive, or optical disk, used to store computer program code and necessary data files. The stored computer program includes: a data acquisition module, a preliminary module, a three-axis search module, and an atmospheric river acquisition module.

[0086] In summary, this invention provides a multi-dimensional adaptive atmospheric river detection method based on BeiDou. Utilizing the precise timing provided by the BeiDou satellite to meteorological observation equipment, the equipment can output accurate real-time meteorological observation data. Starting from the real-time wind field, specific humidity, and relative vorticity meteorological observation data of the meteorological observation equipment, accurate identification of atmospheric rivers is achieved through reasonable threshold definition and efficient logic control. Specifically, based on the traditional threshold morphology algorithm: (1) a threshold field is constructed using Gaussian kernel local smoothing to avoid missed detections due to resolution differences caused by a single fixed threshold; (2) a three-axis search mechanism is proposed to replace the shortest distance assumption, significantly improving the smoothness and length calculation accuracy of the atmospheric river axis; (3) integrated water vapor transport and relative vorticity are incorporated into the discrimination model to avoid ignoring connectivity when position and intensity weights are large, accurately distinguishing tropical cyclones overlapping with atmospheric rivers; (4) a travel time constraint is introduced in the shortest path search to limit the pathfinding range and improve computational efficiency. Therefore, the embodiments of the present invention can not only efficiently and accurately eliminate deviations and misjudgments, but also complete the key trajectory records of atmospheric river water vapor channels, significantly improve the utilization rate and scientific research value of real-time meteorological data, and enhance the disaster prevention and mitigation capabilities of atmospheric river-affected areas.

[0087] Finally, it should be noted that the above specific embodiments are merely representative examples of the present invention. Obviously, the present invention is not limited to the above specific embodiments and many variations are possible. Any simple modifications, equivalent changes, and alterations made to the above specific embodiments based on the technical essence of the present invention should be considered within the protection scope of the present invention.

Claims

1. A multi-dimensional adaptive atmospheric river detection method based on BeiDou, characterized in that, include: S1. Obtain raw meteorological data; S2. The comprehensive water vapor transport is calculated based on the original meteorological data. The comprehensive water vapor transport is judged by Gaussian kernel density smoothing threshold and fixed threshold to obtain the preliminary atmospheric river. S3. Calculate the geometric characteristics of the preliminary atmospheric river using a three-axis search mechanism; S4. Based on the geometric characteristics of the preliminary atmospheric rivers, tropical cyclones are determined, and non-atmospheric rivers are excluded according to the determination results to obtain the final atmospheric rivers.

2. The multi-dimensional adaptive atmospheric river detection method based on BeiDou as described in claim 1, characterized in that, By assessing the comprehensive water vapor transport using a Gaussian kernel density smoothing threshold and a fixed threshold, a preliminary atmospheric river system is obtained, including: The latitude and longitude corresponding to the comprehensive water vapor transport are organized into grid units; By finding all grid cells that exceed a Gaussian kernel density smoothing threshold and a fixed threshold, one or more preliminary atmospheric rivers are formed from continuous grids that meet both thresholds; wherein, the Gaussian kernel density smoothing threshold is obtained by spatially smoothing the composite water vapor transport at a set percentage quantile; the fixed threshold is a set percentage quantile of the composite water vapor transport of all grid cells in the annual atmospheric river detection area.

3. The multi-dimensional adaptive atmospheric river detection method based on BeiDou as described in claim 1, characterized in that, The S3 includes: The largest integrated water vapor transport grid unit in the preliminary atmospheric river is taken as the starting center; The number of grids with integrated water vapor transport in three set directions is calculated based on the starting center, and the initial semicircle search direction is determined based on the number of grids. An initial search is performed based on the initial semicircle search direction. The largest integrated water vapor transport grid cell in the same direction semicircle is found. The search continues with the largest integrated water vapor transport grid cell in the same direction semicircle as the center. The initial search is completed when the largest integrated water vapor transport of the last grid cell is equal to a fixed threshold. After the initial search is completed, a reverse search is performed. The reverse search is the same as the initial search, but the direction of the semicircle is opposite to the direction of the initial search. Connect all the largest integrated water vapor transport grid cells found to obtain the axis of a preliminary atmospheric river, and calculate the geometric characteristics of the axis.

4. The multi-dimensional adaptive atmospheric river detection method based on BeiDou as described in claim 3, characterized in that, The method further includes: The geometric features of the axis include length and width, wherein the length is calculated using the following expression: , lon i and lat i These are the longitude and latitude of the i-th largest integrated water vapor transport grid unit, respectively. and These are the longitude and latitude of the (i+1)th largest integrated water vapor transport grid unit, respectively; The width is the Earth's total surface area of ​​the initial atmospheric river divided by the length of the initial atmospheric river.

5. The multi-dimensional adaptive atmospheric river detection method based on BeiDou as described in claim 1, characterized in that, Tropical cyclone identification based on the aforementioned preliminary atmospheric river geometry includes: Set the relative vorticity threshold; Set a threshold for the number of clusters to reflect the spatial scale of the relative vorticity threshold corresponding to tropical cyclones; When the sliding window detects a cluster that meets the quantity threshold, the average comprehensive water vapor transport value of the preliminary atmospheric river within the corresponding sliding window range is calculated. When the average comprehensive water vapor transport value is greater than a set threshold, the roundness and eccentricity of the corresponding preliminary atmospheric river are calculated, and a first roundness threshold, a second roundness threshold, and an eccentricity threshold are set; among them, preliminary atmospheric rivers with roundness greater than the first roundness threshold, eccentricity less than the eccentricity threshold, and meeting the cluster number threshold are excluded; in addition, preliminary atmospheric rivers with roundness greater than the second roundness threshold are excluded.

6. The multi-dimensional adaptive atmospheric river detection method based on BeiDou as described in claim 1, characterized in that, The comprehensive water vapor transport calculated based on raw meteorological data includes: , , , Where g is the acceleration due to gravity, in meters per second squared; u and v are the eastward and northward components of the wind, in meters per second; q is the specific humidity, in kilograms per kilogram; dp is the pressure difference between two adjacent pressure levels; IVTu is the zonal component of the combined water vapor transport; and IVTv is the polar component of the combined water vapor transport.

7. The multi-dimensional adaptive atmospheric river detection method based on BeiDou as described in claim 1, characterized in that, S1 includes: Specific humidity, wind, and relative vorticity information are periodically observed using fixed-site equipment and uploaded to the BeiDou satellite system. Using mobile observation equipment equipped with BeiDou positioning terminals, the position is dynamically adjusted in real time to capture the best observation area; The system uses meter-level positioning services to track the trajectory of mobile observation equipment in real time and collect raw meteorological data.

8. A multi-dimensional adaptive atmospheric river detection system based on BeiDou, characterized in that, include: The data acquisition module is used to acquire raw meteorological data; The preliminary module is used to calculate the comprehensive water vapor transport based on the original meteorological data, and to judge the comprehensive water vapor transport by using a Gaussian kernel density smoothing threshold and a fixed threshold to obtain the preliminary atmospheric river; The three-axis search module is used to calculate the geometric features of the preliminary atmospheric river through a three-axis search mechanism; The atmospheric river acquisition module is used to determine tropical cyclones based on the geometric characteristics of the preliminary atmospheric river, exclude non-atmospheric rivers according to the determination results, and obtain the final atmospheric river.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the BeiDou-based multidimensional adaptive atmospheric river detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the BeiDou-based multidimensional adaptive atmospheric river detection method according to any one of claims 1 to 7.

Images