Method for automatic recognition of atmospheric vortex horizontal scale and matching with mesoscale convective system

Abstract

The present application belongs to the technical field of atmospheric dynamic structure identification, and discloses a method for automatically identifying the horizontal scale of atmospheric vortex and matching the vortex with a mesoscale convective system, which comprises the following steps: based on the two-dimensional coordinates of the vortex center of each extended height layer of a three-dimensional vortex at time t, a vortex range identification algorithm is used to identify the vortex range of each extended height layer of the three-dimensional vortex at time t; a vortex range smoothing processing algorithm is used to smooth the vortex range at each time t in the vortex life history time sequence; and it is determined whether the range of the mesoscale convective system coincides with the stable vortex range of any extended height layer of the three-dimensional vortex at the same time t. The present application automatically and accurately identifies the range of a large-scale and long-time atmospheric vortex, solves the problem that the traditional vortex identification cannot determine the horizontal range and the real influence scale, realizes the automatic matching and correlation analysis between the vortex and the mesoscale convective system, and constructs an automatic identification process which can be used in real-time or reanalysis business platforms.

Description

Technical Field

[0001] This invention relates to the field of severe convective weather monitoring and atmospheric dynamic structure identification technology, specifically to the automatic identification of horizontal-scale atmospheric vortices and its matching method with mesoscale convective systems. Background Technology

[0002] Currently, severe weather events such as torrential rains and severe convection occur frequently on a small to medium scale. Vortexes and mesoscale convective systems are important weather factors that produce severe weather, and their coupling can produce even more extreme weather phenomena. Most existing vortex identification methods are based on vorticity thresholds. On the one hand, they can only locate the vortex center and there is no method to automatically determine the actual horizontal boundary and scale of the vortex. On the other hand, the identification of mesoscale convective systems (MCS) usually relies on satellite brightness temperature or radar echo connectivity areas, and there is currently a lack of an algorithmic system to automatically determine whether they are related to the dynamic structure of the vortex. This leads to limitations in our understanding of the vortex's influence range, convective organization structure, and precipitation distribution.

[0003] The specific issues are as follows:

[0004] 1. Existing methods for identifying vortex ranges in meteorology and their limitations:

[0005] (1) Vortex field identification: Calculate the Laplace operator of relative vorticity, and identify vortex centers with negative Laplace values ​​and positive relative vorticity. Then calculate the average vorticity within a 20 km radius of the vortex center. and the standard deviation of vorticity When the circular area enclosed by a certain radius ( It is the maximum relative vorticity within the circular region), and When the radius is equal to the radius of the vortex, the vortex boundary is determined.

[0006] This method cannot distinguish between shear lines, upper-level troughs, and vortices, and the number of vortices identified is significantly higher than expected.

[0007] (2) Identification of height field and closed streamlines: Objectively analyze the conventional observation data of each layer from the ground to 500 hPa, draw the flow field and pressure field, manually determine whether there is a closed cyclonic circulation and coordinate with the low-pressure center. If there is, it is identified as a low vortex. The horizontal range of the vortex is determined by the range of the largest closed streamline in each layer from 925 to 500 hPa.

[0008] This method, which relies on manual judgment of closed streamlines, is somewhat subjective and inefficient.

[0009] (3) Potential height field identification: The original coordinate system is moved to a coordinate system with the vortex center as the polar origin. Then, the potential height is interpolated to the 36 directions of the vortex, with 0° representing east, an angular step of 10°, and a radial spatial step of 1 km. The potential height gradient in each direction is found to be less than 0.1 gpm for the first time. A point km⁻¹ is used to determine a closed curve M, representing the horizontal geometry of the vortex, based on points in all directions.

[0010] In this method, the low-pressure center and the vortex center do not match perfectly, and for smaller vortices, the change in the height field is not obvious, which causes errors and difficulties in identification.

[0011] 2. In physical oceanography, methods for identifying vortex extent include the winding angle method: Instantaneous streamlines at each grid point are calculated based on the velocity field, followed by the cumulative change (bending angle) in the direction of each streamline. Streamlines with a cumulative angle |α| > 2π are identified as vortices, corresponding to a closed or spiral curve. Streamlines are then clustered based on their rotation direction and location; those with a center distance less than a threshold (20 km) and the same polarity are considered to belong to the same vortex. Finally, the covariance matrix of all streamline points for a vortex is calculated, using the eigenvalues ​​as the lengths of the major and minor axes of an ellipse, and the eigenvectors representing the direction of the ellipse. This ellipse is considered the extent of the vortex.

[0012] This method works well for identifying ocean eddies with relatively regular shapes, but for atmospheric eddies with varied shapes, the streamlines tracked using this method often include parts that do not belong to the eddy. Since this method identifies eddies using closed streamlines, which is consistent with the manual judgment method, a method for identifying atmospheric eddies was invented based on this method.

[0013] 3. The existing method of matching vortices with mesoscale convection systems is to use a circular region of fixed radius as the range of the vortex and match it with the mesoscale convection system. However, due to the different vortex shapes and scales, there will be significant errors. Therefore, based on the above identification of the horizontal scale of the vortex, it is necessary to further match it with the mesoscale convection system.

[0014] Therefore, there is an urgent need for a method that can automatically extract the horizontal boundary of atmospheric vortex and achieve vortex-MCS matching. Solving this problem is of great significance for improving the accuracy of severe weather forecasts and supporting scientific disaster prevention and mitigation. Summary of the Invention

[0015] To address the shortcomings of existing technologies, this invention provides an automatic identification method for horizontal-scale atmospheric vortices and its matching method with mesoscale convection systems, which can effectively solve the above-mentioned problems.

[0016] The technical solution adopted in this invention is as follows:

[0017] This invention provides an automatic identification method for horizontal-scale atmospheric eddies and its matching method with mesoscale convective systems, including:

[0018] Step S1: Read the time series data of wind field evolution in the study area during the study period;

[0019] Step S2: The study area is vertically divided into several height layers, and the horizontal plane of each height layer is gridded into several grid points;

[0020] Based on the wind field evolution data time series, the evolution characteristics of each three-dimensional vortex at each time t in its life history are identified, forming a three-dimensional vortex evolution characteristic time series; wherein, the evolution characteristics include each height layer of the three-dimensional vortex at time t, referred to as each extension height layer, and the two-dimensional coordinates of the vortex center in the horizontal plane of each extension height layer.

[0021] Step S3: Based on the two-dimensional coordinates of the vortex center of the three-dimensional vortex at each extension height layer at time t, the vortex range identification algorithm is used to identify the vortex range of the three-dimensional vortex at each extension height layer at time t.

[0022] Step S4: Based on the vortex range time series formed by the vortex range at each time t in the same extension height layer during the life history of the three-dimensional vortex, the vortex range smoothing algorithm is used to smooth the vortex range at each time t in the vortex range time series to obtain the stable vortex range of the three-dimensional vortex at each time t in the extension height layer.

[0023] Step S5: Obtain the range of the mesoscale convection system in the study area during the study period;

[0024] Determine whether the range of the mesoscale convection system overlaps with the range of the stable vortex at any extension height layer of the three-dimensional vortex at the same time t; if so, the range of the mesoscale convection system matches the range of the three-dimensional vortex at time t.

[0025] Furthermore, the evolutionary characteristics of each three-dimensional vortex at each time step t throughout its life history were identified by the following method:

[0026] Based on the wind field evolution data time series, identify the vortex characteristics at each altitude layer at each time t within the study period;

[0027] Horizontal linking of vortex features: Determine whether two vortex features at adjacent times and close positions at the same height layer are similar. If they are similar, it is concluded that the two vortex features are the same vortex. Thus, several similar vortex features at a continuous time at the same height layer are obtained, representing the evolution features of the same vortex at that height layer.

[0028] Vertical linking of vortex features: Determine whether two vortex features at adjacent height layers are similar at the same time t. If they are similar, it is concluded that the two vortex features are the same vortex. Thus, several similar vortex features at a continuous height layer at time t are obtained, representing the height layers where the same vortex extends at that time t.

[0029] Furthermore, step S3 specifically includes:

[0030] Step S31, the three-dimensional vortex at time t is located at the vortex center of the extension height layer. ,in, Each 3D vortex has a unique identifier. The floor number for the extended height layer, , The grid index represents the grid point where the vortex center is located, indicating the grid point position in the longitude and latitude directions, respectively;

[0031] Step S32, Track the vortex center 3 The streamlines of the 3-neighborhood's eight adjacent grid points, the streamlines of the vortex center, and the streamlines of its eight adjacent grid points constitute a streamline set. :

[0032] ;

[0033] in: for Abbreviation, representing time t, The extension height layer, numbered as The vortex passes through the grid point Streamlines;

[0034] Step S33, calculate each streamline cumulative angle change value ;

[0035] Step S34, according to each flow line Cumulative angle change value Identify each flow line Is it a closed streamline that rotates counterclockwise? Streamline set The closed streamlines identified by the counterclockwise rotation form a set of closed streamlines. ;

[0036] Step S35: Calculate the set of closed streamlines. The centroids of each closed streamline in the set of centroids constitute the set of centroids. ;

[0037] Step S36: Based on the centroid of each closed streamline, cluster and filter the closed streamlines, retaining those that are aligned with the vortex center. Distance less than Closed streamlines;

[0038] Step S37: Perform edge trimming on each of the remaining closed streamlines, trimming the streamline points located at the edge of each closed streamline, thereby obtaining each edge-trimmed closed streamline;

[0039] Step S38: Based on the closed streamlines after trimming each edge, the vortex center is obtained as follows: The confidence ellipse has three shape parameters: major axis length, minor axis length, and major axis angle; this confidence ellipse is the identified vortex range.

[0040] Furthermore, calculate each streamline. Cumulative angle change value The method is as follows:

[0041] Each flow line Composed of a finite number of discrete streamline points, it can be represented as:

[0042]

[0043] in: streamline The number of streamline points it possesses; Let the first streamline point, the second streamline point, ..., the... be the... Streamline point;

[0044] Cumulative angle change value for:

[0045]

[0046] in: streamline point To streamline point A directed line segment; streamline point To streamline point A directed line segment; It is the angle between two directed line segments.

[0047] Furthermore, according to each flow line Cumulative angle change value Identify each flow line Whether it is a closed streamline rotating counterclockwise, specifically:

[0048] If streamline Cumulative angle change value Not less than streamline It is a closed streamline that rotates counterclockwise.

[0049] Furthermore, edge trimming is performed on each retained closed streamline, removing the streamline points located at the edges of each closed streamline, thus obtaining each edge-trimmed closed streamline as follows:

[0050] Step S371: The remaining closed streamlines constitute a set of closed streamlines. The streamline points of each closed streamline that are retained constitute the streamline point set. :

[0051]

[0052] in: Closed streamline streamline points in , streamline points Longitude and latitude;

[0053] Step S372, set of streamline points The streamline points in the data are analyzed using longitude and latitude coordinates to obtain a set of longitudes representing the longitude distribution. and the set of latitudes representing the latitude distribution. ;

[0054] Step S373, take the longitude distribution quantiles and quantiles As the cutoff boundary in the longitude direction, the latitude distribution is taken as... quantiles and quantiles As the cutoff boundary in the latitudinal direction, cutoff and edge trimming are performed in both the longitudinal and latitudinal directions respectively;

[0055] The streamlined point set after edge clipping Defined as:

[0056]

[0057] This achieves edge clipping.

[0058] Furthermore, the vortex center is The three shape parameters of the confidence ellipse are represented as: major axis length minor axis length and major axis angle .

[0059] Furthermore, a vortex range smoothing algorithm is employed to smooth the vortex range at each time t in the vortex range time series, thereby obtaining the stable vortex range of the three-dimensional vortex at each time t in the extension height layer, specifically as follows:

[0060] Step S41, the vortex range time series is represented as follows: ;in, For the number Three-dimensional vortex at height layer The vortex range at time t; = ; respectively vortex range The corresponding major axis length, minor axis length, and major axis angle of the confidence ellipse;

[0061] For the number The life history of a three-dimensional vortex; numbered as Three-dimensional vortex at height layer In its life history The set formed by each moment in time is represented as ;

[0062] Step S42, for the vortex range The major axis length, minor axis length, and major axis angle are smoothed respectively. The confidence ellipse formed by the smoothed major axis length, minor axis length, and major axis angle is the stable vortex range.

[0063] Will Use elements uniformly The method for smoothing it is as follows:

[0064] element In its life history The resulting time series is represented as ;

[0065] For each element in the time series , The smoothed elements are obtained by performing the following smoothing process. :

[0066]

[0067] in: Given a time threshold; Therefore Number of time intervals extending to the left and right of the center. The set of times, denoted as [ , ], where, if Then let ; This is a floor function; for operations located in the life history... Moments at the middle edge position That is, its left side does not have At any given moment or to its right, there is no [something]. For each time point, the missing value location is determined using a mirroring method; the time set [ , Each moment in the diagram is represented as... Gaussian weights = ; For life history Every moment The corresponding element value, .

[0068] The automatic identification of horizontal-scale atmospheric vortices and its matching method with mesoscale convective systems provided by this invention have the following advantages:

[0069] This invention automatically and accurately identifies the range of large-scale, long-term atmospheric vortices, solving the problem that traditional vortex identification cannot determine the horizontal range and the true scale of influence; it realizes automated matching and correlation analysis between vortices and mesoscale convective systems; and it constructs an automatic identification process that can be used in real-time or reanalysis business platforms. Attached Figure Description

[0070] 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.

[0071] Figure 1 A flowchart of the automatic identification of horizontal-scale atmospheric vortices and its matching method with mesoscale convective systems provided by the present invention;

[0072] Figure 2 An illustration of streamlines being tracked in the ocean and atmosphere;

[0073] Figure 3 The result diagram of the edge clipping effect and confidence ellipse range provided by the present invention;

[0074] Figure 4 This is a diagram showing the final identification effect of the method of the present invention for different height layers and long-lived vortex ranges;

[0075] Figure 5This is a schematic diagram of the height layers of the three-dimensional vortex of the present invention as it extends at each moment t throughout its life cycle. Detailed Implementation

[0076] To make the technical problems solved, the technical solutions, and the beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the invention.

[0077] The present invention provides an automatic identification method for the horizontal scale of atmospheric vortices and its matching with mesoscale convective systems. This method is based on arbitrary wind field data to accurately identify and depict the horizontal scale of atmospheric vortex systems and match them with identified mesoscale convective systems. The present invention automatically and accurately identifies the range of large-scale and long-term atmospheric vortices, solving the problem that traditional vortex identification cannot determine the horizontal range and the actual scale of influence; it realizes automated matching and correlation analysis between vortices and mesoscale convective systems; and it constructs an automatic identification process that can be used in real-time or reanalysis business platforms.

[0078] See Figure 1 The present invention provides an automatic identification method for horizontal-scale atmospheric vortices and its matching method with mesoscale convective systems, comprising:

[0079] Step S1: Read the time series data of wind field evolution in the study area during the study period;

[0080] Step S2: The study area is vertically divided into several height layers, and the horizontal plane of each height layer is gridded into several grid points;

[0081] Based on the wind field evolution data time series, the evolution characteristics of each three-dimensional vortex at each time t in its life history are identified, forming a three-dimensional vortex evolution characteristic time series; wherein, the evolution characteristics include each height layer of the three-dimensional vortex at time t, referred to as each extension height layer, and the two-dimensional coordinates of the vortex center in the horizontal plane of each extension height layer.

[0082] In this step, the evolution characteristics of each three-dimensional vortex at each time t in its life history are identified by the following method: based on the time series of wind field evolution data, the vortex characteristics at each altitude layer at each time t within the study period are identified.

[0083] Horizontal linking of vortex features: Determine whether two vortex features at adjacent times and close positions at the same height layer are similar. If they are similar, it is concluded that the two vortex features are the same vortex. Thus, several similar vortex features at a continuous time at the same height layer are obtained, representing the evolution features of the same vortex at that height layer.

[0084] Vertical linking of vortex features: Determine whether two vortex features at adjacent height layers are similar at the same time t. If they are similar, it is concluded that the two vortex features are the same vortex. Thus, several similar vortex features at a continuous height layer at time t are obtained, representing the height layers where the same vortex extends at that time t.

[0085] like Figure 5 The image shows a schematic diagram of the height layers of a three-dimensional vortex as it extends at each time t throughout its life cycle; Figure 5 In the diagram, "T" indicates the time interval, "L" indicates the level (unit: hectopascals), the numbers represent vortex numbers, and identical shading indicates the same vortex. Figure 5 There are two types of three-dimensional vortices, numbered 1 and 2 respectively. For the three-dimensional vortex numbered 1, it can be seen that at time 1, the height layer of its extension is 950 hPa; at time 2, the height layers of its extension are 950 hPa, 900 hPa and 850 hPa. Figure 5 The information displayed is related to the generation and destruction of the three-dimensional vortex numbered 1.

[0086] Step S3: Based on the two-dimensional coordinates of the vortex center of the three-dimensional vortex at each extension height layer at time t, the vortex range identification algorithm is used to identify the vortex range of the three-dimensional vortex at each extension height layer at time t.

[0087] The main idea of ​​the vortex range identification algorithm in this step is to trace and cluster closed streamlines, trim streamline points located at the edge of the closed streamlines, and calculate the confidence ellipse based on the trimmed closed streamlines as the vortex range.

[0088] Step S31, the three-dimensional vortex at time t is located at the vortex center of the extension height layer. ,in, Each 3D vortex has a unique identifier. The floor number for the extended height layer, , The grid index represents the grid point where the vortex center is located, indicating the grid point position in the longitude and latitude directions, respectively;

[0089] Step S32, Track the vortex center 3 The streamlines of the 3-neighborhood's eight adjacent grid points, the streamlines of the vortex center, and the streamlines of its eight adjacent grid points constitute a streamline set. :

[0090] ;

[0091] in: for Abbreviation, representing time t, The extension height layer, numbered as The vortex passes through the grid point Streamlines;

[0092] Step S33, calculate each streamline Cumulative angle change value ;

[0093] Specifically, each flow line Composed of a finite number of discrete streamline points, it can be represented as:

[0094]

[0095] in: streamline The number of streamline points it possesses; Let the first streamline point, the second streamline point, ..., the... be the... Streamline point;

[0096] Cumulative angle change value for:

[0097]

[0098] in: streamline point To streamline point A directed line segment; streamline point To streamline point A directed line segment; It is the angle between two directed line segments.

[0099] Step S34, according to each flow line Cumulative angle change value Identify each flow line Is it a closed streamline that rotates counterclockwise? Streamline set The closed streamlines identified by the counterclockwise rotation form a set of closed streamlines. .

[0100] In this step, based on each streamline Cumulative angle change value Identify each flow line Whether it is a closed streamline rotating counterclockwise, specifically: if the streamline Cumulative angle change value Not less than streamline It is a closed streamline that rotates counterclockwise.

[0101] Closed streamline collection Represented as:

[0102]

[0103] Step S35: Calculate the set of closed streamlines. The centroids of each closed streamline in the set of centroids constitute the set of centroids. :

[0104]

[0105] Where: closed streamlines The center of mass is This leads to the formation of a set of centroids. .

[0106] Step S36: Based on the centroid of each closed streamline, cluster and filter the closed streamlines, retaining those that are aligned with the vortex center. Distance less than The closed streamlines; in this step, density-based clustering—DBSCAN (Density-Based Spatial Clustering of Applications with Noise)—can be used, and the formula for centroid selection is expressed as:

[0107]

[0108] in: The set of centroids to be preserved. Represents the center of mass To the center of the vortex The distance.

[0109] The formula for closed streamline screening is expressed as follows:

[0110]

[0111] This represents the set of closed streamlines that are retained.

[0112] Step S37: Due to the complex shape of atmospheric vortices, there are often instances where some streamline points are far from the main body of the vortex. To obtain an accurate vortex shape, this invention trims and removes these streamline points. Specifically, the remaining closed streamlines are edge-trimmed, removing the streamline points located at the edges of each closed streamline, thereby obtaining edge-trimmed closed streamlines.

[0113] This step is specifically as follows:

[0114] Step S371: The remaining closed streamlines constitute a set of closed streamlines. The streamline points of each closed streamline that are retained constitute the streamline point set. :

[0115]

[0116] in: Closed streamline streamline points in , streamline points Longitude and latitude;

[0117] Step S372: In order to eliminate streamline points that are far from the vortex range, the streamline point set is... The streamline points in the data are analyzed using longitude and latitude coordinates to obtain a set of longitudes representing the longitude distribution. and the set of latitudes representing the latitude distribution. ;

[0118] Step S373, take the longitude distribution quantiles and quantiles As the cutoff boundary in the longitude direction, the latitude distribution is taken as... quantiles and quantiles As the cutoff boundary in the latitudinal direction, cutoff and edge trimming are performed in both the longitudinal and latitudinal directions respectively;

[0119] The streamlined point set after edge clipping Defined as:

[0120]

[0121] This achieves edge clipping.

[0122] Step S38: Influenced by topography and weather systems at different scales, the vortex resembles an ellipse. Therefore, based on the closed streamlines after trimming each edge, the vortex center is obtained as... A confidence ellipse having three shape parameters: the length of the major axis. minor axis length and major axis angle The confidence ellipse represents the identified vortex range.

[0123] Step S4: Since the vortex range identified in step S3 is random and fluctuates, the vortex range is smoothed in order to reduce the error caused by randomness.

[0124] Based on the vortex range time series formed by the vortex range at each time t in the same extension height layer during the life history of the three-dimensional vortex, the vortex range smoothing algorithm is used to smooth the vortex range at each time t in the vortex range time series to obtain the stable vortex range of the three-dimensional vortex at each time t in the extension height layer.

[0125] This step is specifically as follows:

[0126] Step S41, the vortex range time series is represented as follows: ;in, For the number Three-dimensional vortex at height layer The vortex range at time t; = ; respectively vortex range The corresponding major axis length, minor axis length, and major axis angle of the confidence ellipse;

[0127] For the number The life history of a three-dimensional vortex; numbered as Three-dimensional vortex at height layer In its life history The set formed by each moment in time is represented as ;

[0128] Step S42, for the vortex range The major axis length, minor axis length, and major axis angle are smoothed respectively. The confidence ellipse formed by the smoothed major axis length, minor axis length, and major axis angle is the stable vortex range.

[0129] Will Use elements uniformly The method for smoothing it is as follows:

[0130] element In its life history The resulting time series is represented as ;

[0131] For each element in the time series , The smoothed elements are obtained by performing the following smoothing process. :

[0132]

[0133] in: Given a time threshold; Therefore Number of time intervals extending to the left and right of the center. The set of times, denoted as [ , ], where, if Then let ; This is a floor function; for operations located in the life history... Moments at the middle edge position That is, its left side does not have At any given moment or to its right, there is no [something]. For each time point, the missing value location is determined using a mirroring method; the time set [ , Each moment in the diagram is represented as... Gaussian weights = ; For life history Every moment The corresponding element value, .

[0134] After the above steps, a set of vortex ranges with a certain degree of stability and capable of representing the characteristics of vortex development and evolution is obtained. .

[0135] Step S5: Obtain the range of the mesoscale convection system in the study area during the study period; the range of the mesoscale convection system includes the life history, coverage area and location of the mesoscale convection system.

[0136] Determine whether the range of the mesoscale convection system overlaps with the range of the stable vortex at any extension height layer of the three-dimensional vortex at the same time t; if so, the range of the mesoscale convection system matches the range of the three-dimensional vortex at time t.

[0137] Specifically, precipitation data is used to identify mesoscale convective systems that produce precipitation, and the extent of each mesoscale convective system is determined. Is it related to the vortex range? If they coincide, and the following conditions are met:

[0138]

[0139] It is then assumed that the MCS matches the vortex numbered vor at time t.

[0140] Combination Figure 5 If, during the life history of a three-dimensional vortex, there is a moment and a height at which the vortex range overlaps with that of a mesoscale convection system, the two are considered to be matched.

[0141] This invention provides an automatic identification method for horizontal-scale atmospheric eddies and its matching method with mesoscale convective systems. The main idea is as follows:

[0142] 1. This invention analyzes the cumulative angle change values ​​of streamlines in the center of a vortex and its neighborhood. When the set closure criterion is met and streamline points are clustered, streamlines belonging to the vortex are obtained. Based on this, the streamlines are trimmed at the edges, and the range of the confidence ellipse for retaining the streamline points of the vortex is calculated.

[0143] 2. Based on the vortex life history, different averaging calculations are applied to the vortex range at the same height level. For the vortex range at the same height level at different times with a life history greater than a certain threshold, a Gaussian kernel weighted moving average is applied; otherwise, an arithmetic average is applied to obtain a stable vortex horizontal scale.

[0144] 3. Use precipitation observations to screen and verify the identified mesoscale convective systems;

[0145] 4. Automatic matching between atmospheric vortices and mesoscale convective systems is achieved by determining whether their ranges overlap.

[0146] As an example, such as Figure 2 The image shown is an illustration of streamlines tracing in the ocean and atmosphere. Figure 2 Figure a shows the streamline identification effect of a mesoscale eddy in the ocean (adapted from Ari Sadarjoen and Post, 2000). Figure 2 As can be seen, the counterclockwise rotating streamlines in the ocean are basically closed, with the starting and ending points close to each other. Therefore, when identifying ocean vortex streamlines, it is usually not necessary to trim them.

[0147] like Figure 2 Figure b shows the effect of tracing atmospheric vortex streamlines using the method of steps S31 to S36 in step S3 of this invention; Figure 2 In diagram b, the yellow line represents the fully identified streamlines, and the blue line represents the streamlines that satisfy the condition of cumulative angle change greater than 2π and have undergone clustering. From Figure 2 As can be seen from step S36 of this invention, the vortex streamlines (blue solid lines) identified include parts outside the vortex range in the atmosphere. If a confidence ellipse is calculated for this part of the streamlines, the horizontal scale of the vortex will be too large. Therefore, it is necessary to trim the part that exceeds the range.

[0148] like Figure 3 The image shown is a result of the edge clipping effect and the range of the confidence ellipse. Figure 3 In the diagram, the red dots represent streamlined points retained after edge clipping, and the blue shading represents the range of the confidence ellipse. From... Figure 3 As can be seen, the red streamlined points after edge trimming using the method of this invention are more in line with the shape of a vortex. Figure 3The shaded area of ​​b is a confidence ellipse containing most of the red streamline points, which can be seen to be quite consistent with the vortex range.

[0149] like Figure 4 The image shown is the final identification result of vortexes at different heights and with long lifespans using the method of this invention. The blue shading represents the final identification range; the numbers are vortex numbers. From Figure 4 As can be seen, the identification algorithm of this invention can accurately represent the range of each vortex at different altitudes; and can represent the development and evolution characteristics of the range of long-lived vortices. For example, the major axis direction and scale of the elliptical range of the vortex numbered 1 change with the vortex morphology.

[0150] This invention provides an automatic identification method for horizontal-scale atmospheric eddies and its matching method with mesoscale convective systems. The innovations are summarized as follows:

[0151] 1. This invention tracks the counterclockwise rotating closed streamlines of the vortex center and its surroundings;

[0152] In this invention, the vortex center data is first obtained, and the subsequent process is based on the vortex center data to identify the vortex range. In contrast, in traditional methods, the vortex center is an unknown quantity before tracking the vortex range. Only after the vortex range is identified and matched with other physical quantities is it considered a vortex, and the vortex center is further identified.

[0153] 2. This invention identifies the vortex range of atmospheric vortices and performs edge trimming on streamlined parts that do not belong to the vortex range;

[0154] When traditionally used for identifying the range of eddies in physical oceanography, the edge trimming step is not performed because the range of eddies is closer to a circle than a regular shape.

[0155] 3. For vortices with a lifespan greater than a certain threshold, the present invention performs a Gaussian kernel-weighted moving average on the range of vortices at different times within the same level; for those less than or equal to the threshold, an arithmetic average is performed.

[0156] Traditional methods for identifying eddies using the winding angle method in physical oceanography do not perform this step because they do not perform three-dimensional identification of the eddies. Furthermore, researchers who perform three-dimensional identification of atmospheric eddies do not identify the eddy range, so there is currently no algorithm that can represent the evolution characteristics of the three-dimensional eddy range.

[0157] 4. Based on the above vortex range identification results, this invention matches them with the mesoscale convective system that generates precipitation for subsequent work.

[0158] Traditional methods use a circular region of fixed radius as the vortex range, which does not change with the shape and size of the vortex.

[0159] Compared with existing atmospheric vortex identification methods, this invention introduces a vortex range identification method based on flow field geometry features, achieving automatic detection of vortex closed structures. It can identify 214,587 vortexes at horizontal scales within 20 minutes; simultaneously, it can express the development and evolution characteristics of long-lived vortices at different vertical altitudes. Matching the identified vortices with horizontal scale characteristics with mesoscale convection systems avoids the problems of high subjectivity, non-repeatability, and low efficiency caused by reliance on human experience in existing technologies, enabling the analysis of vortex-convection structure relationships to be carried out within an automated and objective framework.

[0160] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An automatic identification method for horizontal-scale atmospheric eddies and its matching with mesoscale convective systems, characterized in that, include: Step S1: Read the time series data of wind field evolution in the study area during the study period; Step S2: The study area is vertically divided into several height layers, and the horizontal plane of each height layer is gridded into several grid points; Based on the wind field evolution data time series, the evolution characteristics of each three-dimensional vortex at each time t in its life history are identified, forming a three-dimensional vortex evolution characteristic time series; wherein, the evolution characteristics include each height layer of the three-dimensional vortex at time t, referred to as each extension height layer, and the two-dimensional coordinates of the vortex center in the horizontal plane of each extension height layer. Step S3: Based on the two-dimensional coordinates of the vortex center of the three-dimensional vortex at each extension height layer at time t, the vortex range identification algorithm is used to identify the vortex range of the three-dimensional vortex at each extension height layer at time t. Step S4: Based on the vortex range time series formed by the vortex range at each time t in the same extension height layer during the life history of the three-dimensional vortex, the vortex range smoothing algorithm is used to smooth the vortex range at each time t in the vortex range time series to obtain the stable vortex range of the three-dimensional vortex at each time t in the extension height layer. Step S5: Obtain the range of the mesoscale convection system in the study area during the study period; Determine whether the range of the mesoscale convection system overlaps with the range of the stable vortex at any extension height layer of the three-dimensional vortex at the same time t; if so, the range of the mesoscale convection system matches the range of the three-dimensional vortex at time t. Step S3 is as follows: Step S31, the three-dimensional vortex at time t is located at the vortex center of the extension height layer. ,in, Each 3D vortex has a unique identifier. The floor number for the extended height layer, , The grid index represents the grid point where the vortex center is located, indicating the grid point position in the longitude and latitude directions, respectively; Step S32, Track the vortex center 3 The streamlines of the 3-neighborhood's eight adjacent grid points, the streamlines of the vortex center, and the streamlines of its eight adjacent grid points constitute a streamline set. : ； in: for Abbreviation, representing time t, The extension height layer, numbered as The vortex passes through the grid point Streamlines; Step S33, calculate each streamline cumulative angle change value ; Step S34, according to each flow line cumulative angle change value Identify each flow line Is it a closed streamline that rotates counterclockwise? Streamline set The closed streamlines identified by the counterclockwise rotation form a set of closed streamlines. ; Step S35: Calculate the set of closed streamlines. The centroids of each closed streamline in the set of centroids constitute the set of centroids. ; Step S36: Based on the centroid of each closed streamline, cluster and filter the closed streamlines, retaining those that are aligned with the vortex center. Distance less than Closed streamlines; Step S37: Perform edge trimming on each of the remaining closed streamlines, trimming the streamline points located at the edge of each closed streamline, thereby obtaining each edge-trimmed closed streamline; Step S38: Based on the closed streamlines after trimming each edge, the vortex center is obtained as follows: The confidence ellipse has three shape parameters: major axis length, minor axis length, and major axis angle; this confidence ellipse is the identified vortex range.

2. The method for automatic identification of horizontal-scale atmospheric vortices and its matching with mesoscale convection systems according to claim 1, characterized in that, The evolutionary characteristics of each three-dimensional vortex at each time step t throughout its life history were identified by the following method: Based on the wind field evolution data time series, identify the vortex characteristics at each altitude layer at each time t within the study period; Horizontal linking of vortex features: Determine whether two vortex features at adjacent times and close positions at the same height layer are similar. If they are similar, it is concluded that the two vortex features are the same vortex. Thus, several similar vortex features at a continuous time at the same height layer are obtained, representing the evolution features of the same vortex at that height layer. Vertical linking of vortex features: Determine whether two vortex features at adjacent height layers are similar at the same time t. If they are similar, it is concluded that the two vortex features are the same vortex. Thus, several similar vortex features at a continuous height layer at time t are obtained, representing the height layers where the same vortex extends at that time t.

3. The method for automatic identification of horizontal-scale atmospheric vortices and its matching with mesoscale convection systems according to claim 1, characterized in that, Calculate each streamline cumulative angle change value The method is as follows: Each flow line Composed of a finite number of discrete streamline points, it can be represented as: ； in: streamline The number of streamline points it possesses; Let the first streamline point, the second streamline point, ..., the... be the... Streamline point; Cumulative angle change value for: ； in: streamline point To streamline point A directed line segment; streamline point To streamline point A directed line segment; It is the angle between two directed line segments.

4. The method for automatic identification of horizontal-scale atmospheric vortices and its matching with mesoscale convection systems according to claim 1, characterized in that, According to each flow line Cumulative angle change value Identify each flow line Whether it is a closed streamline rotating counterclockwise, specifically: If streamline Cumulative angle change value Not less than streamline It is a closed streamline that rotates counterclockwise.

5. The method for automatic identification of horizontal-scale atmospheric vortices and its matching with mesoscale convection systems according to claim 1, characterized in that, The remaining closed streamlines are edge-trimmed, removing the streamline points located at the edges of each closed streamline, thus obtaining the edge-trimmed closed streamlines as follows: Step S371: The remaining closed streamlines constitute a set of closed streamlines. The streamline points of each closed streamline that are retained constitute the streamline point set. : ； in: Closed streamline streamline points in , streamline points Longitude and latitude; Step S372, set of streamline points The streamline points in the data are analyzed using longitude and latitude coordinates to obtain a set of longitudes representing the longitude distribution. and the set of latitudes representing the latitude distribution. ; Step S373, take the longitude distribution quantiles and quantiles As the cutoff boundary in the longitude direction, the latitude distribution is taken as... quantiles and quantiles As the cutoff boundary in the latitudinal direction, cutoff and edge trimming are performed in both the longitudinal and latitudinal directions respectively; The streamlined point set after edge clipping Defined as: ； This achieves edge clipping.

6. The method for automatic identification of horizontal-scale atmospheric vortices and its matching with mesoscale convection systems according to claim 1, characterized in that, The center of the vortex is The three shape parameters of the confidence ellipse are represented as: major axis length minor axis length and major axis angle .

7. The method for automatic identification of horizontal-scale atmospheric vortices and its matching with mesoscale convection systems according to claim 6, characterized in that, A vortex range smoothing algorithm is used to smooth the vortex range at each time t in the vortex range time series, thereby obtaining the stable vortex range of the three-dimensional vortex at each time t in the extension height layer. Specifically: Step S41, the vortex range time series is represented as follows: ;in, For the number Three-dimensional vortex at height layer The vortex range at time t; = ; respectively vortex range The corresponding major axis length, minor axis length, and major axis angle of the confidence ellipse; For the number The life history of a three-dimensional vortex; numbered as Three-dimensional vortex at height layer In its life history The set formed by each moment in time is represented as ; Step S42, for the vortex range The major axis length, minor axis length, and major axis angle are smoothed respectively. The confidence ellipse formed by the smoothed major axis length, minor axis length, and major axis angle is the stable vortex range. Will Use elements uniformly The method for smoothing it is as follows: element In its life history The resulting time series is represented as ; For each element in the time series , The smoothed elements are obtained by performing the smoothing process using the following formula. : ； in: Given a time threshold; Therefore Number of time intervals extending to the left and right of the center. The set of times, denoted as [ , ], where, if Then let ; This is a floor function; for operations located in the life history... Moments at the middle edge position That is, its left side does not have At any given moment or to its right, there is no [something]. For each time point, the missing value location is determined using a mirroring method; the time set [ , Each moment in the diagram is represented as... Gaussian weights = ; For life history Every moment The corresponding element value, .

Images