Automatic method for forecasting transverse trough-to-vertical type strong cold waves through fluctuation truncation

Abstract

The invention provides an automatic method for forecasting transverse trough-to-vertical type strong cold waves through fluctuation truncation, and belongs to the field of large-range strong meteorological disaster forecasting and forecasting decision information management.The method comprises the steps that potential height data in a 500 hPa key area in a forecasting service is read, and screen coordinates are converted into Cartesian coordinates; analyzing a potential height field contour line; repeatedly splitting fluctuation through truncation, and extracting a truncation point; selecting slot line alternative nodes from the truncation point set by using a vector cross multiplication condition; screening slot line alternative nodes and organizing and constructing a slot line according to the general principle that the slot line faces the west, the south and tends to be straight; and detecting the slope of the slot line fitting line to find the transverse slot, calculating the extension distance in the north-south direction after the transverse slot is found, comparing the extension distance with the north-south extension distance at the next moment, and if the increment exceeds a preset threshold value, judging that the transverse slot starts to rotate vertically, and generating strong cold waves. According to the method, aiming at the transverse groove-to-vertical cold waves with the widest influence range and the most serious disaster, the forecasting automation is realized through technical innovation.

Description

Technical Field

[0001] The present invention belongs to the field of large-scale severe meteorological disaster forecasting, prediction and decision-making information management, and in particular relates to an automatic method for forecasting a strong cold wave from a horizontal trough to a vertical one by wave truncation. Background Art

[0002] A cold wave is a strong cold air activity originating from the polar regions and the frigid zone and invading lower latitudes. In meteorology, it is defined as a large-scale violent outbreak of strong cold air. It does not mean that all cold air invasion processes are called cold waves, but when the minimum temperature is below 4 degrees Celsius and the temperature drops by more than 8 degrees Celsius within 24 hours, or the temperature drops by more than 10 degrees Celsius within 48 hours, or the temperature drops by more than 12 degrees Celsius within 72 hours, this kind of violent activity of strong cold air is called a cold wave.

[0003] On the high-altitude weather map, there are always fluctuations caused by cold and warm air in the westerly belt. The leading edge of the cold air moving southward and eastward from the north is manifested as a trough, also known as a cold trough or a low trough, which is the low point in the fluctuation. The leading edge of the warm air moving northward and eastward from the south is manifested as a ridge, also known as a warm ridge or a high ridge, which is the peak point in the fluctuation. Most of my country's high altitude is in the westerly belt, and the most common troughs and ridges are as follows: Figure 1 As shown, the cold and warm air move from west to east. This trough extending from north to south is also called a vertical trough. In the figure, the thick arrow indicates the direction of movement of the trough.

[0004] In winter, extremely strong cold air accumulates in the polar regions, moving from north to south, and appears as a horizontal trough on the high-altitude map, such as Figure 2 As shown in A. When the extremely northerly and extremely strong cold air, namely trough A, moves southward, it will enter the westerly belt and its direction of movement will change from west to east, becoming vertical trough B. This process is called trough to vertical. A represents the southward cold air trough, and B represents the eastward cold air trough. From A to B is the trough to vertical process, which will cause a strong cold wave.

[0005] The horizontal groove extends very long in the east-west direction, but its length in the north-south direction is very limited. However, after turning vertical, the opposite is true, and its length in the north-south direction increases sharply. Based on this feature, the process of the horizontal groove turning vertical is automatically identified.

[0006] In forecasting business, cold wave processes in the 500hPa potential height field are divided into three types: 1. Small trough development type; 2. Low trough eastward movement type; 3. Horizontal trough to vertical type.

[0007] Among the three types of cold wave processes, the first two are caused by cold air moving from west to east, while the horizontal trough to vertical type is caused by cold air moving from the north to the south. As a result of solar radiation in the Northern Hemisphere, the low-latitude south is warmer and the high-latitude north is very cold; while the first two types generally do not have a large temperature difference in the east-west direction. Therefore, the overall force of cold air moving from west to east is relatively limited; but the cold air moving from the north to the south, that is, the third type of horizontal trough, is extremely powerful. Therefore, the strong cold wave process, that is, the process that causes severe cooling, frost, blizzards, rain and snow in the south, freezing, strong winds, and dust storms in the north and central parts of the country, including the south, is mostly caused by the horizontal trough to vertical type.

[0008] Cold waves are the weather process with the largest impact range, affecting an area far exceeding other weather such as heavy rains and typhoons. In addition, the disasters accompanying cold waves are of many types and high intensity, attracting widespread attention and making them quite difficult to predict.

[0009] Starting from November 2023, in order to improve the accuracy and service level of cold wave forecasts and warnings, cold waves, like typhoons, will begin to be numbered. The importance of cold wave analysis, forecasting, prevention and response has risen to a new level, and the forecasting capabilities and levels must be further improved in an all-round way.

[0010] However, the current traditional cold wave forecasting business requires a lot of manual subjective judgment and experience support in the specific operation process. The feature identification is difficult, not yet automated, intelligent, time-consuming and labor-intensive, and inefficient. This not only increases the workload of forecasters, but also affects the objectivity and consistency of forecast results. Manual identification of cold waves requires forecasters to have rich experience and professional knowledge, as well as a deep understanding of meteorological principles, otherwise it is easy to misjudge or miss the situation; manual identification of cold waves is affected by human factors, such as forecasters' subjective judgment, fatigue, attention, etc., which may lead to unstable and inconsistent identification; manual identification of cold waves is inefficient, and timely and continuous monitoring and early warning require high labor intensity; how to automatically extract key variables in complex weather systems and make correct forecasts and predictions is a major challenge facing cold wave forecasting technology at the business level. In summary, how to improve the accuracy and timeliness of cold wave forecasts, liberate manpower, reduce the impact of cold wave disasters on society, and improve the level of automation are important issues that need to be solved in forecasting business. Summary of the invention

[0011] In view of the above-mentioned deficiencies in the prior art, the present invention provides an automatic method for forecasting strong cold waves of the horizontal trough-to-vertical type through wave truncation, which solves the problems of how to improve the accuracy and timeliness of cold wave forecasts, liberate manpower, reduce the impact of cold wave disasters on society, and how to improve the level of automation. The present invention realizes the automation of forecasting through technological innovation for the horizontal trough-to-vertical type cold waves with the widest impact range and the most serious disasters.

[0012] In order to achieve the above object, the technical solution adopted by the present invention is: an automatic method for predicting a strong cold wave from a horizontal trough to a vertical type by wave truncation, comprising the following steps:

[0013] S1. Read the potential height data in the 500hPa forecast key area in the forecast business and convert the screen coordinates into a Cartesian coordinate system;

[0014] S2. Perform contour analysis on the potential height data, extract candidate nodes of the trough line using the cut wave method according to the analysis results, and construct the trough line organization to identify the trough line;

[0015] S3, calculating the slope of each segmented slot fitting line;

[0016] S4, determine whether the slope is less than 1, if so, a horizontal groove appears, and enter S5, otherwise it is a vertical groove, and calculate the data of the next moment to find the horizontal groove; if there is no horizontal groove at all times, output no horizontal groove, and end the process;

[0017] S5. Mark and number the candidate nodes of the trough line, and calculate the longitudinal north-south length of the transverse trough according to the coordinates of the starting and ending points;

[0018] S6. Determine whether the trough is turning vertical based on the change in its longitudinal north-south length, and complete the forecast of a strong cold wave that is turning vertical.

[0019] Furthermore, the S1 comprises the following steps:

[0020] S101, reading the potential height data in the 500hPa forecast key area in the forecast service, wherein the potential height data is three-dimensional data, including time intervals, latitude and longitude grid intervals;

[0021] S102, selecting a geographical area for the read geopotential height data, calculating the required latitude and longitude ranges, and extracting data in the corresponding area;

[0022] S103, according to the data in the corresponding area extracted, the screen coordinates are converted into a Cartesian coordinate system using the following formula:

[0023] y=height-y*

[0024] Among them, y represents the vertical coordinate in the Cartesian coordinate system, height represents the vertical height of the screen, and y* represents the vertical coordinate in the screen coordinate system.

[0025] Furthermore, the S2 comprises the following steps:

[0026] S201, create a two-dimensional array data_t, and perform contour analysis on the potential height data, and distribute the contour nodes in the list contours according to the elevation level, wherein contours represents a list containing all contour node sets, and the two-dimensional array data_t is used to store the potential height data within the range of 90°N-40°N and 70°E-140°E at time t;

[0027] S202, traversing each contour line in the list contours, and storing the contour line nodes in the list all_points in the form of triplets in order, wherein each contour line contains fluctuations of multiple points;

[0028] S203, extracting the truncation point by using the truncation method, and using the vector cross multiplication value as the positive determination slot line candidate node;

[0029] S204, searching for the minimum value point in the two-dimensional array data_t, and inserting it into the trough candidate node list trough_list;

[0030] S205, based on the list of candidate trough nodes trough_list, the candidate trough nodes are screened according to the general principle that the trough lines are heading west, south, and tending to be straight, to form a trough node organization, and the trough node organization is constructed into a fitting trough list fitting_line;

[0031] S206. Use the convex hull algorithm to determine whether there is a closed contour line. If so, find the minimum position of the closed contour line, reorganize the slot line node list, and complete the identification of the slot line.

[0032] Furthermore, the S201 includes the following steps:

[0033] S2011. In the selected geographical area, select the maximum value Hmax0 and the minimum value Hmin0 of the potential height, respectively, where the 500 hPa potential height field is based on 5000 gpm and contour analysis is performed every 40 gpm;

[0034] S2012, create a two-dimensional array data_t for storing the potential height data within the range of 90°N-40°N and 70°E-140°E at time t;

[0035] S2013. Perform contour analysis on the two-dimensional array data_t and distribute the contour nodes in the list contours according to the elevation levels.

[0036] Furthermore, the S203 includes the following steps:

[0037] S2031, draw a straight line with the starting and ending points of the original multi-wave curve, calculate the distance from each point on the original multi-wave curve to the straight line, and take the point with the largest distance as the first-level truncation point;

[0038] S2032, based on the first-level truncation point, the original multi-wave large curve is divided into two shortest small curves, the starting points and the ending points of the two small curves are respectively connected to form two new straight lines, and the distances from the points of the two small curves to the corresponding new straight lines are calculated, and the point with the largest distance is calculated and found as the second-level truncation point;

[0039] S2033, repeating S2031 and S2032 until the distance of the farthest truncation point is less than a preset first distance threshold, and then the extraction of the truncation point is completed;

[0040] S2034, define the order of the points before and after the cutoff point set: take the current point as P, and its previous point as , the next point is ;

[0041] S2035, calculate the vectors according to the defined front and back points and :

[0042]

[0043]

[0044] S2036, for two vectors and Perform a cross product calculation:

[0045]

[0046]

[0047]

[0048] in, Respectively represent vector The components in the x and y directions, Respectively represent vector Components in the x and y directions;

[0049] S2037. If the cross product calculation result is positive, it is a candidate node for the slot line, and thus the candidate node for the slot line is identified in the truncation point set.

[0050] Furthermore, the S205 includes the following steps:

[0051] S2051, sorting the trough candidate node list trough_list according to the potential height;

[0052] S2052, grouping the sorted candidate trough nodes in the candidate trough node list trough_list according to the contour lines where they are located, and saving the grouping results into the list collected_list;

[0053] S2053, constructing a list now_isotherm_point_index_list for storing the indexes of the currently selected nodes constituting the slot fitting line, wherein the initial value is the index 0 of the first point, that is, the minimum value point in the two-dimensional array data_t;

[0054] S2054, constructing a list fitting_line for storing selected slot line nodes in a dictionary form;

[0055] S2055, traverse each contour line in the list collected_list in turn, starting from index 1, create empty lists each_contour_distance_list and angle_list for each contour line in the list collected_list, set the last node in the current slot line fitting line as start_point, traverse the slot line candidate nodes in each contour line, and calculate the Euclidean distance between each slot line candidate node in the current contour line and the last node start_point in the current slot line fitting line :

[0056]

[0057] Among them, x0 represents the horizontal coordinate of a node each_point in the current contour line, y0 represents the vertical coordinate of a node each_point in the current contour line, x represents the horizontal coordinate of the node start_point in the current slot line fitting line, y represents the vertical coordinate of the node start_point in the current slot line fitting line, and both node each_point and node each_point are candidate nodes for the slot line;

[0058] S2056. Determine the position of node each_point relative to node start_point using the following formula:

[0059]

[0060]

[0061] in, and Indicates the position of node each_point relative to node start_point, each_point_x and each_point_y respectively represent the horizontal and vertical coordinates of node each_point, start_point_x and start_point_y respectively represent the horizontal and vertical coordinates of node start_point;

[0062] S2057, if , that is, when the node each_point is on the left side of the node start_point and the number of points selected to constitute the groove fitting line stored in the list now_isotherm_point_index_list is greater than or equal to 2, the angle between the current node each_point and the last two nodes in the list now_isotherm_point_index_list is calculated, wherein the last two nodes calculated are the latest selected nodes to constitute the groove fitting line stored in the list now_isotherm_point_index_list, and the angle between the three nodes is stored in the list angle_list, the three nodes include the current node and the last two nodes, and the angle is the angle formed by the current node each_point and the last two nodes in the list now_isotherm_point_index_list, wherein the last node in the list now_isotherm_point_index_list is located at the center;

[0063] like , that is, the node each_point is on the left side of the node start_point, and , that is, the node each_point is above the node start_point, then the distance between the node each_point and the node start_point is calculated, and the distance between the node each_point and the node start_point is added to the list each_contour_distance_list; otherwise, the distance between the node each_point and the node start_point is multiplied by the second distance threshold and added to the list each_contour_distance_list;

[0064] like , that is, the node each_point is on the right side of the node start_point, and the distance is set to 99999 according to the meteorological business specifications, added to the list each_contour_distance_list, and the angle is set to 90°, added to the list angle_list;

[0065] S2058. Filter the three situations of S2057 to obtain a fitting slot line list fitting_line.

[0066] Furthermore, the S2058 includes the following steps:

[0067] S20581. Find the candidate node of the trough line closest to the node start_point in the current contour line, and use minize_index to record the index of the minimum distance:

[0068]

[0069] in, Represents the index of the minimum distance, Indicates that the objective function distance( , start_point) takes the index of the minimum value, Indicates the candidate node of the slot line that is calculating the distance with the node start_point in the current contour line. Indicates the index of the candidate node of the trough line closest to the node start_point in the current contour line;

[0070] S20582. Calculate the difference between each angle in the list angle_list and 180° :

[0071]

[0072] in, It means recording and storing the current slot candidate node j and the last two nodes in the node list that have been selected to form the slot fitting line, and the list of angles angle between the three nodes;

[0073] S20583, if and ,but: ;

[0074] in, Indicates that the number of angles in the list angle_list is greater than 2. Indicates that the number of points in the list now_isotherm_point_index_list is greater than 3, Indicates updating the index nearest_flatten_angle_index to point to the candidate slot line node whose angle with the previous slot line is closest to 180°;

[0075] S20584, compare the nearest_flatten_angle_index and the minize_index to see if they are inconsistent. ,and If it is less than or equal to the preset first threshold, the index minize_index is updated to the index nearest_flatten_angle_index;

[0076] S20585, add the coordinates of the node in the current contour line that best fits the groove line fitting line to the fitting groove line list fitting_line, and add its index to the list now_isotherm_point_index_list, update count, add 1 to the number of nodes that currently have been selected to form the groove line fitting line, and perform the next traversal. If the length of the list each_contour_distance_list is 1 and the value exceeds the set second threshold, it means that there is only one point in the current contour line and the distance is the largest, then terminate the loop and the groove line will no longer extend, where count represents the total number of nodes that currently have been selected to form the groove line fitting line, where the node that best fits the groove line fitting line is the node corresponding to the index minize_index;

[0077] S20586. The candidate node information of the groove line corresponding to all the indexes in the list now_isotherm_point_index_list is stored in the list contours_list_with_line to complete the fitting of the groove line.

[0078] Furthermore, the S206 is specifically as follows:

[0079] Calculate the distance between the starting and ending points of each contour line;

[0080] If the distance is less than the set third threshold, the point set array contour_data is determined to be approximately closed; or

[0081] If the distance is equal to the set third threshold, the point set array contour_data is determined to be approximately closed;

[0082] Store the minimum positions of all closed contour lines into the list closed_min_list;

[0083] According to the minimum position of the closed area, the contours_list_with_line node list of the groove fitting line is truncated into multiple sub-lists;

[0084] Insert the minimum position of the enclosed area at the beginning of each sublist, sort the contour line fitting line node list contours_list_with_line according to the height value, and update the fitting groove line list fitting_line.

[0085] Furthermore, the S3 comprises the following steps:

[0086] S301, analyze the list contours_list_with_line to determine whether it includes a segment sublist, if so, proceed to S303, otherwise, proceed to S302;

[0087] S302, calculate the slopes between the nodes of each slot line, take the average value, use the average value as the slope of the slot line fitting line, and save it:

[0088]

[0089]

[0090] in, represents the slope of the slot line fitting line, Representation list The number of data in Representation list The data index in Represents each slot node and its next slot node A list of slopes between ;

[0091] S303, traverse the slot line nodes in each segment sublist in turn, calculate the slopes between the slot line nodes in each segment sublist, and take the average value to obtain the average slope of the segment sublist, and average the average slopes of all segment sublists to obtain the slot line approximation slopes of all segment sublists, and store the slot line approximation slopes of all segment sublists:

[0092]

[0093]

[0094]

[0095] in, Represents the slope of the slot line fitting and line of all segment sublists, m represents the list The number of data in Representation list The data index in a list representing the average slope of the slotted fits for each sublist of segments, Representation list The number of data in Representation list The data index in Represents each segment sublist The slot node in With the next slot node A list of the slopes of the lines between , and Represents the slot line nodes The horizontal and vertical coordinates of and Respectively represent the next slot line node The horizontal and vertical coordinates of .

[0096] Furthermore, the S5 is specifically:

[0097] Traverse the time series and analyze and process the data at each moment;

[0098] If the previous moment There is a horizontal groove, from the previous moment Starting point of horizontal groove and end point The vertical coordinate difference is used to calculate the longitudinal north-south length of the horizontal groove. :

[0099] ;

[0100] By calculating the current time Starting point of horizontal groove and end point The vertical coordinate difference is used to calculate the current longitudinal north-south length of the horizontal groove. :

[0101] .

[0102] Furthermore, the S6 is specifically:

[0103] Traverse each time point and calculate the increase in the longitudinal north-south length of the horizontal groove between the current moment and the previous moment Relative value of:

[0104]

[0105] If the growth If it is greater than the preset growth threshold, it indicates that the horizontal trough turns vertical, and a severe cold wave weather process is predicted, and the start time of the horizontal trough turning vertical is output;

[0106] Otherwise, if the judgment conditions for the horizontal trough to vertical transition are not met at all times, the output is that there is no horizontal trough to vertical transition during the forecast period, and the process ends.

[0107] The beneficial effects of the present invention are:

[0108] (1) The present invention is dedicated to solving the problem of how to improve the accuracy and timeliness of cold wave forecasts, liberate manpower, reduce the impact of cold wave disasters on society, and improve the level of automation. The present invention aims at the horizontal trough-to-vertical cold wave with the widest impact range and the most serious disasters, and realizes the automation of forecasting through technological innovation.

[0109] (2) Based on the 500hPa potential height field, which is the most commonly used in meteorological business, the present invention has developed an automatic method for forecasting the transition from horizontal trough to vertical cold wave through wave truncation. This method belongs to the field of meteorological disaster forecasting, prediction, warning and decision-making information management in natural disaster defense, directly serves emergency management work, and is committed to protecting people's lives, property and social security.

[0110] (3) Previously, cold wave forecasts relied on manual judgment, which not only consumed manpower, but also lacked accuracy. The method for automatically identifying cold waves by observing the process of horizontal trough turning vertical based on meteorological potential data proposed in the present invention is based on the definition of the occurrence of horizontal trough turning vertical, designs relevant judgment logic, and then realizes the automatic judgment of cold waves without manual participation, and the accuracy is greatly improved.

[0111] (4) The present invention selects and constructs continuous slots from a list of candidate slot nodes, including sorting and grouping the slot nodes by height, calculating the Euclidean distance and angle, and according to the general principle that the slot is westward, southward, and tends to be a straight line, that is, the angle tends to be 180°, so as to give priority to points that are close and on the left or below, and ensure the smoothness and continuity of the slot by optimizing the angle. By storing the selected slot node index and coordinates in the fitting line and index list, and terminating the loop when no suitable point can be found, the method effectively constructs a smooth and coherent slot, providing reliable data support for terrain analysis and visualization data analysis. The advantage of the present invention is that it realizes automatic screening and extraction of slot nodes of continuous slots by height sorting, distance calculation and angle optimization, ensuring the smoothness and logical consistency of the slot. Compared with the traditional method that relies on manual judgment or simple threshold screening, this method can more accurately and efficiently identify the most suitable slot nodes, reduce human errors, improve data processing efficiency, and provide more reliable and accurate support for terrain analysis and visualization.

[0112] (5) The present invention traverses the contour data and uses the convex hull algorithm to detect closed contours, determines the minimum position within the area, and truncates and reorganizes the node list of the groove fitting line according to these positions. This method ensures that the points in the closed area are orderly and logical in data processing, thereby improving the accuracy and reliability of contour data processing. Compared with traditional methods, the closed area is processed more accurately, human intervention and errors are reduced, and the continuity and consistency of data processing are ensured, thereby improving the accuracy and reliability of analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0113] Figure 1 A schematic diagram of the movement of cold and warm air in the westerly belt airflow in the background technology.

[0114] Figure 2 It is a schematic diagram of the southward cold air horizontal trough A and the eastward cold air vertical trough B in the background technology.

[0115] Figure 3 The present invention is a flow chart of the method.

[0116] Figure 4 Schematic diagram of Cartesian coordinate system and screen coordinate system.

[0117] Figure 5 This is a diagram illustrating the structure of the contours list.

[0118] Figure 6 This is a diagram illustrating the structure of the list all_points.

[0119] Figure 7 Schematic diagram of the analysis principle of the chopped wave method.

[0120] Figure 8 Schematic diagram of the turning angle at the truncation point.

[0121] Fig. 9 Build a schematic diagram for the organization of the trough line.

[0122] Fig.10 A schematic diagram of the visualization effect of the slot fitting line at time t.

[0123] Fig.11 Another schematic diagram showing the visualization effect of the slot fitting line at time t.

[0124] Fig.12 Schematic diagram of the visualization effect of the slot fitting line when t is multiple time points.

[0125] Fig.13 Schematic diagram of the visualization effect of the slot fitting line at t = 30.

[0126] Fig.14 Visualization of the slot fitting line when t = 0. DETAILED DESCRIPTION

[0127] The specific implementation modes of the present invention are described below to facilitate those skilled in the art to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0128] Example

[0129] like Figure 1 As shown, the present invention provides an automatic method for predicting a strong cold wave from a horizontal trough to a vertical type by wave truncation, and the implementation method is as follows:

[0130] S1. Read the potential height data in the 500hPa forecast key area in the forecast business and convert the screen coordinates into a Cartesian coordinate system. The implementation method is as follows:

[0131] S101, reading the potential height data in the 500hPa forecast key area in the forecast service, wherein the potential height data is three-dimensional data, including time intervals, latitude and longitude grid intervals;

[0132] S102, selecting a geographical area for the read geopotential height data, calculating the required latitude and longitude ranges, and extracting data in the corresponding area;

[0133] S103, according to the data in the corresponding area extracted, the screen coordinates are converted into a Cartesian coordinate system using the following formula:

[0134] y=height-y*

[0135] Among them, y represents the ordinate in the Cartesian coordinate system, height represents the vertical height of the screen, and y* represents the ordinate in the screen coordinate system;

[0136] S2. Perform contour analysis on the potential height data. According to the analysis results, use the cut wave method to extract the candidate nodes of the trough line, and construct the trough line organization to identify the trough line. The implementation method is as follows:

[0137] S201. Create a two-dimensional array data_t, perform contour analysis on the potential height data, and distribute the contour nodes in the list contours according to the elevation level, where contours represents a list containing all contour node sets. The two-dimensional array data_t is used to store the potential height data within the range of 90°N-40°N and 70°E-140°E at time t. The implementation method is as follows:

[0138] S2011. In the selected geographical area, select the maximum value Hmax0 and the minimum value Hmin0 of the potential height, respectively, where the 500 hPa potential height field is based on 5000 gpm and contour analysis is performed every 40 gpm;

[0139] S2012, create a two-dimensional array data_t for storing the potential height data within the range of 90°N-40°N and 70°E-140°E at time t;

[0140] S2013. Perform contour analysis on the two-dimensional array data_t and distribute the contour nodes in the list contours according to the elevation level;

[0141] S202, traversing each contour line in the list contours, and storing the contour line nodes in the list all_points in the form of triplets in order, wherein each contour line contains fluctuations of multiple points;

[0142] S203, using the truncation method to extract the truncation point, and using the vector cross multiplication value as the positive determination slot line candidate node, the implementation method is as follows:

[0143] S2031, draw a straight line with the starting and ending points of the original multi-wave curve, calculate the distance from each point on the original multi-wave curve to the straight line, and take the point with the largest distance as the first-level truncation point;

[0144] S2032, based on the first-level truncation point, the original multi-wave large curve is divided into two shorter small curves, the starting points and the ending points of the two small curves are respectively connected to form two new straight lines, and the distances from the points of the two small curves to the corresponding new straight lines are calculated, and the point with the largest distance is calculated and found as the second-level truncation point;

[0145] S2033, repeating S2031 and S2032 until the distance of the farthest truncation point is less than a preset first distance threshold (in the present invention, it has been experimentally determined that the first distance threshold is 25 for the best effect), and then the extraction of the truncation point is completed;

[0146] S2034, define the order of the points before and after the cutoff point set: take the current point as P, and its previous point as , the next point is ;

[0147] S2035, calculate the vectors according to the defined front and back points and :

[0148]

[0149]

[0150] S2036, for two vectors and Perform a cross product calculation:

[0151]

[0152]

[0153]

[0154] in, Respectively represent vector The components in the x and y directions, Respectively represent vector Components in the x and y directions;

[0155] S2037, if the cross product calculation result is positive, then it is a candidate node for the slot line, and thus the candidate node for the slot line is obtained by concentrated identification at the truncation point;

[0156] S204, searching for the minimum value point in the two-dimensional array data_t, and inserting it into the trough candidate node list trough_list;

[0157] S205. Based on the list of candidate trough nodes trough_list, the candidate trough nodes are screened according to the general principle that the trough extends westward, southward and tends to be a straight line to form a trough node organization, and the trough node organization is constructed into a fitting trough list fitting_line. The purpose of this processing is to construct a trough from the trough node organization. The principle of constructing the trough is to extend westward and southward and tend to be a straight line, that is, the angle is close to 180 degrees. The implementation method is as follows:

[0158] S2051, sorting the trough candidate node list trough_list according to the potential height;

[0159] S2052, grouping the sorted candidate trough nodes in the candidate trough node list trough_list according to the contour lines where they are located, and saving the grouping results into the list collected_list;

[0160] S2053, constructing a list now_isotherm_point_index_list for storing the indexes of the currently selected nodes constituting the slot fitting line, wherein the initial value is the index 0 of the first point, that is, the minimum value point in the two-dimensional array data_t;

[0161] S2054, constructing a list fitting_line for storing selected slot line nodes in a dictionary form;

[0162] S2055, traverse each contour line in the list collected_list in turn, starting from index 1, create empty lists each_contour_distance_list and angle_list for each contour line in the list collected_list, set the last node in the current slot line fitting line as start_point, traverse the slot line candidate nodes in each contour line, and calculate the Euclidean distance between each slot line candidate node in the current contour line and the last node start_point in the current slot line fitting line :

[0163]

[0164] Among them, x0 represents the horizontal coordinate of a node each_point in the current contour line, y0 represents the vertical coordinate of a node each_point in the current contour line, x represents the horizontal coordinate of the node start_point in the current slot line fitting line, y represents the vertical coordinate of the node start_point in the current slot line fitting line, and both node each_point and node each_point are candidate nodes for the slot line;

[0165] S2056. Determine the position of node each_point relative to node start_point using the following formula:

[0166]

[0167]

[0168] in, and It indicates the position of node each_point relative to node start_point, each_point_x and each_point_y respectively indicate the horizontal and vertical coordinates of node each_point, start_point_x and start_point_y respectively indicate the horizontal and vertical coordinates of node start_point, where each_point indicates to examine each candidate slot node in list collected_list one by one and take the candidate slot node being examined.

[0169] S2057, if , that is, when the node each_point is on the left side of the node start_point and the number of points selected to constitute the groove fitting line stored in the list now_isotherm_point_index_list is greater than or equal to 2, the angle between the current node each_point and the last two nodes in the list now_isotherm_point_index_list is calculated, wherein the last two nodes calculated are the latest selected nodes to constitute the groove fitting line stored in the list now_isotherm_point_index_list, and the angle between the three nodes is stored in the list angle_list, the three nodes include the current node and the last two nodes, and the angle is the angle formed by the current node each_point and the last two nodes in the list now_isotherm_point_index_list, wherein the last node in the list now_isotherm_point_index_list is located at the center;

[0170] like , that is, the node each_point is on the left side of the node start_point, and , that is, the node each_point is above the node start_point, then the distance between the node each_point and the node start_point is calculated, and the distance between the node each_point and the node start_point is added to the list each_contour_distance_list; otherwise, the distance between the node each_point and the node start_point is multiplied by the second distance threshold and added to the list each_contour_distance_list;

[0171] like , that is, the node each_point is on the right side of the node start_point, and the distance is set to 99999 according to the meteorological business specifications, added to the list each_contour_distance_list, and the angle is set to 90°, added to the list angle_list;

[0172] Because the slot line nodes with different elevation program numbers come from different contour lines, and different contour lines do not intersect each other, the slot line nodes will not be equal in terms of x or coordinates, that is, there is no situation where dx=0.

[0173] S2058. Filter the three cases of S2057 to obtain a fitting slot line list fitting_line, which is implemented as follows:

[0174] S20581. Find the candidate node of the trough line closest to the node start_point in the current contour line, and use minize_index to record the index of the minimum distance:

[0175]

[0176] in, Represents the index of the minimum distance, Indicates that the objective function distance( , start_point) takes the index of the minimum value, Indicates the candidate node of the slot line that is calculating the distance with the node start_point in the current contour line. Indicates the index of the candidate node of the trough line closest to the node start_point in the current contour line;

[0177] S20582. Calculate the difference between each angle in the list angle_list and 180° :

[0178]

[0179] in, It means recording and storing the current slot candidate node j and the last two nodes in the node list that have been selected to form the slot fitting line, and the list of angles angle between the three nodes;

[0180] S20583, if and ,but: ;

[0181] in, Indicates that the number of angles in the list angle_list is greater than 2. Indicates that the number of points in the list now_isotherm_point_index_list is greater than 3, Indicates updating the index nearest_flatten_angle_index to point to the candidate slot line node whose angle with the previous slot line is closest to 180°;

[0182] S20584, compare the nearest_flatten_angle_index and the minize_index to see if they are inconsistent. ,and If it is less than or equal to the preset first threshold, the index minize_index is updated to the index nearest_flatten_angle_index;

[0183] S20585, add the coordinates of the node in the current contour line that best fits the groove line fitting line to the fitting groove line list fitting_line, and add its index to the list now_isotherm_point_index_list, update count, add 1 to the number of nodes that currently have been selected to form the groove line fitting line, and perform the next traversal. If the length of the list each_contour_distance_list is 1 and the value exceeds the set second threshold, it means that there is only one point in the current contour line and the distance is the largest, then terminate the loop and the groove line will no longer extend, where count represents the total number of nodes that currently have been selected to form the groove line fitting line, where the node that best fits the groove line fitting line is the node corresponding to the index minize_index;

[0184] S20586, storing the candidate node information of the groove line corresponding to all the indexes in the list now_isotherm_point_index_list in the list contours_list_with_line, and completing the fitting of the groove line;

[0185] S206, using the convex hull algorithm to determine whether there is a closed contour line, if so, it is necessary to find the minimum position of the closed contour line area, reorganize the slot line node list, and complete the identification of the slot line, which is specifically as follows:

[0186] Calculate the distance between the starting and ending points of each contour line;

[0187] If the distance is less than the set third threshold, the point set array contour_data is determined to be approximately closed; or

[0188] If the distance is equal to the set third threshold, the point set array contour_data is determined to be approximately closed;

[0189] Store the minimum positions of all closed contour lines into the list closed_min_list;

[0190] According to the minimum position of the closed area, the contours_list_with_line node list of the groove fitting line is truncated into multiple sub-lists;

[0191] Insert the minimum position of the closed area at the beginning of each sublist, sort the contour line fitting line node list contours_list_with_line according to the height value, and update the fitting groove line list fitting_line at the same time;

[0192] S3. Calculate the slope of each segmented slot fitting line. The implementation method is as follows:

[0193] S301, analyze the list contours_list_with_line to determine whether it includes a segment sublist, if so, proceed to S303, otherwise, proceed to S302;

[0194] S302, calculate the slopes between the nodes of each slot line, take the average value, use the average value as the characteristic slope of the slot line, and save it:

[0195]

[0196]

[0197] in, represents the characteristic slope of the groove line, Representation list The number of data in Representation list The data index in Represents each slot node and its next slot node The slope between

[0198] S303, traverse the slot line nodes in each segment sublist in turn, calculate the slopes between the slot line nodes in each segment sublist, and take the average value to obtain the average slope of the segment sublist, and average the average slopes of all segment sublists to obtain the slot line approximation slopes of all segment sublists, and store the slot line approximation slopes of all segment sublists:

[0199]

[0200]

[0201]

[0202] in, Represents the slope of the slot line fitting and line of all segment sublists, m represents the list The number of data in Representation list The data index in a list representing the average slope of the slotted fits for each sublist of segments, Representation list The number of data in Representation list The data index in Represents each segment sublist The slot node in With the next slot node A list of the slopes of the lines between , and Represents the slot line nodes The horizontal and vertical coordinates of and Respectively represent the next slot line node The horizontal and vertical coordinates of .

[0203] S4, determine whether the slope is less than 1, if so, a horizontal groove appears, and enter S5, otherwise it is a vertical groove, and calculate the data of the next moment to find the horizontal groove; if there is no horizontal groove at all times, output no horizontal groove, and end the process;

[0204] S5. Mark and number the candidate nodes of the trough line, and calculate the longitudinal north-south length of the transverse trough from the coordinates of the starting and ending points, which is specifically:

[0205] Traverse the time series and analyze and process the data at each moment;

[0206] The current moment There is a horizontal groove, from the previous moment Starting point of horizontal groove and end point The vertical coordinate difference is used to calculate the longitudinal north-south length of the horizontal groove. :

[0207]

[0208] By calculating the current time Starting point of horizontal groove and end point The vertical coordinate difference is used to calculate the longitudinal north-south length of the horizontal groove. :

[0209] ;

[0210] S6. Determine whether the trough turns vertical based on the change in the longitudinal north-south length of the trough, and complete the forecast of the strong cold wave of the trough turning vertical, which is specifically as follows:

[0211] Traverse each time point and calculate the increase in the longitudinal north-south length of the horizontal groove between the current moment and the previous moment Relative value of:

[0212]

[0213] If the growth If it is greater than a preset growth threshold (through experimental comparison in the present invention, setting the threshold to 0.1 has the best effect), it indicates that the horizontal trough turns vertical, predicts a severe cold wave weather process, and outputs the start time of the horizontal trough turning vertical;

[0214] Otherwise, if the judgment conditions for the horizontal trough to vertical transition are not met at all times, the output is that there is no horizontal trough to vertical transition during the forecast period, and the process ends.

[0215] In this embodiment, the input data of the present invention is high temporal and spatial resolution geopotential height data (unit m² / s² or geopotential meter, abbreviated as gpm), and the geopotential height data is saved in the form of a three-dimensional array, and its arrangement is: time, latitude, longitude (the present invention takes the z_2020.12.27-30 data file as an example for explanation), wherein the time interval is 1 hour (ranging from 2020.12.27T0:00:00 to 2020.12.30T23:00:00), including 144 hours after the forecast time, i.e. 6 days; the latitude grid spacing is 0.25 degrees (ranging from 90°N to 90°S), i.e., 721 latitude points from the North Pole to the South Pole 180°; the longitude grid spacing is 0.25 degrees (ranging from 0° to 360°), i.e., from the prime meridian eastward around the earth, where 0° and 360° are the same point, for a total of 1440 longitude points. Therefore, the three-dimensional organization of the data is 96×721×1440. This can represent the temporal change and spatial distribution of the potential height in the atmosphere. The read potential height data is saved in the array data in floating-point data format.

[0216] In this embodiment, the high-altitude cold air source that affects the process of horizontal trough turning into vertical strong cold wave is located in Siberia, that is, the area north of my country. Therefore, the area for finding the initial horizontal trough is set to 90° to 40° north latitude and 70° to 140° east longitude. Based on this, the geographical area is selected for the array data read. The specific calculation is as follows:

[0217] Latitude calculation:

[0218] The starting latitude of 90° corresponds to row 0. The ending latitude of 40° north corresponds to row number It can be calculated by the following formula:

[0219]

[0220] in, Indicates the target latitude.

[0221] For 40°N, substitute the value 40:

[0222]

[0223] in, Indicates the ending line number.

[0224] Therefore, the latitude ranges from row 0 to row 200, for a total of 201 rows.

[0225] Longitude calculation:

[0226] The column number corresponding to the starting longitude 70° East It can be calculated by the following formula:

[0227]

[0228] in, Indicates the target longitude.

[0229] For 70°E, substitute the value 70:

[0230]

[0231] in, Indicates the starting column number.

[0232] The end column number corresponding to the end longitude 140° It can be calculated by the same formula:

[0233] .

[0234] Therefore, the longitudes are from column 280 to column 560, for a total of 281 columns.

[0235] By calculating the data from row 0 to row 200 in the second dimension of latitude in the array data (a total of 201 rows), and the data from column 280 to column 560 in the third dimension of longitude (a total of 281 columns), the data within the required geographic range is extracted.

[0236] In this embodiment, the traditional coordinate system is a Cartesian plane rectangular coordinate system, in which the horizontal axis to the right is the positive x direction, and the vertical axis upward is the positive y direction. On the computer display screen, the horizontal x-axis is defined the same as the Cartesian plane rectangular coordinate system, but the vertical y-axis is defined opposite to the Cartesian plane rectangular coordinate system, with the positive y direction from top to bottom. This will cause a contradiction when performing vector operations. For example, in the traditional Cartesian plane rectangular coordinate system, when the x-axis is used as a vector cross-multiplied by the y-axis vector, the right-hand rule is followed, and the resulting vector z is positive, which is in the direction from the screen to the outside. However, in the screen coordinate system, the right-hand rule is also followed, and the resulting z vector is negative, which is in the direction pointing to the inside of the screen (such as Figure 4 As shown in Figure 2.1, there is a contradiction between the two. To solve this problem, we use traditional mathematical expressions in subsequent analysis and convert screen coordinates into Cartesian coordinates.

[0237] Since the horizontal coordinates x of the two are the same, no transformation is performed. Instead, the vertical coordinate y* in the screen coordinate system is converted to y in the Cartesian coordinate system. If the vertical height of the screen is height, the conversion formula is:

[0238] y=height-y*

[0239] Among them, y represents the vertical coordinate in the Cartesian coordinate system, height represents the vertical height of the screen, and y* represents the vertical coordinate in the screen coordinate system.

[0240] After the above transformation, the screen coordinate system is converted into a Cartesian coordinate system, ensuring that subsequent operations can be mathematically expressed in the traditional conventional way.

[0241] In this embodiment, a two-dimensional array data_t is created to store the potential height data, i.e., elevation, within the range of 90°N-40°N and 70°E-140°E at time t for contour analysis.

[0242] First, select the maximum value Hmax0 and minimum value Hmin0 of the potential height in the selected area. The 500 hPa potential height field in the business is based on 5000 gpm, and contour analysis is performed every 40 gpm.

[0243] Then, the contours are analyzed for the two-dimensional array data_t, and the nodes on each contour are stored in the list contours. contours is a nested list containing all the node sets on the contours, where each contour set itself is a list containing all the nodes of the contour. Its length is the same as the elevation level value of the contour.

[0244] That is, contours[i] is a list containing contour points, corresponding to the first contours[i][j] is a numpy array representing the first The nodes on the jth contour line in the contour level; each numpy array is (n, 2), where n represents the n numbers on the contour line of contours [i][j], and 2 represents the (x, y) coordinates of a point on the contour line of contours [i][j], that is, the longitude and latitude coordinates of each point, such as Figure 5 As shown, there are i contour line levels.

[0245] This method can be used to represent the distribution of geopotential height in the area of ​​interest, which is helpful for subsequent analysis and judgment of the occurrence and development of cold waves. This processing method improves the readability and intuitiveness of the data, providing support for further meteorological analysis and forecasting.

[0246] From the obtained contour node contours data, traverse each contour in each contour set in turn according to the value of the contour, that is, the elevation level order, and add its nodes to the list all_points. The structure of all_points is a list, in which each element is a list containing multiple points. Each point on the contour is stored in the form of a triple, that is, (x, y, value), which represents the x and y coordinates of the point and the two-dimensional array data_t value created for storing potential height data, so as to extract the node data of the contour. Value represents the potential height data of the point, such as Figure 6 As shown in the figure, the number of contour lines and the number of points contained in the contour lines are involved. Through this method, the contour point data can be systematically managed and processed, so that it can be standardized and orderly in subsequent analysis and application, and the efficiency and accuracy of data processing can be improved. This processing method can not only accurately extract contour point data, but also optimize the data structure, providing a reliable data basis for further analysis, calculation and visualization.

[0247] In this embodiment, at high altitude, the contour lines of the potential height field coincide with the airflow, showing a fluctuating curve. Figure 1 and Figure 2 The air flow fluctuations in the contour are also the fluctuations of the potential height contour. The present invention adopts the wave clipping method to analyze, and disassembles multiple continuous fluctuations into smaller fluctuations one by one until the simplest fluctuation is obtained. For the contour line that needs to be analyzed, the multiple fluctuations on this contour line are continuously intercepted and subdivided. The analysis principle is as follows Figure 7 The truncation method repeatedly truncates the fluctuations until all the fluctuations are disassembled. Figure 7 In the figure, the first full curve is connected from beginning to end as a thick line, and the first-level truncation point is calculated and selected as a large dot. Here, the original curve is divided into two segments, and the beginning and end of each segment are connected with a thinner straight line. The second-level truncation points are calculated and represented by small dots. Repeat this process to complete the disassembly of all fluctuations and obtain many truncation points, that is, the strongest points of fluctuations. As shown in the figure, the first truncation of the curve obtains the farthest point on the fluctuation, which is the first-level truncation point. Here, the original curve is divided into two segments. The second truncation of the curve again obtains the farthest point on the fluctuation, which is the second-level truncation point. Here, the original curve is divided into two segments again.

[0248] In this embodiment, first, a straight line is drawn with the starting and ending points of the original multi-wave curve, and the distance from each point on the original wave curve to this straight line is calculated. The point with the largest distance is where the fluctuation is the strongest, and is taken as the truncation point. Here, the original multi-wave large curve is divided into two shorter small curves. Then the starting point and the ending point of the two small curves are connected respectively to form two new straight lines, and the distance from each point of the two small curves to the corresponding straight line is calculated, and the point with the largest distance is found as the second-level truncation point. Furthermore, the curve is divided again at this truncation point to obtain a shorter small curve. By repeatedly performing this process, the original complex fluctuation can be gradually disassembled and a series of truncation points can be obtained. These truncation points are all the points with the strongest fluctuations in each section of the curve. After obtaining these truncation points, it is necessary to determine their properties to determine whether they are candidate nodes for the slot line.

[0249] In this embodiment, the turning angle at the truncation point is defined as the angle between two adjacent line segments with the point as the intersection point, such as Figure 8 As shown in the figure, Both represent the cutoff point. The angle of rotation is the vector With vector The angle between the two. The counter clock wise (CCW) rule is used to determine the positive and negative of the angle: Figure 8 middle, In clockwise order, the cutoff point The rotation angle is negative (as indicated by the solid arrow); In counterclockwise order, the cutoff point The angle is positive (as indicated by the hollow arrow). According to the rule for determining the positive and negative angles (CCW), and combined with the properties of the ridge and groove, the angle is positive and is a candidate node for the groove line (such as Figure 8 middle ).

[0250] The present invention creates a class function based on the above method to realize the function of extracting candidate nodes of trough lines from contour lines. The initialization method of this class needs to pass in the contour line set, that is, the point set array list all_points and the set contour line interval. The output is a list for storing the extracted candidate nodes of the trough line. The core of this class is to reduce the number of midpoints of a given curve, improve efficiency and keep its original shape as much as possible. The specific implementation has the following steps:

[0251] 1. First, traverse each contour line in the contour line set in turn, where each contour line contains the fluctuations of multiple points.

[0252] 2. Then calculate the distance. Select the point in the fluctuation that is farthest from the straight line connecting the starting point and the end point, that is, the point where the fluctuation is the strongest, as the cutoff point within the fluctuation range.

[0253] 3. Then, divide the current longer fluctuation into two smaller fluctuations at the truncation point.

[0254] 4. Repeat steps 2 and 3 to calculate the distance and continue to truncate the fluctuations, gradually breaking up the longer fluctuations. Recursively split each contour line into points below the first distance threshold (in this invention, it has been experimentally determined that the threshold is 25 for the best effect). The recursive termination condition is that the distance of the farthest truncation point is less than the first distance threshold (25), and return a list of truncation points in each contour line.

[0255] 5. Record each truncation point and determine whether it is a candidate node for the slot line based on the geometric relationship of the following truncation points:

[0256] Calculate the vectors around each feature point and use vector cross product to determine the angle relationship, so as to determine whether the point is a candidate node for the slot line. The steps are as follows:

[0257] 1) Define the front and back points:

[0258] Take the current point as P, and its previous point as , the next point is .

[0259] 2) Calculate vector and :

[0260]

[0261]

[0262] 3) Cross product of two vectors:

[0263]

[0264]

[0265]

[0266] in, Respectively represent vector The weight in x, y, Respectively represent vector The weight in terms of x, y.

[0267] 4) Determine the nature of feature points:

[0268] If the cross product is positive, then:

[0269]

[0270] It means the airflow turns left, that is, the slot line node, such as Figure 8middle .

[0271] 6. According to the result of the cross product and the above-mentioned judgment rules, determine whether the truncation point found by the truncation wave is a candidate node for the slot line, add it to the empty list, return the extracted list of candidate nodes for the slot line, and perform further analysis operations.

[0272] In this embodiment, the global minimum point is found and inserted into the slot candidate node list to form a new slot candidate node list.

[0273] Find the index of the minimum value from the multidimensional array data_t created to store the potential height data. Use the flat index conversion method to convert the index to the coordinates of the corresponding multidimensional array data_t, and insert this minimum point (global lowest point) as a new node to the beginning of the list trough_list. Because the trough line starts from the global minimum potential height, a list of candidate nodes for the trough line trough_list is formed.

[0274] In this embodiment, Figure 1 and Figure 2 It can be seen that a slot line crosses different contour lines. After extracting the candidate slot line nodes from each contour line, it is found that the spatial distribution of these nodes is relatively complex, such as Fig. 9 As shown in Fig. 9 The schematic diagram of the organization construction of the groove line shows the position diagram of the candidate node each_point of the groove line relative to the last node start_point in the current groove line fitting line. Note the complex relationship between the groove line node 3 and nodes 4, 5, 6, and 7. The organization rules must be established to correctly represent the groove line. In the figure, the current node (node ​​5, node 6) and the last two nodes (node ​​2, node 3) stored in the node array that has been selected to form the groove line fitting line, and the angle angle (angle1 and angle2 respectively) between the three nodes are the data stored in the list angle_list.

[0275] In meteorology, trough lines extend from contour lines with low potential height to contour lines with high potential height, or from lower elevations to higher elevations. In general, trough lines are organized and constructed according to the principle that trough lines go westward, southward, and tend to be straight lines.

[0276] Therefore, the obtained list of candidate nodes for the trough line, trough_list, is first sorted according to the third element of each point, i.e., the potential height, so that when performing complex terrain analysis, the trough line nodes with specific height characteristics can be quickly located and processed, providing an efficient data access structure for subsequent processing. Then, in the sorted list trough_list, that is, from a series of contour line data, the following comprehensive analysis is performed using distance and angle: starting from the current node, the most suitable node is selected in the next contour line with a higher elevation, and then the trough line nodes that can form the trough line are screened. The specific steps are as follows:

[0277] First, the sorted list trough_list is used to group the candidate nodes of the trough line into a certain range according to the range of the value of the contour line. By traversing the list of candidate nodes of the trough line trough_list and judging whether its value is in the contour line within the given range, a group list containing close contour lines is gradually constructed, that is, the candidate nodes of the trough line in the sorted list trough_list are grouped according to the contour line where they are located, and the result of this step is saved in the list collected_list for further analysis and processing. count is used as an index accumulator. The list now_isotherm_point_index_list is established to store the list of indexes of the currently selected nodes that constitute the trough line fitting line (trough line nodes), and the initial value is the index 0 of the first point, that is, the global minimum point found. The list fitting_line is used to store the selected trough line nodes (that is, the nodes that constitute the trough line fitting line) in the form of a dictionary, with count as the key and the first point as the initial value. Each contour line in the list collected_list is traversed in turn. Starting from index 1, process each contour line in the list collected_list: create empty lists each_contour_distance_list and angle_list. Set the last node in the current groove fitting line to start_point. Traverse the points in each contour line and calculate the Euclidean distance distance between each node in the current groove candidate node and start_point:

[0278]

[0279] Among them, x0 represents the horizontal coordinate of each_point, a candidate node of a slot line in the current contour line, y0 represents the vertical coordinate of each_point, a candidate node of a slot line in the current contour line, x represents the horizontal coordinate of the node in the current slot line fitting line as start_point, and y represents the vertical coordinate of the node in the current slot line fitting line as start_point.

[0280] Determine the position of node each_point relative to node start_point:

[0281]

[0282]

[0283] Each_point means to examine each candidate node of the slot line in the list collected_list one by one, and take the node being examined. and Indicates the position of node each_point relative to node start_point, each_point_x and each_point_y respectively indicate the horizontal and vertical coordinates of node each_point, and start_point_x and start_point_y respectively indicate the horizontal and vertical coordinates of node start_point. Fig. 9 As shown (node ​​3 is the start_point, node 4, node 5, node 6, and node 7 are each_point respectively).

[0284] It can be seen that it is relatively simple to organize and construct the slot line from node 0 to node 3. However, it is relatively complicated to continue to extend and organize the slot line from node 3, which can be divided into three cases:

[0285] Case A1: If , that is, the node each_point is on the left side of the node start_point, and the number of points selected to form the groove fitting line stored in the list now_isotherm_point_index_list is greater than or equal to 2 (such as Fig. 9 Node 5 and node 6), calculate the current node each_point and the last two nodes in the list now_isotherm_point_index_list, which are the latest nodes selected to form the groove fitting line stored in the list now_isotherm_point_index_list, and the angle angle between the three nodes (such as Fig. 9 angle1 and angle2), and store them in the list angle_list. The specific steps are as follows:

[0286] A1-1. Define vector:

[0287]

[0288]

[0289] in, , and They represent the node each_point, the node start_point and the second to last node of the most recently selected slot fitting line stored in the list now_isotherm_point_index_list respectively.

[0290] A1-2. Calculate the dot product of vectors:

[0291]

[0292] in, , and The nodes are ,node and nodes The coordinates of .

[0293] A1-3. Calculate the modulus of the vector:

[0294]

[0295]

[0296] A1-4. Calculate the cosine of an angle :

[0297] .

[0298] A1-5. Conversion between radians and degrees:

[0299]

[0300] Case A2: If , that is, node each_point is to the left of node start_point, and , that is, node each_point is above node start_point: calculate the distance between node each_point and node start_point, and add the distance to the list each_contour_distance_list (such as Fig. 9 Middle node 5); otherwise: multiply the distance by 0.7 (through experimental comparison in the present invention, setting the second threshold to 0.7 has the best effect, such as Fig. 9 Middle node 6), and then add it to the list each_contour_distance_list.

[0301] Case A3: If , that is, the node each_point is on the right side of the node start_point: according to the meteorological business specifications, the distance is set to 99999, and the angle is set to 90° (such as Fig. 9 midpoints 4 and 7).

[0302] According to the nature of the trough line, only one trough line node is selected on a contour line. Therefore, in the selection of the trough line node, the stored data should be screened according to the above three situations, that is, the node farthest to the east is first filtered out because it does not conform to the principles of meteorology. The specific steps are as follows:

[0303] B1. Find the candidate node of the trough line closest to the node start_point in the current contour line, and use minize_index to record the index of the minimum distance ( Fig. 9 Node 5 is the closest point to node 3, so the index minize_index is node 5).

[0304]

[0305] in, Represents the index of the minimum distance, Indicates that the objective function distance( , start_point) takes the index of the minimum value, Indicates the candidate node of the groove line in the current contour line that is calculating the distance with the start_point. Indicates the index of the point closest to the node start_point in the current contour line;

[0306] B2. Calculate the difference between each angle in the list angle_list and 180 degrees, recorded as :

[0307] in, It indicates that the list of angles angle between the current slot candidate node j and the last two nodes in the node list that have been selected to form the slot fitting line is recorded and stored.

[0308] B3. If and ,but: ;

[0309] in, Indicates that there are at least 2 angles in the list angle_list. Generally, the overall curvature of the slot system is not large and is closer to a straight line, or the angle formed by adjacent points is close to 180°.

[0310] Indicates that there are at least 3 points in the list now_isotherm_point_index_list. Indicates updating the index nearest_flatten_angle_index to point to the point whose angle with the previous groove line is closest to 180°. That is, select the index that minimizes angle_diff[j] (such as Fig. 9 The angle between node 6 and the previous groove fitting line is closer to 180°, that is, a straight line, than node 5, so the index nearest_flatten_angle_index is node 6).

[0311] B4. Compare the nearest_flatten_angle_index and the minize_index to see if they are inconsistent. If: ,and is less than or equal to a preset first threshold (the first threshold is 90), then: . This formula means: if the index nearest_flatten_angle_index is different from the index minize_index and the distance of the nearest_flatten_angle_index point in the list each_contour_distance_list is less than or equal to 90 (the first threshold), then update the list minize_index to the list nearest_flatten_angle_index (such as Fig. 9 If the index nearest_flatten_angle_index is node 6 and the index minize_index is different from node 5, the list minize_index is updated to node 6, that is, node 6 is selected as the point in the current contour line that best fits the groove fitting line). The purpose is to gradually select a more suitable node from a series of node sets. Considering the distance and angle of the nodes, points that are closer and in the vicinity or to the left (west) are preferred to construct a smooth and logically coherent groove fitting line for further analysis.

[0312] B5. Add the coordinates of the node that best fits the groove fitting line in the current contour line to the fitting groove line list fitting_line, and add its index to the list now_isotherm_point_index_list, update count, add 1 to the number of the currently selected groove fitting line nodes (groove line nodes), and perform the next traversal. If the length of each_contour_distance_list is 1 and the value exceeds the second threshold (the second threshold is 100), it means that there is only one point in the current contour line and the distance is extremely large, and it is impossible to find a suitable next point, so the loop is terminated and the groove line is no longer extended.

[0313]

[0314]

[0315]

[0316] in, Indicates the coordinates of the nodes in the current contour line that best fit the groove line fitting line. The index of the node in the current contour that best fits the groove fitting line, Indicates the total number of currently selected slot fitting line nodes (slot nodes).

[0317] B6. Finally, the candidate node information of the groove lines corresponding to all the indexes in the list now_isotherm_point_index_list is stored in the list contours_list_with_line.

[0318] In this embodiment, a special case is that when there are closed contour lines, it is necessary to find the minimum position of the closed contour line area and reorganize the filtered slot line node list.

[0319] In view of this situation, each contour line in the point set array contour_data is traversed in turn to determine whether there is a closed contour line: the mature and basic convex hull algorithm in computational geometry is used here to determine the outermost structure of the point set, which is used to find the smallest convex polygon that can surround all points in a set of points. In the present invention, the distance between the starting point and the end point of each contour line is calculated. If the distance is less than or equal to the set third threshold (the third threshold is set to 1 in the present invention), the point set is considered to be approximately closed. The minimum position of the area of ​​all closed contour lines is stored in the list closed_min_list. Then, according to the minimum position of the closed area, the groove line fitting line node list contours_list_with_line is truncated into multiple sub-lists, and the minimum position of the closed area is inserted at the beginning of each sub-list. Finally, the groove line fitting line node list contours_list_with_line is sorted according to the height value, and the fitting groove line list fitting_line is updated accordingly. This process solves the problem of accurately locating and processing closed areas in contour line data.

[0320] In this embodiment, the generated list contours_list_with_line is analyzed to determine whether it contains a segment sublist representing a closed area, which is specifically:

[0321] C1. If the list does not contain a segmented sublist of a closed area, each slot node in the list will be traversed one by one.

[0322] For each slot node in the list , calculate the slot line node and its subsequent slot line nodes The slope of the straight line between them is calculated, and each slope is stored in the list slope_list in turn.

[0323]

[0324] in, and Represents the slot line nodes The horizontal and vertical coordinates of and Respectively represent the next slot line node The horizontal and vertical coordinates of .

[0325] Next, the average of these slopes is calculated as the characteristic slope of the groove and saved in the variable slope.

[0326]

[0327] in, Representation list The number of data in Representation list The data index in Represents each slot node and its next slot node A list of slopes between .

[0328] C2. If the list contains segmented sublists of closed areas, traverse the slot nodes in each segmented sublist in turn.

[0329] For each segment sublist The slot node in , and similarly calculate its relationship with the next slot line node Calculate the slope of the straight line between the two and store the result in the list slope_list_j.

[0330]

[0331] Then, the average slope of each segmented sublist is calculated, and the average slopes of all segmented sublists are averaged again to obtain the average slope of the entire groove line, and finally the value is stored in the variable slope.

[0332] The average slope of each segmented sublist is:

[0333]

[0334] The average slope of all segment sublists is:

[0335]

[0336] in, Represents the slope of the slot line fitting and line of all segment sublists, m represents the list The number of data in Representation list The data index in a list representing the average slope of the slotted fits for each sublist of segments, Representation list The number of data in Representation list The data index in Represents each segment sublist The slot node in With the next slot node A list of the slopes of the lines between , and Represents the slot line nodes The horizontal and vertical coordinates of and Respectively represent the next slot line node The horizontal and vertical coordinates of .

[0337] In this embodiment, the slope of the calculated slot line fitting line is Make a judgment: if the slope of the slot fitting line is less than 1, it is considered to be a horizontal slot, otherwise it is a vertical slot. According to this standard, analyze the slopes of all slot lines; if there is no horizontal slot in the whole field, analyze the data of the next moment; if there is no horizontal slot at each moment, output no horizontal slot and end, otherwise go to the next step.

[0338] In this embodiment, when observing the trough fitting line only at a single time point, all points in the trough candidate node list trough_list are visualized, and these trough candidate nodes are numbered in sequence, and the trough nodes in the list fitting_line are connected in sequence, and then the trough fitting line is visualized, and the slope of the trough fitting line is marked. The schematic diagram of the visualization effect is shown in the figure below. Fig.10 , Fig.10 This is a diagram showing the visualization of the slot fitting line at time t (the contours_list_with_line list does not contain the segmented sublist of the closed area). Fig.11 This is a schematic diagram of the visualization effect of the trough fitting line at time t (the contours_list_with_line list contains a segmented sublist of the closed area). In the figure, longitude indicates the longitude range is 70° to 140° east longitude (unit: °E), latitude indicates the latitude range is 90° to 40° north latitude (unit: °N), and value indicates the potential height data of the point.

[0339] In this embodiment, when fitting the slot line of multiple time points, the visualization effect of the last time point is used as the background. For other time points other than the last time point, different colors are used for visualization. The specific operation is to use the minimum value point of the slot line fitting line at each time point as the starting point, and use the slope of the slot line fitting line at that time point to draw. In addition, the slope is marked in the format of time t: slope, so as to distinguish the slot line status at different time points. The corresponding visualization result schematic diagram is shown as follows Fig.12 As shown, Figure 10 to Figure 14 The horizontal coordinate is longitude, the unit is °E, the vertical coordinate is latitude, the unit is °N, and value is the potential height of the point.

[0340] In this embodiment, whether the horizontal groove turns vertical is determined by the change in the longitudinal north-south length of the groove line, which is specifically:

[0341] Traverse the time series and analyze and process the data at each moment. The previous analysis shows that "the horizontal trough extends a long distance in the east-west direction, but its length in the north-south direction is very limited. But after turning vertical, the opposite is true, and its length in the north-south direction increases sharply." Based on this, formulate the judgment rules. Traverse each forecast moment. If there is a horizontal trough in the previous moment, but there is no horizontal trough at the current moment t, then detect the next moment.

[0342] If the previous moment There is a horizontal groove, from the previous moment Starting point of horizontal groove and end point The vertical coordinate difference is used to calculate the longitudinal north-south length of the horizontal groove. :

[0343] ;

[0344] And calculate the current time Starting point of horizontal groove and end point The vertical coordinate difference is used to calculate the longitudinal north-south length of the horizontal groove. :

[0345] ;

[0346] Calculate the increase in the longitudinal north-south length of the horizontal groove between the current moment and the previous moment Relative value of:

[0347] .

[0348] If the growth If it is greater than the preset growth threshold of 0.1, it indicates that the horizontal trough turns vertical, and a severe cold wave weather process is predicted, and the start time of the horizontal trough turning vertical is output.

[0349] Otherwise, repeat the above steps to perform detection at the next moment until the detection and analysis of all forecast moments are completed.

[0350] If the horizontal trough turning vertical judgment conditions are not met at all times, the output forecast period will have no horizontal trough turning vertical and the system will exit.

[0351] In this embodiment, the maximum value Hmax0 and the minimum value Hmin of the potential height are selected from the selected area in the present invention, wherein the 500 hPa potential height field is based on 5000 gpm and contour analysis is performed every 40 gpm. The 500 hPa potential height field in the Eurasian range at 6:00 on December 28, 2020 is used for analysis, such as Fig.13 As shown, Fig.13 Visualization of the groove fitting line when t = 30 (the contours_list_with_line list does not contain the segmented sublist of the closed area).

[0352] In this embodiment, the potential height field of the Eurasian 500 hPa weather map at 0:00 on December 27, 2020 is used for analysis. Fig.14 As shown, Fig.14 Visualization of the grooved fit line for t = 0 (the contours_list_with_line list contains a sublist of segments of the enclosed area). Figure 10-Figure 14 The horizontal coordinate is longitude, the unit is °E, the vertical coordinate is latitude, the unit is °N, and t represents time.

[0353] In summary, the present invention utilizes the most commonly used 500hPa potential height field data in meteorological services, and realizes the automatic forecast of the most severe cold wave meteorological disasters affecting my country by automatically distinguishing the vertical transition of horizontal troughs, greatly improving the automation level of business forecasts, and enhancing my country's disaster prevention, mitigation, and relief capabilities for large-scale areas in response to severe cold waves. The technical solution of the present invention not only improves the accuracy and timeliness of cold wave monitoring, but also provides strong data support for meteorological analysis and forecasting, and improves the ability of meteorological disaster warning and emergency response. Compared with traditional methods, this method is direct, convenient, and efficient, and realizes fully automated operation without the need for human participation, thereby significantly reducing the workload of staff and eliminating errors that may be introduced by human subjective judgment. The automation characteristics of the technical solution of the present invention not only improve work efficiency, but also reduce dependence on operators.

Claims

1. An automatic method for predicting strong cold waves from horizontal trough to vertical type through wave truncation, It is characterized in that The following steps are involved: S1. Read the potential height data in the 500hPa forecast key area in the forecast business and convert the screen coordinates into a Cartesian coordinate system; S2. Perform contour analysis on the potential height data, extract candidate nodes of the trough line using the cut wave method according to the analysis results, and construct the trough line organization to identify the trough line; S3, calculating the slope of each segmented slot fitting line; S4, judging whether the slope is less than 1, if so, a horizontal groove appears, and entering S5, otherwise it is a vertical groove, and the data at the next moment is calculated to find the horizontal groove; If there is no horizontal slot at all times, then output "no horizontal slot" and end the process; S5. Mark and number the candidate nodes of the trough line, and calculate the longitudinal north-south length of the transverse trough according to the coordinates of the starting and ending points; S6. Determine whether the trough is turning vertical based on the change in its longitudinal north-south length, and complete the forecast of a strong cold wave that is turning vertical.

2. The automatic method for predicting the transition of a horizontal trough to a vertical strong cold wave by wave truncation according to claim 1, It is characterized in that The S1 comprises the following steps: S101, reading the potential height data in the 500hPa forecast key area in the forecast service, wherein the potential height data is three-dimensional data, including time intervals, latitude and longitude grid intervals; S102, selecting a geographical area for the read geopotential height data, calculating the required latitude and longitude ranges, and extracting data in the corresponding area; S103, according to the data in the corresponding area extracted, the screen coordinates are converted into a Cartesian coordinate system using the following formula: y=height-y* Among them, y represents the vertical coordinate in the Cartesian coordinate system, height represents the vertical height of the screen, and y* represents the vertical coordinate in the screen coordinate system.

3. The automatic method for predicting the strong cold wave from horizontal trough to vertical type by wave truncation according to claim 2, It is characterized in that The S2 comprises the following steps: S201, create a two-dimensional array data_t, and perform contour analysis on the potential height data, and distribute the contour nodes in the list contours according to the elevation level, wherein contours represents a list containing all contour node sets, and the two-dimensional array data_t is used to store the potential height data within the range of 90°N-40°N and 70°E-140°E at time t; S202, traversing each contour line in the list contours, and storing the contour line nodes in the list all_points in the form of triplets in order, wherein each contour line contains fluctuations of multiple points; S203, extracting the truncation point by using the truncation method, and using the vector cross multiplication value as the positive determination slot line candidate node; S204, searching for the minimum value point in the two-dimensional array data_t, and inserting it into the trough candidate node list trough_list; S205, based on the list of candidate trough nodes trough_list, the candidate trough nodes are screened according to the general principle that the trough lines are heading west, south, and tending to be straight, to form a trough node organization, and the trough node organization is constructed into a fitting trough list fitting_line; S206. Use the convex hull algorithm to determine whether there is a closed contour line. If so, find the minimum position of the closed contour line, reorganize the slot line node list, and complete the identification of the slot line.

4. The automatic method for predicting the transition of a horizontal trough to a vertical strong cold wave by wave truncation according to claim 3, It is characterized in that The S201 includes the following steps: S2011. In the selected geographical area, select the maximum value Hmax0 and the minimum value Hmin0 of the potential height, respectively, where the 500 hPa potential height field is based on 5000 gpm and contour analysis is performed every 40 gpm; S2012, create a two-dimensional array data_t for storing the potential height data within the range of 90°N-40°N and 70°E-140°E at time t; S2013. Perform contour analysis on the two-dimensional array data_t and distribute the contour nodes in the list contours according to the elevation levels.

5. The automatic method for predicting the strong cold wave from horizontal trough to vertical type by wave truncation according to claim 2, It is characterized in that The S203 comprises the following steps: S2031, draw a straight line with the starting and ending points of the original multi-wave curve, calculate the distance from each point on the original multi-wave curve to the straight line, and take the point with the largest distance as the first-level truncation point; S2032, based on the first-level truncation point, the original multi-wave large curve is divided into two shortest small curves, the starting points and the ending points of the two small curves are respectively connected to form two new straight lines, and the distances from the points of the two small curves to the corresponding new straight lines are calculated, and the point with the largest distance is calculated and found as the second-level truncation point; S2033, repeating S2031 and S2032 until the distance of the farthest truncation point is less than a preset first distance threshold, and then the extraction of the truncation point is completed; S2034, define the order of the points before and after the cutoff point set: take the current point as P, and its previous point as , the next point is ; S2035, calculate the vectors according to the defined front and back points and : S2036, for two vectors and Perform a cross product calculation: in, Respectively represent vector The components in the x and y directions, Respectively represent vector Components in the x and y directions; S2037. If the cross product calculation result is positive, it is a candidate node for the slot line, and thus the candidate node for the slot line is identified in the truncation point set.

6. The automatic method for predicting the strong cold wave from horizontal trough to vertical type by wave truncation according to claim 2, It is characterized in that The S205 comprises the following steps: S2051, sorting the trough candidate node list trough_list according to the potential height; S2052, grouping the sorted candidate trough nodes in the candidate trough node list trough_list according to the contour lines where they are located, and saving the grouping results into the list collected_list; S2053, constructing a list now_isotherm_point_index_list for storing the indexes of the currently selected nodes constituting the slot fitting line, wherein the initial value is the index 0 of the first point, that is, the minimum value point in the two-dimensional array data_t; S2054, constructing a list fitting_line for storing selected slot line nodes in a dictionary form; S2055, traverse each contour line in the list collected_list in turn, starting from index 1, create empty lists each_contour_distance_list and angle_list for each contour line in the list collected_list, set the last node in the current slot line fitting line as start_point, traverse the slot line candidate nodes in each contour line, and calculate the Euclidean distance between each slot line candidate node in the current contour line and the last node start_point in the current slot line fitting line : Among them, x 0 Indicates the horizontal coordinate of each_point in the current contour line, y 0 Indicates the ordinate of a node each_point in the current contour line, x indicates the abscissa of the node start_point in the current slot line fitting line, y indicates the ordinate of the node start_point in the current slot line fitting line, and both node each_point and node each_point are candidate nodes for the slot line; S2056. Determine the position of node each_point relative to node start_point using the following formula: in, and Indicates the position of node each_point relative to node start_point, each_point_x and each_point_y respectively represent the horizontal and vertical coordinates of node each_point, start_point_x and start_point_y respectively represent the horizontal and vertical coordinates of node start_point; S2057, if , that is, when the node each_point is on the left side of the node start_point and the number of points selected to constitute the groove fitting line stored in the list now_isotherm_point_index_list is greater than or equal to 2, the angle between the current node each_point and the last two nodes in the list now_isotherm_point_index_list is calculated, wherein the last two nodes calculated are the latest selected nodes to constitute the groove fitting line stored in the list now_isotherm_point_index_list, and the angle between the three nodes is stored in the list angle_list, the three nodes include the current node and the last two nodes, and the angle is the angle formed by the current node each_point and the last two nodes in the list now_isotherm_point_index_list, wherein the last node in the list now_isotherm_point_index_list is located at the center; like , that is, the node each_point is on the left side of the node start_point, and , that is, the node each_point is above the node start_point, then the distance between the node each_point and the node start_point is calculated, and the distance between the node each_point and the node start_point is added to the list each_contour_distance_list; otherwise, the distance between the node each_point and the node start_point is multiplied by the second distance threshold and added to the list each_contour_distance_list; like , that is, the node each_point is on the right side of the node start_point, and the distance is set to 99999 according to the meteorological business specifications, added to the list each_contour_distance_list, and the angle is set to 90°, added to the list angle_list; S2058. Filter the three situations of S2057 to obtain a fitting slot line list fitting_line.

7. The automatic method for predicting a strong cold wave that turns from a horizontal trough to a vertical one by wave truncation according to claim 6, It is characterized in that The S2058 comprises the following steps: S20581. Find the candidate node of the trough line closest to the node start_point in the current contour line, and use minize_index to record the index of the minimum distance: in, Represents the index of the minimum distance, Indicates that the objective function distance( ,start_point) takes the index of the minimum value, Indicates the candidate node of the slot line that is calculating the distance with the node start_point in the current contour line. Indicates the index of the candidate node of the trough line closest to the node start_point in the current contour line; S20582. Calculate the difference between each angle in the list angle_list and 180° : in, It means recording and storing the current slot candidate node j and the last two nodes in the node list that have been selected to form the slot fitting line, and the list of angles angle between the three nodes; S20583, if and ,but: ; in, Indicates that the number of angles in the list angle_list is greater than 2. Indicates that the number of points in the list now_isotherm_point_index_list is greater than 3, Indicates updating the index nearest_flatten_angle_index to point to the candidate slot line node whose angle with the previous slot line is closest to 180°; S20584, compare the nearest_flatten_angle_index and the minize_index to see if they are inconsistent. ,and If it is less than or equal to the preset first threshold, the index minize_index is updated to the index nearest_flatten_angle_index; S20585, add the coordinates of the node in the current contour line that best fits the groove line fitting line to the fitting groove line list fitting_line, and add its index to the list now_isotherm_point_index_list, update count, add 1 to the number of nodes that currently have been selected to form the groove line fitting line, and perform the next traversal. If the length of the list each_contour_distance_list is 1 and the value exceeds the set second threshold, it means that there is only one point in the current contour line and the distance is the largest, then terminate the loop and the groove line will no longer extend, where count represents the total number of nodes that currently have been selected to form the groove line fitting line, where the node that best fits the groove line fitting line is the node corresponding to the index minize_index; S20586. The candidate node information of the groove line corresponding to all the indexes in the list now_isotherm_point_index_list is stored in the list contours_list_with_line to complete the fitting of the groove line.

8. The automatic method for predicting a strong cold wave that turns from a horizontal trough to a vertical one by wave truncation according to claim 3, It is characterized in that The S206 is specifically as follows: Calculate the distance between the starting and ending points of each contour line; If the distance is less than the set third threshold, the point set array contour_data is determined to be approximately closed; or If the distance is equal to the set third threshold, the point set array contour_data is determined to be approximately closed; Store the minimum positions of all closed contour lines into the list closed_min_list; According to the minimum position of the closed area, the contours_list_with_line node list of the groove fitting line is truncated into multiple sub-lists; Insert the minimum position of the enclosed area at the beginning of each sublist, sort the contour line fitting line node list contours_list_with_line according to the height value, and update the fitting groove line list fitting_line.

9. The automatic method for predicting a strong cold wave that turns from a horizontal trough to a vertical one by wave truncation according to claim 7, It is characterized in that The S3 comprises the following steps: S301, analyze the list contours_list_with_line to determine whether it includes a segment sublist, if so, proceed to S303, otherwise, proceed to S302; S302, calculate the slopes between the nodes of each slot line, take the average value, use the average value as the slope of the slot line fitting line, and save it: in, represents the slope of the slot line fitting line, Representation list The number of data in Representation list The data index in Represents each slot node and its next slot node A list of slopes between ; S303, traverse the slot line nodes in each segment sublist in turn, calculate the slopes between the slot line nodes in each segment sublist, and take the average value to obtain the average slope of the segment sublist, and average the average slopes of all segment sublists to obtain the slot line approximation slopes of all segment sublists, and store the slot line approximation slopes of all segment sublists: in, Represents the slope of the slot line fitting and line of all segment sublists, m represents the list The number of data in Representation list The data index in a list representing the average slope of the slotted fits for each sublist of segments, Representation list The number of data in Representation list The data index in Represents each segment sublist The slot node in With the next slot node A list of the slopes of the lines between , and Represents the slot line nodes The horizontal and vertical coordinates of and Respectively represent the next slot line node The horizontal and vertical coordinates of .

10. The automatic method for predicting a strong cold wave that turns from a horizontal trough to a vertical one by wave truncation according to claim 1, It is characterized in that The S5 is specifically: Traverse the time series and analyze and process the data at each moment; If the previous moment There is a horizontal groove, from the previous moment Starting point of horizontal groove and end point The vertical coordinate difference is used to calculate the longitudinal north-south length of the horizontal groove. : ； By calculating the current time Starting point of horizontal groove and end point The vertical coordinate difference is used to calculate the current longitudinal north-south length of the horizontal groove. : 。 11. The automatic method for predicting a strong cold wave that turns from a horizontal trough to a vertical one by wave truncation according to claim 10, It is characterized in that The S6 is specifically: Traverse each time point and calculate the increase in the longitudinal north-south length of the horizontal groove between the current moment and the previous moment Relative value of: If the growth If it is greater than the preset growth threshold, it indicates that the horizontal trough turns vertical, and a severe cold wave weather process is predicted, and the start time of the horizontal trough turning vertical is output; Otherwise, if the judgment conditions for the horizontal trough to vertical transition are not met at all times, the output is that there is no horizontal trough to vertical transition during the forecast period, and the process ends.

Images