A method and system for extracting fine vertical structure of atmosphere under mountain complex terrain

Abstract

The application belongs to the technical field of meteorological data extraction and processing, and specifically provides a method and system for extracting fine atmospheric vertical structure under complex mountainous terrain, comprising: collecting multi-source observation data of atmospheric vertical profile under complex mountainous terrain, normalizing the multi-source observation data to obtain normalized multi-source observation data; using the altitude of the data collection point, performing space-time matching on the normalized multi-source observation data to obtain space-time groups; analyzing the normalized multi-source observation data in the space-time groups to construct a multi-source profile merging system; extracting the fine atmospheric vertical structure based on the multi-source profile merging system and the normalized multi-source observation data; arranging the fine atmospheric vertical structures of different space-time groups in time sequence, predicting the change of the fine atmospheric vertical structure, and obtaining the prediction result. The technical scheme of the application is helpful for carrying out fine observation of atmospheric vertical structure, and has important significance for improving the disaster weather defense capability and ensuring aviation safety.

Description

Technical Field

[0001] This invention belongs to the field of meteorological data extraction and processing technology, specifically relating to a method and system for extracting the fine vertical structure of the atmosphere under complex mountainous terrain. Background Technology

[0002] Atmospheric vertical structure refers to the vertical distribution characteristics of atmospheric physical parameters and is a core driving factor influencing atmospheric motion and weather and climate change. For example, the occurrence and development of severe weather events such as typhoons, rainstorms, and severe convection are directly related to the dramatic evolution of atmospheric vertical structure. Strong unstable atmospheric stratification, abundant vertical water vapor transport, and strong vertical upward motion are all important factors that may lead to severe weather events. Therefore, obtaining accurate atmospheric vertical structure is a prerequisite for accurate forecasting.

[0003] Atmospheric vertical structure extraction technology refers to a series of methods that utilize various direct and indirect observation techniques to acquire, analyze, and process data on the vertical distribution of physical parameters (such as temperature, pressure, humidity, and wind) at different altitudes of the Earth's atmosphere. As a core means of obtaining these key physical parameters, the technological level of atmospheric vertical structure extraction technology directly determines the upper limit of its service capabilities in application fields, and it has long been a key research and development direction for global meteorological departments, research institutions, and industries such as aviation and the environment.

[0004] Existing atmospheric vertical structure extraction technologies can be broadly categorized into direct observation and indirect observation techniques based on the observation method. Direct observation techniques involve using sensors mounted on a platform to directly enter different altitudes of the atmosphere and collect real-time information on atmospheric physical parameters. These techniques offer advantages such as high measurement accuracy and intuitive data, and are used in methods like radiosonde and aircraft observation. However, limitations in platform performance and cost result in drawbacks such as narrow coverage, low observation frequency, poor environmental adaptability, and weak data representativeness, making it difficult to meet the observation needs of rapidly changing weather systems. Indirect observation techniques utilize ground-based or airborne platforms equipped with remote sensing equipment to receive signals from the atmosphere's scattering, absorption, and radiation of electromagnetic waves. By leveraging the interaction between electromagnetic waves and atmospheric water vapor, aerosols, and precipitation particles, atmospheric vertical structure parameters are retrieved through echo signal inversion. These techniques offer advantages such as wide coverage, high observation frequency, and low cost. However, the inversion process is complex, data accuracy is affected by various factors, and vertical resolution is relatively low.

[0005] The evolution of atmospheric vertical structure is influenced by multiple physical processes, including dynamics, thermodynamics, and radiative transfer, making it a highly complex nonlinear system. In mountainous regions, the complex terrain leads to dramatic changes in atmospheric vertical structure, making traditional direct observation techniques insufficient. To better obtain information on atmospheric vertical structure evolution under complex mountainous terrain, this paper proposes a technical approach that combines the advantages of both direct and indirect detection methods to extract atmospheric vertical structure data under such conditions. This will facilitate refined observation of atmospheric vertical structure and is of great significance for improving disaster weather preparedness and ensuring aviation safety. Summary of the Invention

[0006] This invention aims to address the shortcomings of existing technologies by providing a method and system for extracting fine atmospheric vertical structure in complex mountainous terrain. This provides more reliable reference data for weather forecasting and improves the effectiveness and practicality of atmospheric vertical structure extraction.

[0007] To achieve the above objectives, the present invention provides the following solution:

[0008] A method for extracting fine vertical atmospheric structure under complex mountainous terrain includes:

[0009] Multi-source observation data of atmospheric vertical profiles under complex mountainous terrain are collected, and the multi-source observation data is normalized to obtain normalized multi-source observation data; wherein, the multi-source observation data includes temperature profiles, wind speed profiles, and humidity profiles;

[0010] By utilizing the altitude of the data collection points, normalized multi-source observation data are spatiotemporally matched to obtain spatiotemporal groupings;

[0011] Normalized multi-source observation data in spatiotemporal groupings are analyzed to construct a multi-source profile merging system;

[0012] Based on the multi-source profile merging system and the normalized multi-source observation data, the fine vertical structure of the atmosphere is extracted.

[0013] By arranging the fine vertical structure of the atmosphere into different spatiotemporal groups according to time series, the changes in the fine vertical structure of the atmosphere are predicted, and the prediction results are obtained.

[0014] Preferred methods for obtaining normalized multi-source observation data include:

[0015] The points on the temperature profile, the wind speed profile, and the humidity profile are respectively named temperature profile point, wind speed profile point, and humidity profile point.

[0016] Perform normalization calculations on the temperature values ​​of all temperature profile points to convert the temperature values ​​into normalized temperature values;

[0017] Perform normalization calculations on the wind speed values ​​of all wind speed profile points to convert the wind speed values ​​into normalized wind speed values;

[0018] Perform normalization calculations on the humidity values ​​of all humidity profile points to convert the humidity values ​​into normalized humidity values;

[0019] Based on the normalized temperature value, the normalized wind speed value, and the normalized humidity value, normalized multi-source observation data are obtained.

[0020] Methods for constructing multi-source profile merging systems include:

[0021] By integrating the normalized values ​​of temperature, wind speed, and humidity at a preset altitude, atmospheric element detection information at the preset altitude is established.

[0022] Based on the atmospheric element detection information, a multi-source profile merging system is constructed.

[0023] Preferred methods for extracting the fine vertical structure of the atmosphere include:

[0024] A two-dimensional coordinate system was established with the normalized multi-source observation data as the X-axis and the altitude of the data collection points as the Y-axis to obtain the multi-source profile merging system.

[0025] All normalized multi-source observation data within the same spatiotemporal group are entered into the multi-source profile merging system according to the altitude of the data collection points to obtain the structural profile; wherein, the structural profile includes a temperature normalized profile, a wind speed normalized profile, and a humidity normalized profile.

[0026] Based on preset monotonic low values ​​and preset monotonic high values, the monotonic range of the structural profile is extracted, and the common intersection of the monotonic ranges of all structural profiles is obtained.

[0027] Based on the common intersection, the range of regular changes is obtained, and the fine vertical structure of the atmosphere is extracted; wherein, the altitude interval corresponding to a range of regular changes is a fine atmospheric structure layer.

[0028] Preferred methods for predicting changes in the fine vertical structure of the atmosphere include:

[0029] Obtain the historical variation range of multi-source historical observation data, and analyze the variation patterns of different atmospheric fine structure layers based on the historical variation range;

[0030] Based on the aforementioned patterns of change, the changes in the fine vertical structure of the atmosphere are predicted according to a time series.

[0031] Preferred methods for obtaining the variation patterns of the atmospheric fine structure layer include:

[0032] The atmospheric fine structure layers are numbered in order from bottom to top, and the points on the structural profile corresponding to the historical regularity range of the preset number are taken as the inner profile points of the layer.

[0033] The difference between the target's historical multi-source observation data corresponding to the inner contour point and the minimum value of historical multi-source observation data in the layer height of the historical change range is obtained; wherein, the layer height is the altitude corresponding to the intersection of any two historical change ranges;

[0034] The difference between the altitude corresponding to the target's historical multi-source observation data at the inner contour point of the calculation layer and the minimum value of the layer height within the historical variation range is obtained;

[0035] Based on the height difference and the profile point difference, a data prediction coordinate system is constructed, and a function regression is performed on the data prediction coordinate system to obtain the data prediction function;

[0036] Based on the data prediction function, the variation pattern of the atmospheric fine structure layer is obtained.

[0037] Preferably, the method for predicting changes in the fine vertical structure of the atmosphere based on the aforementioned change patterns and according to time series includes:

[0038] A turning point prediction coordinate system is constructed based on the minimum value of the stratified height within the historical variation range corresponding to a preset number, and the basic data values; wherein, the basic data values ​​are the historical multi-source observation data of the target corresponding to the lowest elevation point.

[0039] Perform function regression on the aforementioned turning point prediction coordinate system to obtain the turning point prediction function;

[0040] A basic prediction coordinate system is constructed based on time series and historical multi-source observation data of the target.

[0041] Perform function regression on the basic prediction coordinate system to obtain the basic prediction function;

[0042] Substitute the preset prediction time into the basic prediction function to solve for the predicted values ​​of the basic data values ​​of different historical multi-source observation data, and obtain the basic prediction data.

[0043] Substituting the basic prediction data into the turning point prediction function, prediction profile data is obtained, thus completing the prediction of the changes in the fine vertical structure of the atmosphere.

[0044] This invention also provides a system for extracting fine vertical atmospheric structure under complex mountainous terrain, used to implement the method, comprising:

[0045] The data processing module is used to collect multi-source observation data of atmospheric vertical profiles under complex mountainous terrain, and to normalize the multi-source observation data to obtain normalized multi-source observation data; wherein, the multi-source observation data includes temperature profiles, wind speed profiles and humidity profiles.

[0046] The spatiotemporal grouping module is used to perform spatiotemporal matching of normalized multi-source observation data using the altitude of the data collection points to obtain spatiotemporal groups;

[0047] The system construction module is used to analyze normalized multi-source observation data in spatiotemporal grouping and construct a multi-source profile merging system;

[0048] The structure extraction module is used to extract the fine vertical structure of the atmosphere based on the multi-source profile merging system and the normalized multi-source observation data.

[0049] The structural change prediction module is used to arrange the fine vertical structure of the atmosphere in different spatiotemporal groups according to the time series, predict the changes in the fine vertical structure of the atmosphere, and obtain the prediction results.

[0050] Preferably, the data processing module includes:

[0051] A profile point construction unit is used to name the points on the temperature profile, the wind speed profile, and the humidity profile as temperature profile points, wind speed profile points, and humidity profile points, respectively.

[0052] The temperature normalization unit is used to perform normalization calculations on the temperature values ​​of all temperature profile points and convert the temperature values ​​into normalized temperature values.

[0053] The wind speed normalization unit is used to perform normalization calculations on the wind speed values ​​of all wind speed profile points and convert the wind speed values ​​into normalized wind speed values.

[0054] The humidity normalization unit is used to perform normalization calculations on the humidity values ​​of all humidity profile points and convert the humidity values ​​into normalized humidity values.

[0055] The normalized data acquisition unit is used to obtain normalized multi-source observation data based on the normalized temperature value, the normalized wind speed value, and the normalized humidity value.

[0056] Compared with the prior art, the beneficial effects of the present invention are as follows: data constraints on multi-source observation data can convert multi-source observation data into a unified data range. At this time, the multi-source observation data can be simultaneously included in the same multi-source profile merging system for synchronous analysis. Then, based on the monotonicity of different multi-source observation data at different altitudes, the altitude is grouped to divide the fine vertical structure of the atmosphere. At this time, the changes of multi-source observation data within each structure are regular changes, which can better reflect the structural distribution of the atmosphere in the vertical direction. Moreover, the multi-source observation data includes data monitored on mountain tops and valleys. The extracted fine vertical structure of the atmosphere can also reflect the finer vertical structure of the atmosphere between mountain tops and valleys under complex mountain terrain, thus improving the accuracy and effectiveness of atmospheric vertical structure extraction.

[0057] This invention obtains the historical variation range of multi-source observation data and analyzes the variation patterns of different atmospheric fine structure layers based on these historical variation ranges. Based on these variation patterns, it predicts the changes in the atmospheric fine vertical structure according to the time series. The advantage is that it predicts the changes of each multi-source observation data, thereby predicting the future distribution of the atmospheric fine structure layer, providing more reliable reference data for meteorological forecasting, and improving the effectiveness and practicality of atmospheric vertical structure extraction. Attached Figure Description

[0058] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 This is a flowchart illustrating the steps of the method of the present invention;

[0060] Figure 2 This is a schematic diagram of the temperature profile of the present invention;

[0061] Figure 3 This is a schematic diagram of the multi-source profile merging system of the present invention;

[0062] Figure 4 This is a schematic diagram of the temperature normalization profile for spatiotemporal high grouping according to the present invention;

[0063] Figure 5 This is a schematic diagram of all the atmospheric fine structure layers of the present invention;

[0064] Figure 6 This is a schematic diagram of the predicted profile data corresponding to the temperature in S1 of the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0067] Example 1:

[0068] like Figure 1 As shown, a method for extracting the fine vertical structure of the atmosphere under complex mountainous terrain includes:

[0069] Step 1: Collect multi-source observation data of atmospheric vertical profiles under complex mountainous terrain, and normalize the multi-source observation data to obtain normalized multi-source observation data; among them, the multi-source observation data includes temperature profiles, wind speed profiles, and humidity profiles.

[0070] Specifically, ground-based remote sensing equipment is installed in mountainous and complex terrain areas. The ground-based remote sensing equipment includes temperature ground-based remote sensing equipment, wind speed ground-based remote sensing equipment, and humidity ground-based remote sensing equipment. Multi-source observation data of atmospheric vertical profiles are obtained through temperature ground-based remote sensing equipment, wind speed ground-based remote sensing equipment, and humidity ground-based remote sensing equipment.

[0071] A further implementation method for obtaining normalized multi-source observation data includes:

[0072] Points on the temperature, wind speed, and humidity profiles are named temperature profile points, wind speed profile points, and humidity profile points, respectively. Normalization is performed on the temperature values ​​of all temperature profile points to convert them into normalized temperature values. Normalization is also performed on the wind speed values ​​of all wind speed profile points to convert them into normalized wind speed values. Similarly, normalization is performed on the humidity values ​​of all humidity profile points to convert them into normalized humidity values. This embodiment uses temperature profile points as an example to provide a specific normalization calculation process. Normalization is performed on the temperature values ​​of all temperature profile points, using the minimum and maximum temperatures as 0 and 1 points, respectively. A linear fitting method is used to convert the temperature values ​​into normalized temperature values, as calculated below:

[0073] ,

[0074] Where N_value is the normalized value after transformation, H_value is the inversion value at a specific height, Min_value is the minimum profile value, and Max_value is the maximum profile value.

[0075] Similarly, normalization calculations are performed on the wind speed (humidity) values ​​of all wind speed (humidity) profile points to convert the wind speed (humidity) values ​​into normalized wind speed (humidity) values;

[0076] Normalized multi-source observation data are obtained based on normalized values ​​for temperature, wind speed, and humidity. The normalized values ​​for temperature, wind speed, and humidity are all within the range of 0 to 1.

[0077] In specific implementation, such as Figure 2 As shown, for example, in all temperature profiles, the maximum temperature is 40℃ and the minimum temperature is -20℃. If the temperature at a certain point on a certain temperature profile is 0℃, then the normalized value of the temperature at that point is 1 / 3. Similarly, all normalized values ​​of temperature, wind speed, and humidity are calculated. The purpose is to unify the dimensions of the X-axis of the temperature profile, wind speed profile, and humidity profile so that they can be analyzed in the same coordinate system. Normalization is an existing data processing method, and this embodiment will not explain it in detail. The normalized values ​​of temperature, wind speed, and humidity after normalization can still form the temperature profile, wind speed profile, and humidity profile. Only the X-axis is converted from its respective dimensions to a unified 0 to 1.

[0078] Step Two: Utilizing the altitude of the data acquisition points, normalized multi-source observation data are spatiotemporally matched to obtain spatiotemporal groupings. Different observation devices have different spatial and temporal resolutions. To better match observations, cubic spline interpolation is used, with a temporal and spatial resolution of 3 minutes and 50m, to resample and match all observation data, ensuring a uniform spatiotemporal resolution. Based on different vertical altitudes, the observation data are merged and integrated to group the atmospheric vertical profile observations. The multi-source observation data at the highest and lowest altitudes are named "high point data" and "low point data," respectively. High point data at the same time are grouped together and named the "spatial-temporal high group," while low point data at the same time are grouped together and named the "spatial-temporal low group." The spatiotemporal high and low groups at the same time constitute a single spatiotemporal group.

[0079] In practice, the division into spatiotemporal groups is merely to integrate multi-source monitoring data recorded at the same location at the same time.

[0080] Step 3: Analyze the normalized multi-source observation data in the spatiotemporal grouping. Taking a certain altitude as an example, by integrating the normalized values ​​of temperature, wind speed, and humidity at that altitude, establish the atmospheric element detection information for that altitude layer and construct a multi-source profile merging system. Utilize atmospheric observation information from all altitude layers to construct a comprehensive atmospheric vertical profile. For example... Figure 3 As shown.

[0081] Step 4: Extract the fine vertical structure of the atmosphere based on the multi-source profile merging system and normalized multi-source observation data. Based on the merged high-resolution profile, the fine vertical structure of the atmosphere is extracted for different types of vertical structures. A further implementation method for extracting the fine vertical structure of the atmosphere includes:

[0082] All normalized multi-source observation data within the same spatiotemporal group are entered into a multi-source profile merging system according to the altitude of the data collection points to obtain structural profiles. The structural profiles include temperature normalization profiles, wind speed normalization profiles, and humidity normalization profiles. Different colored dashed lines represent the temperature normalization profiles, wind speed normalization profiles, and humidity normalization profiles of the high spatiotemporal group, while different colored solid lines represent the temperature normalization profiles, wind speed normalization profiles, and humidity normalization profiles of the low spatiotemporal group.

[0083] Figure 3 The normalized multi-source observation data along the X-axis includes normalized values ​​for temperature, wind speed, and humidity. Because excessive data can make it difficult to distinguish different vertical profiles in a multi-source profile merging system, this embodiment only uses temperature and humidity profiles as examples. Figure 3 The solid black line represents the temperature normalized profile for the lower spatiotemporal group, the solid gray line represents the humidity normalized profile for the lower spatiotemporal group, the dashed black line represents the temperature normalized profile for the higher spatiotemporal group, and the dashed gray line represents the humidity normalized profile for the higher spatiotemporal group. It should be noted that... Figure 3 The altitude is defined with the lowest point of elevation as zero altitude. Assuming the highest point is 600 meters (0.6 km) higher than the lowest point, the lowest altitude for the normalized profiles of temperature, wind speed, and humidity in the spatiotemporal altitude grouping is... Figure 3 The value is 0.6, meaning that there is no data for the temperature, wind speed, and humidity normalized profiles in the air-space high group below 0.6.

[0084] Based on preset monotonic low and high values, the monotonic range of the structural profile is extracted, and the common intersection of the monotonic ranges of all structural profiles is obtained. In this embodiment, when analyzing any structural profile, it is named the target analysis profile. Specifically, a monotonic low and high value are set, represented by symbols A and B. A and B are initially zero in order of increasing height. B is increased to obtain the part of the target analysis profile within the range [A, B], which is named the sub-line to be analyzed. If the sub-line to be analyzed exhibits monotonicity, B is increased further. If the monotonicity of the sub-line to be analyzed is destroyed, B is stopped from increasing. The range [A, B] at this time is named the monotonic range. Then, the value of A is set to B, and B is increased again. The search for the monotonic range is repeated.

[0085] Based on common intersections, the range of regular variations is obtained, and the fine vertical structure of the atmosphere is extracted; where the altitude interval corresponding to a range of regular variations is considered as a fine atmospheric structure layer. In specific implementation, A and B are only used to find different monotonic ranges. This embodiment takes the temperature normalized profile grouped by spatiotemporal altitude as an example, such as... Figure 4 As shown, A and B are initially zero. Since there is no data for the temperature normalization profile of the spatiotemporal height group below 0.6, A and B start at 0.6. B is then increased. When B increases to 10.9, the temperature normalization profile of the spatiotemporal height group within the range [0.6, 10.9] decreases with increasing X-axis, exhibiting monotonicity. However, when B is increased again, the normalized temperature value at height B is higher than the temperature at height 10.9, which does not conform to the monotonicity of the temperature normalization profile of the spatiotemporal height group within the range [0.6, 10.9] decreasing with increasing X-axis. Therefore, the increase of B is stopped. The range [0.6, 10.9] at this point is a monotonic range. Then A is set to 10.9, and B is increased again to find the next... For each monotonic range, it's important to note that the subsequent range [A, B] must use a left-open, right-closed interval, i.e., (A, B]. Similarly, extract the monotonic range for each structural profile. For example, in this embodiment, there are four structural profiles, labeled α1, α2, α3, and α4. α1 has a monotonic range [0, 10.2], α2 has a monotonic range [0.6, 10.9], α3 has a monotonic range [0, 3.6], and α4 has a monotonic range [0.6, 4.8]. They share a common intersection [0.6, 3.6], meaning [0.6, 3.6] represents a regularly varying range. This range, from 0.6 km to 3.6 km in altitude, constitutes a fine atmospheric structure layer. Finally, all the fine atmospheric structure layers are extracted as follows: Figure 5 As shown, Figure 5 The area between each two horizontal dashed lines in the diagram represents a fine structure layer of the atmosphere.

[0086] A further implementation method for predicting changes in the fine vertical structure of the atmosphere includes:

[0087] Obtain the historical variation range of multi-source historical observation data, and analyze the variation patterns of different atmospheric fine structure layers based on the historical variation range;

[0088] Based on the aforementioned patterns of change, the changes in the fine vertical structure of the atmosphere are predicted according to a time series.

[0089] A further implementation method for obtaining the variation patterns of the atmospheric fine structure layer includes:

[0090] The atmospheric fine structure layers are numbered from bottom to top, and points on the structural profiles corresponding to the historical variation ranges of the preset numbers are taken as intra-layer profile points. Specifically, the atmospheric fine structure layers are numbered from bottom to top and denoted by the symbol Sn, where n is a positive integer and n is the sequence number of S. The m-th historical variation range corresponding to Sn is marked as P(n,m), where m is a positive integer and (n,m) is the sequence number of P. The m in P(n,m) corresponds to the chronological order in ascending order. The minimum value of the target's historical multi-source observation data at the layer height of P(n,m) is marked as T(n,m). The target's historical multi-source observation data is any historical multi-source observation data to be analyzed.

[0091] In specific implementation, from Figure 5 Fifteen fine atmospheric structure layers were extracted, representing 15 different altitude layers with varying patterns of change within the atmosphere. These were numbered S1 to S15. The historical variation range reveals the changes in these 15 altitude layers over historical periods. Even within the same altitude layer, their altitude ranges can differ. For example, in a certain historical variation range, S1 ranges from [0.6, 3.8], while the range of S1 extracted in this embodiment is [0.6, 3.6], a difference of 0.2 km. This indicates that the size of the altitude layer range can vary slightly, but the number of altitude layers typically remains constant. This embodiment uses... Historical multi-source observation data corresponding to the temperature profile are used as target analysis data to illustrate the subsequent analysis process. Assuming there are 1000 historical variation ranges, we can label them as P(1,1) to P(15,1000), where 1≤n≤15, 1≤m≤1000, and the smaller m is, the earlier the historical variation range was recorded. P(n,m) where m is the same but n is different represents different historical variation ranges recorded at the same time. The height corresponding to the intersection of any two historical variation ranges is named the stratification height. In addition, there are two other stratification heights located at... Figure 5 The top and bottom of the middle are actually in Figure 5 The height of each horizontal dashed line corresponding to any structural outline is a layer height. Figure 5 Taking the lowest atmospheric fine structure layer S1 as an example, the monotonicity of S1 is monotonically decreasing. Therefore, the minimum value of the target analysis data is located at the rightmost side of S1. By extracting the value of the temperature normalization profile at the layer height on the Y-axis, we can obtain T(n,m). Each value of m corresponds to a different layer height and a different temperature. T(n,m) corresponds to P(n,m). For example, the range of P(1,1) is [0.6,3.8]. At this time, 0.6 and 3.8 are both layer heights, while T(1,1) is 0.6.

[0092] The difference between the target's historical multi-source observation data corresponding to the in-layer profile point and the minimum value of historical multi-source observation data in the stratified height of the historical variation range is obtained as the profile point difference; where the stratified height is the altitude corresponding to the intersection of any two historical variation ranges; the difference between the altitude corresponding to the target's historical multi-source observation data corresponding to the in-layer profile point and the minimum value of the stratified height of the historical variation range is obtained as the height difference; specifically, the structural profile within the range of P(n,m) is marked as L(n,m), the points on L(n,m) are named in-layer profile points, and the target's historical multi-source observation data value and altitude corresponding to the in-layer profile point are marked as R1 and R2 respectively. R1-T(n,m) is calculated and named as the profile point difference. The minimum value of the stratified height of P(n,m) is obtained and marked as E. R2-E is calculated, and the calculation result is named as the height difference.

[0093] Based on the height difference and profile point difference, a data prediction coordinate system is constructed, with the height difference as the X-axis and the profile point difference as the Y-axis. The profile point difference is entered into the data prediction coordinate system according to the height difference. A function regression is then performed on the data prediction coordinate system to obtain the data prediction function. Based on the data prediction function, the variation pattern of the atmospheric fine structure layer is obtained. Each historical multi-source observation data corresponds to a data prediction function in each Sn, and the variation pattern is formed by the combined results of all the data prediction functions.

[0094] In specific implementation, for example, if the coordinates of a certain inner contour point in the multi-source contour merging system are (0.5, 3), then 0.5 is R1 and 3 is R2. Calculating the contour point difference and height difference is to convert the R1 and R2 of the inner contour point into the difference between the X-axis and Y-axis of the corresponding layer height. In subsequent analysis, R1 and R2 can change with the change of layer height. For example, in this embodiment, the range of S1 extracted is [0.6, 3.6], while the R2 of the inner contour point (0.5, 3) is the height, which is between [0.6, 3.6]. Therefore, T(n, m) corresponding to the inner contour point at this time is the height. The normalized temperature value at 0.6 is 0.58, and the calculated profile difference is -0.08. Simultaneously, the calculated height difference is 2.4. A data prediction coordinate system is constructed, and then the data prediction function is obtained through function regression: Y1 = -0.04 × X1 + 0.024, where Y1 is the profile difference and X1 is the height difference. This data prediction function is only the data prediction function corresponding to the temperature in S1, revealing only the trend of the normalized temperature value in S1 with height. Each normalized multi-source observation data in each Sn has a data prediction coordinate system, ultimately forming the variation law.

[0095] Step 5: Arrange the fine vertical structure of the atmosphere in different spatiotemporal groups according to the time series, predict the changes in the fine vertical structure of the atmosphere, and obtain the prediction results.

[0096] A further implementation method involves predicting changes in the fine vertical structure of the atmosphere based on the aforementioned change patterns and according to a time series, including:

[0097] Based on the minimum value of the stratified height corresponding to the historical regularity range of the preset number and the basic data value, a turning point prediction coordinate system is constructed; where the basic data value is the historical multi-source observation data of the target corresponding to the lowest elevation point (ground height); the turning point prediction coordinate system is subjected to function regression to obtain the turning point prediction function.

[0098] Specifically, the E corresponding to Sn is labeled as Wn, and the lowest point of altitude is named ground height; a two-dimensional coordinate system is established with the basic data value as the horizontal axis and Wn as the vertical axis, named the turning point prediction coordinate system; Wn with the same n is entered into the turning point prediction coordinate system according to the basic data value; function regression is performed on the turning point prediction coordinate system, and the function obtained from the regression is named the turning point prediction function; there is a turning point prediction function for each value of n.

[0099] In specific implementation, Wn is the minimum value within the range of Sn, such as 0.6 in S1[0.6,3.6]. The altitude corresponding to the lowest point is the ground height. The purpose is to predict the vertical change of target analysis data at a certain time in the future based on the target analysis data on the ground. In this embodiment, the historical multi-source observation data corresponding to the temperature profile is still used as the target analysis data as an example. A turning point prediction coordinate system is constructed, and then the turning point prediction function is obtained by function regression as Y2=8.061×X22-9.8671×X2+3.5183, where Y2 is Wn and X2 is the basic data value. Since this embodiment uses the historical multi-source observation data corresponding to the temperature profile as the target analysis data as an example, the basic data value at this time is the temperature normalization value. The turning point prediction function here only reveals the trend of Wn in S1 with the temperature normalization value. There is a turning point prediction function for each normalized multi-source observation data in each Sn. This embodiment will not list them in detail.

[0100] A basic prediction coordinate system is constructed based on time series data and historical multi-source observation data of the target (i.e., any historical multi-source observation data to be analyzed). Function regression is performed on the basic prediction coordinate system to obtain the basic prediction function. Specifically, a two-dimensional coordinate system is established with time series data as the horizontal axis and historical multi-source observation data of the target as the vertical axis, named the basic prediction coordinate system. The values ​​of historical multi-source observation data of the target within the first period are entered into the basic prediction coordinate system according to the chronological order of the time series. Function regression is performed on the basic prediction coordinate system, and the resulting function is named the basic prediction function. Each historical multi-source observation data corresponds to one basic prediction function.

[0101] By substituting the preset prediction time into the basic prediction function, the predicted values ​​of the basic data for different historical multi-source observation data are obtained, thus obtaining the basic prediction data; by substituting the basic prediction data into the turning point prediction function, the prediction profile data is obtained, thus completing the prediction of changes in the fine vertical structure of the atmosphere.

[0102] Specifically, the time to be predicted is named the prediction time. The prediction time is substituted into the basic prediction function to obtain the predicted values ​​of the basic data from different historical multi-source observation data, which are then used as the basic prediction data. The basic prediction data is substituted into the corresponding transition prediction function to obtain different values ​​of Wn. The range between Wn and Wn+1 is the predicted Sn, denoted as Un. The data prediction function corresponding to Un is denoted as Kn. The temperature, wind speed, and humidity values ​​are denoted as Fh in ascending order of h, where h is a positive integer and h is the index of F. The Kn of Fh is denoted as G(h,n).

[0103] like Figure 6 As shown, the curve corresponding to G(h,n) is labeled as L(h,n), and L(h,n) is placed into Un. The lowest value of L(h,n) in Un overlaps with Wn, and finally, different multi-source observation data are predicted as prediction profile data.

[0104] In practical implementation, the basic prediction coordinate system actually analyzes the changes in normalized temperature and humidity values ​​over time to predict these values. Furthermore, it can directly obtain various data from weather forecasts and convert them into normalized data. For example, in this embodiment, the prediction time is 1 hour, meaning the basic prediction data is predicted 1 hour later. The weather forecast predicts a temperature of 36℃ 1 hour later. Historical experience shows that the lowest point at this location is usually 2℃ higher than the forecast temperature, i.e., 38℃, with a corresponding normalized value of 0.77. Therefore, the basic prediction data is 0.77. Substituting X2=0.77 into Y2=8.061×X2²-9.8671×X2+3.5183, we obtain W1 for S1 as 0.7km. Then, based on the turning prediction function of S2, we obtain W2 as 4km. Thus, the predicted range for S1 1 hour later is [0.7, 4]. U1 is [0.7,4]. The data prediction function for the normalized temperature value corresponding to U1 is the same as the data prediction function for the normalized temperature value corresponding to S1. That is, K1 is Y1=-0.04×X1+0.024, and the temperature value is F1. Therefore, G(1,1) is marked as Y1=-0.04×X1+0.024, representing the trend of the normalized temperature value within the fine structure layer of S1. The curve corresponding to G(1,1) is marked as L(1,1). Since U1 is [0.7,4], L(1,1) only extracts the curve portion within the range of [0.7,4]. If it is insufficient, it is extended from the highest point according to the G(1,1) function. The lowest value of L(h,n) in Un overlaps with Wn. This is to match the variation pattern of the data in Sn. Because the layer height has changed, L(h,n) needs to be moved to the corresponding layer height. For example, the predicted profile data corresponding to the temperature in S1 is as follows: Figure 6 As shown, Figure 6 The solid line in the figure represents the predicted profile data corresponding to the temperature. Similarly, by analyzing the predicted profile data of all Fh in different Sn, the distribution of the fine structure layer of the atmosphere one hour later can be obtained.

[0105] Example 2:

[0106] This invention also provides a system for extracting fine vertical atmospheric structure under complex mountainous terrain, used to implement the method of Embodiment 1, comprising:

[0107] The data processing module is used to collect multi-source observation data of atmospheric vertical profiles under complex mountainous terrain, and to normalize the multi-source observation data to obtain normalized multi-source observation data; among which, the multi-source observation data includes temperature profiles, wind speed profiles and humidity profiles.

[0108] The spatiotemporal grouping module is used to perform spatiotemporal matching of normalized multi-source observation data using the altitude of the data collection points to obtain spatiotemporal groups.

[0109] The system construction module is used to analyze normalized multi-source observation data in spatiotemporal grouping and construct a multi-source profile merging system.

[0110] The structure extraction module is used to extract the fine vertical structure of the atmosphere based on the multi-source profile merging system and normalized multi-source observation data.

[0111] The structural change prediction module is used to arrange the fine vertical structure of the atmosphere in different spatiotemporal groups according to the time series, predict the changes in the fine vertical structure of the atmosphere, and obtain the prediction results.

[0112] A further embodiment of the implementation includes a data processing module comprising:

[0113] The profile point construction unit is used to name the points on the temperature profile, wind speed profile, and humidity profile as temperature profile points, wind speed profile points, and humidity profile points, respectively.

[0114] The temperature normalization unit performs normalization calculations on the temperature values ​​of all temperature profile points, converting the temperature values ​​into normalized temperature values; the wind speed normalization unit performs normalization calculations on the wind speed values ​​of all wind speed profile points, converting the wind speed values ​​into normalized wind speed values; the humidity normalization unit performs normalization calculations on the humidity values ​​of all humidity profile points, converting the humidity values ​​into normalized humidity values; and the normalized data acquisition unit obtains normalized multi-source observation data based on the normalized temperature values, normalized wind speed values, and normalized humidity values.

[0115] Example 3:

[0116] This application provides an electronic device that may include a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The memory stores computer-readable instructions, and the processor can call these instructions. When the processor executes a computer-readable instruction, it performs steps as described in a method for extracting fine vertical atmospheric structure under complex mountainous terrain, to achieve the following functions: acquiring multi-source observation data of atmospheric vertical profiles under complex mountainous terrain; converting the multi-source observation data into a unified data range; dividing the multi-source observation data into different spatiotemporal groups; constructing a multi-source profile merging system, inputting the multi-source observation data into the multi-source profile merging system, and extracting the fine vertical atmospheric structure; and based on the fine vertical atmospheric structure of different spatiotemporal groups, arranging them chronologically and predicting changes in the fine vertical atmospheric structure based on the chronological order.

[0117] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0118] Example 4:

[0119] This application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can execute the above-described method for extracting the fine vertical structure of the atmosphere under complex mountainous terrain. The method includes: acquiring multi-source observation data of atmospheric vertical profiles under complex mountainous terrain; converting the multi-source observation data into a unified data range; dividing the multi-source observation data into different spatiotemporal groups; constructing a multi-source profile merging system; inputting the multi-source observation data into the multi-source profile merging system and extracting the fine vertical structure of the atmosphere; and based on the fine vertical structure of the atmosphere in different spatiotemporal groups, arranging them in time sequence and predicting the changes in the fine vertical structure of the atmosphere based on the time sequence.

[0120] Example 5:

[0121] This application also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it performs the steps of the above-described method for extracting fine vertical structure of the atmosphere under complex mountainous terrain to achieve the following functions: acquiring multi-source observation data of atmospheric vertical profiles under complex mountainous terrain; converting the multi-source observation data into a unified data range; dividing the multi-source observation data into different spatiotemporal groups; constructing a multi-source profile merging system, inputting the multi-source observation data into the multi-source profile merging system and extracting the fine vertical structure of the atmosphere; and based on the fine vertical structure of the atmosphere in different spatiotemporal groups, arranging it in chronological order and predicting the changes in the fine vertical structure of the atmosphere based on the chronological order.

[0122] Based on the above description of the embodiments, the embodiments of the present invention can be provided as methods, systems, or computer program products. Based on this understanding, the above technical solutions, in essence or in terms of their contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or certain parts of the embodiments.

[0123] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for extracting fine vertical atmospheric structure under complex mountainous terrain, characterized in that, include: Multi-source observation data of atmospheric vertical profiles under complex mountainous terrain are collected, and the multi-source observation data is normalized to obtain normalized multi-source observation data; wherein, the multi-source observation data includes temperature profiles, wind speed profiles, and humidity profiles; By utilizing the altitude of the data collection points, normalized multi-source observation data are spatiotemporally matched to obtain spatiotemporal groupings; Normalized multi-source observation data in spatiotemporal groupings are analyzed to construct a multi-source profile merging system; Based on the multi-source profile merging system and the normalized multi-source observation data, the fine vertical structure of the atmosphere is extracted. By arranging the fine vertical structure of the atmosphere in different spatiotemporal groups according to the time series, the changes in the fine vertical structure of the atmosphere are predicted, and the prediction results are obtained. Methods for extracting the fine vertical structure of the atmosphere include: A two-dimensional coordinate system was established with the normalized multi-source observation data as the X-axis and the altitude of the data collection points as the Y-axis to obtain the multi-source profile merging system. All normalized multi-source observation data within the same spatiotemporal group are entered into the multi-source profile merging system according to the altitude of the data collection points to obtain the structural profile; wherein, the structural profile includes a temperature normalized profile, a wind speed normalized profile, and a humidity normalized profile. Based on preset monotonic low values ​​and preset monotonic high values, the monotonic range of the structural profile is extracted, and the common intersection of the monotonic ranges of all structural profiles is obtained. Based on the common intersection, the range of regular changes is obtained, and the fine vertical structure of the atmosphere is extracted; wherein, the altitude interval corresponding to a range of regular changes is a fine atmospheric structure layer.

2. The method according to claim 1, characterized in that, Methods for obtaining normalized multi-source observation data include: The points on the temperature profile, the wind speed profile, and the humidity profile are respectively named temperature profile point, wind speed profile point, and humidity profile point. Perform normalization calculations on the temperature values ​​of all temperature profile points to convert the temperature values ​​into normalized temperature values; Perform normalization calculations on the wind speed values ​​of all wind speed profile points to convert the wind speed values ​​into normalized wind speed values; Perform normalization calculations on the humidity values ​​of all humidity profile points to convert the humidity values ​​into normalized humidity values; Based on the normalized temperature value, the normalized wind speed value, and the normalized humidity value, normalized multi-source observation data are obtained.

3. The method according to claim 2, characterized in that, Methods for constructing multi-source profile merging systems include: By integrating the normalized values ​​of temperature, wind speed, and humidity at a preset altitude, atmospheric element detection information at the preset altitude is established. Based on the atmospheric element detection information, a multi-source profile merging system is constructed.

4. The method according to claim 1, characterized in that, Methods for predicting changes in the fine vertical structure of the atmosphere include: Obtain the historical variation range of multi-source historical observation data, and analyze the variation patterns of different atmospheric fine structure layers based on the historical variation range; Based on the aforementioned patterns of change, the changes in the fine vertical structure of the atmosphere are predicted according to a time series.

5. The method according to claim 4, characterized in that, Methods for obtaining the variation patterns of the atmospheric fine structure layer include: The atmospheric fine structure layers are numbered in order from bottom to top, and the points on the structural profile corresponding to the historical regularity range of the preset number are taken as the inner profile points of the layer. The difference between the target's historical multi-source observation data corresponding to the inner contour point and the minimum value of historical multi-source observation data in the layer height of the historical change range is obtained; wherein, the layer height is the altitude corresponding to the intersection of any two historical change ranges; The difference between the altitude corresponding to the target's historical multi-source observation data at the inner contour point of the calculation layer and the minimum value of the layer height within the historical variation range is obtained; Based on the height difference and the profile point difference, a data prediction coordinate system is constructed, and a function regression is performed on the data prediction coordinate system to obtain the data prediction function; Based on the data prediction function, the variation pattern of the atmospheric fine structure layer is obtained.

6. The method according to claim 5, characterized in that, Based on the aforementioned change patterns, methods for predicting changes in the fine vertical structure of the atmosphere using time series include: A turning point prediction coordinate system is constructed based on the minimum value of the stratified height within the historical variation range corresponding to a preset number, and the basic data values; wherein, the basic data values ​​are the historical multi-source observation data of the target corresponding to the lowest elevation point. Perform function regression on the aforementioned turning point prediction coordinate system to obtain the turning point prediction function; A basic prediction coordinate system is constructed based on time series and historical multi-source observation data of the target. Perform function regression on the basic prediction coordinate system to obtain the basic prediction function; Substitute the preset prediction time into the basic prediction function to solve for the predicted values ​​of the basic data values ​​of different historical multi-source observation data, and obtain the basic prediction data. Substituting the basic prediction data into the turning point prediction function, prediction profile data is obtained, thus completing the prediction of the changes in the fine vertical structure of the atmosphere.

7. A system for extracting fine vertical atmospheric structure under complex mountainous terrain, used to implement the method described in any one of claims 1-6, characterized in that, include: The data processing module is used to collect multi-source observation data of atmospheric vertical profiles under complex mountainous terrain, and to normalize the multi-source observation data to obtain normalized multi-source observation data; wherein, the multi-source observation data includes temperature profiles, wind speed profiles and humidity profiles. The spatiotemporal grouping module is used to perform spatiotemporal matching of normalized multi-source observation data using the altitude of the data collection points to obtain spatiotemporal groups; The system construction module is used to analyze normalized multi-source observation data in spatiotemporal grouping and construct a multi-source profile merging system; The structure extraction module is used to extract the fine vertical structure of the atmosphere based on the multi-source profile merging system and the normalized multi-source observation data. The structural change prediction module is used to arrange the fine vertical structure of the atmosphere in different spatiotemporal groups according to the time series, predict the changes in the fine vertical structure of the atmosphere, and obtain the prediction results.

8. The system according to claim 7, characterized in that, The data processing module includes: A profile point construction unit is used to name the points on the temperature profile, the wind speed profile, and the humidity profile as temperature profile points, wind speed profile points, and humidity profile points, respectively. The temperature normalization unit is used to perform normalization calculations on the temperature values ​​of all temperature profile points and convert the temperature values ​​into normalized temperature values. The wind speed normalization unit is used to perform normalization calculations on the wind speed values ​​of all wind speed profile points and convert the wind speed values ​​into normalized wind speed values. The humidity normalization unit is used to perform normalization calculations on the humidity values ​​of all humidity profile points and convert the humidity values ​​into normalized humidity values. The normalized data acquisition unit is used to obtain normalized multi-source observation data based on the normalized temperature value, the normalized wind speed value, and the normalized humidity value.

Images