CN111929687B - Automatic recognition algorithm for characteristics of tornado vortex - Google Patents
Automatic recognition algorithm for characteristics of tornado vortex Download PDFInfo
- Publication number
- CN111929687B CN111929687B CN202010861316.5A CN202010861316A CN111929687B CN 111929687 B CN111929687 B CN 111929687B CN 202010861316 A CN202010861316 A CN 202010861316A CN 111929687 B CN111929687 B CN 111929687B
- Authority
- CN
- China
- Prior art keywords
- shear
- dimensional
- azimuth
- azimuthal
- feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 claims abstract description 16
- 239000000178 monomer Substances 0.000 claims abstract description 16
- 238000002592 echocardiography Methods 0.000 claims abstract description 6
- 230000000875 corresponding effect Effects 0.000 claims description 16
- 230000000717 retained effect Effects 0.000 claims description 10
- 230000002596 correlated effect Effects 0.000 claims description 6
- 230000001419 dependent effect Effects 0.000 claims description 3
- 238000002310 reflectometry Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 6
- 230000010349 pulsation Effects 0.000 description 4
- 238000007619 statistical method Methods 0.000 description 3
- 230000000903 blocking effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000007598 dipping method Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- JTJMJGYZQZDUJJ-UHFFFAOYSA-N phencyclidine Chemical class C1CCCCN1C1(C=2C=CC=CC=2)CCCCC1 JTJMJGYZQZDUJJ-UHFFFAOYSA-N 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- CLOMYZFHNHFSIQ-UHFFFAOYSA-N clonixin Chemical compound CC1=C(Cl)C=CC=C1NC1=NC=CC=C1C(O)=O CLOMYZFHNHFSIQ-UHFFFAOYSA-N 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000005923 long-lasting effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/95—Radar or analogous systems specially adapted for specific applications for meteorological use
- G01S13/958—Theoretical aspects
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses an automatic identification algorithm for characteristics of a tornado vortex, which comprises the following steps: step 1: identifying and suppressing non-meteorological echoes in radar base data, and performing deblurring treatment on radial speed; step 2: using a storm identification tracking algorithm to identify a storm monomer and recording the centroid position of the storm monomer; step 3: estimating azimuth vorticity shear by using a two-dimensional local linear least square difference method; step 4: identifying a one-dimensional azimuthal shear segment using multiple thresholds; step 5: identifying a two-dimensional azimuth shear feature based on a spatial proximity principle; step 6: identifying a three-dimensional azimuthal shear feature by vertical correlation; step 7: associating a three-dimensional azimuth shear feature and a storm monomer; step 8: TVS are identified based on the three-dimensional azimuthal shear characteristic of the storm. The invention can reduce the false alarm rate of TVS identification and improve the hit rate of TVS identification.
Description
Technical Field
The invention relates to the technical field of weather radar systems, in particular to an automatic recognition algorithm for characteristics of a tornado vortex.
Background
One essential feature of supermonomers is the long-lasting deep mesocyclone, which is a small-scale vortex (3-10 km in diameter) closely associated with the updraft and backside downdraft of a strong convective storm, which appears as azimuthally symmetric speed vortices on the radar radial velocity map. Bluish gold vortices are often used to simulate a medium cyclone, i.e. the rotational speed increases linearly with distance in the radius of the medium cyclone core, and decreases inversely with distance outside the radius of the core, so that on the same distance circle of the medium cyclone core, there is a tendency for the radial speed of the radar to increase gradually in a clockwise direction, i.e. from a maximum negative speed to zero and then from zero to a maximum positive speed. Over 90% of the medium cyclones are accompanied by strong hail and strengthIn the case of disastrous weather such as wind and tornadoes, only about 25% of medium cyclones develop into tornadoes. In addition to mid cyclone, a much smaller scale (about 1 km) and faster rotation (0.05 s in velocity azimuth shear) can be identified on the radar radial velocity map closely associated with the tornado -1 ) Is referred to as a tornado swirl feature (TVS: tornado Vortex Signature).
The TVS recognition algorithm used in the new generation weather radar system in China comes from the TVS recognition algorithm in the United states, and little document exists in the automatic recognition algorithm of the tornado vortex characteristics which is independently researched and developed in China.
TVS identification algorithm (hereinafter referred to as 88D TVS) followed by WSR-88D (Weather Surveillance Radar-1988 Doppler) used in the United states business (Crum et al, 1993) and the Tornado detection algorithm (Tornado Detection Algorithm) of the United states Strong storm laboratory (NSSL: the National Severe Storms Laboratory) (hereinafter referred to as NSSL TDA) (Mitchell et al, 1998). Both algorithms are designed based on searching for velocity differences in neighboring orientations equidistant from the radar, and filtering out the radial velocities corresponding to low reflectivity factors.
The 88D TVS algorithm is completed on the basis of a mesocyclone identification algorithm, and the algorithm comprises the following steps:
a. one-dimensional azimuth shear segment: firstly, filtering radial velocity data with the reflectivity factor smaller than a set threshold ZT (15 dBZ), then searching the radial velocity data subjected to deblurring, and searching a distance library with adjacent azimuth angles with velocity values continuously increasing clockwise at the same distance from the radar until the velocity values are not increased any more, wherein a one-dimensional distance library sequence is called a one-dimensional azimuth shear section. The retained one-dimensional azimuthal shear section needs to pass through a low Doppler angular momentum threshold TLM (180 km 2 ·h -1 ) Low shear threshold TLS (7.2 h -1 ) High Doppler angular momentum threshold THM (540 km) 2 ·h -1 ) Or a high shear threshold TLS (14.4 h -1 ) And (5) checking.
b. Two-dimensional characteristics: the azimuth distance and radial distance between one-dimensional azimuth shear segment and all one-dimensional azimuth shear segments that have been categorized as two-dimensional features are calculated and classified as one-dimensional feature when preset azimuth and radial distance thresholds L A (2.2 deg.) and LR (1 km) are met. A two-dimensional feature comprises a one-dimensional azimuthal shear segment that satisfies a predetermined threshold TPV (10) having a ratio of radial dimension to azimuthal dimension that is within a certain distance-dependent threshold (symmetric two-dimensional feature).
c. Vertical correlation and classification: two-dimensional features at different elevation angles are vertically associated, and only two-dimensional features with center heights below a preset threshold TFM (8 km) can be vertically associated. Regarding the two-dimensional feature as circular, taking the one with the larger azimuth and radial dimensions as its diameter, if a small two-dimensional feature falls vertically within a larger two-dimensional feature area, then the 2 two-dimensional features are considered to be vertically related. All vertical correlations from low elevation angles to high elevation angles are completed to form three-dimensional features, and one three-dimensional feature comprises at least 2 symmetrical two-dimensional features.
The three-dimensional feature that is preserved after the 3 steps is mesocyclone.
Tvs: TVS identification is done on the basis of mesocyclone identification. The TVS algorithm calculates the differential shear by searching for the maximum and minimum velocities of each two-dimensional feature in the cyclone, if this differential shear is greater than 20 mS -1 ·km -1 The vortex associated with the above-described shear is identified as a potential TVS. If there are 2 or more elevation angles within one middle cyclone, then that middle cyclone is identified as a TVS.
2) NSSL TDA algorithm
NSSL TDA is an improvement over the 88D TVS algorithm. The 88D TVS algorithm is based on identifying the mesocyclone and then determining whether a TVS is present in the mesocyclone, and NSSL TDA is independent of the mesocyclone algorithm without first identifying the mesocyclone as a precondition. Compared with the 88D TVS algorithm, NSSL TDA recognizes the main characteristics of vortex: a) Searching radial velocity shear between two distance libraries of adjacent azimuth angles at the same distance from the radar; b) The algorithm is not required to first identify the mesocyclone. The algorithm process is as follows:
a. one-dimensional azimuth shear segment: radial velocity data with reflectivity factors smaller than a set threshold ZT (0 dBZ) are filtered out first, and radial velocity data with distance folding and data missing are also not considered. The radial velocity difference between two range bins of adjacent azimuth angles at the same distance from the radar is then found for each elevation scan, the finding process requiring a radius within 150km and a height below 10 km. If the velocity difference is greater than a preset adjustable threshold (e.g., 11 m/s), the velocity pair is stored as a shear segment and some attribute parameters are calculated. This process is repeated until all velocity data within one radar elevation scan is processed and all shear segments exceeding the minimum velocity difference threshold are found.
b. Two-dimensional characteristics: two-dimensional features were constructed using 6 threshold (35, 30, 25, 20, 15, 11 m/s) cycles from high to low. The two-dimensional features for each threshold level are made up of at least 3 shear segments, each shear segment centroid being less than 1 ° from its neighboring shear segment centroid, and the radial distance being less than 500m. The aspect ratio (radial dimension/azimuthal dimension) of all two-dimensional features is calculated, leaving only two-dimensional features with a ratio less than 4. The above process is repeated using a low threshold, which is dropped if there is an overlap of the two-dimensional features of the different thresholds.
c. Three-dimensional characteristics: performing vertical correlation continuity check in the whole body scanning range, wherein one three-dimensional characteristic is at least composed of 2 two-dimensional characteristics, and the three-dimensional characteristics are at most separated by an elevation angle; the horizontal distance between two-dimensional features constituting the three-dimensional features is less than 2.5km. All three-dimensional features consisting of 3 and more two-dimensional features are referred to as three-dimensional vortices. Each three-dimensional vortex is divided into 2 types: TVS and overhead TVS (ETVS). If a three-dimensional vortex satisfies: (1) The minimum strength and thickness criteria, (2) the base expansion to 0.5 degree elevation or a specified height (e.g., 600 m), then the three-dimensional vortex is referred to as a TVS. And is referred to as ETVS if only condition (1) is satisfied.
Defects and deficiencies of the prior art:
1) The 88D TVS algorithm has a false alarm rate FAR for the tornado approaching zero, but the hit rate POD is also very low. The reason is that when the maximum and minimum speed differences in the middle cyclone are larger, the azimuth interval of the middle cyclone can be larger, so that the corresponding azimuth shear is smaller, and the shear standard of the TVS is not met. In fact, the smaller scale vortices contained in the mid cyclone may have a slightly smaller maximum-minimum speed difference than the mid cyclone, but may have a corresponding azimuth pitch that is much smaller than the mid cyclone, so the calculated azimuth cut can meet the TVS criteria, the 88D TVS algorithm cannot identify this type of TVS, and the hit rate POD is low.
2) The 88D TVS algorithm uses a single differential azimuth shear threshold value, if the threshold value is too high, the identified two-dimensional characteristic occupies too small area of a real middle cyclone nuclear area, and if the threshold value is too low, weak shear areas outside the middle cyclone nuclear area are added, so that estimation of azimuth shear is affected, and accordingly the identification result of the TVS is affected.
3) The NSSL TDA algorithm has a higher hit rate POD, but the false alarm rate FAR is also higher. The reason is that the algorithm only uses adjacent azimuth speed difference and differential azimuth shear to construct one-dimensional features and two-dimensional features, and small-scale natural pulsation in the algorithm is common to a radial speed field, and TVS identification is easily affected by the small-scale natural pulsation of the speed, so that the false alarm rate is high. In addition, the differential azimuth shear criterion is also influenced by radial speed quality problems caused by speed deblurring errors or failures, noise, ground clutter, beam partial blocking and the like, so that the false alarm rate is high.
4) Both the 88D TVS algorithm and NSSL TDA algorithm filter out radial velocity data using a low reflectivity factor threshold, while the mid cyclone is a small scale vortex closely associated with the updraft and backside downdraft of a strong convective storm with little precipitation in the region of the strong updraft, the corresponding reflectivity factor may be below the threshold given in the algorithm, resulting in some large radial velocity values being discarded.
5) The NSSL TDA algorithm requires that the TVS contain at least 3 elevation two-dimensional azimuth shear features, which may render some low centroid tornado supermonomers unrecognizable.
6) The NSSL TDA algorithm calculates differential shear using only adjacent azimuth radial velocities, where the maximum minimum velocity may not be the adjacent azimuth at close range, but the differential shear is greater than the maximum adjacent azimuth differential shear, which may result in missed identification of TVS at close range.
Disclosure of Invention
The invention aims to provide an automatic recognition algorithm for the characteristics of a tornado vortex, which can solve the problems in the background technology.
In order to achieve the above purpose, the present invention provides the following technical solutions:
an automatic recognition algorithm for the characteristics of a tornado vortex comprises the following steps:
step 1: firstly, recognizing and suppressing non-meteorological echoes such as ground clutter, super-refraction echoes and noise in radar base data by adopting the prior art, and performing deblurring treatment on radial speed;
step 2: using a storm identification tracking algorithm to identify a storm monomer and recording the centroid position of the storm monomer;
step 3: estimating azimuth vorticity shear using a two-dimensional local linear least squares difference (Linear Least Squares Derivative, LLSD) method;
step 4: one-dimensional azimuthal shear segment identification
The one-dimensional azimuthal shear segment is identified by searching in the radial direction for an azimuthal vorticity shear greater than a predetermined threshold (default class 6, 70, 60, 50, 40, 30, 20, respectively, in 10 units -4 s -1 ) Is a continuous library of (a) is provided. When an azimuth vorticity shear bank greater than a threshold is searched for the first time from the radar in the radial direction, then subsequent azimuth vorticity shear banks greater than the threshold are clustered together until an azimuth vorticity shear bank less than the threshold is searched; if the difference between the azimuth vorticity shear value of the subsequent bin and the threshold value is less than or equal to the azimuth vorticity shear variation threshold value (default value of 4, unit of 10 -4 s -1 ) And the number of consecutive distance bins meeting the condition is less than the break count threshold (default value of 2), then the search for the shear segment continues, otherwise the shear segment expires before the first bin less than the threshold. If the length of the shear segment is greater than the segment length threshold (default value of 0.95 km), the segment is saved, otherwise it is rejected. A distance library recording the beginning and end of each shear segment azimuth of,Maximum and minimum speeds, etc.
Step 5: identifying two-dimensional azimuthal shear features based on spatial proximity principles
When the last radial one-dimensional azimuth shear segment of an elevation angle is searched, the one-dimensional azimuth shear segments are combined into a two-dimensional azimuth shear feature based on the principle of space proximity. The merging of adjacent segments meets 2 criteria: firstly, the azimuth difference of 2 segments is smaller than the azimuth interval threshold (default value is 1.5 degrees), and secondly, the overlapping distance of 2 segments in the radial direction is larger than the segment overlapping threshold (default value is 0.45 km). A two-dimensional azimuthal shear feature contains a number of one-dimensional azimuthal shear segments that must be greater than a segment number threshold (default value of 2) and an area that is greater than a two-dimensional azimuthal shear feature area threshold (default value of 3 km) 2 ). For the searched two-dimensional azimuth shear feature, calculating and storing some feature quantities, including a central position (inclined distance, azimuth and elevation angle), an initial azimuth and inclined distance, a maximum azimuth vorticity shear, a maximum minimum speed, an azimuth and inclined distance where the maximum minimum speed is located, a corresponding differential azimuth shear, a maximum adjacent azimuth speed difference, a corresponding differential azimuth shear and the like.
After two-dimensional azimuthal shear feature identification of all level thresholds is completed, in order to extract the strongest azimuthal shear region information, the low-threshold two-dimensional azimuthal shear feature surrounding the high-threshold two-dimensional azimuthal shear feature needs to be discarded. If the center of the high threshold two-dimensional azimuth shear feature falls within the low threshold two-dimensional azimuth shear feature area, the low threshold two-dimensional azimuth shear feature is discarded, but parametric information recording the two-dimensional azimuth shear feature wrapping all of its level thresholds is added to the retained highest level threshold two-dimensional azimuth shear feature information.
Step 6: three-dimensional azimuthal shear feature identification by vertical correlation
The retained two-dimensional azimuthal shear features are ordered by maximum azimuthal vorticity shear magnitude and then vertically correlated. Each identified three-dimensional azimuth shear feature contains at least 2 two-dimensional azimuth shear features of successive elevation angles. Vertical correlation is an iterative process starting from the lowest elevation angle. First vertically correlating two-dimensional azimuth shear features having a center distance of adjacent elevation angles less than a threshold (default value of 2.5 km), if there are multiple two-dimensional azimuth shear features that can be correlated, then only the one two-dimensional azimuth shear feature with the greatest azimuth vorticity shear is selected for correlation. If there are two-dimensional azimuthal shear features that are not vertically associated after the first vertical association is completed, the search radius is increased to 5km and the above steps are repeated for all two-dimensional azimuthal shear features that are not vertically associated. If there is still two-dimensional azimuth shear feature not vertically associated after the second vertical association is finished, the searching radius is increased to 7.5km, and the above steps are repeated again to perform the 3 rd vertical association. This process forms a three-dimensional azimuth shear feature after all adjacent elevation angles are performed. For the retained three-dimensional azimuth shear feature, a plurality of parameters are calculated and stored, including centroid position (azimuth, chute and altitude), bottom height, top height, maximum azimuth vorticity shear value and its altitude, vertical integral azimuth vorticity, all two-dimensional azimuth shear feature parameter information and the like.
Step 7: correlating three-dimensional azimuth shear features and storm monomers
Storm-independent three-dimensional azimuth shear features are nulled using storm cell locations identified based on reflectivity factor data, i.e., a three-dimensional azimuth shear feature is ultimately preserved when a storm exists within a 30km radius of the center of the three-dimensional azimuth shear feature. The finally retained three-dimensional azimuthal shear feature is associated with the identified storm entity using a distance nearest principle.
Step 8: TVS identification
The TVS identification is completed based on the three-dimensional azimuth shear characteristic statistical analysis results of the identified three-dimensional azimuth shear characteristics and the three-dimensional azimuth shear characteristic statistical analysis results of the single-body tornado examples, and the identification algorithm is limited to a range of 150km with radar as a center and below 6 km. The number of azimuth sequence number intervals n of maximum minimum speed, i.e. n=1.25/(rΔθ), where r is the skew, Δθ is the radar beam width, n=1 when n <1, n=13 when n >13, is estimated using 1.25km as the default TVS size. The identification of TVS needs to meet the following 3 criteria:
(1) The rotational speed of the two-dimensional azimuthal shear feature must be greater than the rotational speed (half of the maximum minimum speed difference) criterion (rv=42r) based on the three-dimensional azimuthal shear feature statistics of the supermono-hull tornadoes -0.215 R is the slant distance, the unit km, RV is the rotation speed, the unit m.s -1 )
(2) And calculating the azimuth sequence number interval of the maximum and minimum speeds of the rotation speed to be less than or equal to n.
(3) The maximum azimuthal vorticity shear of the two-dimensional azimuthal shear feature must be greater than the azimuthal vorticity shear criterion (u a ＝1104.3r -0.577 R is the slant distance, unit km, u a For azimuthal vorticity shear, unit 10 -4 s -1 )。
The rotational speed of the three-dimensional azimuth shear feature and the azimuth sequence number interval number of the corresponding maximum and minimum speeds have a plurality of groups of values, namely the rotational speed calculated by the maximum and minimum speeds in the two-dimensional azimuth shear feature with different level shear thresholds and the azimuth sequence number interval number corresponding to the rotational speed calculated by the maximum adjacent azimuth speed difference in the two-dimensional azimuth shear feature with the lowest level shear threshold and the maximum and minimum speeds of the calculated rotational speed, as long as a group of values simultaneously meet the judgment standards (1) and (2).
When the two-dimensional azimuth shear characteristics satisfying the conditions (1) - (3) below the 6km height are at least 2, the potential TVS exists, and when the height of the two-dimensional azimuth shear characteristics of the bottommost layer of the potential TVS is smaller than 1km or the elevation angle of the two-dimensional azimuth shear characteristics is the lowest scanning elevation angle, the potential TVS is judged to be the TVS, otherwise, the potential TVS is judged to be the overhead TVS; when the two-dimensional azimuth shear characteristics satisfying the conditions (1) - (3) below the 6km height are only 1, the potential TVS is considered to exist when at least one two-dimensional azimuth shear characteristic of adjacent elevation angles reaches the mid-cyclone standard, and when the two-dimensional azimuth shear characteristics satisfying the conditions (1) - (3) are less than 1km or the elevation angle where the two-dimensional azimuth shear characteristics are located is the lowest scanning elevation angle, the potential TVS is judged to be the TVS, otherwise, the potential TVS is judged to be the overhead TVS.
Compared with the prior art, the invention has the beneficial effects that:
1) The quality control of the radar data before identification can reduce the influence of non-meteorological echoes, speed ambiguity and other problems on TVS identification. The radial velocity is filtered without using a low reflectivity factor threshold to avoid removing large radial velocity values in the weak echo or bounded weak echo region (strong updraft region).
2) Compared with finite difference azimuth shear, azimuth vorticity shear calculated by using a two-dimensional local LLSD method is more stable, the influence of speed quality (such as noise, partial wave beam blocking, ground clutter and the like) and small-scale natural pulsation of speed on one-dimensional azimuth shear segment identification can be reduced, the false alarm rate of TVS identification is reduced, and the hit rate of TVS identification is improved.
3) Core vortices (e.g., along a radial line) that may be embedded within a rectangular shear region may be separated using a multi-threshold one-dimensional azimuthal shear segment search method. When the one-dimensional azimuth shear segment is identified, the method allows the azimuth vorticity shear library with a limited number smaller than a set threshold value to be contained, can reduce the influence of small-scale natural pulsation of speed on the one-dimensional azimuth shear segment search, and improves the hit rate of TVS identification.
4) In two-dimensional azimuth shear feature identification, in order to accurately find the strongest azimuth vorticity shear region, the low-threshold two-dimensional azimuth shear feature wrapping the high-threshold two-dimensional azimuth shear feature is abandoned, but parameter information of all the abandoned low-threshold two-dimensional azimuth shear features is reserved. The parameter information is used in TVS identification to obtain a plurality of groups of rotation speeds and corresponding maximum and minimum speed azimuth sequence number intervals, and the hit rate of TVS identification can be improved as long as one group of the parameter information meets one of the criteria of the TVS even if the one group of parameter information meets the criteria of the TVS. Because the azimuth sequence number interval of the maximum and minimum speeds of the TVS can be changed along with the distance of the TVS from the radar, the rotation speeds of different TVSs can be different, and the TVSs with different intensities, sizes and distances from the radar can be dynamically checked by using a plurality of groups of rotation speeds and the azimuth sequence number interval of the maximum and minimum speeds, so that the situation of TVS missing identification caused by judging whether the azimuth shear meets the TVS standard by using the maximum and minimum speed difference of the cyclone in a single threshold value is improved. In addition, since the middle cyclone and the TVS are not all symmetrically distributed along the azimuth (for example, the lower layer is an radial cyclone), if only the speed difference between adjacent azimuth at the same distance is calculated, the maximum rotation speed and azimuth shear cannot be found, and the multiple sets of rotation speeds adopted by the algorithm can compensate for the defect.
5) The advantage of using three sets of search radii for vertical correlation in three-dimensional azimuth shear feature identification is that the more highly-dipping three-dimensional azimuth shear feature is fully identified as much as possible, because super-monomer storms tend to be generated in strong vertical wind shear environments, with more storm dipping. The identified storm monomers are used for carrying out the nulling treatment on the three-dimensional azimuth shear characteristics irrelevant to the storm, and the three-dimensional azimuth shear characteristics and TVS which are mistakenly identified due to the quality problem of radial speed data can be effectively removed.
6) In TVS identification, using 1.25km as the default tornado scroll feature size to estimate the maximum minimum speed azimuth sequence number interval threshold may solve the problem of the variation of the maximum minimum speed azimuth sequence number interval with distance in TVS, because not all TVS's maximum and minimum speeds are located in adjacent azimuth. The rotation speed standard and the azimuth vorticity shear standard of the TVS used in the algorithm are obtained according to statistics of batch tornadoes in China, and the 2 standards are power functions which change along with the distance, so that the hit rate of the TVS identification in China can be improved. The continuous 2 elevation angles are used for meeting the TVS judgment standard or one elevation angle is used for meeting the TVS standard, and the other adjacent elevation angle is used for meeting the medium cyclone standard, so that the hit rate of the low centroid TVS identification can be improved, the false alarm rate of the TVS is reduced by adopting various methods, and the false alarm rate of the TVS is not increased by reducing the vertical stretching thickness standard.
Drawings
FIG. 1 is a flow chart of an automatic identification algorithm for the characteristics of a tornado vortex in the invention;
FIG. 2 shows radial identifications 70, 60, 50, 40, 30 and 20×10 of the present invention -4 s -1 A threshold shear segment map;
FIG. 3 is a schematic view of 3 shear segments according to the present invention;
FIG. 4 is a two-dimensional azimuthal shear feature map of the present invention;
FIG. 5 is a view of 2 two-dimensional azimuthal shear characteristics of the present invention;
FIG. 6 is a vertically correlated two-dimensional azimuthal shear feature map of the present invention;
figure 7 is a scatter plot of cyclone rotational velocity (a) and azimuthal vorticity shear (b) versus distance from radar in a cyclone of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
FIG. 1 shows a flow chart of an automatic recognition algorithm for the characteristics of the tornado vortex, wherein the reflection factor clutter suppression, the radial velocity deblurring and the storm monomers have research bases, and the following modules are specifically described.
1) Directional vorticity shear estimation
Taking a two-dimensional local window with the size of (2N+1) x (2M+1) under the polar coordinate, setting the radial speed of each azimuth-inclined distance library in the window as u i,j (i= -N, … 0,1 … N; j= -M, … 0,1 … M). Taking the window center as a reference (i=0, j=0), taking i as positive along the radial center, taking j as positive along the tangential clockwise direction, and then the azimuthal vorticity shear u of the window center a The method comprises the following steps:
wherein,r is the slant distance (unit km) from the center of the window to the radar, and Deltar is the radial speed library length (unit km),/for the radar>Is the beam width (in radians). w (w) i,j For weight, w should be given to all i and j ij >0，w i,j ＝w -i,j ，w i,j ＝w i,-j The algorithm takes w i，j =1. If there is a cyclonic rotation in the velocity field, u a ＞0，u a The larger the positive azimuthal vorticity shear, the greater.
Since polar radial velocity data is used when using the LLSD method to calculate azimuth vorticity shear, the azimuth resolution varies with distance r from the radar, so does the size of the local window taken. Let the size of the cyclone core be a constant D (default 1.25 km), then the values of N and M are respectively:
wherein (int) denotes rounding. When the azimuth resolution of the radial velocity data is 1 DEG, if M <1, let M=1; if M >13, let m=13.
2) One-dimensional azimuthal shear segment identification
The one-dimensional azimuthal shear segment is identified by searching in the radial direction for an azimuthal vorticity shear greater than a predetermined threshold (default class 6, 70, 60, 50, 40, 30, 20, respectively, in 10 units -4 s - ) Is a continuous library of (a) is provided. As shown in fig. 2, when a bank of azimuth vorticity shears greater than a threshold is searched for the first time radially outward from the radar, then subsequent banks of azimuth vorticity shears greater than the threshold are clustered together until a bank of azimuth vorticity shears less than the threshold is searched; if the difference between the azimuth vorticity shear value of the subsequent bin and the threshold value is less than or equal to the azimuth vorticity shear variation threshold value (default value of 4, unit of 10 -4 s -1 ) And the number of consecutive distance bins meeting the condition is less than the break count threshold (default value of 2), then the search for the shear segment continues, otherwise the shear segment expires before the first bin less than the threshold. If the length of the shear segment is greater than the segment lengthThreshold (default 0.95 km), then the segment is saved, otherwise it is rejected. FIG. 3 shows 3 shear segments (the hatched area represents the saved portion, the brackets indicate the azimuthal vorticity shear value less than the corresponding threshold in 10 -4 s -1 。(a)40×10 -4 s -1 Shear section, (b) 30×10 -4 s -1 Shear section, (c) 20×10 -4 s -1 And (5) shearing the section. ) It can be seen that only those azimuth vorticity shear values above the threshold that meet both the azimuth vorticity variation threshold and the discontinuity count threshold can be included in the saved shear segments.
3) Two-dimensional azimuthal shear feature identification
When the last radial one-dimensional azimuth shear segment of an elevation angle is searched, the one-dimensional azimuth shear segments are combined into a two-dimensional azimuth shear feature based on the principle of space proximity. The merging of adjacent segments meets 2 criteria: firstly, the azimuth difference of 2 segments is smaller than the azimuth interval threshold (default value is 1.5 degrees), and secondly, the overlapping distance of 2 segments in the radial direction is larger than the segment overlapping threshold (default value is 0.45 km). A two-dimensional azimuthal shear feature contains a number of one-dimensional azimuthal shear segments that must be greater than a segment number threshold (default value of 2) and an area that is greater than a two-dimensional azimuthal shear feature area threshold (default value of 3km 2). FIG. 4 shows a threshold of 20X 10 -4 s -1 Comprising 6 one-dimensional azimuthal shear segments. For the searched two-dimensional azimuth shear feature, some feature quantities are calculated and stored.
After two-dimensional azimuthal shear feature identification of all level thresholds is completed, in order to extract the strongest azimuthal shear region information, the low-threshold two-dimensional azimuthal shear feature surrounding the high-threshold two-dimensional azimuthal shear feature needs to be discarded. If the center of the high threshold two-dimensional azimuth shear feature falls within the low threshold two-dimensional azimuth shear feature area, the low threshold two-dimensional azimuth shear feature is discarded, but parametric information recording the two-dimensional azimuth shear feature wrapping all of its level thresholds is added to the retained highest level threshold two-dimensional azimuth shear feature information. FIG. 5 shows 2 two-dimensional azimuthal shear characteristics with a threshold of 40×10 -4 s -1 Is embedded at a threshold of 30 x 10 -4 s -1 Is within the two-dimensional azimuthal shear feature.
4) Three-dimensional azimuthal shear feature identification
The retained two-dimensional azimuthal shear features are ordered by maximum azimuthal vorticity shear magnitude and then vertically correlated. Each identified three-dimensional azimuth shear feature contains at least 2 two-dimensional azimuth shear features of successive elevation angles. Vertical correlation is an iterative process starting from the lowest elevation angle. First vertically correlating two-dimensional azimuth shear features having a center distance of adjacent elevation angles less than a threshold (default value of 2.5 km), if there are multiple two-dimensional azimuth shear features that can be correlated, then only the one two-dimensional azimuth shear feature with the greatest azimuth vorticity shear is selected for correlation. If there are two-dimensional azimuthal shear features that are not vertically associated after the first vertical association is completed, the search radius is increased to 5km and the above steps are repeated for all two-dimensional azimuthal shear features that are not vertically associated. If there is still a two-dimensional azimuthal shear feature that has not been vertically associated after the second vertical association is completed, the search radius is increased to 7.5km and the above steps are repeated again for the 3 rd vertical association (see FIG. 6 for a schematic). This process forms a three-dimensional azimuth shear feature after all adjacent elevation angles are performed. For the retained three-dimensional azimuthal shear feature, a plurality of parameters are calculated and stored.
5) Correlating three-dimensional azimuth shear features and storm monomers
Storm-independent three-dimensional azimuth shear features are nulled using storm cell locations identified based on reflectivity factor data, i.e., a three-dimensional azimuth shear feature is ultimately preserved when a storm exists within a 30km radius of the center of the three-dimensional azimuth shear feature. The finally retained three-dimensional azimuthal shear feature is associated with the identified storm entity using a distance nearest principle.
6) Tornado vortex feature identification and typing
The TVS is identified based on the three-dimensional azimuth shear characteristics of the identified three-dimensional azimuth shear characteristics and the three-dimensional azimuth shear characteristic statistical analysis results of the single super-monomer tornado examples, and the identification algorithm is limited to a range of 150km centered on the radar, and the height is less than 6 km. The azimuth sequence number interval n of maximum minimum speed is estimated using 1.25km as the default tornado scroll feature size, i.e., n=1.25/(rΔθ), where r is the skew (calculated from 5 km), Δθ is the radar beam width, n=1 when n <1, and n=13 when n > 13. The identification of TVS requires that the following 3 criteria be met:
(1) The rotational speed of the two-dimensional azimuthal shear feature must be greater than the rotational speed (half of the maximum minimum speed difference) criterion (see fig. 7 a) based on the three-dimensional azimuthal shear feature statistics of the supermono-hull tornadoes (rv=42r) -0.215 R is the slant distance, the unit km, RV is the rotation speed, the unit m.s -1 )
(2) And calculating the radial sequence number interval number of the maximum and minimum speeds of the rotation speed to be less than or equal to n.
(3) The maximum azimuthal vorticity shear of the two-dimensional azimuthal shear feature must be greater than the azimuthal vorticity shear criterion (see FIG. 7 b) based on the three-dimensional azimuthal shear feature statistics of the supermono-hull tornado (u a ＝1104.3r -0.577 R is the slant distance, unit km, u a For azimuthal vorticity shear, unit 10 -4 s -1 )。
The rotational speed of the three-dimensional azimuth shear feature and the azimuth sequence number interval number of the corresponding maximum and minimum speeds have a plurality of groups of values, namely the rotational speed calculated by the maximum and minimum speeds in the two-dimensional azimuth shear feature with different level shear thresholds and the azimuth sequence number interval number corresponding to the rotational speed calculated by the maximum adjacent azimuth speed difference in the two-dimensional azimuth shear feature with the lowest level shear threshold and the maximum and minimum speeds of the calculated rotational speed, as long as a group of values simultaneously meet the judgment standards (1) and (2).
When the two-dimensional azimuth shear characteristics satisfying the conditions (1) - (3) below the 6km height are at least 2, the potential TVS exists, and when the height of the two-dimensional azimuth shear characteristics of the bottommost layer of the potential TVS is smaller than 1km or the elevation angle of the two-dimensional azimuth shear characteristics is the lowest scanning elevation angle, the potential TVS is judged to be the TVS, otherwise, the potential TVS is judged to be the overhead TVS; when the two-dimensional azimuth shear characteristics satisfying the conditions (1) - (3) below the 6km height are only 1, the potential TVS is considered to exist when at least one two-dimensional azimuth shear characteristic of adjacent elevation angles reaches the mid-cyclone standard, and when the two-dimensional azimuth shear characteristics satisfying the conditions (1) - (3) are less than 1km or the elevation angle where the two-dimensional azimuth shear characteristics are located is the lowest scanning elevation angle, the potential TVS is judged to be the TVS, otherwise, the potential TVS is judged to be the overhead TVS.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, should be covered by the protection scope of the present invention by making equivalents and modifications to the technical solution and the inventive concept thereof.
Claims (10)
1. An automatic recognition algorithm for the characteristics of a tornado vortex is characterized by comprising the following steps:
step 1: firstly, recognizing and suppressing non-meteorological echoes such as ground clutter, super-refraction echoes and noise in radar base data by adopting the prior art, and performing deblurring treatment on radial speed;
step 2: using a storm identification tracking algorithm to identify a storm monomer and recording the centroid position of the storm monomer;
step 3: estimating azimuth vorticity shear by using a two-dimensional local linear least square difference method;
step 4: identifying a one-dimensional azimuthal shear segment using multiple thresholds;
step 5: identifying a two-dimensional azimuth shear feature based on a spatial proximity principle;
step 6: identifying a three-dimensional azimuthal shear feature by vertical correlation;
step 7: associating a three-dimensional azimuth shear feature and a storm monomer;
step 8: based on the three-dimensional azimuth shear characteristics of storm, 3 standards such as azimuth sequence number interval of distance-dependent rotation speed and corresponding maximum and minimum speed, azimuth vorticity shear and the like are used for counting the number of super single tornadoes to identify the tornadoes characteristic TVS.
2. The automatic identification algorithm of the characteristics of the tornado scroll as claimed in claim 1, wherein the azimuthal vorticity shear is estimated using a two-dimensional local linear least squares difference method.
3. The automatic identification algorithm of tornado scroll features as claimed in claim 1 wherein the identification of one-dimensional azimuthal shear segments is searching a continuous pool of azimuthal vorticity shears in a radial direction greater than a preset threshold, wherein a maximum of 2 azimuthal vorticity shear pools within 4 less than the set threshold are allowed to be interspersed.
4. The automatic identification algorithm of the characteristics of the tornado scroll as claimed in claim 2, wherein the default of the preset threshold is classified into 6 levels of 70, 60, 50, 40, 30, 20, respectively, in 10 units -4 s -1 。
5. The automatic identification algorithm of the tornado scroll feature of claim 1 wherein the one-dimensional azimuth shear segments are merged into a two-dimensional azimuth shear feature based on the principle of spatial proximity after the last radial one-dimensional azimuth shear segment of an elevation angle is searched.
6. The automatic tornado scroll feature recognition algorithm of claim 1, wherein the retained two-dimensional azimuthal shear features are ordered by maximum azimuthal vorticity shear magnitude and then vertically correlated, each recognized three-dimensional azimuthal shear feature comprising at least 2 two-dimensional azimuthal shear features of consecutive elevation angles.
7. The automatic identification algorithm of the tornado vortex characteristics of claim 1 wherein a three-dimensional azimuthal shear characteristic and a storm monomer are associated.
8. The automatic tornado scroll feature recognition algorithm of claim 1, wherein the TVS recognition based on the three-dimensional azimuthal shear feature uses 3 criteria of rotational speed, azimuthal sequence number interval of maximum and minimum speeds corresponding to rotational speed, and azimuthal vorticity shear.
9. The automatic recognition algorithm of the tornado vortex characteristics according to claim 1, wherein the rotational speed of the three-dimensional azimuth shear characteristic and the azimuth sequence number of its corresponding maximum and minimum speeds are separated by a plurality of sets of calculated values, as long as one set of calculated values satisfies a criterion.
10. The automatic identification algorithm of the spiral scroll feature of claim 1, wherein the rotation speed and the azimuth vortex degree shear criterion used are distance dependent power functions based on three-dimensional azimuth shear feature statistics of super-monomer spiral, the azimuth sequence number interval criterion of the maximum and minimum speeds corresponding to the rotation speed is calculated using a geometric interval of 1.25km, and is equal to 13 when the azimuth sequence number interval is greater than 13, and is equal to 1 when the azimuth sequence number interval is less than 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010861316.5A CN111929687B (en) | 2020-08-25 | 2020-08-25 | Automatic recognition algorithm for characteristics of tornado vortex |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010861316.5A CN111929687B (en) | 2020-08-25 | 2020-08-25 | Automatic recognition algorithm for characteristics of tornado vortex |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111929687A CN111929687A (en) | 2020-11-13 |
| CN111929687B true CN111929687B (en) | 2023-11-21 |
Family
ID=73305089
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010861316.5A Active CN111929687B (en) | 2020-08-25 | 2020-08-25 | Automatic recognition algorithm for characteristics of tornado vortex |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111929687B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112748434B (en) * | 2021-02-07 | 2023-04-11 | 江西师范大学 | Doppler weather radar speed de-ambiguity method based on polynomial fitting |
| CN112698333B (en) * | 2021-03-24 | 2021-07-06 | 成都千嘉科技有限公司 | Ultrasonic flight time measuring method and system suitable for gas and liquid |
| CN115542326A (en) * | 2022-09-14 | 2022-12-30 | 江苏省气象台 | Isolated downburst identification method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5523759A (en) * | 1992-04-14 | 1996-06-04 | Honeywell Inc. | In flight doppler weather radar wind shear detection system |
| US7486220B1 (en) * | 2006-09-28 | 2009-02-03 | Rockwell Collins, Inc. | Storm top detection and prediction |
| CN106199606A (en) * | 2016-07-20 | 2016-12-07 | 国网河南省电力公司电力科学研究院 | A kind of multi thresholds squall line recognition methods based on radar return 3 d mosaics |
| CN106772387A (en) * | 2016-12-21 | 2017-05-31 | 中国航空工业集团公司雷华电子技术研究所 | A kind of wind shear recognition methods |
| CN106919792A (en) * | 2017-02-24 | 2017-07-04 | 天津大学 | Vortex center automatic identifying method based on high accuracy numerical value Wind Data |
| CN108680920A (en) * | 2018-04-28 | 2018-10-19 | 深圳市雅码科技有限公司 | A kind of hazard weather identification early warning system and method based on dual polarization radar |
-
2020
- 2020-08-25 CN CN202010861316.5A patent/CN111929687B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5523759A (en) * | 1992-04-14 | 1996-06-04 | Honeywell Inc. | In flight doppler weather radar wind shear detection system |
| US7486220B1 (en) * | 2006-09-28 | 2009-02-03 | Rockwell Collins, Inc. | Storm top detection and prediction |
| CN106199606A (en) * | 2016-07-20 | 2016-12-07 | 国网河南省电力公司电力科学研究院 | A kind of multi thresholds squall line recognition methods based on radar return 3 d mosaics |
| CN106772387A (en) * | 2016-12-21 | 2017-05-31 | 中国航空工业集团公司雷华电子技术研究所 | A kind of wind shear recognition methods |
| CN106919792A (en) * | 2017-02-24 | 2017-07-04 | 天津大学 | Vortex center automatic identifying method based on high accuracy numerical value Wind Data |
| CN108680920A (en) * | 2018-04-28 | 2018-10-19 | 深圳市雅码科技有限公司 | A kind of hazard weather identification early warning system and method based on dual polarization radar |
Non-Patent Citations (3)
| Title |
|---|
| A COMPARISON OF HIGH RESOLUTION TORNADO SURVEYS TO DOPPLER RADAR OBSERVED VORTEX PARAMETERS: 2011-2012 CASE STUDIES;James G. LaDue 等;《26th Conf. on Severe Local Storms》;1-13 * |
| 利用CINRAD雷达产品进行风暴外推与预警;马中元;《气象水文海洋仪器》;第27卷(第02期);47-52、56 * |
| 区域暴雨过程中两次龙卷风事件分析;张一平 等;《气象学报》;第70卷(第05期);961-973 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111929687A (en) | 2020-11-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111929687B (en) | Automatic recognition algorithm for characteristics of tornado vortex | |
| WO2021008202A1 (en) | Method for kernel support vector machine target classification based on millimeter-wave radar point cloud features | |
| Zrnić et al. | Automatic detection of mesocyclonic shear with Doppler radar | |
| Yi et al. | Enhancing the accuracy of automatic eddy detection and the capability of recognizing the multi-core structures from maps of sea level anomaly | |
| CN108181620B (en) | Three-coordinate radar trace point quality evaluation method | |
| CN110018461B (en) | Group target identification method based on high-resolution range profile and monopulse angle measurement | |
| CN110261857B (en) | Spatial interpolation method for weather radar | |
| CN102495403B (en) | Radar signal processing method | |
| WO2018000252A1 (en) | Oceanic background modelling and restraining method and system for high-resolution remote sensing oceanic image | |
| CN111931409B (en) | Vortex three-dimensional body state identification and tracking method based on high-resolution ocean mode data | |
| CN112180354A (en) | High-frequency radar target joint detection method by utilizing time-frequency analysis and constant false alarm technology | |
| CN111123205B (en) | Ground reconnaissance radar target point trace aggregation method | |
| CN113570005A (en) | Long-distance ship type identification method based on airborne photon radar | |
| Wang et al. | Characterizing the capability of mesoscale eddies to carry drifters in the northwest Pacific | |
| CN111736156B (en) | Upwind area identification method and device based on weather radar | |
| CN112363144B (en) | Ring-scan radar distance ambiguity and azimuth ambiguity identification method | |
| CN115932779B (en) | Dead tree target detection method based on obstacle avoidance radar distance Doppler integration | |
| CN110412609B (en) | Multi-pulse laser radar target detection method | |
| CN117148358A (en) | Barnes spatial filtering-based Doppler weather radar speed deblurring method | |
| CN115079121B (en) | Multipath target identification method and device for vehicle-mounted radar and storage medium | |
| CN113552563B (en) | Method for analyzing correspondence between vertical measurement information and high-frequency ground wave radar clutter information | |
| CN105260591A (en) | Multi-elevation MIGFA gust front recognition improved algorithm | |
| CN115128571A (en) | Multi-person and non-motor vehicle identification method based on millimeter wave radar | |
| Zrnic et al. | Probing tornadoes with a pulse Doppler radar | |
| CN114594463A (en) | Sea surface small target feature detection method based on combined convex hull |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |