CN-122017781-A - Severe storm monomer splitting identification method based on meteorological radar base data

CN122017781ACN 122017781 ACN122017781 ACN 122017781ACN-122017781-A

Abstract

The invention discloses a severe storm monomer splitting identification method based on weather radar base data, and relates to the technical field of weather radar signal processing and weather modification operation. The method is based on weather radar base data, realizes two-dimensional storm structure identification through multi-threshold connected domain dynamic cutting, and combines spatial association relations of adjacent height layers to construct a three-dimensional storm body, and in the three-dimensional structure construction process, a multiple Flag (Flag) reflecting the complexity of the vertical structure inside the storm body is introduced to characterize one-to-many or many-to-one association relations of two-dimensional structures in different height layers. And the evolution characteristics of the multiple marks in the vertical direction are analyzed, so that the advanced identification of the severe storm monomer splitting process is realized. Further, by combining a storm body time sequence matching and track prediction method, stable tracking and split early warning of storm monomers are realized. The method can effectively avoid adhesion misjudgment of a plurality of adjacent strong convection monomers in two-dimensional projection while retaining the weak echo edge information of the primary hail cloud, and provides reliable technical support for manual hail suppression operation command.

Inventors

LV ZHILIN
LU JIANPING
QIN JIANG
LEI LIANFA
QIN ZHENGYANG
MU TAO
HU XINYUE
WANG XIANG

Assignees

北方天穹信息技术（西安）有限公司

Dates

Publication Date: 20260512
Application Date: 20260203

Claims (7)

1. A severe storm monomer splitting identification method based on meteorological radar base data is characterized by comprising the following steps: Step 1, radar data preprocessing, namely acquiring weather radar base data, and performing space mapping and height interpolation processing on the radar base data to generate three-dimensional radar body sweep data for storm structure analysis and equal-height reflectivity products (CAPPI) of a plurality of height layers; step 2, two-dimensional structure identification, namely setting a plurality of reflectivity thresholds on each equal-high-rise reflectivity product, and adopting a connected domain marking method to identify a two-dimensional storm structure, wherein a first reference reflectivity threshold is used for identifying a connected region, judging whether the connected region internally comprises a plurality of sub-connected regions identified by a second reference reflectivity threshold or not, and cutting and separating the connected region by adopting the second reference reflectivity threshold and a higher reflectivity threshold when the plurality of sub-connected regions exist so as to obtain a plurality of independent two-dimensional storm structures; Step 3, generating a vertical association and a multiple mark of the two-dimensional structure, wherein the vertical association is performed based on the spatial overlapping relation between two-dimensional storm structures of different height layers, and when the two-dimensional storm structures of adjacent height layers have an association relation of one to many or many to one, the multiple mark for representing the complexity of the three-dimensional structure is generated at the corresponding height layer; step 4, constructing a three-dimensional storm body, namely stacking and combining a plurality of two-dimensional storm structures with high layers according to the vertical association result to construct the three-dimensional storm body, and reserving the multiple marks in the three-dimensional storm body; Step 5, storm body time sequence association and splitting discrimination, namely, based on the space and form difference between three-dimensional storm bodies at continuous moments, establishing a matching relation between the storm bodies, and determining continuation, splitting, merging, new generation or extinction states of the storm bodies; and 6, storm path prediction, namely, short-time prediction is carried out on the future movement path of the storm body based on the historical position information of the storm body.
2. The method for identifying severe storm monomers based on meteorological radar base data according to claim 1, wherein said step 1 comprises: The reflectivity data in the weather radar base data are read, and the data are organized and stored according to the radar scanning elevation angle and azimuth angle information; Performing space coordinate conversion on the radar base data, and mapping the radar base data in a polar coordinate form to a rectangular coordinate system under the condition of considering the influence of the earth curvature and atmospheric refraction to generate three-dimensional radar body scan data; performing high interpolation processing on the reflectivity data based on the three-dimensional radar body sweep data to obtain equal-high-layer reflectivity products (CAPI) of a plurality of preset height layers; And preprocessing the equal-high-layer reflectivity product to remove a small-scale isolated echo region, and providing input data for subsequent two-dimensional storm structure identification.
3. The method for identifying severe storm monomers based on meteorological radar based data according to claim 1, wherein said step 2 comprises the following two-dimensional storm structure identification logic: Setting at least two different levels of reflectivity thresholds on a high-rise reflectivity product, including a first reference reflectivity threshold for initially identifying a storm range, and a second reference reflectivity threshold that is higher than the first reference reflectivity threshold; Identifying a communication region by adopting the first reference reflectivity threshold value, and judging whether a plurality of high-reflectivity core regions identified by the second reference reflectivity threshold value exist in the communication region; When a plurality of high-reflectivity core areas exist in the communication area, judging the communication area as a multi-monomer adhesion area, and dividing the communication area based on the second reference reflectivity threshold value and the reflectivity threshold value sequences above the second reference reflectivity threshold value so as to obtain a plurality of independent two-dimensional storm structures; And when a plurality of high-reflectivity core areas do not exist in the communication area, carrying out structure extraction on the communication area based on the first reference reflectivity threshold value and the reflectivity threshold value sequence above, and reserving the communication area as a single two-dimensional storm structure.
4. The method for identifying severe storm monomers based on weather radar based data according to claim 1, wherein said step 3 comprises the following two-dimensional structure vertical association and multiple sign generation process: Vertical association judgment is carried out on the two-dimensional storm structures on different height layers according to the sequence of the height layers, and whether the two-dimensional storm structures belong to the same three-dimensional storm body is determined according to the space adjacent relation between the two-dimensional storm structures of the adjacent height layers; when the two-dimensional storm structures of adjacent high layers meet the preset space association condition, establishing an association relation of the upper and lower two-dimensional storm structures, and inheriting the identification information of the upper two-dimensional storm structure to the lower two-dimensional storm structure; Generating a multiple mark at a corresponding height layer when a single two-dimensional storm structure in one height layer corresponds to a plurality of two-dimensional storm structures in an adjacent height layer or a plurality of two-dimensional storm structures correspond to a single two-dimensional storm structure in an adjacent height layer in the vertical association process, and representing a structure bifurcation or convergence state of a three-dimensional storm body on the height layer; the multiple marks are used for indicating that a multi-monomer structure exists in the three-dimensional storm body in the vertical direction, and when the multiple marks continuously appear in a plurality of height layers or show a change trend extending from a high layer to a low layer, the three-dimensional storm body is judged to have an early sign of splitting.
5. The method for identifying severe storm monomers based on weather radar based data according to claim 1, wherein said step 5 comprises the following steps of timing association and status determination of storm: Based on three-dimensional storm bodies obtained at successive moments, constructing matching cost reflecting spatial position changes and structural differences among different storm bodies, and representing association possibility of storm bodies among adjacent moments; Introducing a virtual matching object for representing new generation and extinction states of the storm body in the storm body matching process, and carrying out joint matching on the real storm body and the virtual matching object under a unified matching frame; And carrying out global optimization solution on the matching cost to obtain an optimal association relation between storm bodies at adjacent moments, thereby determining the continuation, splitting, merging, new generation or extinction states of the storm bodies on a time sequence.
6. The method for identifying severe storm monomers based on weather radar based data according to claim 1, wherein said step 6 comprises a short-time prediction process of storm body movement path as follows: Based on mass center position and motion state information of the storm body at continuous moments, a state model and an observation model of the movement of the storm body are established and used for describing the position change of the storm body in space; according to the deviation between the actual observation result and the prediction result, dynamically adjusting the uncertainty parameter in the state model to realize the self-adaptive correction of the storm body motion state; The influence of the historical observation information is weighted and attenuated by introducing a forgetting factor, so that the prediction process can timely respond to the steering, acceleration or deceleration change in the movement process of the storm body, and a movement path prediction result of the storm body in a short time in the future is obtained.
7. The method for identifying severe storm monomers based on weather radar based data according to claim 1, further comprising the following comprehensive discrimination process of storm monomers splitting and merging events: Based on time sequence association results between storm bodies at adjacent moments, identifying one-to-many or many-to-one association relation of the storm bodies on time sequence, wherein the association relation is used for representing the structural evolution state of the storm bodies; combining multiple marks generated by the three-dimensional storm body in different height layers to analyze the variation trend of the vertical structure in the storm body; when the time sequence association result represents that a storm body has a one-to-many association relationship, and the multiple marks continuously appear or show a change trend extending from a high layer to a low layer in a plurality of high layers, judging that the storm body has a splitting event; When the time sequence association result represents that the storm body has a many-to-one association relationship, determining that the storm body has a merging event.

Description

Severe storm monomer splitting identification method based on meteorological radar base data Technical Field The invention relates to the field of weather radar signal processing and weather modification operation, in particular to a severe storm monomer optimizing identification and splitting discrimination method based on weather radar base data. The method can be used for strong convection weather monitoring, hail cloud identification and artificial hail suppression operation command systems. Background Hail disasters are a strong convection weather process with strong burst and strong destructive power, and widely occur in a plurality of areas in China, and have serious influence on agricultural production, traffic operation, building facilities and safety of an electric power system. Aiming at the defending requirement of hail disasters, artificial influence on the hail prevention operation in weather has become one of the important technical means of current disaster reduction and prevention. In the manual hail suppression operation process, the spatial structure characteristics of the hail cloud body are accurately acquired, the generation, development and evolution states of the cloud body are timely identified, and the method is a key premise for realizing scientific decision making and fine operation command. The weather radar is used as main detection equipment for acquiring the three-dimensional structure information of the cloud body, can provide radar echo data with high time resolution and high spatial resolution, and provides an important information source for the identification, tracking and early warning of strong convection storm. The existing storm monomer identification and tracking method based on weather radar echo mostly adopts a fixed reflectivity threshold value or a structure identification mode based on two-dimensional projection to extract and correlate a time sequence of storm monomers. However, under actual complex weather conditions, the method has the defects that on one hand, the fixed threshold strategy is difficult to consider the identification requirements of the weak echo primary storm and the strong convection mature storm, the primary hail cloud is easy to ignore or miss, and on the other hand, when a plurality of strong convection monomers are adjacent to each other in space or vertically overlapped, the identification mode based on two-dimensional projection is easy to generate monomer adhesion or error combination, so that the accuracy of storm quantity statistics and morphological evolution judgment is influenced. In addition, in the storm development process, the cloud body often accompanies obvious vertical structure change, splitting, merging and other complex evolution behaviors, the traditional method has limited capability of describing storm three-dimensional structure evolution characteristics, and early signs of storm splitting are difficult to capture in time, so that early warning advance of hail suppression operation is insufficient, and scientificity and effectiveness of operation opportunity selection are affected. Therefore, there is a need for a storm monomer identification and splitting discrimination method which can fully utilize meteorological radar-based data, comprehensively consider two-dimensional and three-dimensional structural characteristics of a storm and consider the storm time sequence evolution rule, so as to improve the completeness of hail cloud identification, the stability of storm tracking and the reliability of splitting early warning, thereby providing more accurate and timely technical support for artificial hail suppression operation. Disclosure of Invention The invention aims to provide a severe storm monomer optimizing identification and splitting discrimination method based on meteorological radar base data, which realizes accurate identification, stable tracking and splitting early warning of storm monomers through multi-threshold communication identification, three-dimensional structure reconstruction, multiple index discrimination and storm time sequence association. The method of the invention comprises the following steps: Step 1, preprocessing radar data: The method comprises the steps of obtaining weather radar base data, carrying out coordinate conversion and spatial interpolation processing on the radar base data to construct radar three-dimensional body scanning data, generating equal-high-layer reflectivity products (CAPI) of a plurality of high layers on the basis of the three-dimensional body scanning data, carrying out noise suppression and clutter rejection on the generated radar data, and reserving a convection echo region for subsequent storm two-dimensional structure identification and three-dimensional structure construction. Step 2, two-dimensional storm structure identification: In the equal-high-layer reflectivity product (CAPI), setting a plurality of reflectivity thresholds for radar echo data o