Mesocyclone identification method based on Doppler radar echo images

Abstract

The invention discloses a mesocyclone identification method based on Doppler radar echo images. The mesocyclone identification method comprises the following steps of: filtering reflectivity images with 0.5-degree, 1.5-degree and 2.4-degree elevation angles to acquire an effectively reflectivity area; blending information on the effective reflectivity area into radial velocity images with same elevation angles to acquire radial velocity images of a limited searching area; segmenting the radial velocity images into which the information on the effective reflectivity area is blended, and screening positive and negative velocity center areas in the radial velocity images; configuring all possible velocity pairs through distribution histograms of radial velocity values of the velocity center areas so as to obtain a velocity pair set; and eliminating pseudo velocity pairs from the velocity pair set, and identifying the mesocyclone. By the mesocyclone identification method, the complicity of time and space is reduced, most of false mesocyclones are eliminated, the identification rate is increased, the misstatement rate is reduced, the working efficiency is improved; moreover, the method has the characteristic of not limiting the cyclone direction and structure information of the mesocyclone, so the identification result is more comprehensive.

Description

technical field [0001] The invention relates to the field of identification of central cyclones in meteorology, in particular to a method for identifying central cyclones based on Doppler weather radar echo images. Background technique [0002] Meteorologists have found through a large amount of data statistics that more than 92% of my country and more than 90% of the United States have severe convective storms [1] (such as heavy rain, strong wind, hail and tornado, etc.) will be accompanied by the occurrence and development of mesocyclones, especially mesocyclones generally exist some time before the strong storm strikes. In 1994, Stumpf and Witt proposed a new mesocyclone identification method (E-MDA). They constantly adjusted some important parameter values ​​in the experiment, and then compared the results repeatedly, and finally made a major breakthrough, increasing the identification rate by 12 %about [2] . Then they proposed an improved version (MDA) that combines ...

AI Technical Summary

Problems solved by technology

[0003] The inventor finds that there are following deficiencies in the prior art in the process of realizing the present invention: (1) because the present widely used meteorological monitoring and forecasting software PUP is written according to the weather characteristics of the United States, in actual business operations due to regional different, poor adaptability, low recognition accuracy, high false negative rate and false positive rate; (2) NSSL MDA needs to scan each point on the average radial velocity map with the same distance from the radar one by one. There will be missed or weak reports, and anticyclonic rotation cannot be detected. Scanning each point one by one greatly increases the time and space complexity of the method; (3) due to improper deblurring processing, there will be False mesocyclone; (4) Although almost all vortices can be detected by reducing the threshold parameter, but at the same time, the false alarm rate is increased, and the adjustment of the threshold has objective factors such as man-made and uncertain; (5) Only three-dimensional information is classified, but the identification of mesocyclones by pure rotation, convergent rotation, and divergent rotation is not studied, making the identification results incomplete

Images