PolarGUI: A MATLAB-Based Tool for Polarization Analysis of the Three-Component Seismic Data Using Different Algorithms

Abstract

We present an open-source and MATLAB-based tool with an easy-to-use graphical user interface (GUI) consisting of four polarization analysis approaches: the particle-motion trajectory (a hodogram in a 3D plane), eigenvalue decomposition (EVD) based on the covariance matrix (including two calculation methods), singular value decomposition using principal component analysis, and EVD based on a constructed analytic signal matrix (EVD-ASM). We review the calculation processes and features of the four cited methods. The eigenvalue and eigenvector are applied to obtain the polarization attributes of the three-component (3C) seismic data. Using rose graphs and histograms, the corresponding azimuth and incidence angle are calculated to determine the propagation direction of the seismic wave. Statistical distribution curves of the corresponding rectilinearity and planarity of the waves are also plotted. The polarization analysis GUI can simultaneously analyze two selected data sections in a seismic recording corresponding to P and S waves. We evaluate the performance of these algorithms using real 3C earthquake datasets. Comparison tests indicate that the aforementioned four methods have different time consumption, and the differences between the results of the EVD-ASM and those of the other methods are very small.

Six‐Component Earthquake Synchronous Observations Across Taiwan Strait: Phase Velocity and Source Location

“…Normally, polarization analysis based on 3C translational records is believed to be hard to estimate azimuth without 180° ambiguity. The EVD‐CM method we used for 3C polarization analysis could estimate azimuth under ideal conditions (Li et al., 2021), but it relies on the appropriate time window and frequency band selection of P wave, while many teleseismic events induce much significant surface waves than body waves, which makes 3C polarization less robust than rotational methods based on surface waves in terms of large epicentral distance. Besides, high‐frequency P‐waves are easily affected by the potential complexity of the seismic source and the small‐scale structure under the station, which will increase the probability of azimuth ambiguity.…”

“…Here we use eigenvalue decomposition based on the covariance matrix (EVD‐CM, Li et al., 2021) for 3C polarization analysis. As P waves are polarized in the consistent direction of propagation in a homogeneous medium, we select the records after first arrival of P waves for hodograms and calculating polarization azimuth, as illustrated in Figure 4.…”

Using Fuzzy C-Means Clustering to Determine First Arrival of Microseismic Recordings

“…Power and variance are defined as follows: where is the mean value of and q represents the window length of the cluster analysis. As for the polarization linearity feature , we firstly intercept the 3C microseismic recording with window length q , and construct the covariance matrix M using Equation ( 8 ) [ 48 ]. where x , y , and z represent two horizontal and a single vertical component data, respectively.…”

“…As for the polarization linearity feature L i , we firstly intercept the 3C microseismic recording with window length q, and construct the covariance matrix M using Equation ( 8) [48].…”