eseis (version 0.8.0)

signal_motion: Calculate particle motion parameters

Description

The function calculates from a data set of three seismic components of the same signal the following particle motion paramters using a moving window approach: horizontal-vertical eigenvalue ratio, azimuth and inclination.

Usage

signal_motion(data, time, dt, window, step, order = "xyz")

Value

A List object with rectilinearity (rectilinearity), angularity (polarity), azimuth (azimuth) and incidence (incidence), as well as the corresponding time vector for these values.

Arguments

data

List of eseis objects or matrix, seismic components to be processed. If data is a matrix, the components must be organised as columns.

time

POSIXct vector, time vector corresponding to the seismic signal components. If omitted, a synthetic time vector will be generated. If omitted, the sampling period (dt) must be provided or is taken from the first eseis object in the data list.

dt

Numeric value, sampling period. Only needed if time is omitted or if data is no eseis object.

window

Numeric value, time window length (given as number of samples) used to calculate the particle motion parameters. If value is even, it will be set to the next smaller odd value. If omitted, the window size is set to 1 percent of the time series length by default.

step

Numeric value, step size (given as number of samples), by which the window is shifted over the data set. If omitted, the step size is set to 50 percent of the window size by default.

order

Character value, order of the seismic components. Description must contain the letters "x","y" and "z" in the order according to the input data set. Default is "xyz" (EW-NS-vertical).

Author

Michael Dietze

Details

The function code is loosely based on the function GAZI() from the package RSEIS with removal of unnecessary content and updated or rewritten loop functionality. In its current form, it also uses additional workflows from obspy.signal.polarisation, specifically following the Flinn (1965) approach. It windows the input signals, calculates the covariance matrix and performs a singular values decomposition to use the eigen values and vectors to determine the ratios corresponding to the output values rectilinearity, angularity, azimuth and incidence.

Note that the names of the output objects as well as the calculation routine have changed from the earlier version (V. 0.6.0), to increase computational efficiency and fix a bug in the windowing implementation.

Examples

## load example data set
data(earthquake)

## filter seismic signals
s <- eseis::signal_filter(data = s, 
                          dt = 1/200, 
                          f = c(1, 3))

## convert list of signal vectors to column-wise matrix
s <- do.call(cbind, s)

## calculate particle motion parameters
pm <- signal_motion(data = s, 
                    dt = 1 / 200,
                    window = 500, 
                    step = 250)
                    
## plot function output
par_original <- par(no.readonly = TRUE)
par(mfcol = c(2, 2))

plot(pm$time, pm$rect, type = "b")
plot(pm$time, pm$plan, type = "b")
plot(pm$time, pm$azimuth, type = "b")
plot(pm$time, pm$incidence, type = "b")

par(par_original)