CodaQback: A Simplified Python Code Facilitating Auto-Windowing for Estimating Seismic Coda Attenuation Parameter

Top

1. Introduction

The energy of seismic wave at several distances from earthquake source is rigorously affected by the medium through which wave propagated. Attenuation is one of the vibrant parameters that describe the medium which is mainly a geological medium. It is generally measured by quality factor Q, a dimensionless quantity. By definition, Quality factor gives an estimate of dissipated energy (Aki and Chouet,1975; Knopoff and Hudson, 1964) while spreading through the geological medium. In general, Q varies inversely with attenuation. As such, higher attenuation is observed in seismic with lower Q values.

There are several reports concerning the estimation of Q_C, the quality factor of the coda wave, for different parts of the world. According to these reports, attenuation in crust and lithosphere can be best studied through the use of coda wave analysis (Aki and Chouet, 1975; Aki, 1969; Sato,1977; Padhy and Subhadra, 2010; Hazarika et al., 2009; Biswas et al. a,b, 2013).

The Kopili Fault Zone in Northeast India is an intraplate earthquake source zone and is characterized as a probable sector for a large forthcoming earthquake. Till now, this fault caused two large earthquake and several intermediate earthquakes (Bora et al., 2018). This makes it a possible province for seismic activity as well as socio-economic droughts, and it would be convenient to study attenuation features of this area in order to make it available for other researchers who would be concerned in micro-zonation of this active region.

On other hand, Jin and Aki (1988), stated that the regional estimates of Q_C and its spatial nonconformity are directly connected to the seismicity and tectonics, which acts a key role in hazard investigation and engineering seismology. Bora and Biswas, (2017); Bora et al (2017) and Bora et al (2018) investigated the attenuation mechanism for this Kopili region by using their own Matlab codes.

During estimation of this vital parameter, most of the researchers resorted to customized code or painstaking analysis and subsequent lengthy computation. It is indeed essential that there should be a robust computational tool which will render this whole analysis effortless but reliable. Keeping this objective in mind, the present work aims at designing software for assessment of the coda-wave attenuation entailing local and regional earthquakes.

In this direction, seismological software is available towards the measurement of attenuation of coda waves along with a routine way. To the best of our knowledge, this is the first endeavor of designing a computational tool for estimating Coda Q based on the back scattering model. Consequently, there is no scope of correlating with other existing seismic codes. As for instance, the CodaQ subroutine of SEISAN (Havskov and Ottemoller, 2005) enables the users to attain this parameter. The SGRAPH program (Abdelwahed, 2012) enables the user to estimate the intrinsic and scattering seismic attenuation parameter by using the Multiple Lapse time Window method (Zeng, 1991). Similarly, Predein et al. (2017) recently developed a software package named as CodaNorm, based on the coda-normalization method (Knopoff and Hudson, 1964). This package allows the estimation of the seismic quality factor and its frequency dependence (n) for direct body waves for different central frequency. In our own developed CodaQback package, the initial data used are:

• 1.

  The earthquake waveforms (seismograms) in the PITSA format;

• 2.

  The excel file having all the hypocentral factors;

Top

2. The Single Backscattering Model

We try to measure coda wave attenuation parameter by adopting the single backscattering method proposed by Aki and Chouet (1975). In this model, the amplitude of coda wave (A_C) at a central frequency f for a preferred frequency band and a precise lapse time t measured from the earthquake origin time t₀ can be expressed as:

(1) where, S(f), G(f), I(f) denote the source response, site amplification and instrument response. The geometrical spreading parameter α is measured as unity (constant) in this study as opined by Havskov et al. (1989). According to Havskov et al. (1989), coda waves are considered as backscattered body waves (Aki, 1981 and Aki,1980). The term Q_C denotes the coda wave quality factor which describes the average attenuation of the medium for a predefined area.