In this study, we discriminated the quarry blasts from microearthquakes in the Surendranagar region in Saurashtra provenance of Northwestern India.
3.1 P/S amplitude ratio
The key to the efficacious discrimination of numerous types of seismic sources is the clear observation of the signals released from seismic sources (Kim et al. 1998). One of the most important problems is the correction of the effects of the source to receiver paths in the observed regional signals (Kim and Richards 1996). After removing the free-surface effects from our data, we calculated the P/S ratio for four mutually exclusive frequency bands: 2 to 4, 4 to 6, 6 to 8, 8 to 10, and a common bin from 1 to 10 Hz. The ratio of P/S is calculated at the nearest station SUR and plotted against the epicenter and magnitude of the events. Figure 2 and 3 shows the P/S ratios over five frequency bands. In this figure, it appeared that the best separation is observed in the 8 to 10 Hz range. Since S wave amplitude of the earthquakes is greater than P wave and, P wave is dominant in quarry blasts, hence P/S>1 are quarry blasts and P/S<1 are earthquakes. Kim et al. (1998) found that the best separation between quarry blasts and earthquakes was in 6-8 Hz. We observed that the identified quarry blasts are in the range between 22 and 38 km (Figure 2), additionally, the observed quarry blasts are between the magnitude range of 1.4 and 2.2 (Figure 3).
3.2 Short-Time Fourier transform
After the P/S discrimination, many events are close to the separation line (Figure 2 and 3). If the P/S ratio is close to one, then it indicates that the amplitude of P and S waves is somewhere equal. In these cases, we used Short-time Fourier transform to decide whether these events are earthquakes or quarry blasts. This method is a valuable tool to study the frequency and amplitude content of the entire seismic waveforms (Kim et al. 1994). In quarry blasts, the energy released in a quick time and maximum amplitude occurred at the starting time of the record (Yilmaz et al. 2012). For earthquakes, P wave amplitude is much shorter than S wave and maximum energy observed after the P wave train. The figure 4 and 5 shows the spectrogram of quarry blast and earthquake respectively. From figure 4, we observed that the P wave is dominated in the seismogram. P wave energy is shown in green color and we found no other phase in the spectrogram. Figure 5 shows that the amplitude of the S wave is much larger than those of P wave and represented in red and green color respectively. Considering the above criteria related to maximum released energy, which confirms that the event in figure 4 is quarry blast and figure 5 shows an earthquake.
Additionally, we found Rg phases in both quarry blast and earthquake at lower frequencies (0.5-2.5 Hz). Similar features have been observed by Yilmaz et al. (2012). This indicates that the above events are at a very shallow depth.
3.3 Coda Wave method
The coda Qc has been computed using the single backscattering model proposed by Aki and Chouet (1975). This model clarifies that coda waves are backscattered body waves generated by randomly distributed heterogeneity in the Earth’s crust and upper mantle. Each wavelet is scattered only once before reaching the receiver. Data from both quarry blast and earthquakes sources were processed by an identical procedure to study the possible differences in coda attenuation. Figure 6a, c, g, e shows the measured Qc of earthquakes as a function of lapse time for frequencies 1.5, 3.0, 6.0 and 12 Hz, respectively, while Figure 3b, d, f, h shows the corresponding results for quarry blasts. The resultants Qc־¹ values for all earthquakes were plotted together. The solid line connects the mean points calculated by averaging the individual measurements. By comparing these figures, we found that for the frequencies of 1.5 and 3.0 Hz, Qc־¹ shows significant differences between quarry blasts and earthquakes. The Qc־¹ values at these low frequencies are greater for quarry blasts than for earthquakes at lower frequencies of 1.5 and 3.0 Hz. It is elucidated that in contrast to earthquakes, more surface waves are generated in blasts as they originate from extremely shallow depths. Because the propagation paths of surface waves being confined at superficial depths thus the surface waves attenuate quickly. At t>20 sec for 6 and 12 Hz and at t > 30 sec for 1.5 and 3.0 Hz, body waves become dominant over surface waves resulting in same coda decay rate and path effects for both blasts and earthquakes which further indicates that delayed lapse time is dominated almost certainly by S waves, i.e a single type of body wave, excluding coda waves at lower frequencies which are associated with the surface wave (Su et al., 1991).
3.4 Induced Seismicity
The Qc value of 2.75 (Figure 7c) observed after the discrimination is greater than the threshold value of 1.5. This suggests that the observed earthquake activity is more during the daytime when compared with the after-dark recordings (Figure 7), which indicates that these observed micro-earthquakes may be mining-induced.
Mining induced seismicity is a well-known long studied phenomenon throughout the world (e.g. Gibowicz, 1990; Johnston, 1992). The decrease in normal stress or an increase in the shear stress acting on a fault is responsible for the mining-induced seismicity (McGarr et al., 1999). In this study, we found numerous microearthquakes following the quarry blasts (Figure 8). However, we do not identify these microearthquakes instantaneously/during or just after the passage of the quarry blasts waves i.e. these microearthquakes are observed with some delays. This observation may suggest that the dynamic stress imposed by the quarry blasts may have invoked the faults in the study area. For instance, when a fault (or population of faults) is near to failure, then it is more susceptible to dynamic stress perturbation. In other words, the addition of any small stress perturbation on a critically stressed fault can lead to the occurrence of brittle failure (Brodsky & van der Elst, 2014; Gomberg et al., 1998). The subcritical crack growth model substantiates the aforementioned and puts forward the idea of earthquake nucleation (Atkinson, 1984), which suggests that the intensity of stress at the crack tip (relative to crack size) governs the rate of crack growth. At the initial stage, the crack extends very slowly and then grows rapidly with the unexpected rise of/increase in stresses at the crack tip eventually leading to what is called apparently delayed fault rupture (Atkinson, 1984; Rinne, 2008). However, we cannot completely rule out the involvement of earthquakes that are occurred by natural stress perturbations in the study area (Figure 7c).
3.5 Seasonal and Diurnal observations
India is a land of festivals and Indians give special importance to them. Some of the main festivals which are celebrated include Diwali, Holi, Raksha Bandhan, and Pongal. Indians also hold great regard for National festivals like Independence (15th August) and Republic Day (26th January). We examined the ISR seismic catalog four days before and after these festivals. We found no signals of quarry blasts during these festival days (Figure 9), as quarry blasts are operated only during the working days. Additionally, on the other hand, we also inspected the ISR catalog during the days of the week (Sunday to Saturday). We found the events on weekends also (Figure 7e, f). Hence, we cannot completely rule out the operations of illegal quarry blasts and mining during the holidays (The Indian Express, January 01, 2020; The Telegraph, January 06, 2020). Illegal mining is a perennial problem in India (India TV news, June 18, 2018), however, to operate the mining during the monsoon is extremely difficult. We observed that during the Indian monsoon (June-September) the seismicity decreases and it increases steadily after the monsoon because during the monsoon season the quarries are not operational.