A widely felt Tremor (ML 3.5) of 12 April 2020 in and around NCT Delhi in the backdrop of prevailing

An earthquake of small magnitude (ML3.5) occurred on 12 April 2020 near the east district boundary of NCT, Delhi with maximum PGA for the event observed to be 14.13 gals. A few smaller aftershocks also occurred in the area. The estimated fault plane solution of the mainshock suggests normal faulting with some strike slip component. The focal mechanism corroborates with the NE SW orienting lineaments mapped in the region near the epicenter. The source parameters of the event, namely, seismic moment, stress drop, corner frequency, and source radius are estimated to be 1.15 x 1014N-m, 25.7 bars, 5.7Hz and 300m, respectively. The decay rate of acceleration with epicentral distance suggests a regression relation PGA 1⁄4 474D 1.347, which may be useful for understanding the ground motion in the region. A noise analyses at NDI rock and UJWA soil sites clearly suggest a significant reduction in ambient noise by 10dB in the frequency band (1.0–10.0) Hz at the respective sites, during the COVID-19 lockdown situation. The reductions of the noise level improve the signal to noise ratio substantially at all the seismic stations located in the urban agglomerations, which enabled the recording of clear phases of the event and hence improved the analysis. ARTICLE HISTORY Received 11 May 2020 Accepted 23 July 2020


Introduction
On 12 April 2020 (Sunday), the National Capital Territory (NCT) Delhi and its neighborhood jolted with a sudden shaking caused due to a small magnitude tremor (M L 3.5), which occurred at local time 17:45:02.3 IST (12:15:02.3 UTC) in the National Capital Region (NCR) near the east district boundary of NCT Delhi. It was a time when the entire country including the NCR of Delhi was under lockdown after the unprecedented COVID-19 pandemic outburst. Although the ground motion lasted only for a few seconds, it created panic in the people and caused them to rush out of their homes. On Sunday evening, in countrywide lockdown situation, mostly the people were confined in their houses, also primarily of multi-storeyed type, where they heard a rumbling sound on the passage of the seismic waves. No damage was reported in the region. The mainshock (M L 3.5) was later followed by few mild aftershocks with magnitudes < 3.0; the events of M L 1.3-2.7 occurred during 12.04.2020 to 16.04.2020. These recent Delhi earthquakes were recorded by the national network, which is operated and maintained by National Centre for Seismology (NCS), New Delhi. Evidently, the mainshock of 12 April 2020 was recorded at more than 29 seismic stations. The tremor was widely felt in the region including NCR. Figure 1 shows the epicentres of the mainshock and the subsequent aftershocks, which occurred during April 12-16, 2020. The NCT Delhi lies in the seismic zone IV of the seismic zoning map of India (BIS: 1893 (Part 1), 2002), which corresponds to a peak ground acceleration (PGA) equivalent to 0.24 g (i.e. 240 gals). The region has a well-known history of the felt earthquakes, both from local as well as regional sources  (Verma et al. 1995;Iyengar 1999). Table 1 shows a list of the historical and some significant past earthquakes in and around the Delhi region. A moderate size earthquake of 27 August 1960 was earlier located near the Delhi -Gurgaon border having magnitude M 4.8 (Table 1) but was later revised to M 6.0 (Iyengar 1999 In the present study, we relocated the recent Delhi tremors and also analyzed the best recoded waveforms to understand the focal mechanism, source parameters, strong ground motions and their impact in the region. The results from the analyses and observations are presented in this paper. A similar analysis, however, for the aftershocks couldn't be carried out, as they were recorded at only a few seismic stations. Further, an effort is made to characterize implication of the COVID-19 lockdown silence on the ambient ground noise, which highlights the temporarily improved seismic recoding due the coronavirus shutdown.

Seismotectonics of the Delhi region
A seismotectonic map of the Delhi region is shown in Figure 1. The region is located on a folded crustal ramp represented by quartzite basement rocks of the Delhi super Group, bounded between two regional faults namely, the Mahendragarh Dehradun Sub Surface Fault (MDSSF) to the west of NCT Delhi and the Great Boundary Fault (GBF) to the east. Another important structural element of the belt is the NW -SE trending Delhi Sargodha Ridge (DSR), which is flanked by the Sahaspur and Bikaner Basins to the north and southwest, respectively, and crosses the MDSSF close to Delhi. Another N-S trending fault, known as the Sohna Fault (SF), is running from Sohna to the western part of Delhi. All the tectonic features are found to be quite active and also, they are the possible causative sources of the seismicity in and around the Delhi region (e.g. Hukku 1966;Gupta and Sharda 1996;Bansal et al. 2009;Bansal and Verma 2012;Shukla et al. 2016;Singh 2020;Tripathy-Lang 2020). Moreover, the NE -SW orienting faults and lineaments mapped in the region are found to be consistent with the nodal planes of a few past Delhi earthquakes (Singh et al. 2010). Seismically the most active regions of the Himalaya namely, the Main Central Thrust (MCT), and Main Boundary Thrust (MBT) strongly influence the seismic activity in the Delhi region and also, they are probably the liable to the genesis of the moderate damaging events (NCS Catalog; GSI 2000).

Data analysis and results
Earthquake monitoring in Delhi and surrounding regions is efficiently carried out using a state-of-the-art digital telemetered national network equipped with broadband seismographs (BBS) and Strong Motion Accelerographs (SMA) spread across the region. In the back drop of the prevailing lockdown due to the COVID crisis, the recent earthquake M L 3.5 of 12 April 2020 is an unusually well-recorded tremor in the region due to reduced ambient noise level at the recording stations. In the present study, we relocated the mainshock using well recorded data from 25 seismic stations showing clear P-and S-phases as well as improved signal to noise ratio (SNR). Figure 1 shows the location map of the field seismic stations (BBS and SMA) that were considered in the analysis. Details of the recording stations including the azimuths and epicentral distances are listed in Table 2. Apparently, nine seismic stations are within the radius of 50 km from the epicenter of the mainshock. Two stations are within the distance range 50-100 km, while three stations are located between 100 km and 200 km, and the remaining stations are beyond 200 km. We mention that the BBS and SMA stations are collocated at each site. The well recorded earthquakes of the Delhi region for the last about two decades (January, 2001 -March, 2020) clearly demarcate the zones of seismic activity ( Figure 2). During this period, about 733 earthquakes have been located in and around Delhi (NCS, Catalog); more than 90% events of these are of magnitude < 3.0 having focal depth 15 km. The NCT, however, witnessed about 146 events during the period, including an event (on 25 November 2007) with maximum magnitude M w 4.1 (Singh et al. 2010).

Earthquake location
The mainshock of 12 April 2020 is well recorded by 25 seismic stations in and around the NCT. The event was relocated using the SEISAN software package (Havskov and Ottemoller 1999), and the epicenter is estimated to be (28.791 o N, 77.268 o E) with rms error 0.23 s, and the focal depth is found to be $14.5 km. The accuracy of the location in latitude and longitude was estimated to be about ± 1.0 km and ± 1.4 km respectively; however, the focal depth accuracy was about ± 1.5 km. Details of the focal parameters are listed in Table 3. An aftershock that occurred on 13 April 2020 (M2.7) was also relocated in a similar manner using well recorded data from 9 seismic stations. However, the other mild aftershocks during 13-16 April 2020 could not be relocated as they were recorded at only a few stations.

Focal mechanism
A well constrained focal mechanism of an earthquake of 25 November 2007 (M4.1) within NCT Delhi suggests strike-slip faulting with some normal component (Singh et al. 2010). Two more tremors, viz., 28 April 2001 (M3.4) and 18 March 2004 (M2.6) of NCT, also revealed a similar faulting mechanism having strike slip with some normal component (Bansal et al. 2009). In the present study, we have simulated the recent earthquake of 12 April 2020 for the fault plane solution using the ISOLA waveform inversion technique (Sokos and Zahradnik 2008). Well recorded waveform data from nine BBS stations, which were located within 100 km epicentral distance except KUDL station located at 110 km (Table 2), were used in the analysis. We mention that for three stations, namely, LDR, JMIU and GNR, although they were located within 50 km from the epicenter, the records were too noisy to be used in the analysis. The waveforms with cut-off SNR > 2.0 in the frequency range of interest (0.05-15 Hz) were used in the inversion. Various available velocity models for the study region were tested for the simulation (e.g. Chun 1986;Kayal 2001;Suresh et al. 2008;Kumar et al. 2009;Mitra et al. 2011), and the model of Mitra et al. (2011) was found the most suitable to compute synthetic waveform with high correlation coefficient and DC %. The component-wise best fits between the synthetic and the recorded waveforms in the frequency band 0.06-0.1 Hz are shown in Figure 3(a). Some seismic stations as well as components of the simulated waveforms that showed negative correlation and did not match well with the observed waveforms were excluded from the analysis to get the moment tensor solution. Figure 3(b) shows the estimated fault plane solution of the 12 April 2020 event using a vertical grid search method at various trial depths, and the best solution was achieved for the correlation > 0.5 and DC% > 70, which suggests normal faulting with a strike slip component. Further, the moment magnitude of the event is estimated to be M w 3.5. Table 4 shows the fault plane solution of the 12 April 2020 mainshock, and the final beach ball is depicted in Figure 1. The nature of faulting of the current event and one of its nodal planes oriented NNE -SSW corroborate well with the results of earlier small magnitude earthquakes that occurred in NCT (Bansal et al. 2009;Singh et al. 2010). We also found a nodal plane consistent with orientation of lineaments mapped in the region (Verma et al. 1995). We further suggest that the large number of past earthquakes along the major faults in and around the Delhi region might have stressed the lineaments in the epicentral region, and thus the mainshock might have reactivated the release of stored energy along such lineaments as small magnitude events. However, a detailed study is to be undertaken to understand the sources of these small magnitude tremors in the region, which primarily may include precise estimation of focal mechanism and focal depths of these events Ebel 2020, 2019).

Source parameters
An analysis of the source parameters of the Delhi earthquakes (M L 3.5 and 2.7) using the S-wave recordings at the BBS stations, considering the Brune circular source model for the event (Brune 1970) were performed. The source parameters were obtained separate from the far-field displacement amplitude spectra of the S-waves on both the radial and transverse components. A sample displacement spectrum at the Thakurdwara (TKR) station is depicted in Figure 4. The spectra were corrected for geometrical spreading which decreased the amplitude of the signal with hypocentral distance. The quality factor is Q (f) ¼ 253f 0.80 , which was earlier estimated for the Himalayan arc region (Singh et al. 2004). The mean value of the source parameters of the recent earthquake, namely, the seismic moment, corner frequency, stress drop, and source radius, are estimated to be 1.15 Â 10 14 N-m, 5.7 Hz, 25.7 bars, and 300 m, respectively. The estimated result, therefore, is characterized $ 26.0 bars stress drop for the Delhi region, which is found to be within the stress drop range (10-144 bars) for stable-craton intraplate earthquakes (Kumar et al. 2014, Sairam et al. 2018). Further, the stress drop for the small to moderate earthquakes in the Central United States are found to be in the range of 46-300 bars, with a median of 84 bars (Huang et al. 2017). It is to mention that in general, higher stress drops usually coincide with the intraplate region (Sairam et al. 2018).

Strong ground motions
Strong-motion data from the 12 SMA stations for the Delhi tremors that occurred during 12 À 13 April 2020 were analyzed to understand ground-motion impact in terms of the peak ground acceleration (PGA), peak ground velocity (PGV) and peak ground displacement (PGD) at the recording stations. The stations were spread over the region at variable epicentral distances between 12 km to 104 km. The component-wise observed PGA, PGV, and PGD for the M L 3.5 and 2.7 events are listed in Tables 5 and 6, respectively. The M L 2.7 event being relatively a smaller event, the resulting strong ground motion recorded at only few stations and the maximum observed PGA were found to be quite feeble (< 0.6 gals). A PGA decay curve with epicentral distance for M L 3.5 event is shown in Figure 5, which is found to be the best fit with a regression equation PGA H ¼ 472.89 Ã D À1.373 , where D is the epicentral distance. Evidently, the NDI station being located nearest to the event, with epicentral distance 12 km, shows the horizontal PGA to be 10.19 gals. Despite the site JMIU being located 25 Km away from the event, more than twice the distance of NDI, a slightly higher horizontal PGA 14.13 gals was observed. Such higher ground motion at a relatively farther distance may be attributed to the local site effects. The JMIU station is established in an area underlain with a thick sediment column above the weathered basement rock (Shukla et al. 2007 and2016). Although the PGA value at JMIU is the highest among all the sites, it was yet too feeble to result in damage in the area. We mention that PGA H (Horizontal Peak Ground Acceleration) is estimated using the square root of the sum of the square of the PGA on the NS and EW components. Similarly, PGV and PGD are estimated for the events by integrating the acceleration waveform. Felt reports of the ground shaking in the mainshock were received from people on the official website (www.seismo.gov.in) and APP (RISEQ) of NCS, and have been tabulated in Table 7. A majority of the felt reports were equivalent to Intensity II on the Modified Mercalli intensity scale (MMI) scale ( Figure 6). The two sites located at Rajendra Nagar and Lodhi Road reported relatively higher Intensity, viz., III and IV, respectively. However, the isoseismal map as estimated using a global empirical relationship (Hough 2012) depicts almost the entire NCT under the Intensity IV. Apparently, the entire region has experienced the ground shaking varying between intensity II -IV, which may be attributed to the local site effects (Shukla et al. 2016), types of buildings and construction practices.

Ambient noise
Ambient noise analyses of the continuously recorded ground motion data at NDI station (rock site) and UJWA station (soil site) for a 7-day period, prior to and during the prevailing lockdown condition due to unprecedented COVID-19 crisis, were   . We considered only the Z component records in the analysis as representative of the ground motion to determine the noise power density acceleration spectrum and is measured in dB that is referred to 1 ((m/s 2 ) 2 /Hz). We mention that all the seismic stations were equipped with tri-axial broadband velocity sensors with 120 s period. The power density acceleration spectrum for both the segments (prior to and during the lockdown) at rock (NDI) and soil (UJWA) sites are shown in Figure 7. The Peterson Low Noise Model (LNM) and High Noise Model (HNM) are represented in the dark grey color. The noise spectra in both the segments are well constrained between the LNM and HNM in the case of the rock site (NDI); however, it crossed the HNM in the case of the soil site (UJWA) prior to the lockdown period, indicating it is a noisy site. It is important to mention that the    noise spectrum of the UJWA site during the lockdown period falls well within the upper bracket of the standard noise model i.e. HNM. At both the sites the ambient noise levels were higher than the LNM, but they also were found to be relatively stable as evidenced by the probabilities. Figure 7 shows about 10 dB decrease in the noise level at both the rock as well as soil sites during the lockdown period in the high-frequency range between 1 Hz to 10 Hz. However, a slight decrease in noise level at lower frequencies around 0.1 Hz is observed in both the cases. We mention that such long period signals are mostly related to in the quasi stationary state (Kumar et al. 2012), which is described as approximately time independent out of equilibrium state such as oceanic waves and wind etc. Further, a drop in average ground displacement from 25 nm to 8 nm indicates a better recording of the seismic signals during the lockdown period, even at relatively noisy sites located in the urban agglomerations.

Conclusion
We have analyzed the broadband waveform and strong motion records of the April 2020 earthquakes, which shook the NCT Delhi and its surrounding regions. The epicenter of the event of 12 April 2020 (M L 3.5) was found to be located in the NCR close to the East district boundary of NCT Delhi. The aftershocks (M < 3.0) that occurred during 12-16 April 2020 were primarily located to the SW of the mainshock. The focal mechanism of the mainshock is estimated using a moment-tensor inversion approach, which reveals normal faulting with a strike slip component. Although the nodal planes of the event did not coincide with any of the existing identified faults, they were found consistent with the NE -SW trending lineaments mapped in the region. The past events along the major faults in Delhi and surrounding regions might have stressed the lineaments in the epicentral area, which probably reactivated due to the mainshock of 12 April 2020 causing small magnitude events. However, we suggest a detailed analysis to be undertaken to understand the causative source of these small magnitude events that occurred after the mainshock. The amplitude spectra of the S-waves indicate the seismic moment of mainshock to be 1.15 Â 10 14 N-m. The stress drop during the event is observed close to the lower range of the stress drop (i.e. $30 bars) of the stable continental region in India. We emphasize that the event was felt widely probably due to the reduced mobility in Delhi and surrounding region. Due to the prevailing lockdown situation, the majority of the people were confined to their houses in Delhi and nearby areas where the buildings are mainly taller and are mostly erected over thick soil cover. Hence, the event could be felt by many people in the region. The strong ground motion analyses reveal a maximum peak ground acceleration (PGA) to be 14.13 gals at the JMIU site for the mainshock (M L 3.5) that was felt in the region with maximum intensity IV at Lodhi Road to the SW of the epicenter. However, for the aftershock on 13 April 2020 (M 2.7) the maximum PGA was feeble (< 0.6 gals) and hence it could not be felt widely. The noise analysis of the ground motion prior to and during the nationwide lockdown due to the unprecedented COVID-19 suggests a significant reduction of noise ($ 10 dB) at rock as well as soil sites in the frequency band 1-10 Hz during the lockdown, which probably improved the ground motion seismic recording of the recent tremors of Delhi region. A clear peak of the P-wave in the mainshock record at the Latur seismic observatory, located about 1100 km away, indicates reduced noise level at the recording site. We attribute such clear recording of the phases at far distance to the silence caused due to country wide COVID-19 lockdown.