We studied the variations in spatial and temporal clustering of earthquake activity (during 2001–2013) in the Kachchh seismic zone, Gujarat, India, by precisely relocating 3478 events using a joint hypocentral determination (JHD) relocation technique, and high-quality arrival times of 21032 P- and 20870 S-waves.Temporal disposition of estimated station corrections of P- and S-waves suggests that the fluid flow in the causative fault zone of the 2001 Bhuj mainshock increased during 2001–2010, while it reduced during 2011–2013, due to the healing process associated with the perturbed Kachchh fault zone. We also estimated the isotropic seismic diffusivities from epicentral growth patterns, which are found to bemuch lower than those observed for reservoir-induced seismicity sites in the world. Finally, we analysed the spatial and temporal evolution of this earthquake sequence by solving the diffusion equation of pore-pressure relaxation caused by co- and post-seismic stress changes associated with earthquakes. The value of the isotropic diffusivity is estimated to be 100 m2/s for the Kachchh rift zone. This gives a higher permeability (after a lapse time of 14 years from the occurrence of the 2001 Bhuj mainshock) in comparison to those observed for other intraplate regions in the world. Our results suggest that the observed spatio-temporal migration of seismicity is consistent with the shallow (meteoric water circulationat 0–10 km depths) and deeper (metamorphic fluid and volatile CO2 circulation at 10–40 km depths) fluid flows in the permeable and fractured causative fault zone of the 2001 Bhuj earthquake.

Earthquake source parameters and crustal $Q_{0}$ values for the 138 selected local events of ($\hbox {M}_{\mathrm{w}}{:}2.5{-}4.4$) the 2001 Bhuj earthquake sequence have been computed through inversion modelling of S-waves from three-component broadband seismometer data. SEISAN software has been used to locate the identified local earthquakes, which were recorded at least three or more stations of the Kachchh seismological network. Three component spectra of S-wave are being inverted by using the Levenberg–Marquardt non-linear inversion technique, wherein the inversion scheme is formulated based on $\omega ^{2}$ source model. SAC Software (seismic analysis code) is being utilized for calculating three-component displacement and velocity spectra of S-wave. The displacement spectra are used for estimating corner frequency (in Hz) and long period spectral level (in nm-s). These two parameters play a key role in estimating earthquake source parameters. The crustal ${Q}_{0}$ values have been computed simultaneously for each component of three-component broadband seismograph. The estimated seismic moment (M₀) and source radius (r) using S-wave spectra range from 7.03E+12 to 5.36E+15 N-m and 178.56 to 565.21 m, respectively. The corner frequencies for S-wave vary from 3.025 to 7.425 Hz. We also estimated the radiated energy ($E_{S}$) using velocity spectra, which is varying from 2.76E+06 to 4.07E+11 Joules. The estimated apparent stress drop and static stress drop values range from 0.01 to 2.56 and 0.53 to 36.79 MPa, respectively. Our study also reveals that estimated $Q_{0}$ values vary from 119.0 to 7229.5, with an average $Q_{0}$ value of 701. Another important parameter, by which the earthquake rupture process can be recognized, is Zuniga parameter. It suggests that most of the Kachchh events follow the frictional overshoot model. Our estimated static stress drop values are higher than the apparent stress drop values. And the stress drop values are quite larger for intraplate earthquakes than the interplate earthquakes.

We measure source parameters for the 21 May, 2014 Bay of Bengal earthquake through inversion modeling of S-wave displacement spectra from radial–transverse–vertical (RTZ) components recorded at ten broadband stations in the eastern Indian shield. The average source parameters are estimated using estimates from seven near stations (within epicentral distances $\leq$ 500 km). The average seismic moment and source radius are determined to be 1.0$\times$ 10$^{18}$ N-m and 829 m, respectively, while average stress drop is found to be 76.5 MPa. The mean corner frequency and moment magnitude are calculated to be 1.6 $\pm$ 0.1 and 5.9 $\pm$ 0.2 Hz, respectively. We also estimated mean radiated energy and apparent stress, which are found to be 6.1$\times$10$^{13}$ joules and 1.8 MPa, respectively. We observe that mean $E_{s}$/$M_{o}$ estimate of 5.5 $\times$ 10$^{-5}$ is found to be larger than the global average for oceanic strike-slip events. This observation along with large stress drop and apparent stress estimates explains the observed remarkably felt intensity data of the 2014 event. The full waveform moment tensor inversion of the band-passed (0.03–0.12 Hz) broadband displacement data suggests the best fit for the multiple point sources on a plane located at 65 km depth, with a moment magnitude 6.4, and a focal mechanism with strike 318$^{\circ}$, dip 87$^{\circ}$, and rake 34$^{\circ}$.

We herein present new frequency-dependent coda-Q ($Q_{\rm{c}}$) relations ($Q_{\rm{c}}$=$Q_{0}f ^{n}$) (frequency ranges between 2 and 18 Hz) for three regions of the eastern Indian craton (EIC), viz., the Singhbhum Odisha craton (SOC) and the Eastern Ghat mobile belt (EGMB), comprising the Mahanadi basin and the Chotanagpur granitic gneissic terrain (CGGT). The frequency-dependent coda-$Q_{\rm{c}}$ relations are obtained through the single backscattering model for coda waves ($Q_{\rm{c}}$) of local earthquakes which are recorded on 15 three-component broadband seismograph stations in the regions. In this work, we pay special attention to test the lapse time ($t_{\rm{L}}$) dependency of coda-Q ($Q_{\rm{c}}$) estimates for the three regions. Lapse time signifies the sample area of the coda wave of the study region. Generally, the sample area increases with lapse time. To test the lapse time ($t_{\rm{L}}$) dependency, nine different lapse time windows ($t_{\rm{L}}$) from 10 to 90 s with 10 s interval are considered. On the ground of estimated poor correlation coefficients, only six lapse time windows ($t_{\rm{L}}$) from 40 to 90 s with 10 s interval are considered. Our results suggest more heterogeneity in EGMB than that of the SOC and CGGT region. Estimates of $Q_{0}$ and n for the three regions of EIC (SOC, EGMB and CGGT) are found to be consistent with the results of $Q_{0}$ and n for mildly active less heterogeneous seismic zones in different parts of the world. By assuming entirely intrinsic attenuation characteristics, actual hazard parameters, i.e., extinction distance and anelastic attenuation coefficients are also computed for the three regions. The extinction distance ($L_{\rm{e}}$) provides an idea of the distribution of scatterers in the lithosphere and anelastic attenuation coefficients signify the anelasticity of the medium, i.e., fluid movement and grain distribution. The estimate of extinction distance and attenuation coefficients suggests that for all three study regions, the upper mantle is relatively less heterogeneous and attenuation below 110–126 km depth is also comparatively lower. Coda Q indicates the degree of fracture and heterogeneity in the lithosphere related to seismicity. A higher estimate of $Q_{0}$ values in the Archaean SOC region and the Proterozoic CGGT region is found when compared with that of the sedimentary-rich EGMB. It can be inferred that seismically less active cratons in general comprise high $Q_{0}$ values, whereas the sedimentary-rich EGMB is more attenuative, characterised by a low coda $Q_{0}$ value. Moreover, it is found that the estimated $Q_{0}$ values for CGGT region are a little bit higher than that for the SOC region. This can be explained as a comparatively less disturbed and less heterogonous land mass that is present in the CGGT region as compared to the SOC region, which comprises different minerals, ore bodies, fault scarps and shear zones. The developed $Q_{\rm{c}}$ relation for the EIC region could be useful for the study of hazards and ground motion prediction.

Modelling of earthquake source locations and parameters infers seismogenesis of earthquakes. In this study, we modelled the earthquake source locations through hypocenter location algorithm using the difference in arrival time of P and S waves and source parameters through the Levenberg–Marquardt nonlinear inversion method using S-wave spectra. A total of 340 aftershocks of 2001 Bhuj mainshock (1.8$\leq M_{w}$ <4.3), which have occurred in Kachchh, Gujarat, India from January 2014 to January 2015, are located in this study. Out of 340 aftershocks, digital waveforms of 78 aftershocks (2.2$\leq M_{w}$ <3.9) are used for estimation of the earthquake source parameters. The results obtained from earthquake locations show two clusters of seismicity along the Kachchh Mainland Fault (KMF) and North Wagad Fault (NWF) and three felt events ($M_{w} =\geq 3.0$); one along the Katrol Hill Fault (KHF) ($M_{w} = 3.3$), two along the Banni Fault (BF) ($M_{w} = 3.0; 3.2$). The generation of these three felt events is attributed to the triggering mechanisms caused by the migration of Cuids or the stress pulse generated by the 20 MPa stress drop of the $M_{w} 7.7$ Bhuj earthquake. A marked concentration of events is noticed in 15–30 km depth range, which could be attributed to the presence of a mafic intrusive body, resulting in stress build-up for earthquake generation in this region. The results of source parameters; seismic moment ($M_{0}$), source radius ($r$) and stress drop ($\Delta\sigma$) vary from $1.86 \times 10^{12} \rm{to 3.2 \times 10^{15} N m}$, 146–262 m and 0.04–5.73 MPa, respectively. The maximum stress drop value is estimated to be 5.73 MPa at 24 km depth for the largest studied event of $M_{w} 3.9$. Large stress drops are confined to the 8–33 km depth range, which indicates the probable existence of the base of the seismogenic layer in this depth range. This observed large stress drops could be attributed to stresses induced by crustal maBc intrusive bodies and the presence of aqueous fluids in the lower crust below the region.

Unravelling the anisotropic behaviour of the upper mantle helps to shed light on its present and past deformation processes. In this study, we attempt to explore the seismic anisotropy prevailing within the upper mantle beneath the Kachchh rift zone through shear wave splitting analysis. We have measured the splitting parameters (e.g., fast axis orientation ($\Phi$) and delay time ($\delta t$)) using SKS/SKKS core refracted phases from 112 teleseismic events recorded at NGRI network in the Kachchh region, during 2006–2009 and 2013–2016. The ‘$\Phi$’ and ‘$\delta t$’ estimates vary from $\rm{N}34^{\circ}$ to $\rm{N73^{\circ}E}$ and 0.80 to 1.5 s, respectively. The average vector means of ‘$\Phi$’ and ‘$\delta t$’ for all the stations are found to be $\rm{N(58 \pm 10)^{\circ}E}$ and $(0.99 \pm 0.19)\,\rm{s}$, respectively. Measurements of 59 good SKS/SKKS splitting parameters from 112 earthquakes reveal that the upper mantle is highly anisotropic beneath the Kachchh rift zone with an average fast axis orientation of $\rm{N(58 \pm 10)^{\circ}E}$, which is deviated nearly ($\sim\rm{N18^{\circ}E}$) from the absolute plate motion (APM) direction ($\rm{N40^{\circ}E}$) of the Indian plate in a no-net-rotation reference frame. This deviation of fast axis orientation from APM direction may be attributed to the effect of Kachchh rift zone as well as the presence of structural imprints of the 65 Ma Deccan mantle plume in the study region. And the average delay time of $(0.99 \pm 0.19)\,\rm{s}$ is consistent with the global average (1 s) for continents. Furthermore, the modelled seismic layer thicknesses reveal that anisotropic sources beneath study region are associated with both the lithospheric deformation processes (e.g., 184 Ma African rifting, 88 Ma Madagascar rifting, 65 Ma Deccan mantle plume) as well as asthenospheric flows.

To provide reliable and quick estimation of magnitude for moderate to large earthquakes at regional distances, two magnitude relations specific to the peninsular shield have been proposed based on long-period magnitude ($M_{A}$) and energy magnitude ($M_{E}$), using broadband velocity data of 23 regional events recorded at 18-station seismic network in the state of Telangana and Andhra Pradesh, India. $M_{A}$ is estimated using amplitude of filtered (0.03–0.08 Hz) broadband velocity seismograms, while $M_{E}$ is estimated based on radiated energy using broadband velocity spectra. It is observed that $M_{A}$ for larger events with $M_{w}$ >7.2 saturates, whilst $M_{E}$ does not suffer from saturation even for larger events. Thus, it is apparent that these two magnitude relations can provide magnitude estimates without saturation for all moderate to large regional earthquakes, which, in turn, can provide a homogeneous catalogue for moderate to large regional Indian earthquakes. The data transmission from remote stations to the central server at CSIR-National Geophysical Research Institute (NGRI) is quasi-real-time since it is connected by GPRS and VSAT. Using the proposed region specific magnitude relationships it becomes possible to estimate reliable magnitudes for moderate to large regional Indian earthquakes ($M_{w} \leq 7.2$) within 30 min of the occurrence of an event.