SEISMIC TRAVELTIME ANALYSIS AND ATTENUATION STUDIES FOR INDIAN REGIONS

Abstract

Delhi and its surrounding region have suffered earthquakes since historical times. Understanding of seismicity of these areas is essential as these areas fall in the seismic zone IV and are geologically confined to the Delhi Fold Belt. A total of 376 local earthquakes occurred between January 2001 and December 2012 and recorded by the Seismic Telemetry Network operated by India Meteorological Department (IMD) in and around Delhi region between latitude 270N to 30.50N and longitude 750 E to 790E were located. Seismicity in Delhi region shows occurrence of earthquakes along various portions of the fault systems of the Delhi Fold Belt. Using five initial velocity models, a minimum one dimensional (1D) velocity model is estimated from the travel-time inversion of local earthquakes. For incorporation of data for inversion, only those earthquakes are chosen for each of which a minimum of four P- and four S-phase readings are available. After relocating earthquakes with a 1D velocity models we found that most of the earthquakes beneath Delhi and its surrounding regions occur at depths between near surface and 50 km below the mean sea level. Estimated Moho depth is about 45 km. The estimated body wave velocity model shows that the P wave velocity and S wave velocity in the crust varies from 5.8 km/s to 6.99 km/s and 3.32 km/s to 3.75 km/s respectively. The velocity variation of P wave shows that the crust is made up of three distinct layers. The Vp/Vs ratio as observed from present study shows a variation from 1.68 to 1.76 except the depth interval 45-50 km where the same ratio is 2.0. This shows that a fluid-filled fractured zone is present just above the Moho. Attenuation of seismic wave energy from Delhi and its surrounding regions has been estimated using coda of 190 local earthquakes from the same network. Estimated quality factor (Qc) values are found to be strongly dependent on frequency and lapse time. Frequency dependency of Qc has been estimated from the relationship Qc(f) = Q0f n for different lapse time window lengths. Q0 and n values vary from 73 to 453 and 0.97 to 0.63 for lapse time window lengths of 15 s to 90 s respectively. Average estimated frequency dependent relation is 𝑄𝑐 = 135 ± 8𝑓0.96±0.02 for the entire region for a window length of 30 s, where the average Qc value varies from 200 at 1.5 Hz to 1962 at 16 Hz. These values show that the region is seismically active and highly heterogeneous. To investigate if there is a spatial variation in attenuation characteristics in this region, the entire study region is divided into two sub-regions: area surrounding north-eastern part of Delhi Haridwar Ridge marked as region 2 and area surrounding south-western part of Delhi Haridwar Ridge marked as region 1, according to the geology of the area. It is observed that at smaller lapse times both regions iv have similar Qc values. However, at larger lapse times the rate of increase of Qc with frequency is larger for Region 2 compared to Region 1. This is understandable, as it is closer to the tectonically more active Himalayan ranges and seismically more active compared to Region 1. The difference in variation of Qc with frequencies for the two regions is such that at larger lapse time and higher frequencies Region 2 shows higher Qc compared to Region 1. For lower frequencies the opposite situation is true. This indicates that there is a systematic variation in attenuation characteristics from the south-west (Region 1) to the north-east (Region 2) in the deeper part of the study area. This variation can be explained in terms of an increase in heat flow and a decrease in the age of the rocks from south to north. Records of 77 local earthquakes with a magnitude range 2.8–4.5, recorded using 8 Broadband stations operated by IMD from 2011 to 2013, were analyzed to estimate the attenuation characteristics of the NE India region using coda of the earthquakes. It is seismically one of the most active zones in the world. Estimated Qc values are strongly dependent on frequency and lapse time. The entire study region is divided into three subregions: Shillong plateau, Mikir hills and central part of Indo-Burma Ranges to investigate if there is a spatial variation in attenuation characteristics in the region. Estimated average frequency dependencies of coda wave attenuation for 30s window length are 𝑄𝑐 = 135 ± 7𝑓0.99±0.03 for Shillong plateau, 𝑄𝑐 = 109 ± 7𝑓1.10±0.03 for Mikir hills and surrounding Brahmaputra River Valley and 𝑄𝑐 = 90 ± 2𝑓1.04±0.02 for the central part of Indo-Burma Ranges. The depth variations of the Qc, Q0 and n values are also examined. It is observed that the rate of increase of Q0 with depth is not uniform for all the sub-regions. The central part of Indo-Burma Ranges has the smallest Q0 and the largest n values at all depth levels amongst the three sub-regions. These results indicate that the central part of Indo-Burma Ranges is the most attenuative, seismically active and heterogeneous in nature. However, this region has smaller Qc values than the other two sub-regions for all window lengths up to the 6 Hz. This means at lower frequencies the subsurface beneath this area is more attenuative compared to the other two sub-regions. Similar trends are observed at 8, 10 and 12 Hz, up to 45 s window lengths. For window lengths ≥ 55 s, the central part of Indo-Burma Ranges has higher Qc values at 10 and 12 Hz compared to the Shillong plateau. Qc values are lower for Shillong plateau compared to the other two regions for window length  55s at 10 and 12 Hz, which corresponds to depth levels  90 km. Such a complicated variation in Qc values is a manifestation of complex nature of tectonic regime in NE India. The frequency-dependent attenuation of P and S waves has been estimated using the extended coda normalization method for both Delhi and its surroundings, and NE India. A v total of 84 local earthquakes for Delhi with a magnitude range of 2.0 to 4.2 recorded by five broadband stations operated by IMD were analyzed for this study. For NE India 82 local earthquakes with a magnitude range of 2.8 to 4.5 recorded by six broadband stations operated by IMD were used. The P, S and coda-wave spectra are analyzed for 5 seconds windows at the central frequencies of 1.5, 2, 3, 4, 6, 8, 10 and 12 Hz. Both Qp and Qs values increase with increasing frequency for Delhi and NE India. Qp increases from 60 at 1.5 Hz to 493 at 12 Hz and Qs increases from 133 at 1.5 Hz to 1250 at 12 Hz for Delhi region whereas for NE India Qp and Qs increases respectively from about 50 and 116 at 1.5 Hz to 496 and 1519 at 12 Hz. The obtained relations for P and S wave are 𝑄𝑝 = 42 ± 2𝑓0.99±0.02 and 𝑄𝑠 = 97 ± 19𝑓1.04±0.09 respectively for Delhi region whereas these relations for NE India are 𝑄𝑝 = 31 ± 3𝑓1.12±0.02 and 𝑄𝑠 = 86 ± 12𝑓1.16±0.06. For the NE India region, Qp0 and Qs0 values are less compare to Delhi and its surrounding region. It indicates that NE India is more attenuative compare to Delhi region at 1 Hz. The results obtained for average Qp and Qs in this study for both region show good agreement with those for other studies carried out for tectonically active regions, with high n (~1.0) values and low Q0 (<200) values, which are characteristic of active regions in terms of seismicity and tectonics. The obtained ratio of Qs/Qp for both the regions is >1 for the entire analyzed frequency range which indicates the presence of a high degree of subsurface heterogeneities in the study region. The obtained average value of Qs/Qp for NE India is 2.87 whereas this ratio is 2.43 for Delhi region which indicates that the study regions are partially saturated with fluids.