Significant lateral heterogeneities in the velocity structures of the crust and upper mantle beneath the plateau are noticed. The lower velocities in the shallower layers are associated with the near surface geology. The velocity reduction near the surface may result from a combination of high sediment thickness, high pore pressure and active strain. In profile GH, the low velocity zone continues along the Oldham Fault and Barapani Thrust. In the uppermost mantle, a low seismic velocity and σ are quite prominent beneath the plateau. The uppermost mantle is not as seismically active as those in the lower crust. The occurrence of earthquakes in the upper crust inferred that high velocity zones are stress accumulators in the strong heterogeneity medium. However, seismic activity along these two faults is not noticed significantly during the study period, and it may be inferred that the movement along these faults might be generating compressional stress in the region, thus gyrating the local faults and lineaments to be active and generating seismic activity of smaller magnitudes at shallower depths. The high velocity bodies beneath the plateau, where mostly seismicity is located, indicate that the seismogenic zone that shows brittle behaviour is thicker in this area than other similar tectonics of the regions. A robustness test for this anomaly, shows a good recovery of the input structure. This plateau is situated over the buckled-up part of the Indian lithosphere (Raoof et al., 2017). At depths of approximately 35 to 50 km, there are narrow low velocities and σ below the plateau (Figs. 3-5). This could indicate that the top of the buckled-up part cracked and was partially undertrusted by low-velocity materials. The low velocity towards the northern part of the region may not put up much resistance to underthrusting leading to less seismic activity there. Within the Shillong Plateau region, N-S trending low Vp, Vs and high-Vp/Vs zones (Figs. 3 -5) could be associated with the Dudhnoi faut and Oldham fault and several lineaments present in this area (Bilham and England, 2001). It is also noticed that an approximately 200 km long E-W Dauki fault is well imaged, indicating high –Vp and -Vs towards the western region and low-Vp and Vs on the eastern sides (Figs. 3 and 4). This evidence of a subducting Indian plate is noticeably imaged as high velocities and low σ. It is striking that these seismic velocity structure variations in the crust are localized and isolated. The great earthquakes that occur in the upper crust suggest that a strong anisotropy is developed in response to the shear processes responsible for the occurrence of great earthquakes in that sector of the Shillong Plateau. The high velocity zones mark areas that are lacking in seismicity. Beyond 60 km, we have no better resolution of the assimilated seismic structures because of the reduced spatial distribution of crisscrosses among seismic rays at deeper depths (Fig. 8). Despite the high tectonic complexity of the region, some correlation may be highlighted between geological units and velocity structures. In some of the profiles, the high velocities and low Vp/Vs ratio penetrate the upper crust and dip northwards, where the thickness of the crust varies vigorously. Consequently, the low σ perturbations suggest that the low value could have been due to fewer pores and competent rock, which can store stress accumulation to a greater extent. On the other hand, the higher values of σ in the lower crust or in some patches may suggest relatively weaker rocks of the subsurface, which can undergo metamorphism processes at deeper layers under suitable pressure and temperature or the presence of aqueous pore fluids (Nakajima et al., 2001).This high σ ratio may have facilitated the process of brittle failure at depth by relaxing the seismogenic zones with the release of accumulated stress in metamorphic rocks. The seismic velocity profiles along the A-B C-D and E-F at depths of 5-10 km is conspicuously characterized by a low velocity zone, which is due to the presence of sedimentary sequences deposited in the marine environment. The vertical dispositions of both BT and OF are marked up to deeper layers through which a train of distinct low velocity zones emanates from the deeper layers towards the surface, which may have been behaving as a fluid on the time scale of deformation where earthquakes occur in the depth range of 35 km to 45 km. We believe that our results are more authentic in sense that we analysed the data collected from dense seismic networks, which in turn provided an opportunity to conduct fine 3-D seismic imaging of the region. Our σ estimates at these depth ranges also provide a good constraint on our interpretations. Furthermore, the OF is more or less flat at depth toward the aseismic semibrittle zone below the seismogenic layer (Figs. 3 - 5). Flatten in the mid crust region and detach at the top of the aseismogenic layer. Beneath the plateau, the thickness of the crust is over ~45 km around the Moho, and a high velocity zone obviously exists. In comparison OF and DF have a strong impact on the velocity structures in the middle and lower crust. As the crust thickens beneath the plateau, pressure increases can trigger metamorphic reactions at lower crustal levels; in particular, evidence for transformation to eclogite in the Indian plate provides support for theories about mass transfer from the lithosphere to the underlying mantle (Henry et al., 1997; Beck and Zandt, 2002; Monsalve et al., 2008). The low σ, high conductivity and earthquakes are consistent with the presence of fluids with high pore pressure in highly fractured materials. As Lemonnier et al., 1999 suggest, fluids can originate from dehydration reactions in the Indian crust under thrusts beneath the Himalaya. In this study Moho configuration is clearly resolvable. The Moho configuration obtained by receiver function analysis for the study region varies between 38 km and 40 km (Mitra et al., 2005).
The plateau is characterized by positive Bouguer and isotstaic anomalies, which imply denser and thicker crust beneath the plateau (Verma et al., 1976). Several faults are encountered in the complexity zones. As a result, crustal structures are complicated, and strong lateral heterogeneity exists in the crust and upper mantle. It is also interesting to note that very few earthquakes are located in the depth range of 30-60 km, which mainly occurred in the high Poisson’s ratio zone. This again suggests that earthquakes in this depth range are associated with ductile parts of underneath rock materials. We may interpret that the upper mantle materials act as jelly, as proposed in the “jelly sandwich” model of the seismogenic zone, which is anticipated by Chen and Molnar (1993) for continental crustal seismicity. Poisson’s ratio (σ) is more diagnostic of crustal and subcrustal rock properties and has a very close bearing on the seismogenic strengths and material properties of insitu rock beneath the study area and its adjoining areas. It is important to mention a high σ suggests weak and less competent rock materials associated with fractures/cracks. The fractures and cracked volume of rocks in the tectonically complex zone may be partially or fully saturated at the subsurface layers, which require detailed geotectonic information about the area for compressive interpretation.
The plateau is associated with several seismotectonic faults, which may have caused fracturization of underlying rocks. This region where the great 1897 Shillong earthquake occurred indicated that the earthquake in the region contributed to the relaxation of high ambient stresses were locally concentrated within rheological heterogeneities and the relatively competent and crystalline parts of rock materials beneath the study area. This region contains highly fractured and faulted granitic bodies at the surface (GSI, 2000).The Shillong Plateau may accommodate significant strain due to the compression of faults from all sides.
A closer examination of the distribution of the P- and T -axes suggested that although the pattern seems quite variable, there is a consistent picture of an overall northeast and southwest compression reflecting a contraction with the Indian plate at a rate geodetically estimated to be 4 ± 6 mm/yr (Bilham and England, 2001). Fig. 6 shows a prevailing strike-slip environment in the Shillong region with both P and T-axes oriented subhorizontally. The distribution in Fig. 7 also inferred a more variable pattern along the OF, BF and KF faults, perhaps owing to the strong heterogeneities along faults (Fig. 2a, b). The seismotectonic setting of the Shillong plateau is summarized in Fig. 7. The pop-up tectonics is confirmed by present seismic tomography study. This plateau shows high-velocity bodies < 40 km depth where 76% seismicity is observed, indicating a seismogenic zone. The brittle behaviour is thicker in this area than what is observed by many researchers in other parts of NE India.
Detailed information on subsurface structures is essentially required as inputs for evolving a comprehensive model for earthquake risk mitigation for a region. The generation of a shallow crustal earthquake could be controlled by a deep process in the lower crust and upper mantle. From this point of view, it is vital to investigate the detailed structure and processes of the lower crust and upper mantle to clarify the seismogenesis and reduce earthquake risk. It is insufficient to refer only to the surface description of spatial features to predict the seismic potential of a region. Large damaging earthquakes occurred in décollement zones that leave very little surface evidence of faulting that can be used to identify past earthquakes. We infer that the use of detailed 3-D seismic tomography may offer potential information on Pop - Up tectonics beneath the Shillong Plateau to unravel what and how the genesis of such large damaging earthquakes caused. This study may help in evolving a comprehensive earthquake hazard mitigation model for a region.
We present robust P- and S-seismic wave velocity images of the Shillong Plateau, which shed important insight into the seismogenic and rheological behaviours of the study region. The uppermost crust is clearly revealed as a low V and low Vp/Vs layer down to a depth of 10 km due to fractures and cracks and the low permeability of the layer. The geological faults are well imaged in the cross sections. The DF fault northwards and is associated with high-velocity zones. We noticed a high velocity at the depth of 20-38 km. The thickened crust across the collision zone could be associated with a pressure increase that can trigger metamorphic reactions at lower crustal levels. It is inferred that some zones appear favourable to accumulate differential strain due to the prevalence of contrasting velocity and Vp/Vs ratio at the boundary layers, which can generate micro earthquakes once the shearing strength of rocks is suppressed by the differential stain accumulation.