Individual, stacked and gather radial P-RFs (Figs. 2a-l, S2-S12) show a clear negative arrival at 2–3 s after direct P (corresponding to the low velocity MHT at 10–20 km depths), and two clear positive arrivals at 4–7 s (corresponding to first Moho interface at 35–55 km depth) and 7–9 s (corresponding to the second deeper Moho interface at 55–65 km depth). Similar observations were also made for the Lhasa, Tibet, by Li et al.29. Their study has detected two positive arrivals on radial PRFs at 7 and 10 s, respectively, referred to as a “Moho doublet” by Kind et al.12. Yuan et al.11 interpreted the ~ 7s phase (corresponding to 60 km depth) as the base of the upper crust or top of the under-thrusting Indian lower crust while the ~ 10s phase (corresponding to 80 km depth) has been interpreted as the Moho phase30. Thus, our modelled positive peaks of Ps phases at 4–7 s and 7-9s suggest the presence of a Moho doublet below the UK Himalaya, which could be interpreted as the evidence for the broken Indian Moho due to the huge indentation of the Indian plate onto the Eurasian plate at 55 Ma. This is further evidenced by the relatively more deformed crust in the north of the study area as marked by red dotted areas in Figs. 4b-f, 6b-i. Additionally, there could be densification and subsequent delamination of the Indian lower crustal material into the mantle12.
Our CCP images along five NE-SW profiles very clearly map large lateral heterogeneities in the crustal structure across the Himalayan collisional front, showing a highly folded and deformed crust-mantle boundary below the UK Himalaya (Figs. 4a-f). The mapped Moho structure reveals a strong topography locally dipping toward east i.e. perpendicular to the direction of the Indian plate motion. This observation supports the idea that the lower crustal deformation in the Lesser UK Himalaya is decoupled from the underlying Indian mantle lithosphere as also observed in Lhasa through the imaging of P- and S- receiver functions by Li et al.29. A Moho doublet is also modelled below the Lhasa as modelled by us here below the Lesser UK Himalaya12,29. The presence of two clear Moho phases (Pm1 and Pm2) on the gather radial PRFs at 40 out of 52 stations further supports the idea of the presence of Moho doublet underlying the UK Himalaya (Figs. 3a-p; see Supplementary Figs. S8a-p, S9a-p).
Our all five NE-SW profiles (i.e. AA’, BB’, CC,’ DD’ and FF’ as shown in Fig. 4a) begin in the Siwalik Himalaya, pass through Lesser Himalaya and end in the region just north of the Vaikrita. Thus, they do not go through Tethyan and Tibetan Himalayas. Along AA’ profile, crustal thickness decreases from 55 km in the SH to 42 km in LH and then to 35 km below the region just north of VT where a double Moho is also detected below the main rupture zone of the 1803 Mw7.6 Uttarakhand and 1991 Mw6.8 UK events (Fig. 4b). Several intra-crustal conversions and ductile lower crust are also being imaged. A low velocity MHT (10–20% drop in Vp and Vs, 10–15% increase in Vp/Vs) between 10 and 20 km depth is also mapped where the 1991 UK event also occurred (Fig. 4b). Moreover, dual Moho boundaries are evident at the depths of 33 and 50 km in the region outer lesser Himalaya (OLH) and VT (i.e. below the hypocentral zone of the 1803 Uttarakhand and 1991 UK main events). Along BB’ profile, the crustal thickness decreases from 54 km in the SH to 35 km below the region just north of VT where a double Moho is detected. White dotted lines mark the MHT and Moho (Fig. 4c). Along CC’ profile, the crustal thickness reduces from 52 km in the SH to 35 km in the LH where a double Moho is also detected. And, then it increases toward north to 56 km depth (Fig. 4d). Along DD’ profile, the crustal thickness decreases from 52 km in the SH to 42 km in the LH and then it shows a marked undulation with a steep dip toward north below the region just south of the MT. Finally it reaches 56 km below the region north of VT (Fig. 4e). Along FF’ profile, the crustal thickness increases from about 30 km in the SH to 50 km in the ILH where a double Moho between 30–50 km is also mapped (Fig. 4f). And, then it dips toward the north probably showing the subducted Indian plate.
Our two WNW-ESE profiles (i.e. GG’ and HH’ as shown in Fig. 5a) run parallel to the strike of the Himalayan collisional front, thus, they pass through all different kinds of geological formations in the UK Himalaya. Along with GG’s profile, the crustal thickness increase from 40 km in the western end of the UK Himalaya to about 50 km in the region between MT and VT and then it decreases to 36 km below the region between the epicentral locations of the 1803 Uttarakhand, 1991 Uttarkashi (UK) and 1999 Chamoli (CH) events where a double Moho is also mapped (Fig. 5b). Afterwards, it increases to 48 km below the epicentral zone of the 1999 CH event and then it decreases to 42 km in the region just west of the 1999 CH MS location and then it fluctuates between 35–45 km beneath the remaining eastern part of the UK Himalaya (Fig. 5b). Along with HH’s profile, the crustal thickness reduces from 48–50 km in the SH to 45 km below Tehri and it increases to about 60 km below the region of new Lansdowne klippe where a double Moho at 40–60 km is also mapped (Fig. 4c). Recently, the H-K stacking of PRFs has shown a Moho depth of 43 km at Lansdowne31 while a Moho depth of 48 km has been modelled through the joint inversion of PRFs and group velocity dispersion data of Rayleigh waves6. Such a variation in Moho depth estimates at Lansdowne could be attributed to the presence of a double-Moho structure as revealed by our present study (Fig. 5c). Then it reduces to 40 km in the ILH and it further reduces to 35 km below the Almora klippe and then it gradually increases to 40 km at the eastern end of the UK Himalaya. From the CCP image along with the GG' profile, a low-velocity MHT is also mapped at 10–20 km depths on which both 1991 and 1999 events occurred (Fig. 5c). From the depth-sections of PRF image, and Vp, Vs and Vp/Vs tomograms along AA’, CC’, and GG’ profiles (Figs. 5–6), we also observe that the hypocentral zones of the 1803 Uttarakhand, 1991 UK and 1999 CH are characterized by large negative normalized PRF amplitude, low Vp, low Vs and high Vp/Vs indicating the probable presence of aqueous/metamorphic fluids within the MHT at 10–20 km depths, which might be providing triggering effect for nucleating earthquakes on the main Uttarakhand rupture plane (MURP) on the MHT.
In the UK Himalaya, the H-K stacking of radial P-RFs has shown marked lateral variations in Moho depths and crustal Vp/Vs values along the strike of the Himalayan collisional front, implying strong changes in the Moho geometry and properties31. Similar changes in crustal thickness and composition have also been observed in Lhasa, Tibet, by Kind et al.12. And, they interpreted this sudden change in Moho depths and Vp/Vs values in terms of strong changes in Moho topography beneath the Lhasa, Tibet. Our modelling maps a strong Moho topography with an eastward dipping along the Himalayan frontal arc (Figs. 4b-f, 5b-c, 6b-c) with marked local changes in Moho depths. This dip of Moho topography toward the direction perpendicular to the convergence direction of the Indian plate suggests that the deformation in the lower crust is decoupled from the upper mantle of the Indian plate, as also observed in Lhasa, Tibet29.
The most significant finding of our modelling is the detection of a distinct double Moho structure (crustal thinning) below the hypocentral zones of the 1803 Uttarakhand, 1991 UK and 1999 CH events, which might play a key role in accumulating large crustal stresses/strains on the MHT resulting in vulnerable locales for the nucleation of moderate to large earthquakes. The presence of two clear Moho phases (Pm1 and Pm2) on the gather radial PRFs at 40 out of 52 stations further supports the idea of the presence of Moho doublet underlying the UK Himalaya (Figs. 3a-p; see Supplementary Figs. S8a-p, S9a-p). Similar nature of the PRF has also been observed in Lhasa, Tibet, where a Moho doublet has also been modelled through the CCP stacking of radial PRFs by Li et al.29. The depth sections of the PRF image and velocity tomograms suggest a low-velocity zone (large negative normalized PRF amplitude, 10–20% drop in Vp and Vs, 10–15% increase in Vp/Vs) on the MHT coinciding with the hypocentral zones of the 1803 Uttarakhand, 1991 Uttarkashi and 1999 Chamoli main events (Figs. 5b,d,f,h, 6b-g). Thus, we can infer the presence of metamorphic/aqueous fluids in the hypocentral zones of the above-mentioned three M ≥ 6 events suggesting that these earthquakes were triggered by the fluid flows / high pore-fluid pressure within the MHT. Our modelling also detects a ductile lower crust where only 10% of earthquakes are occurring, which suggests less accumulation of strain energy within the ductile lower crust.