Introduction

The crust-mantle boundary (Moho), lithosphere-asthenosphere barrier (LAB), 410-km (d410), and 660-km (d660) discontinuities are the Earth's first-order discontinuities1. A substantial change in velocity and density (acoustic impedance) across the Moho and LAB boundaries has been imaged below practically everywhere in the world using receiver function imaging or tomography2,3. The top and bottom of the Mantle Transition Zone (MTZ) are defined by deeper Mantle limits viz., 410-km and 660-km discontinuities, respectively4. Seismic velocity jumps occurred at depths of 410 km and 660 km due to phase transitions from olivine to -spinel, -spinel to bridgmanite, and ferropericlase5. Helffrich and Wood6 demonstrated that at the d410 discontinuity, olivine transforms into wadsleyite, and at the d660 boundary, ringwoodite transforms into bridgmanite and magnesiowüstite, implying that the MTZ discontinuities are related with olivine phase transitions in the mantle. Thus, the depth and type of these discontinuities have been observed to be changed by the altered mantle above and inside the MTZ as a result of any geodynamic activity. The MTZ narrows in the presence of mantle plumes, whereas cold slabs thicken it. Continental slabs over the MTZ will raise d410 and d660. However, dehydration melting triggered by an ascending ambient mantle from the high-water-solubility transition zone to the dry mantle above d410 may result in unique behaviour7. The temperature and compositional properties of the upper mantle, as well as the development of the Himalayas, can be explained using receiver function imaging of d410 and d660 lateral fluctuations. LVL-410, a distinct low-velocity layer over the 410 discontinuity, has been observed in various sites on Earth, indicating that it is common. Toffelmier and Tyburczy8 modelled a marked conductivity layer at the 410 discontinuity, which supports LVL-410. The hydrous MTZ’s partial melting has been linked to the LVL-410, supporting the transition zone water filter hypothesis that the 410-km discontinuity dehydrates, as observed in various parts of the world7,9,10,11,12.

Because receiver functions are very sensitive to changes in acoustic impedance (a product of velocity and density) across interfaces, these major barriers have been mapped globally by examining and simulating seismological conversion phases along these discontinuities. These investigations also discovered relief of less than 20 km and 30–40 km at 410-km and 660-km discontinuities13. According to the IASP91 velocity model14, the global average arrival times for 410 and 660 km discontinuities are 44 and 68 s after direct P-arrival on the radial PRF, respectively, which corresponds to 67° distances. Several studies in India have identified these discontinuities14,15,16. Oreshin et al.16 demonstrated that difference timings measured at numerous Indian sites varied from 24.0 s in the south to 25.5 s in the Himalayas near the Main boundary thrust. The differential time between the 410-km and 660-km discontinuities in the Dharwar craton is 24 s, which is equivalent to the average differential arrival time in a homogeneous flat earth model, implying less MTZ thickness fluctuations below the craton17. Lessing et al.18 proposed that the LVL-410 is beneath the Eastern Indian craton in the Indian Ocean.

When India collided with Asia between 50 and 55 million years ago19,20, geodynamic processes formed the Tibetan-Himalayan orogen. Since the Neotethys Ocean closed, the collision has resulted in crustal shortening, subduction, and material ejection over 3600 km length. North–south crustal shortening is responsible for 1400–1700 km of this convergence21,22, with deep mantle subducted slabs accounting for a significant amount. Several seismic tomographic studies have shown the downwelling Indian lithosphere, including surviving Tethyan slabs in the deep mantle23. Replumaz et al.24 believe that continental subduction in the Himalaya could account for around 45% of Asian convergence. Tomographic imaging has discovered Indian slabs at MTZ depths beneath Central Tibet25,26,27,28,29,30. The loss of subducting mantle lithosphere in Hamalaya has been linked to delamination and convective removal caused by gravitational instability31,32,33. Over the last two decades, hundreds of seismic stations have been constructed in the Himalayan collision zone to conduct scientific research. A heated mantle34,35,36 and subducted slabs with a thicker MTZ have been detected in different parts of the Tibet-Himalaya collision zone. Several profiles in central Tibet reveal a downwelling Indian plate25,37. P-wave triplicated waveform analysis of a Himalaya-central Tibet profile38 revealed a stagnant slab in the MTZ with a high P wave speed of d660. Tseng and Chen39 discovered no high Vs atop d660 with the same profile as central Tibet's high Vp. Vp and Vs anomalies indicate anhydrous, unconnected mantle lithosphere. A significant deepening of d410 and d660 in receiver function pictures under central and eastern Tibet revealed an anomalous zone with high temperature and low mantle velocity35. They also discovered low-angle subduction in the Indian lithosphere beneath the Himalayan-Tibet collision zone. Current receiver function images28 and P wave seismic triplication studies40 reveal a cold and hard subducted Indian lithosphere that elevates d410 and d660 beneath the central Tibetan terranes. Due of the hydrous partial melt and high temperature, high-resolution triplicated P waveforms revealed a low velocity layer on top of d410 and a dispersed MTZ10. Seismological investigations using techniques such as seismic tomography and receiver function analysis have revealed important information about the structure and behaviour of the mantle transition zone beneath several sectors of the Himalayan orogeny41,42. Tseng and Chen39 and Kumar et al.42 investigated the mantle transition zone beneath the Tibet and north-western Himalaya, respectively, revealing significant variations in seismic velocities and discontinuity depths, which help to reveal the intricate tectonic processes underlying the collision of the Indian and Eurasian plates. Saikia et al.36 studied the nature of MTZ below the eastern Himalaya, using the analysis and imaging of the P-receiver functions to investigate the mantle structure. Similarly, Chamoli et al.43 studied the Garhwal-Kumaun Himalaya, using deep seismic reflection methods to investigate the crust and upper mantle structure. These investigations have shown differences in the thickness and composition of the crust and upper mantle throughout distinct portions of the Himalayan orogen. Furthermore, Kumar et al.44 conducted research on the upper mantle velocity structure beneath the north-western Himalaya, offering light on the tectonic processes that drive crustal deformation and mountain development. Dubey et al.45 used advanced seismic tomography methods to map the detached slabs of Indian lithosphere beyond 600 km beneath the Burmese Arc. Their findings revealed the eastward escape of Tibetan-lithospheric material between the Eastern Himalayan Syntaxis and Sichuan Basin. This observation calls into question earlier subduction theories and emphasises the Himalayan tectonic regime's dynamic nature. Furthermore below the Sikkim Himalaya and surrounding regions, seismic anomalies observed within the MTZ point to the presence of old subducted oceanic crust, providing important information about the tectonic development of the Himalayan orogeny42. Rai et al.46 found that the depth of the d410 discontinuity is typical beneath Precambrian terrains and thickened by ~ 10–15 km in the Ganges basin and the Himalaya, indicating gradual cooling or thickening of the Indian lithosphere towards its northern boundary. However, the upper mantle discontinuities and MTZ are poorly known due to limited seismic ray coverage beneath the research region and a scarcity of seismic data from the Himalaya and its foredeep. A dense coverage of numerous places is required to produce a coherent depiction of the upper mantle discontinuities beneath India and Tibet.

The Council of Scientific Industrial Research—National Geophysical Research Institute, located in Hyderabad, India, has been conducting seismic monitoring in the Uttarakhand Himalayan region since 2017. This monitoring is carried out using a comprehensive network of 76 three-component broadband seismographs47. This network has provided high-quality 3-component digital broadband waveforms for around 1000 teleseismic events with magnitudes ranging from Mw5.5 to Mw8.0 (47; Fig. 1a–c). These events occurred between 2017 and 2023 and had epicentral distances between 30° and 90°. The waveforms were used to construct radial and transverse P-receiver functions (PRFs). This study utilises the Common Conversion Point (CCP) stacking method and depth migration of stacked radial PRFs to image the mantle discontinuities at depths of 410-km and 660-km beneath the Uttarakhand Himalaya. An effort has been made to interpret the modelling results in relation to active geodynamic processes occurring under the region.

Figure 1
figure 1

(a) Station location map of the Kumaon—Garhwal (KG) Himalayan region. Filled red triangles mark the location of broadband stations. Two medium size blue-filled circles mark the epicentral locations of the 1991 Mw6.8 Uttarkashi and 1999 Mw6.5 Chamoli earthquakes. The solid black line represents major faults. MT Munsiari Thrust, VT Vaikrita Thrust, MBT Main Boundary Thrust, MFT Main Frontal Thrust, RT Ramgarh Thrust, MHT Main Himalayan Thrust. SH, LH, AK, LK and MZ mark Siwalik Himalaya, Lesser Himalaya, Almore klippe, Lansdown Klippe and MCT zone, respectively. (a) is generated using the Generic Mapping Tool (GMT) software version 661(https://doi.org/10.1029/2019GC008515). The elevation data used in generating GMT plot is obtained from the open source Digital Elevation Model (DEM) (https://asterweb.jpl.nasa.gov/gdem.asp). Locations of five NE-SW (viz., AB, CD, EF, GH and IJ) are shown by filled thick white lines) while four WNW–ESE profiles (viz., KL, MN, OP and QR) are shown by filled thick black lines. The inset shows Indian subcontinent where our study area is marked by a red square area, (b) Tectonic depth cross-section (Valdiya, 1980) across the NE-SW CD profile, whose location is shown in (a), and (c) Epicentral plot of 133 teleseismic events of Mw 5.0–7.8, whose broadband data from the Uttarakhand network, are used for our P-receiver function study. A filled red triangle marks the center of our network (Lat. 79°, Long. 30°).

Full size image

Geology and tectonics of the area

Four north-dipping thrust fault systems—the Himalayan frontal thrust (MFT), main Boundary thrust (MBT), main Central thrust (MCT), and South Tibetan detachment (STD)—divide the Himalayan geology into four major geological provinces: the Siwalik Himalaya (SH), Lesser Himalaya (LH), Higher Himalaya (HH), and Indo-Tsangpo suture zone (ITSZ). The MFT divides the SH and Indo-Gangetic Plain. Siwalicks in the SH are mostly middle Miocene to late Pleistocene48. MCT limits LH's northern boundary, while MBT limits its southern boundary (Fig. 1a,b; also Supplementary Fig. S1). THE STD distinguishes HH and ITSZ, while the MCT distinguishes HH and LH. Between the MCT and STD is central crystalline, while between the STD and ITSZ is Paleozoic Muth granite, Nilgiri limestone, Kanchan Shale, and ophiolites49. Indian plate subducts north of ITSZ. The upper Cretaceous Dras Island complex north of ITSZ suggests an Island arc and Indian plate subduction50,51. MCT zone is between Munsiari thrust (MT) and Vaikrita thrust (VT)50,52. At mid-crustal depths, all significant thrusts merge into a north-dipping low angle plane known as the main Himalayan thrust (MHT) (Fig. 1b), where the India plate's northward convergence creates the highest strain energy, which is released during earthquakes of varying magnitudes. The Delhi-Haridwar ridge (HDR), Faizabad ridge (FZR), and Munger-Saharsa ridge have also split the Himalayan frontal arc (see Supplementary Fig. S151). According to several research, the DHR extends below the Higher Himalaya51,53.

Earthquake data and computation of P-receiver functions

The seismic network of CSIR-NGRI consisting of 76 three-component broadband seismographs (Fig. 1a) recorded several hundreds of teleseismic events from 2017 to 2023, which are of very good quality as shown by the background noise spectra at different stations by Cook et al.47. We employed CCP stacking of radial PRFs and network information to measure thicknesses of 410-km discontinuity (d410) and 660-km discontinuity (d660) in the Uttarakhand Himalaya. High-resolution 24-bit Reftek-130 recorders, broadband sensors, and GPS time tracking are included at each station. Stations gather 50 samples/second broadband data continuously. The network recorded 1000 good teleseismic events of moment magnitude (Mw) ranging from 5.5 to 8.0 (back azimuth between 38° and 309° and ray parameters from 0.047 to 0.077 s/km; Fig. 1c), allowing us to compute 2000 radial and 2000 transverse P-receiver functions (PRFs) using 200 iterations of Ligorria and Ammon54’s time domain deconvolution. After extracting P-waveforms from 3-component ZNE seismograms spanning -5 to 90 s, we employ the SAC software transfer command to conduct instrumental correction using poles and zeros. Back-azimuth and SAC software rotate windowed ZNE seismograms to RTZ. Finally, we deconvolve Z from R and T using Ligorria and Ammon54’s time domain deconvolution to calculate the radial and transverse PRFs (see Supplementary Figs. S2–S4). The observed smaller amplitudes on the transverse indicate insignificant effect of the anisotropy and dipping Moho structure on our radial PRFs, which are used for modelling d410 and d660 phases. PRFs are estimated with a Gaussian width factor (‘a’) of 1.5 (f = 0.75 Hz). We also selected radial PRFs that recreated 90% of the signal energy on the radial component when convolved back with the vertical trace but had negligible amplitudes on the corresponding transverse PRFs (please see Supplementary Figs. S2–S4).

Using 3000 three-component waveforms from 52 of 76 broadband stations and the above criteria, we can obtain 2000 good individual radial PRFs (with minimum transverse PRF amplitude). The transverse PRFs that correspond to our chosen radial PRFs have significantly less amplitude fluctuation with back-azimuths, indicating that anisotropy and Moho dipping have less impact on our PRFs (please see Supplementary Figs. S2–S4). Radial PRFs also contain crustal multiples (PpPms and PpSms + PsPms) from the crust-mantle boundary (see Supplementary Figs. S2–S4). The arrivals of d410 and d660 conversions (marked by thick green dotted lines) on the stacked radial PRFs at 53 stations along two WNW-ESE profiles are shown in Fig. 2a–d. On the stacked radial PRFs, d410 and d660 conversions arrive at 42.5–46 s and 69–72 s, respectively. Negative arrivals from LVLs atop d410 and d660 mantle discontinuities are shown by thick dotted green lines in Fig. 3a–e.

Figure 2
figure 2

Time vs distance plots of radial PRFs along two NNE-SSW profiles (a) OP and (b) QR, Green dotted lines mark the arrivals of 410-km (d410) and 660-km (d660) conversions while black solid line marks the average line of arrivals of conversions from d410 and d660. This figure is generated using the Funclab software version 1.8.259 (https://doi.org/10.1785/gssrl.83.3.596) and MATLAB software version 9.4.0.813654 (R2018a).

Full size image
Figure 3
figure 3

Depth-migrated sections of stacked radial PRFs along three NE-SW profiles (a) CD, (b) GH, (c) IJ, and two WNW-ESE profiles, (d) OP, and (e) QR. Black thick doted lines mark the positive peaks associated with 410-km and 660-km Mantle discontinuities, respectively. Thick black dotted lines mark the low-velocity layer atop 410-km boundary. LVLs atop d410 and d660 are shown by LVL-410 and LVL-660 and marked by thick green dotted lines. This figure is generated using the Funclab software version 1.8.259 (https://doi.org/10.1785/gssrl.83.3.596) and MATLAB software version 9.4.0.813654 (R2018a).

Full size image

Results and discussions

We present here results from the analysis of the Mantle Transition Zone (MTZ) structure beneath the Uttarakhand Himalaya, through the analysis and migration of stacked radial PRFs and CCP images along five NE-SW and four WNW-ESE profiles across the region. Our modelling results are presented in the following:

Figure 2a–d show the occurrences of 410-km and 660-km conversions on stacked radial receiver functions (PRFs) for 52 stations along two WNW-ESE trending profiles, OP and QR. The locations of these profiles are illustrated in Fig. 1a. According to the IASP91 velocity model14, the typical arrival times for d410 and d660 in the global reference are 44 and 68 s, respectively. The OP profile indicates that the mean arrival times for d410 and d660 are (44.85 ± 1.46) s and (71.25 ± 1.24) s, respectively (Fig. 2a,b). Compared to the global reference times, the calculated arrival times for the d410 phase were seen to be somewhat delayed by around 0.85 s, while for the d660 phase, the delay was 3.25 s. This indicates the presence of a thicker mantle transition zone (MTZ) beneath the OP profile in comparison to the global reference model14. The QR profile includes simulated arrivals of d410 and d660, which have durations of (44.09 ± 1.15) s and (70.99 ± 1.35) s, respectively. Therefore, the d410 phase arrives late by 0.09 s in comparison to the global average, while the d660 phase is predicted to arrive 2.99 s later than the global average. This observation also indicates that the MTZ underneath the QR profile is slightly thicker compared to the OP profile (Fig. 2c,d). The average arrival times of d410 and d660 in the Uttarakhand Himalaya are estimated to be (44.47 ± 1.33) s and (71.08 ± 1.29) s, respectively. This indicates that the d410 layer is slightly deeper while the d660 layer is reasonably deeper in comparison to the global reference velocity model. Figure 3a–e display the depth-migrated sections of stacked radial PRFs along three NE–SW and two WNW-ESE trending profiles. In Fig. 3a–e, we observe significant negative arrivals associated with LVLs located above the d410 and d660 discontinuities. These LVLs are shown by thick green dotted lines. The LVL-410 is located at depths ranging from 350 to 390 km, with an average depth of (374 ± 8) km. The identification of LVL-410 suggests the existence of hydrated MTZ caused by partial melting or fluids related to the previous subduction episode. On the other hand, the LVL-660 is estimated to be situated between 600 and 640 km depths, with an average depth of (629 ± 12) km. The detection of LVL-660 indicates the presence of wet oceanic sediments at the bottom of the MTZ due to paleo subducted oceanic lithosphere55,56 below the Lesser Uttarakhand Himalaya.

The migration of PRF stacks at depths ranging from 350 to 750 km along three NE-SW and two WNW-ESE profiles (CD, GH, IJ, OP, and QR; locations shown in Fig. 1a) clearly delineates the nature of the upper (410 km discontinuity, d410) and lower (660 km discontinuity, d660) depth boundaries of the Mantle Transition Zone (MTZ) (Fig. 3a–e). The 410-km discontinuity shows a large variation of 10–30 km below the OP profile (Figs. 1a, 3d), as well as a smaller variation of 5–10 km below the CD, GH, and QR profiles (Fig. 3a,b,e), and an almost flat (~ 410 km) behaviour below the IJ profile in the easternmost part of the study region (Fig. 3c). On average, below the study region, the 410-km discontinuity is shallower than the worldwide reference depth of 410-km, but the 660-km discontinuity is deeper. The 660-km discontinuity shows a minor variation of 0–5 km below the GH and QR profiles (Fig. 3b,e), but a significant variance of 10–15 km below the CD and OP profiles (Fig. 3a,d), and a modest variation of 5–10 km below the IJ profile in the easternmost section of the research region (Fig. 3e). Note that, with the exception of the OP profile, d660 obviously drops to the northeast. As a result, the MTZ thickness is dictated by the depth difference between the d410 and d660 discontinuities. The simulated mean MTZ thickness ranges from 260 km beneath the QR profile to 290 km below the OP profile. Figure 1a shows that QR is the southernmost WNW-ESE profile and OP is the northernmost WNW-ESE profile, indicating a distinct NE-dipping MTZ below the Lesser Himalaya in Uttarakhand. However, when we move east–west, we detect a thicker MTZ of 275 km below the GH profile in the centre of the research region. However, the westernmost profile CD predicts a mean MTZ thickness of 270 km, whilst the easternmost profile IJ shows a mean MTZ thickness of 265 km. The variation in MTZ thicknesses suggests that the average MTZ thickness grows towards NE below the Uttarakhand Himalaya (Fig. 3a–e), which could reflect a gradual cooling or thickening of the Indian lithosphere near its northern boundary46.

In a summary, our analysis and depth migration of stack radial PRFs identify lateral variations in depths of the 410-km and 660-km mantle discontinuities, extending from 375 to 425 km (with a mean of (403 ± 9) km) and 650 to 680 km (with a mean of (667 ± 5) km) (Fig. 3a–e). Modelled MTZ thickness spans from 260 to 275 km, with a mean of (272.0 ± 11.5) km (Fig. 3a–e). Our modelling shows a low-velocity layer spanning a 410-km discontinuity (LVL-410), with depths ranging from 350 to 390 km (mean: (374 ± 8) km) (Figs. 3a–e, 4a–d). We also model an LVL above the d660 discontinuity (LVL-660) between 600 and 655 km depth (mean: 629 ± 12 km) (Fig. 3a–e).

Figure 4
figure 4

2-D images obtained from the CCP stacking of radial PRFs along five NE-SW profiles, whose locations are shown in Fig. 1. PRF imaging along the profile (a) AB, (b) CD, (c) EF, (d) GH and (e) IJ. Black doted lines mark the positive peaks associated with 410-km and 660-km Mantle discontinuities, respectively. Thick blue dotted lines mark the low velocity layer atop 410-km (LVL-410) and 660-km (LVL-660) boundaries. Filled green lines mark the 410-km and 660-km depths. This figure is generated using the Funclab software version 1.8.259 (https://doi.org/10.1785/gssrl.83.3.596) and MATLAB software version 9.4.0.813654 (R2018a).

Full size image

The results of our CCP stacking of radial PRFs along five NE-SW and four WNW-ESE profiles are displayed in the Figs. 4a–e and 5a–d. The Ps/P amplitudes differ significantly for the 660-km and 410-km discontinuities (Figs. 4a–e, 5a–d). As we go from NNE to SSW along profiles AB, CD, EF, GH, and IJ, we find a raised d410 (~ 10 km) along the IJ profile in the eastern half of the study region. On average, the d410 is shallower (with a fluctuation of 5–10 km) than the worldwide reference of 410 km beneath the study region (Fig. 4a–e). The CD profile reveals a virtually flat d410 at 400 km depth, which is shallower than the worldwide reference. Whereas d410 is determined to be around 410 km below the AB, EF, and GH profiles, with a little fluctuation of 5–10 km. The LVL-410 is estimated to be 10–35 km thick on average across all five NE-SW profiles. In contrast, the d660 is expected to be deeper than the global standard of 660 km. The d660 is estimated to be nearly flat below the CD and GH profiles, with a modest variation of 5–10 km below the NE end of both profiles, where its thickness ranges from 660 to 670 km. The d660 is observed to be nearly flat at a depth of 665 km below the IJ profile. Below the GH and EF profiles, d660 varies by 10–15 km in thickness, ranging from 660 to 680 km deep. The average d660 is deeper than typical (with a range of 5–15 km) under the study area (Fig. 4a–e). The MTZ thickness varies between 240 and 280 km below the various NE-SW profiles, with a higher MTZ thickness towards the NE (Fig. 4a–e).

Figure 5
figure 5

2-D images obtained from the CCP stacking of radial PRFs along four WNW-ESE profiles, whose locations are shown in Fig. 1a. PRF imaging along the profile (a) KL, (b) MN, (c) OP and (d) QR. Black doted lines mark the positive peaks associated with 410-km and 660-km Mantle discontinuities, respectively. Thick blue dotted lines mark the low-velocity layer atop 410-km (LVL-410) and 660-km (LVL-660) boundaries. Filled green lines mark the 410-km and 660-km depths. This figure is generated using the Funclab software version 1.8.259 (https://doi.org/10.1785/gssrl.83.3.596) and MATLAB software version 9.4.0.813654 (R2018a).

Full size image

The upliftment in d410 decreases as we travel along various WNW-ESE profiles, with the same shallower-than-usual d410 observed at depths ranging from 350 to 390 km below the research area (Fig. 5a–d). On the western part of all the WNW-ESE profiles, d410 is shallowed to a depth of 400 km, while on average, d410 shows an undulation of 10–20 km below the eastern part of all the profiles, with a preferred dip towards ESE, and the depth of d660 varies from 650 to 670 km below all the profiles. An uplifted d660 (~ 640 km) is found along the KL and MN profiles in the northern part of the investigation region. Underlying the KL and MN profiles, d660 reveals a maximum observed undulation of 10–20 km below the region between 75 and 200 km along the profiles, but the OP and QR profiles show the same undulation between 75 and 250 km. Most notably, we see a large positive impedance (Ps/p) anomaly below the 660-km discontinuity between 75 and 200 km along the OP and QR profiles in Uttarakhand’s Lesser Himalayan region, which could indicate imprints of some subducted blocks. In contrast, we see a negative impedance (low velocity) region below the d660 between 50 and 150 km along the KL and MN profiles in the north, along the Munsiari and Vaikrita thrusts. The undulations in d660 are larger than those in d410, and it is often deeper than the global average beneath the various WNW–ESE profiles (Fig. 5a–d). The MTZ thickness varies between 245 and 290 km below the various WNW-ESE profiles, with MTZ thickness increasing towards the NE (Fig. 5a–d).

Negative conversions are observed on d410, particularly between 29.3° to 31.0°N latitude and 78.0°–80.5°E longitude (Figs. 4a–e, 5a–d). The depths connected with these conversions range from 350 to 390 km (with a thickness of 10–20 km), with a lateral span of around 150–300 km. The area consists mostly of Uttarakhand's lesser Himalaya. The magnitudes of negative conversions vary among profiles (AB, CD, FE, GH, and IJ in Fig. 4a–e; KL, MN, OP, and QR in Fig. 5a–d), with greater magnitudes detected along CD, KL, OP, and QR compared to d410 magnitudes (Figs. 4, 5). This layer with negative conversions represents a low-velocity layer above the d410 discontinuity, which has been simulated in several regions around the world and is thought to be caused by partial melting of carbonate-bearing silicates accumulated above the MTZ, influenced by local slab dynamics11,12. Minerals in the mantle transition zone have a water solubility of approximately 1–3%, which is significantly higher than that of minerals in the upper and lower mantles, which is below 0.1%57. When water from the mantle transition zone (MTZ) enters the upper mantle, it plays an important role in facilitating the occurrence of partial melting due to a decrease in melting temperature. As a result, the dense partial melt from the upper mantle settles on top of the d410 layer. Geophysical investigations by Huang et al.58 and geological research by Bercovici and Karato7 have both confirmed this. All of the profiles show a low-velocity layer with negative arrivals above the d410 discontinuity (Figs. 3, 4, 5, 6). Previous research has found similar properties of 410-LVL in Delhi, Bastar craton, DVP, and EDC16,27,37. Singh et al.27 proposed that a 'wet' MTZ is a possibility, but a mid-mantle discontinuity caused by plume-upper mantle contact is a more likely explanation for LVL. Oreshin et al.16 argued that LVL may be caused by subduction-related fluids or mantle upwelling, which is consistent with our observations of LVL atop d410 in the Uttarakhand Himalaya. In the Tethyan oceanic subduction zone, LVL atop d410 has been described as partial melting caused by hydrous and/or other volatile materials upwelling from the MTZ, induced by slab subduction on a regional scale10. Underlying all the considered profiles, we notice another continuous negative signal at a depth of 600–640 km (with a thickness of 20–30 km), which has different features than the deeper d660 along all profiles (Figs. 4, 5, 6). Previous studies have interpreted the negative phase atop the d660 as a low-velocity layer caused by the buildup of ancient subducted oceanic material or higher water content at the MTZ's base55,56.

Figure 6
figure 6

Schematic model of the Mantle Transition zone below the Lesser Himalaya in Uttarakhand based on our modelling results.

Full size image

In order to confirm the existence of low velocity layers located above the d410 and d660 boundaries, we employ Kennett57's propagator matrix method, more especially Ammon's respknt code, to generate artificial P, Sv, and Sh values. The calculation is performed using a velocity structure that is one-dimensional and has a crustal thickness of 45 km. This structure was obtained by combining radial PRFs and Rayleigh wave group velocity dispersion data from the TRNR station located at coordinates 39.64°N, 78.98°E. The maximum depth of the deeper velocity model was restricted to 800 km by utilising the IASP91 reference velocity model14 (see to Supplementary Fig. S5a). Ligorria and Ammon54's time domain deconvolution technique is employed to separate the vertical component (P) and radial component (Sh). The separation allows for the creation of radial P-wave receiver functions (PRFs) for horizontal slowness (p) values ranging from 0.04 to 0.08 s/km (refer to Supplementary Fig. S5a–f). Initially, we utilised velocity models that did not include LVLs at the 410 and 660 discontinuities to compute synthetic PRFs, as depicted in Supplementary Fig. S5a. Afterwards, synthetic PRFs are produced using slowness values that range from 0.04 to 0.08 s/km. The positive arrivals associated with the d410 and d660 discontinuities are noticed without any additional lobes (see Supplementary Fig. S5b–f). This suggests that side lobes are not generated during the computing procedure of PRFs. The range of slowness values is seen in consistent alterations in the magnitudes and occurrences of the positive peaks linked to the d410 and d660 discontinuities (refer to Supplementary Fig. S5b–f). To accurately reproduce the significant changes linked to the LVL-410 and LVL-660, we introduce a real background noise prior to the direct P-arrival on the observed broadband seismograms (refer to Supplementary Fig. S6a–f). In addition, we constructed PRFs with background noise from real data for five different slowness values to quantify the uncertainty of the radial PRF. This was achieved by adding real background noise to both radial and vertical components of the observed broadband seismograms. Subsequently, the time-domain deconvolution method of Ligorria and Ammon54 is used to model the synthetic radial PRFs for five distinct horizontal slowness values, specifically 0.04, 0.05, 0.06, 0.07, and 0.08 s/km (refer to Supplementary Fig. S6b–f). We arrange the synthetic radial PRFs for five slowness values specified earlier, as depicted in Supplementary Fig. S6b–f. The synthetic PRFs exhibit distinct conversions originating from the upper (d410) and lower (d660) boundaries of the Mantle transition zone (MTZ). The conversions are depicted as d410 and d660 in Supplementary Fig. S6b–f. The Supplementary Figs. S5b–f and S6b–f demonstrate that the conversions d410 and d660 have a maximum variability of 1 to 1.5 s across five different horizontal slowness values. This observation indicates that the negative conversions mentioned above d410 and d660 are real, as demonstrated by our constructed synthetic PRFs (see Supplementary Figs. S5a–f and S6a–f).

By applying the common conversion point (CCP) stacking method to radial P-wave receiver functions, we were able to identify variations in the depths of the 410-km and 660-km mantle discontinuities. These variations ranged from 395 to 420 km (with a mean of 406 ± 8 km) for the 410-km discontinuity, and from 645 to 675 km (with a mean of 659 ± 10 km) for the 660-km discontinuity, within the depth range of 350 to 750 km. The estimated thickness of the mantle transition zone (MTZ) varies from 240 to 265 km, with an average of 255 ± 7 km. The model we used also detects a layer with low velocity above the 410-km discontinuity. This layer has a thickness of 16 ± 7 km and is located at a depth range from 350 to 385 km, with a mean depth of 369 ± 13 km. This suggests the presence of a hydrated mantle transition zone in the area. For further details, refer to Figs. 4a–e and 5a–d. In addition, we illustrate the presence of a low-velocity layer, which is approximately 19 ± 9 km thick, located above the d660 discontinuity. This layer is found at depths ranging from 590 to 640 km, with an average depth of 620 ± 16 km. These findings are depicted in Figs. 4a–e and 5a–d. This layer could be linked to the deposition of ancient subducted oceanic material or an increased water content at the lower part of the mantle transition zone, as shown in previous studies conducted in South Africa and China58,59. A schematic representation of the Mantle transition zone beneath the Lesser Himalaya in Uttarakhand (Fig. 6) has been constructed using the aforementioned modelling outcomes. The model depicts a d410 with a shallower depth, a d660 with a deeper depth, an LVL positioned above the d410, and an LVL positioned above the d660.

Conclusions

The study provides a comprehensive analysis of the structure of the Mantle Transition Zone (MTZ) beneath the Uttarakhand Himalaya. Through the examination of stacked PRFs at 52 stations along five NNE-SSW and four WNW-ESE trending profiles, we have calculated the average arrival times of d410 and d660 conversions to be (44.47 ± 1.33) s and (71.08 ± 1.29) s, respectively. The arrival time for d410 is delayed by 0.47 s in comparison to the expected time of 44 s according to the IASP91 global reference velocity model. Whereas, the arrival time for d660 is delayed by 3.08 s compared to the predicted global reference time of 68 s. The presence of these arrivals suggests the presence of a slightly deeper or undisturbed flat layer at a depth of 410 km and a deformed and noticeably deeper layer at a depth of 660 km in the region. We also see significant negative arrivals associated with LVLs located above the d410 and d660 discontinuities.

The analysis of stacked radial PRFs and CCP PRF images at depths of 350 and 750 km along five NNE-SSW and four WNW-ESE profiles in the region reveals a shallow 410-km boundary. The average depth of this boundary ranges from 395 to 420 km, with a mean depth of (406 ± 8) km. The estimated depth of the 660-km discontinuity ranges from 645 to 675 km, with an average depth of 659 ± 10 km. The Mantle Transition Zone (MTZ) has a gradual inclination towards the north-northeast, with a thickness ranging from 240 to 265 km and an average thickness of 255 ± 7 km. Our model provides detailed information about the characteristics of the mantle boundaries at depths of 410 km and 660 km beneath the Lesser Himalayan region. It demonstrates that the mantle transition zone becomes thicker as we move towards the NE. Additionally, we have identified two LVLs located above d410 and d660. The LVL-410 has a depth range of 350 to 385 km, with an average depth of 369 ± 13 km. The mean thickness of this layer is estimated to be 16 ± 7 km beneath the region. The presence of LVL-410 implies the potential existence of a hydrated mantle transition zone below the region, perhaps caused by partial melting or the presence of fluids connected to subduction above the d410 boundary. A second LVL-660 is simulated in depths ranging from 590 to 640 km, with an average depth of 620 ± 16 km. The thickness of this layer is determined to be 19 ± 9 km beneath the area. This layer may be associated with the accumulation of ancient subducted oceanic material or an enriched water content at the base of the mantle transition zone, akin to discoveries made in South Africa and China. The aforementioned observations offer substantial evidence for the existence of imprints from ancient subduction slabs beneath the Uttarakhand Himalayas.

Methodology

Common Conversion Point (CCP) Structural Imaging and depth-wise migration of stacked radial PRFs

In this study, we use the Funclab software60 to perform CCP imaging of radial PRFs, using Dueker and Sheehan61’s methodology to coherently stack p-to-s phase conversions for generating a 2-D image of impedance contrast at depth. We model CCP images at depths ranging from 0 to 800 km using five NE-SW and four WNW-ESE profiles, the locations of which are depicted in Fig. 1a. The CCP images are created using a stacking volume of 30 km. Both Cladwell et al.2 and the Funclab manual59 explain CCP stacking methods for radial PRFs. We used the Funclab software to stack CCPs with a bin width of 0.3°, a bin spacing of 3 km, a depth increment of 5 km, and a maximum depth of 800 km. Note that our study area is located between latitudes 29° and 31.5° and longitudes 77.5° and 81°. Initially, we build a one-dimensional velocity model, with the deeper part (from the Moho depth down to 800 km) constrained by the IASPEI velocity model14, and the crustal part (down to the Moho depth) derived from the Mandal et al.62 1-D velocity model. We utilised the above-mentioned velocity model to stack radial PRFs at depths ranging from 0 to 800 km, with five NE–SW and four WNW–ESE profiles (Fig. 1a). We also used the Funclab software to construct depth-wise migration of stacked radial PRFs between 350 and 750 km depths along these profiles, allowing us to clearly determine the nature of the 410-km and 660-km discontinuities. Figures 4a–e and 5a–d show CCP of radial PRF images along these profiles at depths ranging from 350 to 750 km. In the present work, we calculate PRFs with a Gaussian width factor (a) of 1.5 (f < 0.75 Hz) and a wavelength of 5.3 for an average crustal shear velocity of 4.0 km/s. Thus, the computed PRFs with a = 1.5 can discern layers as thick as 2.65 km. The nominal vertical resolution at the Moho may be 4 km for 1-D PRFs and 1.7 km for 2-D CCP stack2. As a result, CCP imaging can provide more reliable and accurate estimates of layer thickness. Thus, we present here our final modelling results for the CCP stacking of radial PRFs.