Comparison of Crustal Stress and Strain Fields in the Himalaya–Tibet Region: Geodynamic Implications

Abstract

The Himalaya–Tibet region represents a complex region of active deformation related to the ongoing India–Eurasia convergence process. To provide additional constraints on the active processes shaping this region, we used a comprehensive dataset of GNSS and focal mechanisms data and derived crustal strain and stress fields. The results allow the detection of features such as the arc-parallel extension along the Himalayan Arc and the coexistence of strike-slip and normal faulting across Tibet. We discuss our findings concerning the relevant geodynamic models proposed in the literature. While earlier studies largely emphasized the role of either compressional or extensional processes, our findings suggest a more complex interaction between them. In general, our study highlights the critical role of both surface and deep processes in shaping the geodynamic processes. The alignment between tectonic stress and strain rate patterns indicates that the crust is highly elastic and influenced by present-day tectonics. Stress and strain orientations show a clockwise rotation at 31°N, reflecting deep control by the underthrusted Indian Plate. South of this boundary, compression is driven by basal drag from the underthrusting Indian Plate, while northward, escape tectonics dominate, resulting in eastward movement of the Tibetan Plateau. Localized stretching along the Himalaya is likely driven by the oblique convergence resulting from the India–Eurasia collision generating a transtensional regime over the Main Himalayan Thrust. In Tibet, stress variations appear mainly related to changes in the vertical axis, driven by topographically induced stresses linked to the uniform elevation of the plateau. From a broader perspective, these findings improve the understanding of driving crustal forces in the Himalaya–Tibet region and provide insights into how large-scale geodynamics drives surface deformation. Additionally, they contribute to the ongoing debate regarding the applicability of the stress–strain comparison and offer a more comprehensive framework for future research in similar tectonic settings worldwide.

1. Introduction

The spatial comparison of crustal stress and strain patterns is one of the most compelling approaches in a geodynamically complex framework. It provides useful constraints on driving forces and highlights those regions which accumulate stress and encourage seismic deformation. This approach has been widely applied in the literature to describe crustal behavior and deformation models [1,2,3,4,5,6]. For instance, Chang et al. [6], studying Taiwan Island, found that both stress and strain fields perceptively reflect the long-term collision mechanism and that inherited structure can play a role in redirecting the stress–strain orientation. In Japan, Townend and Zoback [5] found that observed geodetic velocities are mainly driven by the crustal collision rather than the subduction zone locking; Palano et al. [2] applied the comparison to the Gibraltar Orogenic Arc, finding a significant coherence between well-constrained stress indicators and GNSS-derived strain, therefore suggesting that deep dynamic processes control both crustal stress and surface deformation. Other studies have been conducted both in geothermal-active regions [7] and in volcanically active areas [4], providing information on the seismicity distribution and deformation behavior of these areas.

Himalaya–Tibet (Figure 1) represents a complex region of active deformation related to the long-lasting and still active India–Eurasia convergence process [8]. Although the Indian-Eurasian convergence is a puzzling geodynamic process that generates active deformation and seismicity across different geographical areas, such as the Kachchh region [9,10,11,12], most of the crustal deformation is currently concentrated along the Himalayan Arc, driven by the high convergence rate between the Indian and Eurasian Plates [13]. In the central sector of the Himalayan Arc (Nepal), the convergence is predominantly arc-normal [14,15], while to the east and west, a significant arc-parallel component of deformation has been observed both by geological and geodetic data [16,17,18,19,20,21]. Despite the ongoing N–S convergence, southern and central Tibet are also experiencing significant E–W extension. This complex structural configuration, characterized by the concurrent presence of active strike-slip and normal fault systems traversing Tibet [22,23], hampers effective kinematic zonation in the region that still lack definitive characterization.

Due to its complexity, different models have been proposed for the ongoing deformation (at least for the last few Ma) of the Himalaya–Tibet region: the topographic loading [24], the indenter-escape tectonic [25], the thinning model [21], the oblique convergence [26], and the channel flow [27].

We employ the latest and most comprehensive geodetic crustal motion data, along with a novel estimation of the seismological stress and geodetic strain rate fields. Using these results, we aim to furnish a critical bridge among current geodynamic models proposed for the Himalaya–Tibet region, providing therefore an empirical evaluation of their applicability through a detailed stress–strain comparison. Such a detailed comparison will offer a refined perspective on ongoing tectonic processes, enhancing understanding of the driving forces behind crustal deformation in complex tectonic settings.

Figure 1. Geological background of the Himalaya–Tibet region. (a) Main terranes, suture zones and principal tectonic lineaments on the study region (modified from Li et al. [28]). KSZ—Kunlun Suture Zone, JSZ—Jinsha Suture Zone, BSZ—Bangong-Nujiang Suture Zone, YSZ—Yarlung-Tsangpo Suture Zone, STD—South Tibetan Detachment, MFT—Main Frontal Thrust. (b) Active and main faults from Zelenin et al. [29]: blue: thrust; red: normal fault; green: strike-slip fault. MCT—Main Central Thrust, MBT—Main Boundary Thrust, MFT—Main Frontal Thrust. The terrain model is from GEBCO [30], accessed on 31 October 2024.

Figure 1. Geological background of the Himalaya–Tibet region. (a) Main terranes, suture zones and principal tectonic lineaments on the study region (modified from Li et al. [28]). KSZ—Kunlun Suture Zone, JSZ—Jinsha Suture Zone, BSZ—Bangong-Nujiang Suture Zone, YSZ—Yarlung-Tsangpo Suture Zone, STD—South Tibetan Detachment, MFT—Main Frontal Thrust. (b) Active and main faults from Zelenin et al. [29]: blue: thrust; red: normal fault; green: strike-slip fault. MCT—Main Central Thrust, MBT—Main Boundary Thrust, MFT—Main Frontal Thrust. The terrain model is from GEBCO [30], accessed on 31 October 2024.

2. Background

2.1. Geodynamic Evolution

Tibet (i.e., the region in central Asia located between the Tarim Basin to the north and the Indian Plate to the south) has an extensive geological history [8]. The current geological setting of Tibet resulted from the occurrence of sequential geodynamic events such as microcontinent collision, the building of orogenic belts, and the formation of island arcs, at least since the early Paleozoic [31]. Nowadays, from north to south, the Kunlun (KUT), Songpan-Ganze (SGT), Qiangtang (QT), Lhasa (LHT), the Tethys Himalaya (THT), and the Indian Himalaya (INT) terranes represent the pronounced horizontal heterogeneity of the plateau [19] (Figure 1a). These terranes are separated by suture zones, including the Kunlun (KSZ), Jinsha (JSZ), Bangong-Nujiang (BSZ), and Yarlung-Tsangpo sutures (YSZ) (Figure 1a). All these suture lines are the result of a succession of subductions, from the closure of the Paleo-Tethys in the Late Triassic (Kunlun and Jinsha suture zones) up to the closure of the Bangong-Nujiang Ocean, the northern branch of the Neo-Tethys, with the formation of the Bangong-Nujiang suture zone in the Late Cretaceous [32].

Since the Late Cretaceous, India began its northward drift, separating Gondwana. During this process, the Tethys Himalaya Terrane moved away from India, leading to the opening of the North Indian Sea. Both India and the Tethys Himalaya moved towards the Eurasian plate at a rate of approximately 15–18 cm/yr until they collided around 60–55 Ma [8]. Around 50 Ma, the convergence rates significantly slowed to 6 cm/yr, marking the onset of the continent–continent collision phase [33]. This latest phase of shortening has been accommodated by the continuous underthrusting of the Indian shield beneath the Tethyan Himalaya and the Lhasa Terrane [34], as also evidenced by recent seismic observations [35] which imaged a distinct cold body beneath the Lhasa Terrane. The underthrusting also explains the uplift in Tibet during the last 40 Ma [36,37]. However, other hypotheses have been suggested [38,39,40].

Following the India–Eurasia collision, the deformation fronts progressively shifted southward from the YSZ, giving rise to the imbricated fold-and-thrust belt of the Himalayas. Southward, Tibet is bounded by the South Tibetan Detachment (STD), which represents the lower boundary of the stacked Tethyan Himalaya sequence. The STD is a lithospheric shear zone in which both reverse and normal faulting episodes have been recorded [39,40,41]. The detachment acted as a shortening structure during the Eocene [42], meanwhile the normal faulting was probably active during the Miocene [43]. Different hypotheses have been proposed for the reactivation of the STD with normal faulting features [44] such as gravitational collapse [45], wedge extrusion [46], channel flow [47], or tectonic wedging/duplexing [48,49].

In the southern part of the STD, the Indian Himalaya Terrane is affected by a complex thrust system formed by the Main Central Thrust (MCT), the Main Boundary Thrust (MBT), and the Main Frontal Thrust (MFT). While activity on the MCT started around ~22–18 Ma and may have remained episodically active up to the present day [50,51], the MBT developed from 11 Ma until the Pleistocene. The MFT defines the southernmost margin of the orogenic arc and represents the near-surface expression of the deep basal décollement (Main Himalayan Thrust; hereinafter, MHT). The MHT separates the Indian shield from the Indian Himalaya terrane and is currently acting as the gently inclined basal detachment with a ramp-flat geometry underlying this complex structure [16]. The MHT appears to be segmented by subsurface transverse tectonic lineaments and faults, with normal to strike-slip features in regions with low seismic coupling values [50]. These regions are marked at depth by the presence of subsurface paleo-basement ridges, such as the Delhi–Haridwar (DHR), the Faizabad (FR) and the Munger–Saharsa (MSR) [52].

2.2. Seismotectonic Framework

Past seismic activity highlights the Himalaya–Tibet region as an area of high seismic risk, with numerous moderate-to-large earthquakes (6 < M_w < 7) distributed across the region [53]. By inspecting the USGS catalog (https://earthquake.usgs.gov, accessed on 31 May 2024), ~15,000 earthquakes with M_w > 2.5 spanning the 1960–March 2024 time interval have been selected (Figure 2). Despite being one of the most seismically active regions in the world, instrumental hypocenters in the area are poorly constrained due to the lack of local seismic networks, especially in Tibet. Instrumental seismicity occurs within approximately 100 km of the Main Himalayan Thrust (MHT) downdip, with notable seismic gaps in Bhutan [54]. In more detail, it is sparsely distributed in a clustered pattern along the collisional arc, following some of the strike-slip lineaments and in a N–S aligned pattern around a few normal faults. It is also primarily concentrated within the upper 30–35 km of the crust, though intermediate-depth (>40 km [55]) seismic events have been locally recorded beneath the Himalayas [30]. In the last century, Himalaya has experienced four large seismic events (M_w > 7) such as the M_w 8.9 1950 Assam earthquake [56,57], the M_w 7.8 and M_w 7.3 2015 Gorkha [58,59], and the M_w 7.6 2005 Kashmir ones [60].

The southern border of Tibet is currently experiencing a marked E-W crustal stretching, which is accommodated by N-S-oriented normal [61] and active strike-slip faults [61,62]. These active faults are associated with widespread seismicity with M_w > 6, e.g., the M_w 6.2 2004 and the M_w 6.3 2005 Zhongba earthquakes [63], the M_w 6.3 2008 Damxung earthquake [64], and the M_w 6.1 2020 Nima earthquake [65], also including notable seismic sequences. Strike-slip faults, predominantly oriented NW-SE or NE-SW, have generated several significant seismic events, including the M_w 7.7 1951 Beng Co earthquake [66], the M_w 6.9 2014 Yutian earthquake [67], and the M_w 7.6 2001 Kunlun earthquake [68].

Figure 2. Instrumental seismicity (M_w > 2.5) from 1960 to 2024 (USGS catalog; https://earthquake.usgs.gov (accessed on 31 May 2024)), colored according to focal depth with a rainbow color scale. Yellow and red stars are seismic events with 6 < M_w ≤ 7 and M_w > 7, respectively. Earthquakes before 1960 (1950 Assam and 1951 Beng Co) are from Coudurier-Curveur et al. [56] and Li et al. [66], respectively. Active and main faults from Zelenin et al. [29]: blue: thrust; red: normal fault; green: strike-slip fault; gray: unknown kinematic fault. The terrain model is from GEBCO [30], accessed on 31 October 2024.

Figure 2. Instrumental seismicity (M_w > 2.5) from 1960 to 2024 (USGS catalog; https://earthquake.usgs.gov (accessed on 31 May 2024)), colored according to focal depth with a rainbow color scale. Yellow and red stars are seismic events with 6 < M_w ≤ 7 and M_w > 7, respectively. Earthquakes before 1960 (1950 Assam and 1951 Beng Co) are from Coudurier-Curveur et al. [56] and Li et al. [66], respectively. Active and main faults from Zelenin et al. [29]: blue: thrust; red: normal fault; green: strike-slip fault; gray: unknown kinematic fault. The terrain model is from GEBCO [30], accessed on 31 October 2024.

2.3. Geodynamic Models

Several geodynamic models have been proposed to explain the complex fragmentation and kinematic heterogeneity of the Himalaya–Tibet region. All models primarily aim to interpret the synchronous coexistence of (i) a roughly N-S striking compression, as resulting from the collisional process between Indian and Eurasian plates, with (ii) an E–W extension, related to a general eastward motion of Tibet.

N-S trending normal faults in Tibet are commonly thought to represent the active deformation due to topographic loading [24,69,70]. Collapse occurs in response to the isostatic compensation in the highly thickened Tibetan crust (~70 km [71]). The topographic loading model is only able to explain extension in the Tibet region, without fully accounting for the eastward motion of all the Tibetan terranes and the distributed active strike-slip faults.

Tapponier and Molnar [25] explain the eastward motion of Tibet by suggesting an indenter-escape tectonic model. According to this model, the weak Tibetan lithosphere is tectonically pushed eastward because of the N-S convergence of the highly rigid lithosphere of India and the Tarim Basin (Eurasia). The indenter block was too buoyant to subduct, so lithospheric accommodation is mainly achieved by the eastward extrusion of Tibet. This model well explains the clockwise rotation of eastern Tibet [72] and the bending of faults such as the Jiali Fault [73]. However, the indenter-escape tectonic model assumes that deformation is mainly confined to the horizontal plane, neglecting vertical deformations, which would be relevant, especially in the Himalaya–Tibet region where the crust is the thickest in the world. Furthermore, such an explanation is insufficient as it considers extensional tectonics isolated from the compressive forces arising from the ongoing India–Eurasia convergence. These compressive forces would be crucial in driving the uplift of the plateau from the Paleogene up to the present, as testified by the ongoing uplift inferred by recent leveling and GNSS measurements [37].

A primary piece of information for enhancing the understanding of the active geodynamics of the Himalaya–Tibet region is the precise location of the Indian Plate beneath Tibet. A gently northward-dipping surface has been presented by different authors [35,74,75,76,77], revealing similar seismic characteristics of the layers up to 31°N. The presence of the underthrusted Indian crust has led to the proposal of two additional models closely related to the active dynamics.

The thinning model [21] suggests that the subsequent extension reflects the thinning of the upper crust in response to thickening due to the underthrusted Indian Plate. This model has two significant limitations: (i) extension does not cease at 31°N but also affects the Qiangtang Terrane and (ii) while extension should theoretically activate progressively from south to north, it appears to have started fairly uniformly across the plateau during the Miocene [69,78].

Beyond that, other authors have proposed the oblique convergence model [26,79,80,81] which also assumes the underthrusting of India but explains the extension through the relative movement between the Indian and Eurasian plates. McCaffrey and Nábělek [26] observed that the former plate moves as a whole towards the north, evidencing that the convergence is perpendicular only in a limited sector of the Himalayas, having its contact zone with the Eurasian Plate in a semicircular shape. Such a feature highlights that in the other sectors, the overall convergence is also partitioned into a parallel component. In such a context, the E-W extension would result from a basal shear caused by the Indian oblique underthrusting beneath Tibet. However, this model is unable to properly explain the deformation northward of the underthrusted Indian Plate, as similarly observed for the thinning model. Additionally, this model requires higher deformation values moving west and south where, according to some authors (e.g., [31]), the convergence is less perpendicular and shallower, with the maximum rates occurring in the western Himalayan region.

An additional attempt to explain the complexity of deformation was the channel flow model by Bird [27]. This author was the first to theorize, through numerical models, the possible presence of a deep zone in the Tibetan crust hosting a planar channel flow. Beaumont et al. [82], using thermal-mechanical models, explain that the effect of surface denudation could be due to crustal channel flow which, as also noted by Colpey and McKenzie [83], causes the Himalayan Arc to expand both radially and circumferentially. Rey et al. [84] suggested that the critical thickness required to weaken the lower crust would be 50 km, raising the question of how a real thickness of ~70 km [71] could be attained. In this context, the observed uplift may be difficult to explain since the channel flow model is typically considered a consequence of uplift rather than its cause [84].

3. Data

3.1. GNSS Data

As a first step, raw observations of more than 150 continuous GNSS stations encompassing the Himalaya–Tibet region were collected from UNAVCO (https://www.unavco.org (accessed on 31 May 2024)). After a preliminary check, 37 stations were removed because they were characterized by a limited temporal coverage of raw data (less than 2.5 years), while the remaining stations were processed by using the GAMIT/GLOBK v.10.7 software [85], following the standard approach described in Billi et al. [86], in order to estimate a consistent set of positions and velocities in a fixed ITRF14 reference frame [87]. To enhance the spatial density of the geodetic velocity field across the studied area, the obtained solution has been integrated with the solutions reported in the recent literature [88,89,90]. Published solutions were aligned with our solution by determining the Helmert transformation parameters that minimize the root mean square of velocity differences at common sites (e.g., [91]). All sites displaying significant inconsistencies with neighboring stations were removed from the unified velocity field, resulting in a final dataset of approximately 1142 GNSS sites. The resulting velocity field (Figure 3) has been rotated to an Eurasian reference frame by using the Euler pole parameters reported in Altamimi et al. [92].

Figure 3. Thin red head arrows represents the GNSS horizontal velocities and 95% confidence ellipses in an Eurasian reference frame [92]. The terrain model is from GEBCO [30], accessed on 31 October 2024.

Figure 3. Thin red head arrows represents the GNSS horizontal velocities and 95% confidence ellipses in an Eurasian reference frame [92]. The terrain model is from GEBCO [30], accessed on 31 October 2024.

3.2. Seismicity and Focal Mechanisms Catalog

Mechanical stress acting in the Earth’s crust can be estimated by inverting earthquake focal mechanisms. Therefore, to provide a reliable estimation of the crustal stress in the study area, a dataset of FMSs has been compiled. In a first step we collected all available FMSs over the Himalaya–Tibet area by inspecting the following sources:

• GCMT (https://www.globalcmt.org/ accessed on 31 May 2024);
• ISC (http://www.isc.ac.uk/iscgem/index.php accessed on 31 May 2024);
• World Stress Map [93] (https://www.world-stress-map.org/download accessed on 31 May 2024).

Multiple solutions for the same event have been removed, selecting the solution having the least difference concerning the hypocentral parameters reported in the USGS seismic catalog (Figure 2).

In the second step, data from the literature have been integrated into the general catalog by preferring the solutions coming from local seismic networks or computed by using refined velocity models [52,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111]. Moreover, since the procedure of the waveform-inversion FMSs is generally more stable and robust than solutions obtained by first-motion polarity (e.g., [112,113]), the former was selected when both types of solutions were available for the same earthquake.

The final catalog contains ~1300 FMSs related to M_w > 4 earthquakes that occurred in the June 1933–March 2024 period (Figure 4). About ~80% of the collected FMSs have been obtained by waveform inversion, while the remaining has been computed by first-motion polarity and are related to the earthquakes that occurred before the advent of the digital epoch in the 1980s. In the study area, most of the FMSs have focal depths less than 30 km; ~8% have focal depths up to 80 km. Among these last solutions, the most remarkable ones are those related to the 2011 Sikkim seismic sequence [114] that occurred beneath the MHT, and those related to the Burma active subduction and the Pamir–Hindu Kush lithospheric subduction [115].

Figure 4. Lower hemisphere, equal area projection for FMSs with M > 4 spanning the June 1933–February 2024 time period. FMSs are colored according to kinematic classification [116]: blue: thrust faulting (TF); light blue: thrust with strike-slip component (TS); green: pure strike-slip (SS); orange: normal faulting with strike-slip component (NS); red: normal faulting (NF); gray: unknown kinematics (UK). Other features are as Figure 2. The terrain model is from GEBCO [30], accessed on 31 October 2024.

Figure 4. Lower hemisphere, equal area projection for FMSs with M > 4 spanning the June 1933–February 2024 time period. FMSs are colored according to kinematic classification [116]: blue: thrust faulting (TF); light blue: thrust with strike-slip component (TS); green: pure strike-slip (SS); orange: normal faulting with strike-slip component (NS); red: normal faulting (NF); gray: unknown kinematics (UK). Other features are as Figure 2. The terrain model is from GEBCO [30], accessed on 31 October 2024.

4. Method

4.1. Geodetic Strain Rate Field

The horizontal strain rate pattern and associated uncertainties over the study area have been calculated using the method described by Shen et al. [117]. This algorithm employs discretized geodetic observations and applies a least-squares procedure to interpolate velocities on an optimally regular grid. Interpolated velocities are then used to derive strain rate values by assuming the medium to be elastic. After some preliminary tests, we chose a value of 18 as the weighting threshold, according to the inter-stations distance of the GNSS network [118] and the comparative analysis conducted by Nucci et al. [119].

This study examines the minimum and maximum stretching, ε_Hmin and ε_Hmax, respectively, as eigenvalues of the computed strain rate tensor, alongside their corresponding azimuths. The two eigenvectors are assumed to be orthogonal to each other and oriented at 45° relative to the azimuth of the maximum shear strain. Additionally, a crucial parameter is the dilatation strain rate, which quantifies the rate of the areal change as the sum of the ε_Hmin and ε_Hmax values.

The adopted grid nodes are evenly spaced at intervals of 0.25° × 0.25°. The distribution of stations across the area is irregular (Figure 3); therefore, the results for these regions, such as north–central Tibet and the Tarim Basin, must be interpreted keeping this aspect in mind. Achieved results are reported in Figure 5.

4.2. FMS Formal Inversion

The stress tensor over the study area has been calculated by using the FMSs database (see Section 3.2) fusing of the STRESSINVERSE code (M_w > 4 earthquakes) [120] as the input. This code calculates the stress tensor parameters assuming: (i) the tectonic stress is uniform and homogeneous in the considered volume, (ii) earthquakes occur on pre-existing faults, and (iii) the slip vector points to the direction of shear stress on the fault [120,121,122]. This method can determine the 3D stress tensor, defining trend, and plunge of the three axes, σ₁, σ₂, and σ₃ (with σ₁ ≥ σ₂ ≥ σ₃, under the compression-positive stress convention), and the dimensionless stress ratio, R [123]:

$$\mathrm{R}={\displaystyle \frac{{\sigma }_1-{\sigma }_2}{{\sigma }_1-{\sigma }_3}}$$

(1)

The stress ratio R varies from 0 to 1.0; when R = 0, σ₁ and σ₂ are equal in magnitude, while R = 1 means that σ₂ and σ₃ have the same values.

To ensure more accurate estimations of stress ratio values, the STRESSINVERSE code selects the optimal nodal plane [120] by quantifying the fault instability [124], following the Mohr–Coulomb failure criterion [125]. To properly evaluate the sensitivity of the inversion procedure to friction parameters, needed in the instability constraint, the inversion was run repeatedly with friction ranging from 0.4 to 1 in steps of 0.05. Thus, the code performs a set of iterations aimed at determining the orientation of the stress axes and the estimation of the stress ratio, considering all possible nodal plane combinations, until the stress parameters converge toward the best model [126]. In this way, the instability of the nodal planes of all FMSs is calculated, selecting those most susceptible to shear faulting.

The inversion has been performed on a moving 2° × 2° equally spaced grid with a 1° overlap on each side. The inversion is only performed in cells containing at least 8 FMSs. Based on the measured stress field geometry and stress ratio, the so-called maximum sub-horizontal compressive stress, σ_Hmax, is also calculated [127] (see Figure 6a).

5. Results

5.1. Strain Rate Field

Strain rate results are reported in Figure 5 as ε_Hmax and ε_Hmin values along with the associated azimuthal pattern (i.e., arrows and trajectories). Despite some very localized artifacts, our results align well with previous solutions [89,128], providing, moreover, an improved resolution of the deformation field thanks to the high spatial density of the geodetic velocity field across the studied area.

The ε_Hmax strain parameter has positive values with very small variability over most of the study area. The largest positive values (up to 30 nanostrain/yr) are observed in the entire of Tibet, while values ranging from 0 to ~25–30 nanostrain/yr can be observed along six localized N-S oriented strips along the Himalayan Arc. Some small patches with negative values (~25 nanostrain/yr) can be observed close to Arunachal, the Shillong plateau, and the Himachal regions. The ε_Hmax trajectories outline a general E-W trend, gradually rotating, moving eastward. Along the Himalayan Arc, the ε_Hmax trajectories pass from a roughly WNW-ESE attitude in the western Himalaya to E-W and N-S attitudes in central and eastern Himalaya (Figure 5a).

The ε_Hmin values effectively depict the general contraction of the study areas, with low values (ca. −25 nanostrain/yr) mainly characterizing the Tibetan region and high values along the Himalayan Arc (up to −100 nanostrain/yr) and at the northern margin of Tibet, along the Altyn-Tagh and Kunlun faults (up to −40 nanostrain/yr). The ε_Hmin trajectories show a general NNE-SSW trend over most of the study area. Such a pattern is roughly orthogonal to the general attitude of the Himalayan Arc. A marked rotation is observed in proximity to the Arunachal area, where the ε_Hmin trajectories follow the clockwise rotation of the orogenic belt, aligning NE-SW. Another interesting feature is the gradual rotation of the ε_Hmin trajectories observed in the Tibetan Plateau. In the southern part of the plateau, the ε_Hmin trajectories show a predominant N-S orientation, while north of the 31°N parallel, the trajectories gradually rotate clockwise about 30°. The horizontal dilatation rate map (sum of ε_Hmin and ε_Hmax) is dominated by the contractional pattern (up to −120 nanostrain/yr) along the Himalayan orogen, clearly evidencing the ongoing India–Eurasia convergence. Such a general contraction concentrated within 100 km north of the MFT indicates the active contraction of the Indian Himalaya above the MHT. In the Tibetan region, the areal change values are close to zero or show small local variations, indicating no significant change in areal strain. However, given that ε_Hmax and ε_Hmin have similar (not zero) values, it can be inferred that crustal deformation in the Tibetan region occurs prevalently as shear strain, in a generally constant area strain in the map view [129]. Such a feature aligns well with the presence of abundant strike-slip faults, responsible for several large earthquakes. The dilatation rate is close to zero in the Tarim Basin where, in contrast to Tibet’s results, ε_Hmin and ε_Hmax have very small or close to zero values, suggesting a small/null deformation of the region. Estimated strain rate parameters over this area would be poorly constrained because of the lack of GNSS stations; however, the small/null deformation within the Tarim Basin is also confirmed by geological data, clearly documenting no active deformation since the onset of the India–Eurasia convergence [130,131]. Moreover, both the southern and northern margins of the Tarim Basin show negative areal changes that appear related to the presence of active faults/structures, such as the Altyn Tagh fault and the Tianshan orogeny, respectively.

5.2. Stress Field

The estimated stress parameters over the study area are reported in Figure 6 (see Supplementary Material for additional details). As already mentioned, most of the collected FMSs are related to earthquakes located in the uppermost part of the crust, within the first 30–35 km. Therefore, our stress results must be considered as representative only of the shallow crust, being the average crustal thickness of ~70 km [71] beneath most of the study area. Figure 6a shows achieved results for each inverted cell in terms of stereoplot representations and stress trajectories. As mentioned previously, the stress tensor inversion has been performed only for those cells containing at least eight FMSs. Only a few cells contain eight FMSs, and their resulting stress parameters agree with the ones obtained for the neighboring cells, evidencing acceptable reliability of the estimations. Each stereoplot shows the retrieved principal stress (σ₁, σ₂ and σ₃) 3D spatial pattern along with the associated confidence limits, while the stress trajectories refer to the σ_Hmax pattern.

Along the Himalayan Arc, the σ₁ stress axis is roughly sub-horizontal (estimated plunge in the 0–35° range), pointing toward a rough N-S orientation, while the σ₂ and σ₃ axes represent, alternatively, the vertical stress. In this region, only the σ₁ axis is well constrained by the inversion, while the σ₂ and σ₃ axes sometimes show a partial overlap of their estimated 95% confidence limits, suggesting that they are approximately equal in magnitude (R > 0.75). As a result, there are alternating sectors of the Himalayan Arc where σ₂ is sub-horizontal, like in most of the central Himalaya, and areas where it becomes vertical, such as in the Sikkim region, Himachal, and Bhutan. However, the σ_Hmax pattern along the Himalayan Arc effectively reflects the primary compression arising from the India–Eurasia convergence.

In Tibet, the best constrained stress axis is σ₃, which is always sub-horizontal with a W-E to NW-SE average strike. σ₁ and σ₂ axes are alternatively sub-vertical, with partial overlap of their estimated 95% confidence limits; therefore, they are approximately equal in magnitude (R < 0.25) and their possible flipping over the entire Tibet (Figure 6a) is admittable. Indeed, the flipping in the position of the σ₁ and σ₂ axes is consistent with the main features of active faults in the region (Figure 2). Tibet is currently deformed by both normal faults, which account for the vertical σ₁, and strike-slip faults, coherent with sub-horizontal σ₁ and σ₃.

Areas such as the central Tibetan region or the transition zone between the central Himalaya and Tibet are characterized by FMSs with heterogeneous kinematics, i.e., strike-slip, normal, and reverse faulting in the former and thrust (southward) and normal (northern) faulting in the latter, respectively. Despite these complex patterns, the achieved results are quite similar to the ones obtained for the neighboring cells, also evidencing an acceptable reliability for these estimations.

The σ_Hmax trajectories show a perpendicular trend with respect to the Himalayan Arc, mainly in the westernmost part, in the easternmost part of Nepal, and the Assam region (Figure 6a). The σ_Hmax on the Tibetan Plateau shows a pattern predominantly oriented along the N-S attitude till latitude 31°N. To the north, it rotates clockwise at approximately 30° and aligns NE-SW. This main direction can be observed northward, up to the Kunlun Fault. To the southeast, the σ_Hmax trajectories depict a clockwise rotation, passing from NW-SE to N-S and NE-SW above the Burma subduction system.

Figure 6b provides the spatial distribution of the estimated stress ratio (R) values. As previously mentioned, the stress ratio R varies from 0 (σ₁ = σ₂) to 1 (σ₂ = σ₃); therefore, in the following, we defined three main groups with 0.25 < R < 0.75, R ≥ 0.75 and R ≤ 0.25. Regions with 0.25 < R < 0.75 are sparsely distributed throughout the study area, generally lacking a distinct pattern and representing transitional values for the other two groups. Regions with R < 0.25 can be recognized across a large sector of Tibet, where, as pointed out before, σ₁ and σ₂ axes flip among them, evidencing that the crustal deformation is accommodated by the coexistence of active strike-slip and normal faults.

Regions with R > 0.75 are the Himalayan Arc, the area north of the Qaidam Basin and the northwest border of the Tarim Basin. In all these zones, the σ₁ axis is sub-horizontal, while σ₂ and σ₃ can alternatively represent the vertical stress, therefore evidencing the coexistence of the strike-slip and reverse faulting regimes.

5.3. Stress–Strain Comparison

The comparison of geodetic strain and seismological stress is carried out by comparing the azimuthal pattern of the ε_Hmin (shortening axis) and the σ_Hmax (maximum horizontal stress) (Figure 7). Strain and stress parameters have been estimated on grids with different sizes which, however, share several nodes; therefore, the comparison of their azimuthal patterns has been performed by considering the estimated values at the shared nodes. A remarkable feature of this comparison is that, in 90% of cases, the estimated angular discrepancy is less than 15°, even in those regions where strain and stress axes show spectacular rotations (Figure 7). Considering uncertainties associated with both strain and stress fields (see Supplementary Material), the methodology used, the type of data, and their spatial distribution, these cases could be considered almost concordant. Based on this last consideration. This alignment is consistent with a macroscopically elastic mechanical response of the sampled upper crust, in which deformation tends to occur in the same direction as the applied stress. However, we acknowledge that this interpretation relies on a simplified model of isotropic elasticity, which may not apply in the presence of elastic anisotropy or more complex behaviors such as plasticity or viscoelasticity. At a regional scale, this correlation between the principal axes of stress and strain rate has been reported in previous studies [5,132]. Moreover, moderate-to-large azimuthal discrepancies can be recognized along the western (Himachal region) and eastern sectors of the Himalaya and along the NW border of the Tarim Basin, evidencing either a significant heterogeneity of the upper crust or the superposition of different stress sources.

We also observe that along the Himalayan Arc, as well as the southern sector of the Tibetan Plateau, stress and strain azimuthal patterns are roughly sub-parallel to the predicted India–Eurasia relative motion (Figure 7). At a northern of latitude 31°N, the stress–strain azimuthal pattern shows a significant oblique relationship with the predicted one. We therefore suggest that the forces arising from the convergence process control both crustal stress and surface strain, at least along the southern sector of the orogenic belt. The observed perturbation of this general pattern would be related to local complexities and/or to the superposition of secondary stress sources. Furthermore, the significant discrepancies between the observed stress and the predicted one northward of the 31° parallel provide significant evidence of a different driving mechanism that is actively shaping the ongoing tectonic setting of the Tibetan area.

6. Discussion and Conclusions

The compiled and analyzed database of GNSS data and earthquake focal mechanism solutions (see the Supplementary Material for details) allow an enhanced understanding of the kinematic zonation across the Himalaya–Tibet region. In particular, the computed GNSS velocity field, referred to the Eurasian reference frame [92], shows a fan-shaped feature (Figure 3) along the whole Himalayan Arc coupled with a large-scale clockwise rotation around the eastern sector of the Himalaya (Figure 4). Both features have been already detected by previous studies focusing on the area [88,134,135,136,137]. The fan-shaped feature along the Himalayan Arc effectively depicts the ongoing convergence between Indian and Eurasian plates, with rates decreasing northward, while the large-scale clockwise rotation appears related to deep processes involving the subducting lithosphere [88]. This computed GNSS velocity field has been used to estimate the horizontal strain-rate fields (Figure 5). The achieved results highlight a contractional pattern along the Himalayan orogeny, related to the India–Eurasia convergence. Tibet is prevalently deforming due to shear strain, coupled with a small fraction of crustal stretching, in agreement with Liu et al. [138]. Deformation in the inner part of the Tarim Basin is null, despite being currently trapped between the Altyn Tagh Fault and the Tianshan orogeny southward and northward, respectively. In such a context, the Tarim Basin behaves as a rigid block, as already suggested by Chang et al. [130] and Yang and Liu [131].

The reverse FMSs located along the Himalayan Arc, effectively suggests active thrusting occurring along the MHT (Figure 4). Reverse FMSs are also located along a roughly E-W oriented belt, extending from Pamir and Tian Shan to the west, up to the Qaidam Basin to the east. All these FMSs highlight the intense deformation accounted by the reverse faults, therefore evidencing that most of the convergence between Tarim and Eurasia is absorbed farther north, along the southern margin of the Kazakh platform, in central Asia [139]. FMSs with strike-slip features dominated the Tibetan Plateau, while numerous FMSs with normal faulting can also be observed. Abundant FMSs with strike-slip features can also be observed eastward of the Burma subduction zone. The achieved stress field is in agreement with previous studies [140] and highlights that along the Himalayan Arc, the σ₁ stress axis is roughly sub-horizontal with a prevailing N-S orientation, while the σ₂ and σ₃ axes represent, alternatively, the vertical stress, being both approximately equal in magnitude (R > 0.75). Conversely, the adjacent Tibetan region is experiencing both strike-slip and normal faulting, with the σ₃ being always sub-horizontal, with average strikes ranging from W-E to NW-SE, and the σ₂ and σ₁ axes approximately equal in magnitude (R < 0.25). The computed σ_Hmax trajectories (Figure 6a) show a main N-S orientation in both the Himalayan Arc and the southern part of Tibet, while at the latitude 31°N, the σ_Hmax orientations rotate toward a NE-SW strike. To the southeast, the σ_Hmax trajectories describe a large clockwise rotation, with orientations ranging from NW-SE to N-S and NE-SW above the Burma subduction system.

The azimuthal patterns of the geodetic shortening axis (ε_Hmin) and the seismological horizontal stress (σ_Hmax) have been compared by considering their estimates at the shared nodes of the computational grids. Over most of the study area, ε_Hmin and σ_Hmax are nearly parallel. Such a general alignment implies that the crust exhibits significant elasticity, as it is contracting along the σ_Hmax direction.

Novel stress–strain maps (Figure 5 and Figure 6) provide new insights into the active geodynamic models suggested to explain the complex geodynamic setting of the Himalaya–Tibet region. We observed that stress and strain are roughly sub-parallel to the predicted motion of the India–Eurasia convergence, which therefore represents the main mechanism in controlling both crustal stress and surface strain. Along the orogenic belt, the general pattern is locally modulated by some significant perturbations; considering available geological information along with our stress results we infer that these perturbations would be related to crustal lateral density changes, strength contrasts and active faults. Evidence of crustal lateral density changes and strength contrasts along the orogenic arc have been provided by local tomographic studies [141,142,143]. Moreover, along the Himalayan orogenic arc, the magnitude of ε_Hmax reveals several regions of arc-parallel stretching, consistent with the presence of N-S-oriented normal faults [144] and reoriented strike-slip faults [145,146]. Such a feature, coupled with the fan-shaped pattern of both ε_Hmin and σ_Hmax along the orogeny, aligns well with the oblique convergence model proposed by McCaffrey and Nábělek [26].

Along Tibet, we inferred small stress ratio values (R ≤ 0.25), which suggests that at the local scale, σ₁ and σ₂ can act as the σ_Hmax. Given the close relationship between stress and strain, it can be inferred that the magnitude of the σ_Hmax axis is elastically proportional to the magnitude of the ε_Hmin axis, leading to a uniform distribution across Tibet (Figure 5c). This uniformity supports the idea that any variations in stress magnitude are more likely to occur along the vertical axis, rather than on the horizontal axis. Therefore, we propose that variation in the vertical stress is related to topographic loading, which would effectively modulate the stress related to the convergence process.

At the latitude of 31°N, our results show a ~30° rotation of the ε_Hmin and σ_Hmax axis. Such a boundary coincides at depth with the location of the northern extent of the underthrusted Indian Plate, as suggested by geophysical data [35,147]. At the surface, the boundary is not geologically well-defined; the normal and strike-slip faults [61,62,148] are quasi-homogeneously distributed at the scale of the whole Tibet. We therefore suggest that southern Tibet deforms in response to the stress field arising from the regional India–Eurasia convergence and local topographic loading, in agreement with the progressive underthrusting of the Indian Plate beneath southern Tibet. Conversely, in northern Tibet, the regional stress field related to the India–Eurasia convergence is largely modulated by the superposition of other stress sources.

Among the models proposed to explain the recent geodynamic evolution of the Tibetan region, the indenter-escape tectonic [25] model assumes that deformation is mainly confined to the horizontal plane and implies that compressive forces must be transmitted across plate boundaries, which in turn implies that the rate of the escaping must be equal to or less than the convergence rate. Moreover, this model also implies the eastward extrusion of Tibet. Such features can be easily observed on the GNSS-based velocity field (Figure 4), but limited toward the north of 31°N. Therefore, the indenter-escape tectonic model provides a plausible explanation of the motion of northern Tibet. However, it is unable to explain the vertical motion; therefore, other stress sources acting on the vertical component, such as the topographic loading, must be considered. Indeed, the combined action of both indenter-escape tectonic and topographic loading would provide a reasonable mechanism to properly explain the ongoing geodynamic evolution of this region.

An additional attempt to explain the complex deformation of the Himalaya–Tibet region was the channel flow model proposed by Bird [27]. This author hypothesized the presence of a deep zone in the Tibetan crust, hosting a planar channel flow which allowed the Himalayan Arc to expand both radially and circumferentially. However, our stress and strain trajectories do not support the channel flow model.

In conclusion, we support the combination of different geodynamic models, refining the model proposed by Copley et al. [149]. South of the 31°N latitude, compression is driven by basal drag from the underthrusting Indian Plate, while northward, escape tectonics dominate, where the eastward extrusion of the Tibetan Plateau becomes evident. All over Tibet, we can see no difference in terms of E-W stretching, leading to the idea that the normal faulting regime is due to topographic loading, given the uniform elevation of the plateau [24,149,150].

Geodetic and seismological data collected and analyzed in this study are provided as Supplementary Material so that they can be used as input for additional studies. Future efforts should point toward an improved estimation of the crustal vertical motion, making this pattern a significant key element of the complex geodynamic processes of the Himalaya–Tibet region.