The XSHF is an active block boundary fault developed on the southeastern margin of the QTP. It is a part of the Xianshuihe-Xiaojiang fault system, which consists of several other great faults, such as the Ganzi Yushu, Anninghe, Daliangshan, and Xiaojiang fault zones (Figure 1A) (King et al., 1997; Zhang, 2013). It acts as the southern boundary of the Bayan Har Block and the eastern boundary of the Sichuan-Yunnan Block (Allen et al., 1991; Chevalier et al., 2018). The tectonic stress field of this region is highly influenced by the northeastward intracontinental subduction of the Indian Plate and the clockwise rotation of the Sichuan-Yunnan Block, which resulted in the left-lateral strike-slip movement of the XSHF (Xie et al., 1995; Sheng et al., 2022).

The XSHF is slightly arc-shaped towards the northeast, with a total length of approximately 400 km, a strike of 40°–50° NW, and a dip of 55°–80° (Figure 1B) (Zhang G. et al., 2004; Bai et al., 2018). The fault zone is not continuous but incorporates several segments and can generally be divided into northwest and southeast sections by the Huiyuan Temple Basin. The northwestern section of the fault is developed in a slightly overlapping left-stepped echelon, and it can be divided into three segments from northwest to southeast: Luhuo segment (LHS), Daofu segment (DFS), Qianning segment (QNS). The southeastern section diverges through the Huiyuan Temple Basin to form multiple secondary faults, i.e., Zheduotang segment (ZDS)/Kanding segment (KDS)/Yalahe segment (YLS), the fault trace overlaps with Moxi segment (MXS) in Kangding area, and connects with the Anninghehe and Daliangshan fault zones near Shimian County (Figure 1B).

The slip rate is one of the most significant indicators for evaluating the activity, accumulated strain energy, and seismic potential of a fault (Molnar and Tapponnier, 1978; Zhang et al., 2008; Liu et al., 2021). High slip rate indicates that the fault has strong activity and high accumulated strain energy (Li et al., 2018; Li et al., 2019), which can be used to determine fault stability and seismic hazard (Li et al., 2020). Several researchers have studied the Holocene slip rate of the XSHF (Table 1). The results show that the fault exhibits a shallow creep slip behavior, and the slip rate gradually increases from northwest to southeast along the fault (Zhang et al., 2018; Li et al., 2020; Qiao and Zhou, 2021; Li et al., 2022a; Zhang et al., 2022). Qiao and Zhou (2021) found a decrease in the strike-slip rate for the northwest segment. They inferred that the branch faults may have been locked, with a locking depth of approximately 7.6–18.5 km. The high creep-slip rate of the southeastern segment of the XSHF suggests high seismic hazard (Li et al., 2020), while the yearly decrease in slip rate and deeper locking in the northwestern segment of the XSHF also suggests possible seismic hazard (Qiao and Zhou, 2021).

As a boundary fault in the eastern QTP, the XSHF has observed strong seismic activities in the past centuries (Zhang et al., 2003). Since 1700, a total of nine earthquakes with Ms ≥ 7.0 and more than twenty earthquakes with Ms ≥ 6.0 have occurred along the fault (Wen et al., 2008; Bai et al., 2018). The seismic activity of the XSHF has been characterized by obvious periodic occurrences, that is, two active periods and one quiet period since 1700 (Papadimitriou et al., 2004; Wen et al., 2008; Zhang et al., 2016). The first active period lasted 116 years, from the Kangding Ms>6.5 earthquake in 1700 to the Luhuo Ms7.5 earthquake in 1816. The second active period lasted 129 years since the Huiyuan Temple Basin Ms7.0earthquake in 1893 (Wen et al., 2008). During these periods, several characteristic earthquakes occurred in the region, such as the 1816 (Ms7.5), 1923 (Ms7.3), and 1973 (Ms7.6) earthquakes in Luhuo and the 1725 (Ms7.0), 1786 (Ms7.75), and 1955 (Ms7.5) earthquakes in Kangding (Figure 2).

We analyzed the distribution characteristics of in-situ stress surrounding the fault zone. The horizontal and vertical principal stress components with respect to the buried depth are shown in Figure 3. The results indicate that, in general, the maximum horizontal principal stress σ H dominates the stress field close to the XSHF at depths shallower than 1400 m, having a stress magnitude of 1.7–53.5 MPa. We further investigated the stress fields of study areas I and II to understand the effects of different locations around the fault. In study area I, the in-situ stress magnitudes are very discrete at shallow depths, especially at depths shallower than 400 m (Figure 3B). According to the Andersons’ fault stress model (Anderson, 1951), the stress state in each area determines the fault types, namely, strike-slip, reverse, and normal types. At depths greater than 400 m, the vertical stress component gradually becomes the intermediate principal stress, indicating that the fault type at greater depths is mainly strike-slip. The southeastern segment of the XSHF in study area I is located at the intersection of the Y-shaped tectonic zone, which is one of the most active areas in the world, resulting in an extremely complex shallow crustal stress state. In study area II, the in-situ data were obtained mainly within a depth of 250 m (Figure 3C). The stress field characterized ( σ H > σ h > σ V ) by the in-situ stress data indicates an inverse faulting stress regime. However, the seismic source mechanism solution reflects a normal fault stress state around the step over between the Yushu fault and XSHF with left-lateral strike-slip motion (Zhou et al., 1983; Wen et al., 2008). This indicates a complex stress state near the fault tip, coupled with topography and geomorphology at shallow depths (e.g., Liu et al., 2005; 2011). Previous studies have reported a similar phenomenon in other pull-apart basins, for example, a reverse fault stress state exists in the basin formed by step over Haiyuan strike-slip fault (Lin et al., 2017; Lei et al., 2021).

The slope of the regression line shows that the change gradients of σ H , σ h , and σ V in study areas I and II are 3.23 MPa, 2.57 MPa, and 2.68 MPa and 6.47 MPa, 4.68 MPa, and 2.61 MPa per 100 m, respectively. The magnitude of the maximum horizontal principal stress component in study area I is 3.3564 MPa, which is greater than that in study area II (2.8931 MPa), indicating that the horizontal tectonic stress near the Earth’s surface is stronger in the southeastern section.

To further analyze the relationship between the horizontal and vertical stresses at different depths around the fault zone, the lateral pressure coefficient K a v proposed by Brown and Hoek (1978) was adopted (Eq. 1), as shown in Figure 4. The results indicate that the lateral pressure coefficient is relatively discrete at shallow depths. As the buried depth gradually increases, K a v gradually converges to approximately 1.03, which is slightly greater than 1.0, indicating that the horizontal stress dominates the stress field within the depth considered in this study. Previous studies have shown that the lateral pressure coefficient K a v of the Sichuan-Yunnan block tends to be 0.95 within a buried depth of 1200 m (Zhang et al., 2022) and that of the middle China North-South Seismic Belt tends to be 0.628 within a buried depth of 749 m (Yang et al., 2012). A comparison of K a v between the XSHF and the other two zones indicates that the XSHF and the adjacent areas withstand stronger horizontal compression. The stress state of the southeastern segment of the fault zone (Study area I, Figure 4B) is more complex than that of the northwestern segment (study area II, Figure 4C), which may be mainly due to the intense tectonic activities and the steeper and higher terrain in study area I. The Earth’s surface processes have substantial effects on the stress field within a certain buried depth close to the Earth’s surface. K a v = σ H + σ h 2 σ V (1) where K a v is the lateral pressure coefficient; σ H is the maximum horizontal principal stress (MPa); σ _V is the vertical stress (MPa); σ h is the minimum horizontal principal stress (MPa).

To analyze the change characteristics of the in-situ stress at different locations around the fault zone, it is necessary to ensure that the buried depths of the in-situ stress measurement points are at the same level. Therefore, we converted the in-situ stresses at different depths to those at the Earth’s surface using Eq. 2 (Brady and Brown, 2006). We obtained the values of maximum shear stress τ _m at each measuring position using Eq. 3, which are equal at either buried depth according to Eq. 2 (Jaeger et al., 2009). σ H o = σ H − k γ h (2) τ m = σ 1 − σ 3 2 (3) where k and γ denote the coefficient of lateral pressure and gravity (kN/m³) respectively; h is depth (m); σ H , τ m , σ 1 , and σ 3 are the maximum horizontal principal stress, maximum shear stress, and maximum and minimum principal stress (MPa), respectively; σ H o is the maximum horizontal principal stress (MPa) at the Earth’s surface.

In study area I, the fault segment is composed of four branched faults, the ZDS, KDS, YLS, and MXS, from southwest to northeast (Figure 1). We found that the maximum horizontal principal stress presents a hump shape on both sides of the fault zone, which can be depicted as a double-hump-type stress field (Figure 5A). There is a low stress zone (hump valley) between YLS and ZDS; as the distance increases from the YLS and ZDS, the stress magnitude gradually increases on both sides and reaches a peak (hump), then decreases and eventually stabilizes. We calculated the maximum shear stress and obtained its average magnitude at different buried depths for each site. As the fault distance increases, the magnitude of the maximum shear stress also shows a similar “double hump” trend (Figure 5B). However, the stress field around the LHS in study area II is not as complex as that in study area I. This indicates that the maximum horizontal principal stress exhibits a “single hump” distribution feature. The stress magnitude reaches a hump near the fault zone, decreases gradually, and then tends to be consistent with the far-field stress field with an increase in the fault distance (Figures 5A, B). The results indicate that the stress always increases near the fault zone, generally within the relatively intact host rocks neighboring the fracturing rocks. The double and single humps near the fault zone primarily stem from the width of the fault zone. A wider fault zone leads to a reduction in stress in the fractured rock mass (Sibson, 1980; Weng et al., 2020).

We interpolated the in-situ stress using the inverse distance weighting method of ArcGIS and obtained the stress contours for study areas I and II (Figures 6A, B). In study area I, a stress concentration (14.3–33.6 MPa) occurs at the southwestern side of the fault tip (the red area of the ZDS in Figure 6A), while a stress relaxation (4.5–9.1 MPa) occurs at the northeastern side of the fault tip (the green area of the YLS in Figure 6A). Such stress distributions mainly result from the movement of the left lateral strike-slip fault zone. The stress concentration and relaxation areas around the southeastern fault zone correspond to the Minya Konka and Yalahe valleys, respectively (Wen and Bai, 1985), which may indicate the possible effect of the sinistral strike-slip on the formation of the geomorphology around the fault tip (Figure 6C) (Guo et al., 2016). In study area II, a significant stress variation is observed around the fault tips, reflecting the influence of the fault zone on the surrounding stress field.

At locations near active faults, the stress orientation is often deflected and inconsistent with the regional stress field, and it is important to understand the stress disturbance around faults, which is still less understood (Zoback, 1992; Chang et al., 2010). We plotted the in-situ maximum horizontal principal stress orientations around the fault zone and the three stress zones using a rose diagram (Figure 7). The results showed that the dominant stress orientation in Baryan Har stress zone (B217)、Longmengshan-Songpan stress zone (B218) and Sichuan-Yunnan stress zone (B219) are 20–40°NE, 280–320°NWW and 290–310°NWW, respectively (Figure 7A). The dominant stress orientation near the northwest tip of the fault is approximately 50°–70° NEE, holding a large intersection angle with the fault. The stress orientation on the two sides of the fault zone is generally in the NW direction, which is in accordance with the fault strike. The dominant stress orientation near the southeastern tip of the fault is NE, which is nearly perpendicular to the fault. Overall, the orientation of the maximum horizontal principal stress exhibits an elliptical shape around the XSHF, which is significantly different from the dominant stress orientation of the three neighboring stress zones. (Figure 7). This type of stress distribution has been reported in several studies (Yale et al., 1994; Scholz, 2000; Su, 2002; Yale, 2003; Zhao et al., 2013). In the elliptical-shaped stress field, the stress orientations exhibit a deflection of approximately 30° between the two sides of the fault in study area I, and the largest deflection occurs around the southeastern tip, reaching nearly 90°. Such a deflection of the nearby stress orientation may result from the extreme heterogeneity in the mechanical properties of the fault zone, fractured rocks, and nearby undisturbed host rock mass (Hudson and Cooling, 1988; Hardebeck and Hauksson, 1999; Provost and Houston, 2001; Su et al., 2003). The in-situ stress analysis indicates that the local stress field around the XSHF is highly disturbed, which may be influenced by a series of factors, such as fault properties, fault spatial distribution, and tectonic activity intensity.

High ground stress tends to drive fault slip and trigger earthquakes (Elsworth et al., 2016). Near the LHS, ZDS, YLS, and MXS, the σ H and τ m are relatively high compared with the far-field stress. The stress concentration around the tips or intersections of fault segments provides the requisite conditions for earthquakes, i.e., study areas I and II. In fact, over half of the historical earthquakes in the XSHF with Ms ≥ 6.0 occurred in these areas (Figures 6A, B).

The sliding potential of a fault is defined by the ratio of the shear stress τ to the normal stress σ acting on the fault (Eq. 4; Morris et al., 1996). When the shear stress τ reaches its sliding resistance, the fault slips, converting the accumulated strain energy into kinetic energy and thereby causing crustal deformation and even earthquakes (Tanaka et al., 1981; Wang et al., 2012). According to Jaeger et al. (2009), the ratio μ m of the maximum shear stress τ m to the maximum normal stress σ m can be expressed as a function of the friction coefficient μ (Eq. 5). The pore water pressure P 0 is also considered because it can weaken the effective normal stress (Chang et al., 2010). τ = μ σ = ⁡ tan ⁡ φ σ (4) μ m = τ m / σ m − P 0 = σ 1 − σ 3 / σ 1 + σ 3 − 2 P 0 = μ / 1 + μ 2 1 / 2 (5) where σ 1 and σ 3 are the maximum and minimum principal stresses acting on the fault plane, respectively; P 0 denotes the pore pressure, which can be approximately represented by water column static pressure, i.e., P 0 = ρ h in this study, where ρ and h denote the density of water and buried depth, respectively.

The larger μ m signifies a higher degree of stress accumulation, and when the stress level along a fault reaches a threshold value, the stress is released through fault activity, and then μ m decreases to maintain the stability of the crust (Zhang et al., 2022). Townened (2000) analyzed boreholes drilled data and found that the stress magnitudes are consistent with Coulomb frictional-failure theory for coefficients of friction of 0.6–1.0 such as are measured experimentally in laboratory settings (Byerlee, 1978), and noted that the stress accumulation degree is high when μ m between 0.5 and 0.7, and that faults are susceptible to sliding instability when μ m exceeds 0.7 (Townened and Zoback, 2000). However, some studies have indicated that the friction coefficient of fault rocks can weaken to approximately 0.2 due to special composition of the fault gauge (Qin et al., 2014). and Li et al. (2018) showed that when μ m is between 0.2 and 0.5, the stress intensity is low and the chance of fault reactivation is small, and when μ m is less than 0.2, the fault is in a stable state. Accordingly, we calculated using Equation 5 that the range of μ m to cause fault activity and evaluate the degree of stress accumulation and stability of the fault zone (Table 3).

It is important to note that the present evaluation approach does not consider the spatial relationship between the fault strike and the axis of the maximum principal stress; therefore, the slip trend obtained represents the maximum magnitude, which may lead to an overestimation of the seismic hazard. In addition, the depths of in-situ stress measurements are limited to 2000 m, whereas earthquakes mostly occur at depths of approximately 10 km. The high stress at shallow depths always reflects the high stress at greater depths. Thus, we hypothesize that the stresses at shallow depths (<2000 m) and great depths (∼10 km) are coupled with each other (Li et al., 2023).

We calculated the effective normal stress and shear stress based on the measured stress data and then evaluated the degree of stress accumulation at each measuring position (Figure 8). On this basis, we discuss the degree of stress accumulation at different fault segments of the XSHF, i.e., Luhuo segment, Daofu-Qianning segment, Zheduotang- Kangding- Yalahe segment, and Moxi segment from northwest to southeast.

Dynamic models depicting the uplift of the southeastern QTP mainly include the lateral extrusion of rigid blocks (Tapponnier et al., 2001), lower crustal flows (Royden et al., 1997), and continuous deformation of the entire lithosphere (Molnar and Tapponnier, 1975; Houseman and England, 1993). The lateral extrusion of rigid blocks refers to the boundary force driven by the uplift deformation of the QTP resulting from the India–Eurasia collision. The rigid block is compressed and uplifted due to the movement of the strike-slip boundary faults in the southeastern QTP (Tapponnier et al., 1982; Peltzer and Saucier, 1996; Peltzer and Tapponnier, 1998; Calais et al., 2006). The lower crust flow model is driven by gravity generated by the crustal thickening of the central QTP. The weak viscous layer moves outward from the central plateau through the lower crust as a channel with uniform thickness and causes crustal uplift around the strong Sichuan and Tarim Basins (Block and Royden, 1990; Clark and Royden, 2000). Because different dynamic models correspond to stresses caused by different sources, the stress distributions can provide the information for clarifying the dynamic models for the plateau crust.

The in-situ stress field shows that the horizontal stress is always the dominant stress around the XSHF (Figure 3). The stress state in the region is characterized by strike-slip with thrust, which is consistent with previous studies that the sinistral strike-slip and thrust movements are the main crustal deformation processes in the eastern margin of the QTP (Molnar and Tapponnier, 1975; Tapponnier et al., 2001; Li et al., 2022b). Further, when the rigid crust constrains the strike-slip fault movements, the adjacent crustal rock mass is laterally shortened by translational compression (Harland, 1971; Sanderson and Marchini, 1984), followed by compensatory tectonic uplift (Lowell, 1972; Wilcox et al., 1973; Bartlett et al., 1981). Meanwhile, the study indicated that strain rate field shows highly localized shear strain along the Xianshuihe fault zone (Li et al., 2019). The results of terminating crustal extrusion near the margin of southeastern Tibetan Plateau are consistent with the decreasing slip rates along the Xianshuihe fault zone (Zheng et al., 2017), where strain localization along the major strike-slip faults ends and transfers to crustal compression and thickening. The geomorphological characteristics of the rapidly uplifting Minya–Konka fault block in this area are consistent with this process (Wen and Bai, 1985). Similar conclusion has been reached by studies in northeastern Tibetan Plateau (Li et al., 2018). Moreover, the geometric spatial distribution of the XSHF is similar to the geometric shape of the slip line in the plane indentation plasticity problem (Figure 10), and the SE–NW oriented regional horizontal principal stress direction distribution may also be related to the indentation effect caused by the India–Eurasia collision (Tapponnier and Molnar, 1976; Sylvester, 1988). The above analysis shows that the study area is controlled by strong horizontal tectonic movement caused by the India–Eurasia collision. Based on the characteristics of the stress field, the tectonic uplift of the southeastern margin of the QTP appears to be more consistent with the rigid lateral compression dynamic model.