Keywords

5.1 Introduction

On April 20, 2013, an unexpected Mw 6.7 earthquake occurred in Lushan County at the southern side of Longmen Shan (LMS). The hypocenter of this earthquake was located approximately 80 km southwest of the 2008 Wenchuan earthquake (Fig. 5.1). Both earthquakes resulted in a large number of casualties and losses [36, 37]. Some studies have suggested that the seismogenic fault was a high-angle listric-reverse fault that is probably linked to the Shuangshi-Dachuan fault (F4) [4]; however, a field investigation did not find coseismic surface ruptures [16] and did not support the idea that F4 is the seismogenic fault. Seismic and geological interpretations suggest that the seismogenic fault is a gently dipping frontal blind thrust at an approximate angle of 30° [17, 33]; however, the relocated aftershocks and focal mechanism solution indicate that the seismogenic fault dips at ~7° to the northwest [9, 41]. In addition, other geological studies have suggested that the Lushan earthquake was caused by thrusting and detachment folding in the frontal propagation belt of the LMS [21] and that the seismogenic fault was the Dayi fault (F6) in the frontal mountain [18]. Moreover, geodetic measurements, including GPS and leveling data obtained during the earthquake, suggest a maximum coseismic displacement in the Lushan syncline area [10, 14]. Furthermore, the coseismic deformation of the Lushan earthquake is not consistent with the tectonics and geomorphology of the shallow sediments and as such there is an ongoing debate regarding the location, geometry, and kinematics of the Lushan seismogenic fault [22].

Fig. 5.1
figure 1

The major fault system and earthquakes in the southern LMS thrust belt. The black solid lines represent the major faults and the dashed lines represent the blind or concealed faults

Full size image

Presently, there are many uncertainties concerning the disastrous earthquakes in the LMS area [39]. For example, calculations of the Coulomb stress variations suggest that the previously accumulated stress was not completely released by the Wenchuan Mw 7.9 and Lushan Mw 6.7 earthquakes [6, 24]. Previous studies have warned that the LMS fault zone is active and presents a high risk [1] and thus, following a devastating seismic event, it is critical to determine where and when the next big earthquake will occur. Therefore, it is increasingly important to urgently survey active faults and assess potential seismic risks in the LMS area. Solving these problems is critical to understanding the seismogenic fault and seismotectonics of the Lushan earthquake and then to estimate the potential seismic hazard along the eastern margin of the Tibetan Plateau.

Legend: 1. Coseismic surface rupture of the Wenchuan earthquake; 2. Major faults; 3. Blind faults; 4. Aftershocks of the Wenchuan Mw 7.9 earthquake; 5. Aftershocks of the Lushan Mw 6.7 earthquake; F1. Longdong fault; F2. Wulong fault; F3. Xiaoguanzi fault; F4. Shuangshi-Dachuan fault; F5. Xinkaidian fault; F6. Dayi blind fault; F7. Sansuchang fault; SGB: Songpan–Ganze Block; LMS: Longmen Shan; SCB: Sichuan Basin; LQS: Longquan Shan.

5.2 Geological Setting

The approximately 500 km long and 30–50 km wide LMS thrust belt forms the eastern margin of the Tibetan Plateau (Fig. 5.1). The belt defines a sharp topographic transition between the Songpan–Ganze Block (SGB) and the Sichuan Basin (SCB) [2], which is a known seismically active area with complex intracontinental deformation [7, 15].

The LMS area has a long and complex tectonic history and its three main deformation stages since the Paleozoic have been identified [5, 19]. A number of extensional events occurred in the western SCB and LMS during the Paleozoic to Middle Triassic [23] that resulted in the formation of normal faults. Since the Mesozoic, the LMS thrust belt has had two major periods of contractional deformation, specifically in the late Triassic and the Cenozoic [13]. During the late Triassic, the tectonic setting changed from a rifted passive margin to a foreland in response to the final closure of the Paleo-Tethys and the continent–continent collisions of the North China, South China, and Qiangtang blocks [2]. During the Cenozoic, tectonism in the LMS was reactivated as a result of the India–Asia continental collision [30] and manifested in thrusting, dextral shear, uplifting, and exhumation [8].

The LMS area currently comprises a series of northeast-trending thrust faults [2, 19, 22]. The Wenchuan earthquake occurred in the central LMS and produced two large surface ruptures along the major faults (Fig. 5.1). The southern LMS thrust belt is distinguishable by the Longdong fault (F1), the Wulong fault (F2), the Xiaoguanzi fault (F3), and the Shuangshi-Dachuan fault (F4). The deformation in the western SCB is characterized by imbricated thrusts and fault-related folds (Fig. 5.2). The Lushan earthquake occurred in the southern LMS, but field investigations have not revealed any evidence of surface rupture due to this earthquake [16, 37].

Fig. 5.2
figure 2

Geology and location of artificial seismic reflection profiles in the research area. Legend: 1. Cenozoic; 2. Cretaceous; 3. Jurassic; 4. Triassic; 5. Permian; 6. Devonian; 7. Silurian; 8. Ordovician; 9. Cambrian; 10. Sinian; 11. Proterozoic; 12. Complex rocks; 13. Major faults; 14. Thrusts; 15. Blind faults; 16. Artificial seismic reflection profile; 17. Deep seismic reflection profile; 18. Drilling wells. SYF: Shiyang fault

Full size image

The outcrops of the southern LMS are mainly Mesozoic strata (Fig. 5.2). This area is characterized by Baoxing complex rocks. Surface geological and drilling data indicate that the shallow sedimentary cover of the southern LMS contains mainly Sinian–Middle Triassic shallow marine clastic and carbonate rocks and late Triassic non-marine clastics (Fig. 5.3). An early Proterozoic crystalline basement lies under this sedimentary cover.

Fig. 5.3
figure 3

Histogram of the regional stratigraphy of the LMS and the western SCB

Full size image

There is an important shallow-level detachment (D1) that developed within the lower-Middle Triassic gypsum layers [13, 22]. This shallow detachment fault slips toward the foreland basin and forms the LQS anticline [20]. In addition, a potential deep-level detachment (D2) exists in the pre-Sinian crystalline basement beneath the southeastern region of the Tibetan Plateau [31, 38]. Multiple detachments and superimposed deformation events are characteristic features of the southern LMS.

5.3 Shallow Structural Deformation

There are three typical seismic reflection profiles in the Lushan earthquake area (Figs. 5.2 and 5.4). We can determine the shallow geological and structural characteristics of the study area from the seismic interpretation of these profiles. The Dayi blind fault (F6) developed on the shallow D1, which is located in the Mid–Lower Triassic strata (Fig. 5.4a), but the Xinkaidian fault (F5) and the Shiyang fault (SYF) developed on a shallower detachment in the Jurassic strata (Fig. 5.4b, c).

Fig. 5.4
figure 4

Geological and structural interpretation of the seismic reflection profiles A, B, and C. The yellow area represents the Mid–Lower Triassic and detachment layer (D1). The red solid lines represent the faults and the black dashed lines represent the axial surfaces

Full size image

In profiles A and B (Fig. 5.4a, b), there is a large anticline beneath the Lushan area. This anticline deformed the D1 and the Mesozoic strata. A previous study interpreted this as a fault-bend fold, and our seismic interpretations were consistent with this previous model [13]. The anticline was formed by a fault-bend folding (FBF) and developed a black thrust. The lower detachment D2 was only inferred from the axial surfaces and structural deformations.

Although the shallow sedimentary and structural deformation of the southern LMS does not reveal the information directly [22], it provides important constraints on the seismotectonics of the 2013 Lushan earthquake.

5.4 3D Coseismic Fault Model

Previous studies have used aftershocks from large earthquakes to model faults by fitting the three-dimensional (3D) surfaces to earthquake clusters using the Gocad software [3]. The approximately 2,000 events (Fig. 5.5a) of the Lushan earthquakes were relocated using the double-difference method [9]. Parallel to the strike of the LMS thrust belt (Fig. 5.5b), most of the hypocenters were confined to a depth range between 7 and 19 km. In addition, the northwest-dipping earthquake cluster intersected the southeast-dipping earthquake cluster and both formed a Y-shape. The relocated aftershocks suggest the presence of at least two coseismically active faults. According to the distribution of the aftershocks, the 3D geometry of the coseismic faults was mapped successfully [21].

Fig. 5.5
figure 5

Imaging the 3D geometry of the coseismic faults using the SKUA-Gocad software. a Map view of the aftershock hypocenters. b View parallel to the strike; the strike direction of the aftershocks is ~55°. c 3D images of the ruptured fault surfaces modeled based on the aftershock hypocenters after smoothing with the Gocad DSI algorithm. d Depth-contoured surfaces of the ruptured faults. The coordinate system is WGS 84 UTM 48 in the northern hemisphere. Approximately 2,000 relocated aftershocks were used (M ≥ 1.0)

Full size image

The aftershock hypocenters of the Lushan earthquake and the 3D images of the seismogenic fault surfaces were produced using the SKUA-Gocad software (Fig. 5.5c). We used the aftershock hypocenter locations of the Lushan earthquake to image the active faults in detail. Figure 5.5c, d shows the same view direction with a 15° plunge angle and a 35° strike.

The 3D depth-contoured surfaces illustrate the detailed geometry of faults f1 and f2 (Fig. 5.5d). The surfaces of these faults are inhomogeneous and contain multiple wave bends. Fault f1 is the major seismogenic fault of the Lushan earthquake and has a 45°–50° dip, whereas fault f2 is a branching backthrust of f1.

5.5 Seismic Interpretation and Quantitative Analysis

Profile E–E′ is a high-resolution petroleum industry seismic reflection profile between the southern LMS and western SCB (Fig. 5.6). This profile crosses the major fault-folding belt, and provides information on the shallow sediments and structures from the subsurface down to a depth of ~10 km. The Lushan syncline lies between the F4 and Xinkaidian fault (F5) (Figs. 5.1 and 5.2) and the hypocenter of the Lushan earthquake is beneath the F4 at a depth of 12–13 km [9, 41].

Fig. 5.6
figure 6

Geological map of the southern LMS with the locations of the artificial seismic reflection profile, drilling wells, and GPS and leveling stations described in this study. Legend: 1. Quaternary; 2. Paleogene; 3. Cretaceous; 4. Jurassic; 5. Triassic; 6. Paleozoic; 7. Proterozoic and complex rocks; 8. GPS stations and measurements; 9. Leveling stations and measurements; 10. Drilling wells; 11. Petroleum industry seismic reflection profile; 12. Scale for GPS data [14]; and 13. Scale for leveling data [10]

Full size image

In this seismic profile (Fig. 5.7), the labels on each layer denote certain seismic reflectors in the sediment rather than stratigraphic boundaries. The seismic reflectors in the Lushan syncline area are clear and continuous, suggesting that no thrust faults extend into the syncline; however, the partially enlarged portion of the seismic profile clearly shows the deformation features on top of the aftershock cluster (Fig. 5.7E1). The seismic reflectors of the Lower Paleozoic units are discontinuous and may suggest the presence of a fault that cuts across the Lower Paleozoic and some Upper Paleozoic layers [21]. The location of this fault is consistent with the location of the 3D-mapped seismogenic fault f1 (Fig. 5.5). Obviously, it has not broken through the Upper Triassic strata and is possibly a blind fault.

Fig. 5.7
figure 7

Comprehensive analysis of the high-resolution seismic reflection profile E–E′

Full size image

The seismic reflection profile E–E′ is ~11 km away from the Lushan Mw 6.7 mainshock. Section E1 displays a detailed seismic reflection image of the deformation features and strong seismic reflectors. It shows aftershocks for a range of ~4 km. The green arrows indicate the characteristics of the same layers. Section E2 shows the quantitative angle measurement of the fault f3. The geometry of faults f1 and f2 was simplified from the 3D images.

FBF: Fault-bend folding (the simple shear FBF model II was modified from [40]). D1: shallow-level detachment; D2: possible deep-level detachment. The black dashed lines represent the axial surfaces and some speculated faults are shown as red dashed lines.

The detachments at the various levels may play a prominent role in the LMS and western SCB (Fig. 5.3). Some shallow detachments are parallel to the sedimentary layers and lead to shallow structural deformation [20]. Some detachments have propagated into the western SCB and produced fault-related folds during the Cenozoic, such as the Longquan Shan (LQS) anticline and Xiongpo anticline [12, 17, 33]. The seismic interpretation of profile E–E′ suggests that a shallow D1 occurs in the Mid–Lower Triassic gypsum beds. Some Mesozoic- and Cenozoic-age thrusts have developed on this detachment in the western SCB (Figs. 5.3 and 5.7E2). According to the S-wave velocity structure and structural analysis, a deep D2 is inferred to exist in the basement rocks at a depth of ~20 km [32]; however, the D2 may not be a regular horizontal surface at an accurate depth. Based on the distribution of the Lushan aftershocks and the seismic interpretation (Figs. 5.5 and 5.7), the D2 was determined to be located at a depth of approximately 19 km in this profile.

Between detachments D1 and D2, there is a large anticline beneath the F5 (Fig. 5.7), which is similar to the three shallow seismic profiles (Fig. 5.4). The f1 and f2 faults were simplified as two straight lines from the 3D images (Fig. 5.5). A main thrust fault f3 cuts the basement and merges with D1. We measured the dips of the layers and calculated the dip of fault f3 using the classical FBF theory [28]. The accumulated slip of fault f3 is ~3.6 km (Fig. 5.7E2). The seismic profile indicates that the f3 fault cuts the layers beneath D1 with a cutoff angle (θ3) of 12°. The cutoff angle θ2 is ~21° and fault f3 likely converges with fault f1 at a depth of ~12 km [21]. There is no evidence to show that the basal layer slides shorten and thicken above the ramp. The D2 detachment layer undergoes contraction and has a simple bedding-parallel surface with no basal fault slip but the layers above the ramp have no bedding-parallel surface. The geometry of fault f3 and the analysis of the axial surfaces suggest that the structure of f3 is consistent with the shear FBF model, which is characterized by long, gentle back limbs that dip less than the fault ramp [29].

Geodetic crustal displacement data is important for the study of active tectonics [27, 34], and can provide independent constraints for fault kinematics, particularly when no surface rupture features are available [14]. The GPS and leveling data collected during the 2013 Lushan earthquake indicate that the maximum coseismic deformations occurred in the Lushan syncline area (Fig. 5.8). In particular, at GPS station LS05 located ~15 km south of the epicenter an ~68 mm horizontal displacement and ~84 mm uplift were observed (Fig. 5.8E3). The angle of the coseismic deformation was approximately 50°, which is similar to the dip of fault f1. The leveling data suggest that the maximum coseismic displacements appeared at station D35 (Fig. 5.8E4) with an uplift of ~198.4 mm [10].

Fig. 5.8
figure 8

Quantitative analysis of the geometry and kinematics of the seismotectonics of the 2013 Lushan earthquake

Full size image

Section E3 shows the geological and structural interpretation of the seismic section in Fig. 5.7E2. The coseismic deformation obtained from the GPS data is shown at the top. The simple shear FBF model II was modified from [29]. Section E4 shows the geometry and kinematics model of seismogenic fault f1. The coseismic deformation of the leveling data is displayed above fault f1. Layer 1 indicates the deformation and the total uplift area, which are similar to the coseismic deformation of the leveling data.

In this study, we mapped the area of the uplift resulting from the coseismic deformation and found that it has broad limb rotation in the back limb (Fig. 5.8E4). The back-dip (δb) measured from the seismic reflection profile is 4° and the ramp dip (θ1) is 47°. According to the simple shear FBF model and relevant formulae [26, 29], the shear angle (αe) is ~5°. The instantaneous kinematics and total uplift of the coseismic deformations indicate that the seismotectonics of the Lushan earthquake correlate with the simple shear FBF model [40], which successfully explains the back limb uplift [21]. This result suggests that the upper crust in the study area has a brittle deformation zone above ~10 km and a brittle–ductile translation or shear zone between 10 and 20 km. This brittle–ductile translation zone is consistent with the results of previous studies [32, 38].

5.6 Seismotectonics of the Lushan Earthquake

A deep seismic reflection profile D–D′ was collected following the 2013 Lushan earthquake (Fig. 5.2). This seismic profile shows the image of a deep structure beneath the Lushan syncline (Fig. 5.9). The seismic reflectors suggest that the Triassic and some Upper Paleozoic layers are roughly continuous [21], which is consistent with the other profiles (Figs. 5.4 and 5.7). However, the Lower Paleozoic layers exhibit a prominent offset above the mainshock of the Lushan earthquake, which suggests the presence of a fault cutting the Lower Paleozoic and basement rocks.

Fig. 5.9
figure 9

Deep artificial seismic reflection profile D–D′ and the coseismic faults of the Lushan earthquake

Full size image

The focal mechanism of the 2013 Lushan earthquake suggests that the seismogenic fault f1 is a thrust fault (Fig. 5.1). However, the deep seismic profile D–D′ shows that it is a typical normal fault, as there are thin layers in the footwall of f1 but thick sediments in the hanging wall between the Upper Paleozoic and Lower Paleozoic units (Fig. 5.9). These syndepositional features formed in response to the growth of fault f1. Based on the tectonic evolution of the LMS [19, 23], we suggest that fault f1 was a normal fault during the Paleozoic (Fig. 5.9D1).

The seismogenic fault f1 involves positive inversion tectonics. It is just starting to form (Fig. 5.9D2); however, as it has not broken through the postrift strata, as positive inversion structural evolution models suggest [35], and has barely caused any structural deformations. Fault f3 is the main fault with a slip at ~3.6 km (Fig. 5.7E2), which suggests that it formed and tectonically inverted during the Late Cenozoic.

In this figure, the seismically active f1 and f2 faults are simplified as two straight lines. Section D1 shows the coseismic faults and the geological model of the 2013 Lushan earthquake, suggesting that fault f1 was a normal fault prior to the Mesozoic period. Section D2 shows the schematic of a classical positive inversion structure. The stratigraphic sequences “a”, “b”, and “c” are prerift, synrift, and postrift, respectively (modified from [35]).

Previous studies demonstrated the formation of several normal faults prior to the Late Triassic period in the LMS area [2, 13]. Since the Cenozoic, some pre-existing syndepositional normal faults facilitated the structural inversion that transformed them into thrust faults in a compressional regime [5]. In the LMS area and the western SCB, several of these pre-existing normal faults were structurally inverted and transformed into active thrust faults in a contractional setting [7, 30].

Clearly, the seismotectonics of the 2013 Lushan Mw 6.7 earthquake are different from the seismotectonics of the 2008 Wenchuan Mw 7.9 earthquake (Fig. 5.10). The Wenchuan earthquake ruptured the upper crust with two main faults in the LMS [36], but the seismogenic fault f1 responsible for the 2013 Lushan earthquake is a blind fault that was reactivated during the Late Cenozoic [22]. It also differs from the 2016 Mw 6.4 Meinong earthquake in Taiwan, which was triggered by a blind thrust fault with fault propagation folds [11]. Both faults f1 and f3 were developing but did not break through the shallow detachment D1. Therefore, we suggest that the Lushan earthquake activity on the seismogenic fault f1 is a hidden earthquake [25].

Fig. 5.10
figure 10

A 3D geological and structural model showing seismotectonics of the Lushan earthquake. Legend: 1. Cenozoic; 2. Cretaceous; 3. Jurassic; 4. Upper Triassic; 5. Mid–Lower Triassic; 6. Upper Paleozoic; 7. Pre-Cambrian and basement; 8. Major faults; 9. Main detachments; 10. Axial surfaces

Full size image

It is known that many active and blind faults exist in the frontal LMS and western SCB [33]; furthermore, several scientists believe that the accumulation of strain energy in the southern LMS area was not completely released by the 2013 Lushan event [6]. The seismotectonics of the 2013 Lushan earthquake show that inversion tectonics could trigger a strong earthquake, and these were an important factor in the development of the southern LMS during the Late Cenozoic. Furthermore, the blind fault led to a hidden earthquake, which suggests that the frontal ramp of these reactivated faults poses a potential seismic risk. With the ongoing compression of the Tibetan Plateau toward the western SCB during the Late Cenozoic [42], the level of seismic risk in the LMS and western SCB area is increasingly uncertain.

5.7 Conclusions

The 2013 Mw 6.7 event was a hidden earthquake that was ruptured by an unknown blind thrust fault in the frontal Longmen Shan (LMS). Three-dimensional imaging of the aftershocks was used to identify two planar faults that form a Y-shape with the major seismogenic fault f1 dipping to the northwest at 45°–50° at a depth of 7–19 km. Seismic interpretations suggest that the seismogenic fault is a typical blind fault that did not penetrate into the overlying Mesozoic and Cenozoic units and it is not a Shuangshi-Dachuan fault (F4) or a frontal Dayi buried fault (F6). Deep seismic data illustrate the syndepositional characteristics of the seismogenic fault. Our study further demonstrated that the seismogenic slip occurred on a reactivated pre-existing normal fault older than the Triassic age that triggered earthquakes during the Late Cenozoic.

There are two major detachments that control the structural deformation of the upper crust in the LMS and western Sichuan Basin (SCB), resulting in multiple superimposed deformation events. Based on the structural and coseismic deformation, we argue that the main seismotectonics of the Lushan earthquake are consistent with the simple shear fault-bend folding model. Clearly, the 2013 Lushan earthquake was not a characteristic earthquake. Due to the strong compression of the Tibetan Plateau toward the SCB, the early-period normal faults were activated after inversion and triggered the Lushan earthquake. The concealed and reactivated faults increase the potential risks related to earthquakes in the LMS and the western SCB.