Double rupture event in the Tianshan Mountains: A case study of the 2021 Mw 5.3 Baicheng earthquake, NW China

1 Introduction

At 21:14 universal time coordinated on March 23, 2021, an Mw 5.3 earthquake occurred in Baicheng County, Xinjiang Uyghur Autonomous Region, northwestern China. The epicenter was located at 41.82° N and 81.16° E and had a focal depth of 10 km, according to the United States Geological Survey (USGS). The earthquake epicenter was located in the Baicheng Basin, 64 km west of Baicheng County in the southern Tianshan Mountains. The Baicheng Basin is a structural depression between the southern Tianshan tectonic belt to the north and the Qiulitag anticline to the south and is located over a series of nearly east–west-trending fold structures. Owing to the collision of the Indian and Eurasian plates, which causes extensive deformation in Eurasia, the landforms in western China are characterized by alternating mountains and basins [1]. The northern margin of the Tarim Basin is the Tianshan arc ridge. A large number of northward-thrusting faults formed between the ridge and basin, which have caused a series of thrust earthquakes in these areas. According to the results of the global positioning system (GPS) velocity field at the Baicheng earthquake epicenter and in the surrounding area, the horizontal GPS velocity on the northern margin of the Tarim Basin is ∼15 mm/year; however, the velocity near the Qiulitag Fault and Kumugeremu Fault Zone is ∼7 mm/year, indicating that the region absorbs a horizontal velocity of ∼8 mm/year [2]. The long-term velocity of the Qiulitag Fault Zone is only 1–4 mm/year based on geological surveys [3], which is far less than the GPS velocity. It has been speculated that a series of parallel fold and thrust belts located at the intersection of the Tarim Basin and South Tianshan Mountains absorbs most of the horizontal velocity [4]. The Baicheng Mw 5.3 earthquake occurred at the intersection of the Tarim Basin and the foothills of the South Tianshan Mountains, which has the largest GPS horizontal shortening rate in the region.

Large-scale faults are located on both sides of the Baicheng Basin: the Kumugeremu and Kasangtuokai faults to the north and the Qiulitag Fault to the south, all of which have been active since the late Pleistocene [5]. Based on the focal mechanism solutions of historical earthquakes with magnitudes >5 since 1976, the earthquakes that occur in this region are mainly thrust events [6,7] (Figure 1). A magnitude 5.7 earthquake occurred on January 5, 1987, and two magnitude 5.1 earthquakes occurred on September 26, 1995, and July 28, 1998, near the epicenter of the Baicheng earthquake, both of which were thrust events. The focal depths of the two M 5.1 earthquakes reached 40 km, whereas the 1987 M 5.7 earthquake had a focal depth of 15 km, indicating that the seismogenic environment in the region is complex [8]. The Global Centroid Moment Tensor (GCMT) and other agencies obtained the source parameters of the March 23, 2021, Baicheng earthquake (Table 1), but the results differed. Investigating the seismogenic structure of earthquakes is an effective method for understanding the blind faults in this region and is important for understanding the tectonic units around the foothills of the South Tianshan Mountains and Tarim Basin.

Figure 1

Map of the Baicheng Mw 5.3 earthquake and aftershock locations, tectonic features, and SAR data range. Inset shows the footprint of the Sentinel-1A ascending track image with the track number (red polygon), Baicheng earthquake mainshock (large green star), Baicheng earthquake foreshock (small green star), Baicheng earthquake aftershocks (colored circle sizes proportional to magnitude), historical earthquakes during 1976–2022 (colored beach balls, circle sizes proportional to magnitude), and all earthquake marks colored according to focal depth. The blue rectangle in the top inset indicates the location of the area shown in the main map.

Interferometric synthetic aperture radar (InSAR) has the advantages of high spatial and temporal resolutions, low operational costs, and wide geographic coverage compared to the global navigation satellite system (GNSS). Crustal deformation data with a high spatial resolution can also be obtained without ground control points, and important constraining parameters can be provided for slip distribution inversion. InSAR has been widely used for geologic hazard monitoring since the beginning of the twenty-first century, owing to the development of synthetic aperture radar (SAR) satellites and processing technology. Researchers have introduced this technology into seismic research, particularly in recent years, including high spatial resolution InSAR deformation field acquisition [9,10,11], seismogenic structure analysis [12,13,14], slip distributions and seismogenic fault plane analysis [15,16], and the locking characteristics of active faults [17,18]. Inversion of source slip distributions requires reliable observations, and seismic waveforms, GNSS, and InSAR can provide the appropriate data constraints for inversion. InSAR has a higher spatial resolution than GNSS and can provide continuous surface deformation coverage. GNSS stations are sparse in the region around the epicenter of the Baicheng earthquake, and no GNSS data were available for the region near the epicenter prior to the event. Therefore, InSAR provides the best data source for the focal slip distribution inversion of the Baicheng earthquake.

Previous studies of co-seismic deformation have mostly focused on earthquakes with M ≥ 6, as the surface deformation caused by strong earthquakes is relatively large in most cases. We performed a field survey after the Baicheng earthquake and observed a 4 km surface rupture length at the epicenter. Surface ruptures caused by faulting are rarely reported for M < 6 earthquakes; thus, the Baicheng earthquake provides a unique opportunity to improve seismic hazard knowledge associated with moderate events [19,20,21,22]. Yao et al. [23] studied the seismogenic structure of the 2021 Mw 5.3 Baicheng earthquake using InSAR data and unmanned aerial vehicle images and observed two deformation centers on the interferometric images. They proposed that the Baicheng event was a left-lateral strike-slip earthquake caused by the rupture of a buried compressional salt–related structure. However, they did not discuss the reason for the two adjacent deformation centers observed at the surface. In this study, we used two sets of Sentinel-1A satellite data (before and after the Baicheng earthquake) to obtain the deformation field using differential InSAR technology, from which we inverted the slip distribution. The characteristics of the seismogenic structure were also analyzed based on the InSAR and surface rupture data.

2 Data and methods

3 Results

3.1 Interferometric displacement

This study only provides the LOS deformation of the ascending data owing to data availability. The deformation interferogram indicates that the surface deformation caused by the Baicheng earthquake had an elliptical distribution, with a long axis of ∼20 km (east–west) and a short axis of ∼10 km (north–south). The long axis of the deformation field is approximately in the SWW direction, which is consistent with the westward extension of the Kumugeremu Fault. The deformation field is also symmetrically distributed on the northern and southern sides of the fault extension. The density and number of interference fringes indicate fault deformation, and the denser interference fringes represent a larger spatial gradient of surface deformation [44]. The interferogram (Figure 2a) includes approximately three color-cycle changes on the northern side of the fault, indicating that the northern side of the fault was displaced by ∼10 cm in the LOS direction. The southern side of the fault includes 2.5 color cycles, indicating that the southern side of the fault was displaced by ∼9 cm. The deformation stripes were also divided into eastern and western sections. The spatial range of the stripes on the eastern side was larger than that on the western side, and the color cycles had the same pattern, indicating that the deformation caused by the earthquake exhibited two adjacent displacement areas in the east and west.

Figure 2

Surface deformation after the earthquake. (a) Ascending InSAR interferogram. (b) Ascending line-of-sight (LOS) deformation. (c) LOS A–A’ profile. (d) LOS B–B’ profile.

Two profiles located perpendicular to the fault strike (A–A’ and B–B’, black dashed lines in Figure 2b) were generated in the eastern and western sections of the deformation area to more intuitively analyze the surface deformation caused by the Baicheng earthquake. Both profiles pass through the surface deformation area on both sides of the fault. Profile A–A’ (Figure 2c) is located in the western section, in which the northern side of the fault was displaced by 10 cm in the LOS direction, whereas the southern side was displaced by 2 cm in the opposite direction. Profile B–B’ (Figure 2d) is located in the eastern section, in which the northern side of the fault was displaced by 9 cm in the LOS direction, whereas the southern side was displaced by 10 cm in the opposite direction. Both profiles changed rapidly at the fault trace, indicating that deformation was strongest near the fault, which is consistent with the results of the surface rupture field surveys.

3.2 Slip distribution

Based on the inversion results (Figure 3), the slip distribution of the Mw 5.3 Baicheng earthquake was consistent with the surface deformation characteristics. Two slip centers were observed at 8 and 2 km in the dip direction. The maximum slip occurred at the western slip center (0.45 m), and the eastern slip center underwent 0.4 m of slip. The inversion results of the slip distribution indicate that a certain amount of slip occurred at the surface. The rupture zone at the surface was 4 km long and exhibited characteristics of a left-lateral strike-slip event with a dip-slip component. The maximum vertical dislocation was 0.12 m and the horizontal dislocation was 0.1 m, according to the field surveys (Figure 4). The amount of slip and the geographic location of the inverted distribution that extended to the surface obtained herein were similar to the field survey results, indicating that the inversion results are reliable. The final inversion results indicate that the seismogenic fault had strike, dip, and rake angles of 248°, 70°, and 22°, respectively; a maximum slip of 0.45 m; and a moment magnitude (Mw) of 5.7 derived from the geodetic data. The deformation epicenter was located at 41.79° N and 81.13° E. The surface LOS deformation simulated by the model is shown in Figure 5, the results of which are in agreement with the InSAR deformation results. The data–model correlation was as high as 99.3%, and the average residual was 4 mm.

Figure 3

Slip distribution of the Baicheng Mw 5.3 earthquake inverted from InSAR LOS displacements.

Figure 4

Field survey photos of Baicheng Mw 5.3 earthquake deformation. (a) A cross-section of the surface rupture with a vertical dislocation of 12 cm and horizontal dislocation of 3 cm. (b) Concrete pavement displacement caused by the earthquake (horizontal displacement of 10 cm). (c) Concrete pavement displacement caused by the earthquake (vertical displacement of 32 cm). (d) A field photo of the epicenter.

Figure 5

Surface deformation results from the inversion of InSAR LOS displacements. (a) Observed surface deformation in the LOS direction. (b) Simulated surface deformation inverted from InSAR LOS displacements. (c) Residuals of the observed and simulated surface deformation.

4 Discussion

The results indicate that the 2021 Baicheng earthquake was a left-lateral strike-slip rupture event with a small dip-slip component. The moment magnitude obtained herein was higher than those obtained by the USGS and GCMT. A magnitude 4.9 foreshock occurred on March 14, and the acquisition time of the primary SAR image was one day before the foreshock. The acquisition time of the secondary SAR image was two days after the Baicheng mainshock; thus, the foreshock and post-seismic deformation were present in the interferograms used to invert the slip distribution. The higher magnitude obtained herein may also be due to the sensitivity of InSAR to shallow events. As the largest rupture zone obtained from the inverted slip model was located at ∼2 km in the dip direction, some enhancement of the slip values may have occurred in the inversion results.

The geometry of the seismogenic fault obtained from the InSAR data confirmed that it was a left-lateral strike-slip fault with a dip-slip component, which was corroborated by the surface rupture field survey results. Thus, this event occurred along a fault with a strike of 248° and a dip of 70° between the Tarim Basin and the southern Tianshan Mountains.

The most prominent feature of the Baicheng Mw 5.3 earthquake was the presence of two independent surface deformation centers. Several other events have exhibited this feature, including the 1999 Mw 7.6 İzmit et al. [45], 2004 Mw 6.0 Parkfield [46], and 2010 Mw 6.9 Yushu earthquakes [47]. The slip model inverted from the InSAR results also contained two rupture zones in the source area that were connected to each other. The western rupture center contained the maximum slip, but had a smaller distribution area. This result may be due to the fact that the interferometric map of the earthquake had two deformation centers. Nine days before the Mw 5.3 earthquake (March 14, 2021), an Mw 4.9 foreshock (41.88° N and 81.18° E) occurred, according to the USGS. The surface deformation data used for the inversion were obtained from two SAR images taken during March 13–25, which included deformation from both the March 14 foreshock and the March 23 mainshock. To determine whether the foreshock and mainshock caused two separate deformation areas at the surface and the influence of the foreshock on the deformation field, we selected another two SAR data pairs acquired on March 20 and April 13, 2021, from which we obtained the surface deformation that excludes the effect of the Mw 4.9 foreshock (Figure 6). The results indicate two deformation centers at the surface, which is similar to the deformation patterns caused by both the foreshock and mainshock. Therefore, we inferred that the Mw 5.3 Baicheng earthquake simultaneously formed two adjacent deformation fields at the surface.

Figure 6

Interferograms obtained from InSAR data. (a) March 13 and March 25 acquisition times. (b) March 20 and April 13 acquisition times.

Petroleum exploration in the study area has yielded many seismic sections and boreholes, as well as structural studies. To analyze the seismogenic structural characteristics of this earthquake, we collected the results [48] of seismic reflection profiles across the epicenter region (Figure 7). The epicenter is located west of the Kuqa fold belt, and the main structural feature on the northern side of the area is a southward-dipping monocline above a northward-dipping reverse fault in the Paleozoic basement. In contrast, a large number of salt structures are developed on the southern side of the region. Gypsum-salt rocks occur along the Qiulitag reverse fault zone on the surface of the north wing of the Miskantak anticline. Salt structures are also located between the Cretaceous and Neogene strata in this area. The deepest rupture of the Baicheng earthquake was located at the junction of the buried fault’s Triassic and Cretaceous strata in front of the South Tianshan Mountains, according to a shallow structural cross-section of the Kuqa fold belt. Paleo-modern Kumgarmu Group gypsum-salt strata with a maximum thickness of >3 km overlie the Cretaceous strata. The gypsum-salt strata have strong plasticity and flow easily, leading to gaps and small faults that can be filled by salt rocks, which require large breakthrough pressures for rupture to occur [49]. When an earthquake rupture occurs and the seismic wave passes through tight gypsum strata, the brittle plastic rock strata suddenly rupture, which likely produces a second rupture center near the source area and increases the intensity of the rupture. The slip distribution results obtained herein also indicate that the rupture depth on the western side was only 2 km, which is also the depth of the gypsum-salt rock strata. This process can explain why the Mw 5.3 earthquake produced such large-scale surface ruptures and deformation fields.

Figure 7

Shallow tectonic profile of the 2021 Mw 5.3 Baicheng earthquake.

The horizontal velocity field derived from GNSS data (Figure 8a) indicates that the 15 mm/year average velocity in the Tarim Basin was reduced to 7 mm/year in the earthquake epicenter area. The thrust fault zone between the Tarim Basin and South Tianshan Mountains absorbs the northward velocity. There are a number of differences in the GPS velocities between the epicenter and surrounding areas: the direction of the GPS-derived velocity east of the epicenter is more westward than that on the western side, indicating that shear strain accumulation occurs in this region. Shear strain rate calculations (Figure 8b) indicate that the maximum horizontal principal strain is north–south throughout the entire study area, suggesting that a north–south compressive stress state exists between the Tarim Basin and the South Tianshan Mountains.

Figure 8

(a) GNSS interseismic velocities around the epicenter. The black dashed line indicates the profile across the seismic fault. The focal mechanism shows the mainshock location. (b) Strain rates estimated from the GNSS interseismic velocities in (a). Maximum shear strain rates are shown in the background color, and principal strain rates are shown as vector pairs. (c) South–north and (d) west–east GNSS interseismic velocities within 150 km on either side of the profile.

To visualize the slip rate of the seismogenic fault, a profile perpendicular to the fault strike was plotted across the epicenter (black dashed line in Figure 8a). A total of 14 GPS monitoring points within 150 km on both sides of the profile were projected onto the profile, and the south–north (Figure 8c) and west–east (Figure 8d) GPS velocity components were plotted. Obvious shortening was observed in the GPS velocities near the epicenter, and the influence of the seismogenic fault on the regional GPS velocities and strain states was considerably larger than those of the other surrounding faults.

5 Conclusions

The deformation field of the Mw 5.3 Baicheng earthquake was calculated using Sentinel-1A satellite data. Combined with the field survey data, the geometric characteristics of the seismogenic fault and source slip distribution were obtained through inversion. The following conclusions were obtained:

1. The seismogenic structure of the Baicheng earthquake was a left-lateral strike-slip motion with a small dip-slip component. Based on the fault parameters and location, we inferred that the seismogenic fault is a blind fault with strike and dip angles of 248° and 70°.

2. The maximum slip was 0.45 m, which occurred 2 km beneath the surface in the dip direction. Another slip center was located to the east of the maximum slip center, which underwent 0.4 m of slip at a depth of ∼8 km in the dip direction. A moment magnitude of Mw 5.7 was obtained from geodetic and surface rupture data, and the source rupture extended to the surface. The combined co-seismic and post-seismic deformation from the foreshock and mainshock was the main reason for the higher magnitude obtained from our inversion. The surface slip obtained from the inversion was 0.1–0.3 m, which is consistent with the surface rupture field survey results.

3. The Mw 5.3 Baicheng earthquake was a shallow event, and the source rupture extended to the surface. Moreover, the surface deformation field and slip distribution obtained via inversion indicate that the earthquake produced two rupture centers in the epicenter area.

4. The analysis of inter-seismic GPS deformation characteristics indicates that slight shear strain accumulation occurs in the epicenter region, which may be responsible for the strike-slip features observed in our slip distribution model.