Late Quaternary activities of the Xinbaoan–Shacheng fault in the North China Plain

Abstract

The Xinbaoan–Shacheng Fault is located in the northeastern Huai-Zhuo Basin, where the eastern Jin–Ji–Meng Plateau tectonic zone intersects the Zhangjiakou–Bohai earthquake belt at the North China Plain. As a major active structure, it shows significant Late Pleistocene (Q_p3) activity, yet its slip rate and strong-earthquake recurrence behavior remain poorly constrained. In this study, satellite remote-sensing interpretation, paleoseismic trenching, detailed geological–geomorphic mapping, and Late Pleistocene chronological analysis were integrated to reconstruct its deformation features in the late Quaternary. The fault preserves a composite coseismic rupture scarp 32 km long, striking NW40°–65° and dipping southwest at ~ 70°, with a height of 1.5–3 m, and the statistical magnitude-displacement empirical relationship indicates that it corresponds to a historical earthquake of M 6¾-7. Trench evidence indicates normal-faulting kinematics, with the most recent surface-rupturing event occurring in the late Holocene, plausibly associated with the 1720 M7 Shacheng earthquake. Deformation–age relationships yield a vertical slip rate of 0.06 ± 0.01 mm/a (R² = 0.97) and a strong-earthquake recurrence interval of 3.39 ± 0.61 ka, characterized by alternating long and short cycles. These results provide robust quantitative constraints on the seismic behavior of the Xinbaoan–Shacheng Fault and offer important implications for regional tectonic evolution and seismic-hazard assessment in the Huai-Zhuo Basin and the whole North China Plain.

Introduction

When evaluating future earthquake hazards along active fault zones, the recurrence interval and slip rate are among the most critical quantitative parameters. In regions lacking historical seismic documentation and modern instrumental records, paleoseismic investigations—encompassing fault activity, slip rate, coseismic displacement, and earthquake recurrence—provide an especially effective means of assessing seismic behavior and forecasting future hazards^{1,2,3,4,5,6,7,8,9}.

The Huai-Zhuo Basin is located in the northeastern segment of the Shanxi rift system (Fig. 1a), at the intersection of the eastern Jin–Ji–Meng Plateau tectonic zone and the Zhangjiakou–Bohai strong-earthquake belt. This region is characterized by pronounced tectonic deformation and occupies a structurally important position in North China^10,11. The basin is dominated by NNW-directed extension and is regarded as both a key strong-earthquake monitoring area and a high seismic-hazard zone in China^12,13. Previous studies indicate that the Huai-Zhuo Basin lies within a zone of intense deformation and frequent seismic activity, jointly controlled by the Shanxi rift system and the Zhangjiakou–Bohai seismic belt^10,11,12,13. Paleoseismic investigations further suggest that Late Pleistocene faulting in this region exhibits temporal clustering and has generated large earthquakes, including the 1720 M7 Shacheng event^12,14,15,16.

Investigating the recurrence of strong earthquakes requires study areas where active faults have repeatedly produced large events and where geological records are sufficiently preserved. The Huai-Zhuo Basin satisfies these conditions: it is situated at the intersection of major tectonic domains in North China, has experienced multiple strong earthquakes, and preserves well-developed Late Pleistocene fault scarps and sedimentary sequences suitable for paleoseismic analysis. Although limited trenching at Liangtiantun and the Brickworks site has identified several paleoearthquake events along the Xinbaoan–Shacheng Fault^12,14,17, existing studies remain sparse in terms of spatial coverage and chronological precision, leaving the long-term seismic behavior of this fault insufficiently constrained.

Previous research in the Huai-Zhuo Basin has established the basic tectonic framework and recognized the Xinbaoan–Shacheng Fault as an active structure exhibiting late Holocene surface deformation (Fig. 1b). Paleoseismic trenching has provided initial constraints on paleoearthquake timing; however, these results are limited by sparse trench distribution, variable stratigraphic interpretations, and insufficient chronological control. Consequently, estimates of slip rate and earthquake recurrence interval remain highly uncertain, underscoring the need for integrated geomorphic analysis, improved trench documentation, and more robust chronological constraints.

Understanding the recurrence of strong earthquakes on active intraplate faults requires geological settings that preserve clear and datable records of past seismic events. The Huai-Zhuo Basin satisfies these conditions. Located at the junction of the Shanxi rift system and the Zhangjiakou–Bohai seismic belt, the basin has experienced multiple strong historical earthquakes, including the 1720 M7 Shacheng event. Reexamination of historical records confirms that this earthquake caused severe damage and exhibited a well-defined intensity distribution in the Shacheng–Huailai region, establishing it as a robust historical reference for regional seismic activity. When integrated with geological and paleoseismic evidence, this event provides an important benchmark for evaluating long-term earthquake recurrence¹⁸. In addition, the Xinbaoan–Shacheng Fault preserves well-developed Late Pleistocene fault scarps and thick sedimentary sequences, offering favorable conditions for high-resolution paleoseismic trenching and chronological reconstruction. Beyond their implications for seismic hazard assessment, the resulting constraints on slip rate and paleoearthquake chronology also contribute to understanding regional crustal deformation, tectonic evolution, Pleistocene geomorphology, and the behavior of intraplate active faults. Collectively, these factors make the Huai-Zhuo Basin an effective natural laboratory for investigating strong-earthquake recurrence.

Fig. 1

Main active tectonic features of the Huai-Zhuo Basin and adjacent areas.

Geological Structural Diagram of the Zhangjiakou-Bohai Tectonic Zone (Modified from Han¹⁹, F1: Shangyi-Pingquan Fault; F2: North Boundary Fault of Huai’an Basin; F3: Zhangjiakou Fault; F4: North Boundary Fault of Tianzhen-Yanggao Basin; F5: Zuosuo Bao-Songzhikou Fault; F6: South Boundary Fault of Huai’an Basin; F7: North Boundary Fault of Yangyuan Basin; F8: South Boundary Fault of Xuanhua Basin; F9: North Boundary Fault of Huai-Zhuo Basin; F10: North Boundary Fault of Yanfan Basin; F11: South Boundary Fault of Yangyuan Basin; F12: South Boundary Fault of Yuguang Basin; F13: Xinbaoan–Shacheng Fault; F14: Sunzhuangzi-Wulonggou Fault; F15: Zijinguang Fault; F16: Shizhuang Fault; F17: Nankou Mountain Front Fault; F18: Huangzhuang-Gaoliying Fault; F19: Sanggan River Fault.

All maps are presented in the WGS 84 coordinate system using the UTM Zone 50 N projection. Elevation data are referenced to the WGS 84 ellipsoid. Base satellite imagery from Google Earth Pro (version 7.3; https://www.google.com/earth/versions/). DEM data were obtained from the 91 Weitu platform (http://www.91weitu.com/). Map preparation was performed using CorelDRAW Graphics Suite 2018 (https://www.coreldraw.com/).

This study characterizes the Late Pleistocene activity of the Xinbaoan–Shacheng Fault using an integrated approach that combines remote-sensing interpretation, paleoseismic trenching, detailed geomorphic and geological mapping, and optically stimulated luminescence (OSL) dating²⁰.

Beyond seismic hazard assessment, this study has broader implications for understanding the tectonic behavior of faults in the North China Plain. NE-trending faults in this region have traditionally been interpreted as predominantly normal faults associated with extensional deformation. However, increasing geological and geophysical evidence suggests that many of these faults exhibit more complex kinematics, potentially involving a strike-slip component. The study area is located at the terminal segment of a major NE-trending fault, where the kinematic transition remains poorly constrained. Determining whether deformation at the fault tip is dominated by normal faulting, strike-slip motion, or a combination of both is critical for refining models of fault segmentation, stress transfer, and regional deformation patterns. The trench observations and structural characteristics documented in this study therefore provide new constraints on the kinematics of NE-trending faults and contribute to a more comprehensive understanding of intraplate tectonics in the North China Plain.

Regional geological tectonic background

Geologically, the Huai-Zhuo Basin is located at the intersection between the northeastern end of the Shanxi fault–depression zone and the northwestern end of the Zhangjiakou–Bohai strong-earthquake zone²¹, and lies within the eastern segment of the Yanshan intracontinental orogenic belt¹². It is an intermontane basin situated between Huailai and Zhuolu counties in northwestern Hebei Province, to the northwest of Beijing. Several NE– to NNE-trending active normal faults have been identified within or along the margins of the basin. The Xinbaoan–Shacheng Fault defines the northern and northwestern boundary of the basin, accompanied by other major structures such as the northern margin fault of the Huailai–Zhuolu (Huai-Zhuo) Basin and the northern boundary fault of the Yanfan Basin^10,11,12. Together, these faults accommodate regional crustal extension and exert strong control over the distribution of seismicity.

The basin margins are predominantly composed of alluvial and floodplain deposits, and dextral shear stress at depth during the Late Cenozoic contributed to the development of normal faults in the shallow crust^21,22,23. Consequently, the Huai-Zhuo Basin exhibits a unique tectonic environment characterized by active deformation, high seismicity, and long-term crustal strain accumulation.

The Xinbaoan–Shacheng Fault is situated near the junction of Shanxi, Hebei, and Inner Mongolia and serves as a principal boundary fault of the Huai-Zhuo Basin. The fault initiates at Liangjiazhang on the western side of Xinbaoan in Huailai County, Hebei, and extends southeastward through Xinbaoan, Liangtiantun, Huailai County, and Tumu Railway Station, terminating at Baying Village northwest of the Guanting Reservoir, where it intersects the northern boundary fault of the Yanfan Basin (Fig. 2). The total length of the structure is approximately 32 km. It strikes NW40°–65° and dips southwest at ~ 70°. Previous studies indicate that the fault has been active since the Late Pleistocene (Q_p3), although most of its surface trace is draped by unconsolidated sediments²².

Geomorphologically, the northwestern segment of the Xinbaoan–Shacheng Fault develops along the front edge of a Late Pleistocene alluvial-fan system at the foothills of the Donghuangzhuang and Zao’erkou Mountains. North of Dongyuan Village, the fault forms a prominent surface scarp 2–3 km in length that may displace Late Pleistocene–Holocene strata¹². Fault activity is inferred to have produced a vertical offset of approximately 4.3 m on the original alluvial-fan surface¹², and a scarp approximately 2 km long and 4.5 m high is still preserved today²². Previous investigations also report a marked sediment-thickness contrast across the fault: Pleistocene deposits on the hanging wall are 360–400 m thick, reaching up to 500 m locally, whereas those on the footwall are only 80–130 m thick, resulting in a sedimentary difference of ~ 280 m across the structure¹².

Research methods and data processing

In this study, Google Earth imagery, fault-displaced geomorphic markers, and field surveys were integrated to obtain precise elevation information and to characterize the geometry of the linear scarp and its surrounding terrain, thereby delineating the spatial extent of deformation along the western side of the fault (Fig. 2, key work area). High-resolution remote-sensing imagery formed the basis for the initial interpretation. Fine digital elevation models (DEMs) were generated from Google Earth Pro stereo satellite imagery (MAXAR data), which has a nominal horizontal resolution of approximately 5–10 m. Hillshade enhancement and terrain-slope filtering were performed in ArcGIS 10.7 to highlight subtle geomorphic discontinuities and scarp-related breaks.

However, extensive anthropogenic surface modification, such as land leveling, agricultural reshaping, and minor construction, has obscured the natural morphology of the scarp in several segments. For this reason, the DEM was used primarily for qualitative geomorphic interpretation, whereas quantitative measurements of vertical separation were obtained directly from RTK-GPS surveys. The DEM supported the extraction of slope shading, contour refinement, and visualization of potential deformation zones associated with fault activity.

To further validate the geometric characteristics and spatial continuity of the fault, field investigations were carried out along its trace, focusing on areas northwest of Xinbaoan Railway Station, Liangtiantun, and the vicinity of Dongbali Village (Fig. 2). Paleoseismic trenching was conducted to confirm the scarp morphology and subsurface deformation. Trenching sites were selected at locations where fault traces are clear and Late Pleistocene deposits are well preserved. The principal trench (TC01) is located north of Xinbaoan Railway Station, oriented perpendicular to the fault strike. After excavation, the trench walls were cleaned and mapped using a 1 × 1 m grid. Detailed documentation of stratigraphy, fault–stratum relationships, filling wedge, and pre-scarp deposits was carried out, and overlapping photographs (≥ 85% overlap) were collected to create a continuous photomosaic for stratigraphic interpretation.

The selection of trench sites was based on multiple factors, including the presence of well-developed scarps, visibly offset geomorphic surfaces, and relatively continuous Late Pleistocene sediments suitable for exposing deformation features and collecting OSL samples. OSL samples were collected under strict light-safe conditions to prevent secondary light exposure. At each sampling location, the trench wall was freshly cleaned, and the sampling position was immediately shielded using opaque black plastic sheets. Samples were extracted horizontally using stainless-steel tubes, which were sealed at both ends upon removal and subsequently wrapped in aluminum foil and black plastic bags. These procedures ensured that the sampled sediments were not exposed to sunlight during collection, handling, or transport.

Laboratory preparation was conducted under subdued red-light conditions. The outer portions of each sample, which may have been affected by incidental light exposure, were removed prior to analysis. Quartz grains were isolated following standard preparation procedures, including wet sieving to obtain the target grain-size fraction, treatment with HCl and H₂O₂ to remove carbonates and organic matter, and HF etching to eliminate feldspar contamination and remove the alpha-irradiated outer layer. The purity of the quartz extracts was checked prior to measurement to ensure suitability for OSL analysis.

Equivalent doses (De) were determined using the single-aliquot regenerative-dose (SAR) protocol. Appropriate preheat and cut-heat temperatures were applied following established laboratory practice. Sensitivity changes during measurement were monitored and corrected using test doses within the SAR sequence. Recycling ratios and recuperation tests were routinely evaluated to assess the reliability of the De determinations.

Environmental dose rates were calculated from measured radionuclide concentrations combined with estimates of long-term water content. The total age uncertainty incorporates contributions from De determination, dose-rate estimation, and water-content variability. Although some samples exhibit scatter in De values, all ages are reported with their associated uncertainties, and the resulting OSL ages are stratigraphically consistent within the trench sequence, providing a robust chronological framework for paleoseismic interpretation.

High-resolution satellite imagery was first used to identify fault-related landforms, followed by trench excavation across key fault scarps. Stratigraphic relationships together with OSL ages were then used to identify paleoearthquake events and constrain their timing. The results were compared with previously documented events from the Liangtiantun trench¹², allowing a refined assessment of recent fault activity. Based on cumulative displacement measurements and geochronological constraints, we further quantified the recurrence interval of strong earthquakes and the Late Pleistocene slip rate.

Accessibility and safety conditions were also considered. These locations coincided with fault segments previously inferred as active, allowing direct comparison with earlier paleoseismic results, such as those from the Liangtiantun trench¹².

Paleoseismic event recognition within the trench relied on several diagnostic criteria, including variations in stratigraphic color, compactness, gravel content, and sedimentary structure. Boundaries between stratigraphic units were identified based on changes in composition and color, whereas paleoearthquake events were inferred from features such as fissure orientations, filling wedge, and gravel-layer geometries^24,25,26,27.

Fig. 2

Seismotectonic map of the Huai-Zhuo Basin.

F1: Xinbaoan–Shacheng Fault; F2: Southern Margin Fault of the Xuanhua Basin; F3: Northern Margin Fault of the Huai-Zhuo Basin; F4: Sanggan River Fault; F5: Northern Margin Fault of the Yanfan Basin; F6: Southern Margin Fault of the Yanfan Basin; F7: Sunzhuangzi–Wulonggou Fault; F8: Shizhuang Fault.

All maps are presented in the WGS 84 coordinate system using the UTM Zone 50 N projection. Elevation data are referenced to the WGS 84 ellipsoid. Base satellite imagery from Google Earth Pro (version 7.3; https://www.google.com/earth/versions/). DEM data were obtained from the 91 Weitu platform (http://www.91weitu.com/). Map preparation was performed using CorelDRAW Graphics Suite 2018 (https://www.coreldraw.com/).

Results

Tectonic and geomorphic features along the Xinbaoan–Shacheng Fault

Field survey points are mainly concentrated in the northwestern (Fig. 3a-c) and southeastern (Fig. 3d) segments of the Xinbaoan–Shacheng Fault. The study area is characterized by hilly-basin transitional landforms, with elevations ranging from approximately 300 to 600 m, and is predominantly underlain by alluvial and floodplain deposits. Field investigations indicate that the fault scarp near the Xinbaoan Railway Station is particularly well developed (Fig. 3b). The scarp trends NE and is located along the geomorphic transition between the foothills and the basin floor.

Several fault-related geomorphic features were identified near the Xinbaoan Railway Station, Dongbali Village, and Liangtiantun, from which four representative landforms were selected for detailed analysis (Figs. 2 and 3). Although the fault scarp shown in Fig. 3a appears subtle in plain view, its presence is clearly expressed by pronounced local relief. In particular, the undulating scarp profile is evident when referenced to adjacent man-made features, such as noticeable elevation differences along the side wall, which emphasize the vertical component of fault-related deformation. These geomorphic expressions provide clear evidence of surface deformation associated with fault activity.

Geomorphic scarp heights were quantified using RTK-GPS topographic profiles collected along mapped transects, which represent cumulative surface relief at the geomorphic scale. RTK-GPS measurements indicate that the alluvial-fan surface in Fig. 3b is vertically offset by approximately 0.6 m. Similarly, the terrace surface shown in Fig. 3d exhibits a vertical displacement of about 1.2 m, expressed as a distinct step between the hanging wall and footwall. Additional geomorphic indicators—including the steeper scarp face, accumulation of coarse gravels at the scarp toe, and alignment of the break-in-slope—further corroborate these vertical offsets.

Fig. 3

Geomorphic features along the Xinbaoan–Shacheng fault. a Dongbali area; b Xinbaoan Railway Station; c southern side of the trench at Xinbaoan; d vicinity of Tumuzhen Town.

All maps are presented in the WGS 84 coordinate system using the UTM Zone 50 N projection. Elevation data are referenced to the WGS 84 ellipsoid. Photographs were taken by the authors.

Paleoseismic evidence and fault deformation revealed by trenching and geomorphic analysis

Along the fault trace, scarps with varying heights are developed, with maximum vertical displacements of approximately 1.2 m (Fig. 3d) and minimum offsets of around 0.3 m (Fig. 3c). Such along-strike variations in scarp height suggest that the fault has likely experienced multiple strong surface-rupturing earthquakes. Based on satellite image interpretation and detailed surface geological mapping, the TC01 trench was excavated across a well-preserved scarp near the Xinbaoan Railway Station (Fig. 4). In addition, two previously reported trenches—the TC02 trench at the brickworks and the TC03 trench at Liangtiantun, located approximately 5 km from TC01¹²—These trenches were excavated in earlier investigations and document fault-related deformation that confirms the position and Late Quaternary activity of the fault along strike. Although detailed stratigraphic logs and new chronological data from TC02 and TC03 are not presented here, their locations and observations are used to support the continuity of the active fault trace and to constrain the spatial extent of faulting. In this study, paleoseismic interpretation and quantitative analysis are therefore focused on the TC01 trench, while TC02 and TC03 serve as supplementary evidence for fault activity at the regional scale.

Based on detailed geomorphic interpretation, the trench TC01 is located at the margin of an alluvial fan developed along the Xinbaoan–Shacheng Fault. The recognition of the alluvial fan geometry is critical for interpreting fault-related landforms and stratigraphic relationships at the trench site. Alluvial fan surfaces commonly act as geomorphic markers that are sensitive to tectonic deformation, and offsets preserved along fan margins or fan surfaces can provide valuable constraints on cumulative fault displacement²⁸. In addition, fan-margin settings are favorable locations for paleoseismic trenching because they tend to preserve thick, rapidly deposited sediments capable of recording surface-rupturing events, while also being susceptible to erosion and reworking following fault motion.

In the TC01 area, the position of the trench at the fan margin explains the observed stratigraphic architecture, including scarp-front accumulation and localized erosion of uplifted units. Mapping the alluvial fan and associated gullies therefore provides an essential geomorphic framework for distinguishing tectonically induced deformation from depositional or erosional processes, and strengthens the interpretation of the trench stratigraphy and paleoearthquake record along the Xinbaoan–Shacheng Fault.

Fig. 4

Geomorphological unit map of the Xinbaoan area.

Base satellite imagery from Google Earth Pro (version 7.3; https://www.google.com/earth/versions/). DEM data were obtained from the 91 Weitu platform (http://www.91weitu.com/). Map preparation was performed using CorelDRAW Graphics Suite 2018 (https://www.coreldraw.com/).The TC01 trench exposes a normal fault (F1) striking 213° and dipping 60°, indicating clear extensional deformation. The upper strata on the hanging wall show a distinct downward displacement, whereas the tops of the lower units exhibit evidence of erosion. Seven stratigraphic units (U1-U7) were identified in the TC01 trench and are numbered from top (youngest) to bottom (oldest). In the exposed sequence, U1 corresponds to the surface humus layer. Fault F1 offsets the light-yellow loess of U3 through the brown loess of U6. Although the basal contact of U6 is not clearly defined, this unit is unequivocally displaced by F1 and displays a measurable fault separation.

Stratigraphic thickness variations across the fault zone provide important evidence for fault-related deformation. Unit U5 exhibits progressive thinning toward the hanging wall, which is interpreted to reflect post-depositional erosion associated with uplift during fault activity. This interpretation is supported by the presence of gravel layers at both the top and base of U5, indicating erosion and reworking of the upper part of the unit. In contrast, Unit U3 shows thinning toward the footwall, which is interpreted as the result of compaction related to scarp-front accumulation. Together, these asymmetric thickness patterns across the fault zone indicate deformation consistent with fault movement and support the paleoseismic interpretation.

To ensure internal consistency, epoch terminology has been standardized throughout the manuscript. The older event is assigned to the Late Pleistocene (Q_p3), whereas the younger event occurred during the Holocene. The most recent earthquake is inferred to have occurred between the deposition of Units U2 and U1, as deformation is clearly observed in U2, while no structural disturbance is present in the overlying U1.

The stratigraphic framework exposed in the TC01 trench (Fig. 5) provides essential constraints for identifying deformation features and distinguishing successive paleoearthquake events. To further constrain the timing of these events and to quantify the fault slip history, independent chronological control is required. For this purpose, optically stimulated luminescence (OSL) samples were collected from key stratigraphic units and deformation-related deposits to determine depositional ages and to bracket the occurrence times of the identified paleoearthquakes.

Fig. 5

Profile of Trench TC01 north of Xinbaoan Railway Station.

To constrain the timing of paleoearthquake events, estimate slip rates, and determine the recurrence interval of strong earthquakes, optically stimulated luminescence (OSL) samples were collected from key stratigraphic units exposed in the trench. In the field, OSL samples were obtained using stainless-steel tubes (approximately 5 cm in diameter and 20 cm in length) inserted horizontally into loess layers to minimize moisture loss and prevent light exposure. A total of seven OSL samples were collected and analyzed at the New Structural and Chronology Laboratory of the Disaster Prevention Technology Institute, yielding depositional age estimates summarized in Table 1.

Although minor age inconsistencies are observed among some closely spaced samples, these differences are interpreted to reflect analytical uncertainties inherent to OSL dating, rather than true stratigraphic age reversals^29,30. All measured ages overlap within their respective uncertainty ranges, and the observed offsets do not affect the relative stratigraphic order of the units or the overall chronological framework used for paleoearthquake identification. Consequently, these uncertainties do not compromise the interpretation of fault activity or the reconstruction of seismic history within the resolution of the available data.

By comparing the ages of successive stratigraphic units and applying the sequential limiting method³¹, we constrained the timing of the most recent fault activity, delineated the deformation windows of individual paleoearthquake events, and estimated both the recurrence interval of strong earthquakes and the Late Pleistocene slip rate of the Xinbaoan–Shacheng Fault.

Table 1 OSL Dating Results of the TC01 Sample from the Xinbaoan–Shacheng Fault Zone, North of Xinbaoan Railway Station.

Table 2 Stratigraphic Description of Trench TC01 North of Xinbaoan Railway Station, Xinbaoan–Shacheng Fault Zone.

The TC01 trench is located north of the Xinbaoan Railway Station (40.441°N, 115.444°E; Fig. 5) within a relatively stable depositional environment. The sediments consist mainly of loess and fine sand, interbedded with minor gravel layers. The trench is oriented nearly perpendicular to the strike of the surface scarp. It has an effective length of ~ 10 m, a depth of ~ 1 m, and a maximum excavation height of ~ 3 m. The exposed profile reveals stratified loess, clay, and gravel deposits with distinct color and textural differences. Based on variations in color, mineral composition, and sedimentary structures, six stratigraphic units and one filling wedge were identified, along with pre-scarp deposits. The corresponding trench sketches are shown in Figs. 5 and 6, and the detailed stratigraphic descriptions are summarized in Table 2.

Fig. 6

Enlarged details of Trench TC01 profile north of Xinbaoan Railway Station.

The stratigraphic division of the TC01 trench is based on systematic variations in sediment color, grain size, compactness, and gravel content. Unit boundaries were identified through distinct shifts in these characteristics, such as transitions from massive yellow loess to laminated fine sand or gravel-rich layers. These criteria provide a reliable basis for distinguishing primary depositional units from colluvial deposits or deformation-related strata. A slight age inversion was observed between the upper and lower OSL samples from Unit U4. Minor discrepancies of this type are common in OSL dating of loess and fine-sand deposits and may arise from incomplete bleaching prior to burial, limited post-depositional mixing, or localized soil disturbance. The observed difference falls within analytical uncertainty and therefore does not affect the established stratigraphic succession or alter the interpretation of paleoearthquake events^32,33,34.

Integrating evidence from stratigraphic deformation, filling wedge, pre-scarp deposits, and fault-related fractures, the TC01 trench records one fault and two paleoearthquake events, indicating extensional deformation and long-term fault activity. The timing of these events is primarily constrained by the ages of the youngest deformed units and the oldest overlying undisturbed strata. Following a unified naming convention, the paleoearthquake events are referred to as the older event and the younger event, listed in chronological order from oldest to most recent.

The younger event represents the most recent paleoearthquake and occurred after the deposition of Unit U3, as fault F1 clearly displaces U3 and the lower portion of Unit U2 in the trench exposure. However, no deformation is observed in the upper deposits of U2 or in Unit U1. Therefore, the younger event is constrained to have occurred after the accumulation of U3 but before the deposition of the upper part of U2 and Unit U1. The vertical displacement associated with this event is approximately 30 cm.

The older event is more clearly expressed in the stratigraphic record. It occurred after the deposition of Unit U5 and before Unit U4, as fault F1 produces a vertical displacement of approximately 180 cm in Unit U5. The presence of a filling wedge adjacent to fault F1 provides additional evidence for this event and confirms its association with significant surface-rupturing deformation.

Analysis of fault activity characteristics in TC01 trench

The Xinbaoan trench exposes a fault zone characterized by a relatively steep deformation belt, with fault dips of approximately 60°–65°, indicating activity since the Late Pleistocene. Materials within the deformation zone are generally soft and exhibit a pronounced preferred orientation (Fig. 6a). Clay-rich materials within the fault zone are interpreted to have formed during coseismic rupture, when pre-existing clay layers were entrained, smeared, or folded into extensional spaces created by fault movement, reflecting intense deformation during fault slip.

Stratigraphic relationships show that fault F1 offsets Units U6 and U5 and propagates upward into Units U4, U3, and U2, but does not rupture through the ground surface (Fig. 6b). This geometry indicates that the most recent fault activity occurred after the deposition of U3, while the overlying part of U2 and Unit U1 remained undeformed. These relationships provide key constraints on the timing of the latest paleoearthquake event.

Within the deformation zone, fault-related coarse-grained deposits containing gravels of variable size are developed adjacent to the fault (Figs. 6c–d). These deposits display poor sorting and massive structures, consistent with rapid deposition associated with strong fault movement. Their occurrence indicates that significant deformation events took place after the deposition of Unit U5.

Unit U7 occurs immediately adjacent to the fault zone and is characterized by a wedge-shaped geometry, massive internal structure, and absence of primary stratification. These features indicate rapid infilling into an open space generated during fault rupture. U7 is therefore interpreted as a filling wedge, formed by post-event sediment infill, rather than a colluvial wedge produced by gravitational scarp collapse³⁵. In contrast, Unit U2 consists of relatively fine-grained sediments with gentle bedding and laterally continuous geometry. It is interpreted as scarp-front accumulation, resulting from gradual sediment deposition at the base of the fault scarp following surface deformation.

Reliability test of the sedimentary-paleoseismic sequence recognition in TC01 trench

To test the reliability of the recognition of the two paleoearthquake events described above, this study attempted to reconstruct the sedimentary deformation evolutionary process of the strata in the TC01 trench, which can be divided into six stages^36,37.

• a. The lowest layers—light-colored loess (U6) and dark brown loess (U5)—were deposited nearly horizontally (Fig. 7a).

• b. The older paleoearthquake event occurred under normal-faulting conditions along fault F1, which displaced the U5 and U6 strata. As a result, the U5 layer adjacent to the fault underwent scarp formation and gravitational collapse, leading to the development of a filling wedge. This colluvial deposit subsequently continued to accumulate under the influence of aeolian processes and other surface agents, and was later modified by erosion and surface leveling (Fig. 7b).

• c. The brown loess (U4) and gravel (U3) layers were deposited on top of U5 and U6 by wind, slope washing, and other factors (Fig. 7c).

• d. The younger paleoearthquake event occurred under normal-faulting conditions along fault F1, resulting in displacement of the U3–U6 strata (Fig. 7d).

• e. Surface loess (U1) was deposited, covering U2 (Fig. 7e).

Fig. 7

Reconstruction of paleoearthquake history of Trench TC01.

Strong earthquake recurrence interval and slip rate

The paleoearthquake record revealed by the TC01 trench on the Xinbaoan–Shacheng Fault indicates that at least two paleoearthquake events have occurred since the Late Pleistocene. The older event is constrained to a time window of approximately 18.73 ± 0.62 ka to 20.69 ± 1.05 ka. Interpretation of this older event is supported by stratigraphic evidence indicating post-event uplift and erosion of the hanging wall of Unit U5. Specifically, the presence of a gravel-rich layer at the top of U5, together with an abrupt change in sedimentary facies, suggests erosion of the upper part of the unit prior to subsequent deposition. Such erosional features are consistent with surface uplift associated with fault movement and provide a reasonable explanation for the observed stratigraphic truncation, thereby supporting the inference of an early faulting event recorded in the trench sequence.

This study focuses on estimating the Late Pleistocene slip rate of the Xinbaoan–Shacheng Fault based on stratigraphically constrained displacements observed in both the hanging wall and footwall. Field observations show that Unit U3 is dominated by gravel and coarse-grained sediments, which limits the reliability of direct laboratory dating. Nevertheless, stratigraphic relationships indicate that the younger paleoearthquake event occurred after the deposition of Unit U3, whereas Unit U2 represents pre-scarp accumulation. These relationships provide key constraints for bracketing the timing of the younger event and for reconstructing the fault slip history.

Unit U3 is composed predominantly of coarse gravel deposits, and no suitable fine-grained quartz material was available for direct OSL dating. To provide a first-order temporal reference for stratigraphic discussion, we therefore estimated the depositional age of U3 using a sedimentation-rate–based simulation, while explicitly acknowledging its inherent uncertainty. Two OSL-dated stratigraphic horizons that bracket U3 were used as boundary conditions, together with measured stratigraphic thicknesses obtained from the trench log.

Assuming a simplified scenario of constant vertical accumulation between the two dated horizons, the sedimentation rate (r) was calculated as the ratio between stratigraphic thickness (Δz) and the age difference (Δt) of the bounding units. To propagate dating uncertainty, the OSL ages of the two bounding horizons were treated as normal distributions defined by their reported mean values and 1σ errors. A Monte Carlo approach was applied to generate an ensemble of sedimentation rates, and for each realization the depositional age of U3 was estimated based on its stratigraphic position relative to the dated horizons. The reported age of U3 and its uncertainty correspond to the mean value and standard deviation of the resulting simulated age distribution.

It is emphasized that this simulation assumes continuous deposition and does not account for potential depositional hiatuses or erosional surfaces. Given the observed stratigraphic discontinuity between Units U2 and U3, the constant-rate assumption may be locally violated. Accordingly, the simulated age of U3 is not treated as a precise depositional age or as a primary constraint on paleoearthquake timing. Paleoearthquake events in this study are constrained mainly by stratigraphic cross-cutting relationships, with the simulated U3 age serving only to illustrate the plausible temporal range permitted by the available chronological data.

Assuming that sedimentation rates in the region did not vary substantially on millennial timescales, the vertical distance between sampling horizons U2 and U4 was measured based on stratigraphic correlation^34,38,39. Using the depositional ages of these two dated units, an average sedimentation rate of approximately 0.09 mm/a was obtained. Application of this rate yields an estimated depositional age of 6.41 ± 0.59 ka for Unit U3, which provides an approximate temporal bracket for the younger paleoearthquake event. Consequently, the younger event is constrained between 6.41 ± 0.59 ka and 2.34 ± 0.08 ka. It should be noted that the inferred age of U3 represents a first-order approximation derived from an average sedimentation rate between two dated horizons. Although chronological control is limited, this estimate is considered reasonable at the millennial scale because Units U2–U4 display similar lithofacies and show no evidence of a major erosional hiatus^40,41 Owing to its coarse-grained texture and lack of suitable quartz, direct OSL dating of U3 was not feasible; therefore, the estimated age of 6.41 ± 0.59 ka is used only as supportive chronological information consistent with trench stratigraphy, rather than as a precise numerical constraint on the younger event.

Based on displacement–age data from the TC01 trench together with previously published results from the Liangtiantun and Brickworks trenches¹², the Late Pleistocene slip rate of the Xinbaoan–Shacheng Fault is estimated at 0.06 ± 0.01 mm/a. This value was obtained by linear regression of vertical displacement versus age data (Fig. 8; n = 11), with the reported uncertainty corresponding to the 97% confidence interval of the regression slope and accounting for both displacement measurement errors and OSL age uncertainties³⁸.

Fig. 8

Slip rate fitting diagram for the Late Pleistocene of the Xinbaoan–Shacheng fault.

Field measurements indicate vertical displacements of 0.5 ± 0.1 m for Unit U3, 0.8 ± 0.3 m for Unit U4, 0.9 ± 0.1 m for Unit U5, and 1.1 ± 0.2 m for Unit U6. Vertical displacement at the TC01 trench was quantified based on detailed trench logging and stratigraphic correlation. Measurements were referenced to the basal boundary of Unit U5, which is laterally continuous across the fault zone and provides a stable stratigraphic marker, whereas the upper part of U5 on the hanging wall was affected by post-event erosion and does not preserve a reliable datum. Vertical separation was measured as the perpendicular distance between correlated basal boundaries on the hanging wall and footwall, using tape measurements and leveled reference lines on the trench wall. Multiple independent measurements were performed to assess repeatability, yielding a cumulative vertical displacement of 0.9 ± 0.1 m for Unit U5 at TC01. This displacement represents cumulative offset recorded by the stratigraphic unit, rather than slip from a single paleoearthquake event.

Combining the long-term slip rate (0.06 ± 0.01 mm/a) with the mean recurrence interval (3.39 ± 0.61 ka) implies an average vertical displacement of approximately 0.20 m per event. Assuming a fault dip of ~ 65°, this corresponds to a maximum net coseismic displacement of ~ 0.36 m. To evaluate the consistency between slip rate, recurrence interval, and earthquake magnitude, we applied commonly used empirical magnitude–displacement scaling relationships of the form.

$${\text{M }} = {\text{ a }} + {\text{ b}}\cdot{\text{log D}}$$

(1)

where D represents the maximum coseismic fault displacement. Substituting the estimated event-scale displacement yields an inferred earthquake magnitude of approximately 6¾, which is consistent with the revised magnitude interpretation adopted in this study⁴².

Coseismic displacement is expected to be heterogeneous along strike and between events. For example, the approximately one-third of a meter of displacement locally observed at the TC01 trench likely represents a near-maximum displacement for a single event, whereas larger offsets recorded by Units such as U5 reflect cumulative displacement from multiple events. Accounting for this distinction reconciles event-scale displacement with the long-term slip rate and recurrence interval, supporting characteristic earthquake magnitudes in the M 6–7 range, consistent with empirical displacement–magnitude relationships for earthquakes of about M 6½–7. The apparent discrepancy between long-term average slip and individual event displacement likely reflects temporal clustering and variable strain accumulation, a behavior commonly observed in intraplate normal faults^43,44,45.

Discussion and conclusion

Comparison with previous research results

Because the Xinbaoan–Shacheng fault zone is largely concealed, quantitative kinematic analysis of its Late Pleistocene activity has long been a challenge in studies of active structures in this region. By integrating results from the Xinbaoan trench (TC01) with previous investigations by Ran¹² at the Liangtiantun and Brickworks trench profiles (Figs. 9 and 10), this study confirms that the Xinbaoan–Shacheng fault is an active structure striking NE and dominated by normal-faulting kinematics.

Compared with the paleoearthquake sequence proposed by Ran¹², the differences identified in this study can be attributed to several factors. First, the TC01 trench is located closer to the principal fault scarp, where deformation features such as fault planes, stratigraphic offsets, and colluvial deposits are better preserved. Earlier trench sites were situated farther from the main displacement zone, which may have limited their ability to record smaller or younger events. Second, the stratigraphic framework established here benefits from improved correlation among depositional units, allowing clearer discrimination between primary sediments and reworked or colluvial deposits. Third, chronological constraints are strengthened by a higher density of OSL samples and more clearly defined age brackets, thereby reducing uncertainty in event timing. Together, these improvements yield a refined paleoearthquake history and underscore the enhanced resolution and novelty of the present study relative to previous work. Similar conclusions regarding fault activity and kinematics have also been reached by other studies using independent approaches⁴⁶. A comparison between the present results and earlier studies is provided in Table 3.

Table 3 Comparison of Paleoearthquakes in the Xinbaoan–Shacheng Fault Trench with Previous Research Results.

With respect to the classification of paleoearthquake sequences, Ran¹² identified one paleoearthquake event in the Liangtiantun profile, with the most recent event dated to 7.619 ± 0.095 ka (Fig. 9). At the Brickworks trench, three paleoearthquake events were recognized, occurring at 8.159 ± 0.54 ka, 16.025 ± 0.615 ka, and 23.205 ± 0.905 ka (Fig. 10). The TC01 trench investigated in this study is located approximately 1.5 km from the Brickworks trench and about 4.5 km from the Liangtiantun profile¹². Two paleoearthquake events were identified at TC01, with the most recent event constrained to 4.38 ± 2.03 ka. Compared with the Liangtiantun results, the TC01 trench records more recent paleoearthquake activity (Table 3), further refining the Late Holocene seismic history of the Xinbaoan–Shacheng fault zone.

Previous studies by Ran¹² on the Xinbaoan–Shacheng Fault have investigated its Late Quaternary activity from multiple perspectives, including fault geometry and surface expression based on geomorphic interpretation and remote sensing, paleoseismic evidence from trench excavations, and chronological constraints mainly derived from OSL dating. Collectively, these studies demonstrate that the fault has remained active since the Late Pleistocene and has experienced multiple paleoearthquake events, allowing first-order estimates of slip rate and recurrence interval to be proposed. Historical earthquake records, particularly those related to the AD 1720 Shacheng earthquake, have also been widely used to explore possible correlations between strong historical earthquakes and fault activity.

Historical documents indicate that the Huailai–Yanqing Basin and its surrounding region have experienced multiple destructive earthquakes since historical times, reflecting the long-term activity of the regional fault system. The earliest well-documented event occurred in AD 294, when an earthquake of approximately M6 struck eastern Yanqing, representing the first clearly recorded strong earthquake in the Huailai–Yanqing Basin⁴⁷. Subsequent major events include the AD 1337 Huailai earthquake (M 6½), with an epicenter near 40.4°N, 115.7°E and a maximum intensity of VIII⁴⁸, and the AD 1484 Yanqing earthquake (M 6¾), which reached intensities of VIII–IX and produced widespread damage across the Beijing–Juyongguan–Xuanfu region⁴⁹.

The most destructive event in this sequence is the AD 1720 Shacheng earthquake (M 6¾), which reached a maximum intensity of IX and caused severe damage in the Shacheng–Huailai area¹⁸. This earthquake is regarded as the latest large-magnitude event affecting the basin. Since then, no earthquakes of magnitude M ≥ 6 have been recorded in the region. Instrumental observations from the Huailai digital seismic network (operational since 1995) further indicate that seismicity since 1970 has been dominated by small to moderate earthquakes (M ≥ 2.0), with activity mainly concentrated at depths of 4–15 km near intersections of NE- and NNW-trending faults, highlighting the long recurrence interval of large earthquakes on the Xinbaoan–Shacheng Fault.

Taken together, the historical earthquake record indicates that several strong earthquakes have affected the Huailai–Yanqing Basin since historical times. However, among these events, the AD 1720 Shacheng earthquake shows the closest spatial correspondence to the TC01 trench location, both in terms of its inferred epicentral area and the distribution of high macroseismic intensities. The TC01 trench is situated along a key segment of the Xinbaoan–Shacheng Fault that lies within the core damage zone of the 1720 event, as documented in historical sources. In contrast, earlier large earthquakes, although significant in magnitude, are either spatially offset from the trench site or less well aligned with the investigated fault segment, while later events were of substantially smaller magnitude. Therefore, despite the chronological uncertainty inherent in the near-surface stratigraphy, the spatial coincidence between the TC01 trench and the 1720 earthquake source area provides the primary basis for correlating the most recent paleoseismic event observed in the trench with the AD 1720 Shacheng earthquake.

Despite these advances, several key limitations remain. Chronological constraints on individual paleoearthquake events–especially the most recent event–are commonly restricted to broad maximum or minimum age bounds, making it difficult to uniquely correlate trench-recorded events with specific historical earthquakes.

Fig. 9

Paleoearthquake trench profile at Liangtiantun site (Ran¹²).

Fig. 10

Paleoearthquake trench profile at Brickworks site (Ran¹²).

Fig. 11

Comparison of paleoearthquake events since the Late Pleistocene in the Xinbaoan–Shacheng fault zone (Ran¹²).

E1–E6 represent paleoearthquake events identified along the Xinbaoan–Shacheng Fault, ordered from oldest (E6) to youngest (E1). Using the sequential limiting method, six paleoearthquake events were recognized along the fault (Fig. 11). Event E1 is most plausibly associated with the 1720 M7 Shacheng earthquake. Event E2, identified in the Xinbaoan trench, displaced multiple sedimentary units but is overlain by brown silt dated to 2.34 ± 0.08 ka, indicating that the event occurred prior to this time. Applying a uniform sedimentation-rate approach⁴⁵, the age of Event E2 is estimated at 4.35 ± 2.05 ka. Event E3, constrained by both the Liangtiantun and Brickworks trenches, occurred at 7.889 ± 0.27 ka. Event E4, exposed in the Brickworks trench, was dated to 16.025 ± 0.615 ka. Event E5, recorded in the Xinbaoan trench, displaced two stratigraphic units and produced a filling wedge, with an estimated age of 19.21 ± 0.48 ka. Event E6, also identified in the Brickworks trench, was constrained to 23.205 ± 0.905 ka.

Ran¹² reported an average paleoearthquake recurrence interval of 7.64 ± 0.46 ka for the Xinbaoan–Shacheng Fault. Building upon Ran’s interpretation of fault geometry, kinematics, and geomorphic expression, the present study integrates newly excavated trench data with previously published trench records to reconstruct a six-event paleoearthquake sequence¹⁴. Differences between the event sequence proposed here and that of Ran¹² can be attributed to three main factors. First, the TC01 trench is located closer to the principal fault scarp, where deformation structures and filling wedge are more clearly preserved, whereas earlier trenches were situated farther from the main displacement zone and may not have captured smaller or younger events. Second, improved stratigraphic resolution in this study enables more reliable discrimination between primary depositional units and reworked or colluvial materials. Third, the OSL chronology employed here benefits from denser sampling and better control of bleaching conditions, resulting in more robust age constraints^32,33,34. Together, these factors explain the discrepancies with previous interpretations and highlight the enhanced resolution and novelty of this study.

Linear regression analysis indicates that the recurrence intervals between events E1–E2–E3 and between E4–E5–E6 are both approximately 3.39 ka (Fig. 12), where the third and fourth paleoearthquakes correspond to sub-events “a” and “b” of Event E3. In contrast, the absence of paleoearthquake evidence between approximately 9 ka and 15 ka may reflect a local sedimentary hiatus, a phase of tectonic uplift limiting deposition, or a period of reduced seismic activity, any of which could account for the lack of preserved records. Alternatively, no paleoearthquakes may have occurred during this interval. If so, this pattern supports Ran’s¹² interpretation that fault segments with limited independence tend to exhibit greater variability in recurrence behavior, commonly characterized by alternating long and short recurrence intervals.

Fig. 12

Fitting diagram of average recurrence interval in the Late Pleistocene for the Xinbaoan–Shacheng fault.

Conclusion

Through high-resolution remote sensing interpretation, large-scale geological and geomorphic mapping, paleoseismic trench investigations, and Pleistocene chronological analyses, this study reconstructs the latest activity of the Xinbaoan–Shacheng Fault and constrains its Q_p3 slip rate and strong-earthquake recurrence interval. The principal findings are summarized as follows:

• (1) Fault geometry and coseismic surface rupture.

The Xinbaoan–Shacheng Fault preserves a discontinuous but well-defined coseismic surface-rupture zone, striking NW40°–65° and dipping southwest at roughly 70°, and the statistical magnitude-displacement empirical relationship indicate that it corresponds to a historical earthquake of M 6¾−7, which is consistent with the revised magnitude raised by the previous studies.

• (2) Late Pleistocene–Holocene activity and possible linkage to the 1720 M7 Shacheng earthquake.

Since the Q_p3, the fault has undergone significant deformation characterized by extensional kinematics. The most recent activity continued into the Late Holocene and may be associated with the 1720 M7 Shacheng earthquake. Although OSL ages from TC01 only provide upper bound of < 2.34 ± 0.08 ka for the youngest event, multiple independent indicators—including fault-strike compatibility, rupture-segment length, and the macro-intensity distribution of the 1720 event—support the plausibility that the 1720 earthquake ruptured this fault. However, the available chronological constraints remain insufficient for a conclusive correlation, and additional shallow OSL or radiocarbon sampling will be required to verify the source-fault relationship.

• (3) Paleoearthquake sequence and slip rate.

Paleoseismic trenching reveals multiple deformation events since the Q_p3, yielding an average recurrence interval of 3.39 ± 0.61 ka. The recurrence pattern appears irregular, exhibiting alternating long and short intervals. Based on cumulative displacement and chronological constraints, the Late Pleistocene vertical slip rate of the fault is estimated at 0.06 ± 0.01 mm/a (R² = 0.97).