To determine tectonic activity and constrain the slip rates of the faults, the offset strata and landform surfaces must be dated using suitable methods. It is difficult to apply radiocarbon 14C dating in the Hexi Corridor because of the lack of organic material. OSL dating is a reliable method to date fine-grained sand, silt and loess deposits if they have been completely bleached. However, the fine-grained sediments often accumulate on the footwall adjacent to the fault scarp. In contrast, there are usually no loess deposits on the hanging wall along the Yinwashan and Xinminpu faults. Quartz-rich pebbles and gravels on its surface provide materials that allow use of 10Be surface exposure dating. Hence we choose OSL dating to constrain the age of fine-grained sediments at the footwall, and 10Be exposure dating to determine the age of faulted landform surfaces.
In this study, all the OSL samples were collected in stainless steel circular tubes with length of 20 cm, hammered into the fine-grained sand or loess. The samples were processed in the State Key Laboratory of Earthquake Dynamics, Institute of Geology, China Earthquake Administration. Extraction and purification of the quartz grains followed the laboratory protocols for Chinese loess under subdued red light (Wang et al., 2005; Forman, 1991; Lu et al., 1988). The samples were processed with 30% HCl and 37% H2O2 to remove carbonates and organic material. Then the 4–11 μm poly-mineral size fraction was separated by hydrostatic precipitation and 30% H2SiF6 was used repeatedly to remove feldspar and isolate the quartz component. The equivalent dose was determined by following the simplified multiple aliquot regenerative-dose (SMAR) method (Wang et al., 2005). The environmental dose rate was also calculated based on the concentrations of U, Th, and K with the contribution of cosmic rays and water content (Aitken, 1998). The results of the OSL dating of samples are shown in Table 1.
Sample Depth(m) U-238(Bq/kg) Th-232(Bq/kg) K-40(Bq/kg) Doserate (Gy/ka) Equivalentdose (Gy) OSLage (ka) YMOSL-6 0.8 34.0±4.2 44.2±0.5 584.6±9.9 3.4±0.3 21.0±1.2 6.2±0.4 YMOSL-14 1 34.5±4.5 27.0±0.4 657.8±11.0 3.3±0.3 54.3±2.0 16.3±0.7 YMOSL-11 0.5 25.1±4.3 25.3±0.4 457.1±8.3 2.5±0.2 9.3±0.6 3.8±0.3 YMOSL-12 0.8 25.7±3.7 30.1±0.4 482.5±8.3 2.6±0.3 11.6±0.4 4.6±0.2
Table 1. Result of OSL analysis for the layers in the trench
For 10Be exposure dating, we used the amalgamation method proposed by Anderson et al. (1996). At every site, a large number (> 50) of cm-sized, quartz-rich pebbles on the terrace surface were obtained. Additionally, a sample (YWS-1) in the active channel was collected to calculate the quantity of inherited 10Be concentration before the deposition the pebbles on the surface of the terraces. We think this sample to represent an inherited 10Be concentration similar to the induced inheritance in the profile (Brown et al., 1998; Anderson et al., 1996). The samples collected in Xinminpu fault were from terraces of the Baiyang River. To calculate exposure age, we used the published value of (3.32±0.57)×105 atmos/g as the inherited 10Be concentration (Liu et al., 2019). Using standard chemical procedures, quartz extraction and BeO preparation were completed at the State Key Laboratory of Earthquake Dynamics, China Earthquake Administration (Zheng et al., 2013a; Hetzel et al., 2002; Kohl and Nishiizumi, 1992). The AMS (Accelerator Mass Spectrometry) analysis to determinate the 10Be/9Be ratios were carried out at CEREGE (Centre Européen de Recherche et D'Enseignement des Géosciences de L'Environnement) in France. We used a low-elevation, high-latitude production rate of 4.3 atoms a-1 g-1 SiO2, which had been adjusted for latitude and elevation (Lal, 1991), to ultimately estimate the surface ages based on the concentrations (Granger et al., 2013). The calculations were all executed with the Microsoft Excel calculator Cosmocalc (Vermeesch, 2007). Table 2 depicts the 10Be dating results.
Age (ka) YWS-1 97.893 81 39.885 23 1 943 - 26.153 0.305 2.04 0.647 2.1 4.91±0.16 - - YWS-2 97.893 13 39.884 96 1 945 18.87 28.291 0.304 2.03 1.35 3.57 9.55±0.26 4.64±0.42 24.6±2.2 YWS-3 97.892 49 39.885 43 1 946 18.89 25.691 0.308 2.06 1.83 4.74 14.54±0.38 9.64±0.54 51±2.9 YWS-4 97.894 19 39.884 18 1 948 18.91 26.744 0.305 2.03 2.01 6.62 15.12±0.50 10.2±0.67 54±3.5 XMP-1 97.719 04 39.942 95 1 767 16.67 36.165 0.297 1.99 6.12 10.6 33.55±0.58 30.23±1.15 181.3±6.9 XMP-2 97.718 01 39.943 49 1 769 16.69 42.111 0.307 2.05 4.53 8.63 21.98±0.42 18.67±0.99 111.8±5.9
Table 2. 10Be and exposure age data used in this study
The Yinwa Shan is an uplifted mountain in the Jiuxi Basin with its highest elevation ~2 195 m, and surrounded by alluvial fans with elevation of ~1 800–1 900 m, and mainly consists of Ordovician greyish-green metamorphic rock (Bureau of Geology and Mineral Resources of Gansu Province, 1989). The NNW-trending Yinwashan fault, with a length of ~25 km, lies along the eastern side of the Yinwa Shan (Fig. 2). At its northwestern end, the fault has developed in the eroded Cretaceous platform. It is difficult to determine its activity here because of the absence of Quaternary sediments. In its middle portion, the fault forms a topographic boundary between the Yinwa Shan and alluvial fans (Fig. 3a). A reverse fault outcrop with dip angle of ~53° was found in this segment (Fig. 3b). The hanging wall is weathered Ordovician greyish-green metamorphic detritus, while the footwall is Late Pleistocene light-grey loose alluvium, whose age is estimated to be about 22.4±2.6 ka by TL dating (Chen et al., 2005). The fault offsets this stratum and is covered by Holocene alluvium which is approximately 8.6±1.6 ka (Chen et al., 2005). To the southeast, the fault is developed in the alluvial fans derived from the Hei Shan. A series of north facing scarps can be found in satellite images and field observations (Fig. 3c). A trench, dug across the scarp, revealed a reverse fault with a dip angle of ~65°, with Cretaceous fuchsia mudstones and sandstones are thrust over Pleistocene alluvial gravels (Fig. 3d). The fault offset and colluvial wedge are overlain with loess and surface soil. One optically stimulated luminescence (OSL) sample collected at the base of the colluvial wedge in the footwall yielded a depositional age of 16.3±0.7 ka. An OSL sample from the loess cover yields an age of 6.2±0.6 ka, suggesting the latest earthquake occurred sometime before this time.
Figure 2. Geological map of the study area and its adjacent region, showing the distribution of bedrock lithology (geological data simplified from Bureau of Geology and Mineral Resources of Gansu Province, 1989).
Figure 3. Photographs of features along the Yinwashan fault and sections across it. (a) Fault troughs along the Yinwashan fault where it neighbors the Yinwa Shan; (b) natural fault section near the Yinwa Shan revealing a reverse fault dipping at ~53°; (c) fault scarps developed on the alluvial fan; (d) trench across the fault scarp on the alluvial fan reveals a reverse fault with dip angle of 65°, black solid circles denote the location of OSL samples.
The Xinminpu fault, striking at ~300° in the middle part of the Jiuxi Basin, extends for ~20 km from Qingquan in the west to the Xinminpu Village in the east (Fig. 2). A natural outcrop near the Xinminpu Village reveals a low angle (~25°) thrust dips SW (Fig. 4a). The hanging wall block is comprised by Neogene mudstones and sandstones, while the footwall is comprised by Late Pleistocene light-grey loose gravels, showing oriented rearrangements and drag deformation. The geomorphic surface is displaced by about 2–3 m in vertical (Fig. 4a). To the west, a series of springs occur along the north facing fault scarps. The new fault scarps are often developed at the base of the old scarps (Fig. 4b). Surface rupture is still preserved along the new scarps (Fig. 4c), which is associated with a historical earthquake in AD 1 785 with a magnitude of 6.8 (He et al., 2010). A trench across the scarp reveals Neogene brick-red mudstones and sandstones are thrust over Holocene alluvial gravels (Fig. 4d). Three thrusts are found on the west trench wall. The thrusts dip SW with a dip of ~28°. Two paleoseismic events can be distinguished in the trench. The lower two thrusts offset the silt soil layer, and is covered by the loose soil at the surface. Two OSL samples obtained from the silt soil layer suggested the earlier event occurred after ~3.8–4.6 ka (Fig. 4d). The latest earthquake event is the top thrust offsets the loose soil near the surface and connects to the surface rupture produced by the latest earthquake (Fig. 4d).
Figure 4. Photographs of feature along the Xinminpu fault and sections across it. (a) Natural fault section on the Xinminpu fault to the east of Xinminpu village; (b) fault scarps developed on the alluvial fan; (c) surface ruptures are still preserved along the new fault scarp; (d) trench across the new fault scarp at the foot of old scarp reveals a thrust fault with a dip of 28°. Black solid circles denote the location of OSL samples.
The study site for the Yinwashan fault is situated at the alluvial fan at the south of the Hei Shan (Fig. 2). The alluvial fans have been incised by ephemeral streams. The hanging wall of fault continues to rise since the activity of fault, resulting in a NE-facing fault scarp with several fluvial terraces of different heights on the hanging wall (Fig. 5a). On the footwall, it was covered by the most wide-spread active alluvial fans. Combining satellite image interpretations with field investigations, three level terraces can be identified at this site, marked as T1, T2, and T3 going from the youngest to the oldest (Fig. 5b). The fault scarp of the T3 terrace is the most obvious (Fig. 5c), Cretaceous sandstones and mudstones having been exposed because of the erosion of its surface. To constrain the vertical displacement on the three terrace surfaces, three topographic profiles (P1 to P3) were obtained using differential GPS (Fig. 5b). We projected the measured profiles onto a profile perpendicular to the fault (Directions are 46°, 51° and 26° from P1 to P3, respectively), and the vertical offsets on every profile were calculated by fitting straight trend lines to the hanging wall as well as the footwall. The displacements for T1, T2 and T3 are 2.1±0.1, 3.0±0.2, and 5.2±0.3 m, respectively (Fig. 5d). As the footwall has been buried by the younger alluvial fan deposits, these vertical displacements are minimum values, and therefore, the estimated slip rates will also be minimum values. We use the vertical offset of the terrace surfaces from P1 to P3 (2.1±0.1, 3.0±0.2, and 5.2±0.3 m) and corresponding abandonment ages (24.6±2.2, 51±2.9, and 54±3.5 ka) to determine the vertical component of the slip rate. Vertical slip rates show mostly two values 0.1–0.09 and 0.06 mm/yr. The smaller rate is about a half of bigger one. The bigger slip rate could be more reasonable than the smaller slip rate since the height of fault scarp is a minimum value. Therefore, we obtain a vertical rate of 0.09±0.01 mm/yr for the fault (Fig. 6). Trenches and natural exposures show that reverse fault dips 50°–65° to the southwest (Figs. 3b, 3d). Considering uncertainties of dip angle, we use 60°±10° to estimate a horizontal shortening rate of 0.05±0.03 mm/yr.
Figure 5. Maps and topographic profiles for the Yinwashan fault study site. (a) Detail of satellite imagery (data from Google Earth), a clear fault scarp cuts the alluvial fan and terraces formed by downcutting streams; (b) geomorphological interpretation of the imagery showing abandoned alluvial surfaces, fault trace, sample locations (stars), locations of topographic profiles, and the location and orientation of the photograph in panel (c); (c) field photograph of the fault scarp for the T3 terrace; (d) topographic profiles across the fault scarp at the locations shown in panel (b) determined using differential GPS.
Figure 6. Average vertical slip rates on the Yinwashan and Xinminpu faults, which were estimated from vertical offsets and their corresponding abandonment ages.
The study site for the Xinminpu fault was situated at the west of Qingquan on an alluvial fan (Fig. 2). The faulted landform features are similar to those of the Yinwashan fault. The terraces in the hanging wall are separated from a younger alluvial fan in the footwall by a NE-facing fault scarp (Figs. 7a, 7b). Three level terraces are preserved on the hanging wall, denoted as T1 to T3 from youngest to oldest (Fig. 7b). Surface ruptures are found developed at the base of scarps in T3 (Fig. 7c). Fresh faces have been preserved, suggesting that it formed relatively recently. The T1 terrace is covered by loess with thickness of ~1 m. As previous study indicates that the loess accumulation in the western Qilian was initiated in Holocene (Küster et al., 2006; Stokes et al., 2003), the age at the bottom of the loess is much younger than the abandonment age of the fluvial terraces. We did not collect OSL or 10Be sample for this terrace. Based on the fault scarp heights of T2 and T3 (Fig. 7d; 12.6±0.6, 17.7±0.6 m, directions are 44° and 56° for P2 and P3) and the corresponding abandonment ages (111.8±5.9, 181.3±6.9 ka) obtained from 10Be dating, we estimate the vertical slip rate of the Xinminpu fault to be 0.1±0.02 mm/yr (Fig. 6). Using the estimated dip of 25°±5° revealed by trenching and natural exposures (Figs. 4a, 4b), the horizontal shortening rate is determined to 0.23±0.06 mm/yr.
Figure 7. Maps and topographic profiles for the Huoshaogou study area of the Xinminpu fault. (a) Satellite image near Huoshaoggou Village (data from Google Earth); (b) geopmorphological interpretation showing abandoned alluvial surfaces, fault trace, sample locations (stars), the location of topographic profiles, and the location and orinetation of the photograph in panel (c); (c) field photograph of the fault scarp for the T3 terrace. New scarp developed in the front of the old scarp. Fresh fault scarps are preserved along the new trace, suggesting that it formed relatively recently; (d) topographic profiles across the fault scarp at the locations shown in panel (b) determined using differential GPS.