Quartz Extraction

The quartz purification experiment was conducted at the Institute of Geology, China Earthquake Administration in Beijing, China, following the procedures described by Liu et al. (2010) and Wei et al. (2020). All the samples were separated into the 75–145 μm size fractions using wet sieving. The 75–145 μm size fraction has been identified as the most susceptible to optical bleaching during transportation (Voinchet et al., 2015). Each sample was treated with 30% H₂O₂ for 24 h and with 40% HCl for 24 h to remove organic material and carbonates, respectively. After washing and cleaning with distilled water, the samples were subjected to heavy liquid separation (2.7 g/cm³) to remove heavy minerals, and mechanical magnetic selection to remove magnetic minerals. In order to remove feldspars and etch quartz grains, the remaining part of the sample was then treated with 40% HF for 40 min. The samples were then placed again into 40% HCl for >4 h to subtract fluorides produced by the former HF treatment. Subsequently, the samples were washed with distilled water to neutralise the acid and dried at low temperature (~43°C) for over 24 h. X-ray diffraction analytical results revealed that the purified quartz from 11 samples had high purity ranging from 97.19% to 99.99%.

Gamma-Ray Radiation Dose

Quartz grains were dated by the ESR method using the multiple-aliquot additive dose method of Ti-related centres (Toyoda et al., 2000; Duval and Guilarte, 2015). Irradiation was performed using the ⁶⁰Co source of the laboratory of Peking University, Beijing, China, with a dose rate of 42.25 Gy/min. Each sample was divided into 10 aliquots for irradiation at the expected doses of 0 Gy, 237.65 Gy, 433.1 Gy, 863.7 Gy, 1677.5 Gy, 2662.5 Gy, 3798 Gy, 4887 Gy, 6512.5 Gy and 8811.5 Gy, calibrated using reference standard alanine dose tablets.

ESR Measurements

The ESR measurements were conducted using a BRUKER EPR041XG X-band spectrometer cooled to 77 K with liquid nitrogen in a finger-dewar at the State Key Laboratory of Earthquake Dynamics, Institute of Geology, China Earthquake Administration, Beijing, China. ESR measurements of the Ti–Li centre were performed with a microwave power of 5 mW and modulation amplitude of 0.16 mT. For the Ti–Li centre, ESR signal intensity was measured from the top of the peak at g = 1.979 to the bottom at g = 1.913 (Fig. 3; Rink et al., 2007; Tissoux et al., 2007; Liu et al., 2010; Duval and Guilarte, 2015). Given the anisotropy of the ESR signal intensity, each aliquot was measured at six angles (~60° interval) in the cavity, with the mean value being calculated to reduce the variation caused by the anisotropy of quartz.

Fig 3.

ESR spectrum showing the method used to determine the ESR signal intensity of the Ti–Li centre. Red arrows represent the measurement positions and g values of the Al and Ti–Li centres. ESR, electron spin resonance.

Equivalent-Dose Determination

In the present work, SSE functions were applied to the Ti–Li centre data to determine the equivalent dose (Woda and Wanger, 2007; Duval and Guilarte, 2015; Wei et al., 2020). The goodness of fit of these functions was assessed using the error coefficient and adjusted R-square. Equations of the fitting functions and all DRCs of all samples are shown in Fig. 4.

Fig 4.

Dose–response curves of the quartz Ti–Li centre for samples collected from terraces T3–T5 of the upper Han River. ESR, electron spin resonance.

Dose Rate Calculation

Contents of uranium (U) and thorium (Th) were measured using inductively coupled plasma–mass spectrometry, and potassium (K) was measured using inductively coupled plasma–optical emission spectrometry at Chang’an University, Xi’an, China. Water content was measured by weighing the samples, which were taken from the same location as the dated samples, before and after drying. The cosmic ray dose rate was estimated for each sample as a function of depth, altitude and geomagnetic latitude according to Prescott and Hutton (1994). Total dose rates were calculated using the conversion factors of Guérin et al. (2011), and the results are provided in Table 1.

Distribution of Terraces

Field investigation and digital elevation model terrain analyses show that five terrace levels are widely distributed in the Han River study area (Fig. 1). The elevations of the sampled terraces in each of the three studied profiles are given in Table 1. The higher terraces T3–T5 can be recognised as flat landforms distributed continuously over long distances. Each terrace is composed of coarse-gravel layers and fine-sand lenses, which are partly covered by loess layers (Fig. 2). Those fluvial–lacustrine deposition characteristics are similar to the terrace sedimentary of the upper Jinsha River (Wei et al., 2023). The fine loess is not suitable for extracting quartz for ESR dating; therefore, sand from lenses in the fluvial–lacustrine sediments was collected for ESR dating.

ESR Dating Results

Wei et al. (2020) showed that the Ti–Li centre ESR signal intensity of present-day Yangtze River fluvial sediments was zeroed by sunlight before burial. The sediment of the Han River, which is a major tributary of the Yangtze River, undergoes sediment transportation mainly by saltation (Sun et al., 2020), which favours zeroing of the ESR signal. Therefore, we consider that the quartz ESR signal of the Han River terrace samples was zeroed. In addition, Fig. 4 shows that the quartz ESR signal intensity increases markedly with increasing γ-ray dose. All of these data indicate that our ESR ages are reliable (Table 1). The ESR ages and uncertainties for T3, T4 and T5 of the Wufeng terrace profile are 417 ± 48 ka, 627 ± 97 ka and 1105 ± 182 ka, respectively; those for T3, T4 and T5 of the Qingqu terrace profile are 405 ± 41 ka, 622 ± 77 ka and 1129 ± 121 ka, respectively; and those for T3, T4 and T5 of the Yunyang Basin terrace profile are 422 ± 67 ka to 401 ± 59 ka, 621 ± 52 ka and 1102 ± 60 ka to 1099 ± 117 ka, respectively.

Formation Ages of Terraces T3–T5 of the Upper Han River

We compared our ESR ages with previously determined OSL dating results for T1–T2 and paleo-magnetic dating results for the higher-level terraces. Pang et al. (2015) determined an age of 55–25 ka for T1, and Zhu et al. (2022) estimated an age of 220–180 ka for T2. Van Buren et al. (2020) determined an age of 693–625 ka for T3. Results of stratigraphic correlation have given age estimates for T4, T5 and T6 of 1.05–0.78 Ma, ~1.5 Ma and 1.82 Ma, respectively (Sun et al., 2017a,b; Zhu et al., 2023).

The results of our field investigation (see Fig. 2) show that the fluvial–lacustrine sediments at the same altitude of various terraces directly overlie bedrock; therefore, the samples collected are of ancient terrace deposits. In summary, our dating results show that the ESR central ages of terrace quartz sediments for each of T3, T4 and T5 in the three studied terrace profiles are highly consistent, in the ranges of 422–401 ka for T3, 627–621 ka for T4 and 1129–1099 ka for T5.

Mechanism of Formation of Higher Terraces of the Upper Han River

The evolution of fluvial systems is highly sensitive to climate change, including the formation and preservation of terraces (Vandenberghe, 2003; Winsemann et al., 2015). In Fig. 5, we compare the formation ages of the upper Han River terraces with the deep-ocean oxygen isotope curve, which is a proxy for paleo-climatic conditions. This figure shows that it is only T3 and T4, but it is likely for all the dated terraces of the upper Han River to have formed during the transition from a glacial to interglacial period. This correspondence is consistent with rivers originating from mountain systems in temperate regions (Hatten et al., 2012). Indeed, changes in monsoon climate and vegetation have also resulted in changes in river runoff and sediment volume in the past (Zhang et al., 2023). During glacial–interglacial transitions, the moderate sediment supply would have led to partial exposure of the riverbed and strengthened undercutting, allowing terrace systems to form (Wang et al., 2015, 2021).

Fig 5.

Correspondence between the ages of upper Han River terraces and climate change recorded by marine δ¹⁸O.

The age range (central value range) for the formation of terrace T5 determined in this study is 1129–1099 ka (i.e., slightly younger than ~1.2 Ma), which corresponds to the early stage of the mid-Pleistocene transition (MPT), when the periodicity of Pleistocene glacial cycles changed from ~41 ka to ~100 ka. It is possible that the small amplitude and short duration of climatic cycles before this time disfavoured the formation and preservation of terraces prior to the formation of T5. This interpretation is supported by studies of terrace sediments in other locations, which have indicated the terrace formation since the MPT has been closely related to the 100 ka climatic cycle (Bridgland et al., 2008). In addition, the increase in amplitude and duration of climatic cycles since the MPT may have led to phases of more intense river erosion and accumulation. This might explain why terraces were formed and preserved after 1.2 Ma (but not before) in the upper section of the Han River.

Uplift of the eastern Qin Mountains since 1.2 Ma has also influenced the development of the Han River terraces (Liu, 2004; Sun et al., 2016). High rates of tectonic uplift are inferred to have led to substantial elevation differences between river terrace levels and amplified the influence of climate change on the sedimentation–erosion process of the river, leading to the development and preservation of flights of terraces (Wang et al., 2015). In summary, after 1.2 Ma, the rates of uplift and erosion increased, and terrace systems were formed in the upper Han River, with cyclic climatic transitions amplifying these effects on terraceformation.