Luminescence Dating of Holocene Fluvial Sediments from the Daluze Area in the North China Plain

Abstract

Optically stimulated luminescence (OSL) dating is an important method for determining the ages of late Quaternary sediments. However, partial bleaching of quartz in fluvial sediments remains a challenge, with debates on grain-size effects in different sedimentary environments. The aim of this paper is to explore the bleaching degree and its influencing factors of different grain-size quartz in fluvial sediments from the Yanchi section in the Daluze area, North China Plain. According to sedimentological methods and grain size analysis, lacustrine and fluvial layers were identified, and the ages of sediments were determined by OSL and ¹⁴C methods. The key findings are as follows: (1) Fine-grained quartz can be better bleached than coarse/medium-grained quartz for early–middle Holocene fluvial sediments. (2) The OSL method can yield reliable ages for early–middle Holocene fluvial sediments, while it overestimates these for late Holocene fluvial sediments. This probably results from variations in sediment sources and hydrodynamic conditions. (3) The dating results show that there are three fluvial activity periods in the Daluze area: 10.8~10.2 ka, 5.3~4.7 ka, and after 1 ka. This paper provides a reliable chronological framework for the evolution of regional sedimentary environments and offers references for luminescence dating of fluvial sediments in similar environments.

1. Introduction

Accurate chronological determination of fluvial sediments is essential for understanding the development and evolution of Holocene fluvial geomorphology. Radiocarbon (¹⁴C) dating is one of the primary methods for dating Holocene sediments and is generally reliable for lacustrine deposits. However, its application to fluvial sediments is challenged by the scarcity of organic material, carbon reservoir effects, and re-deposition processes [1,2,3,4]. In contrast, optically stimulated luminescence (OSL) dating utilizing quartz and feldspar grains, which are commonly found in fluvial sediments, offers a direct means of dating the last exposure of mineral grains to sunlight prior to burial. This makes OSL widely applicable to the dating of fluvial sediments [5,6,7,8,9,10,11,12]. Nevertheless, a major limitation in applying OSL to fluvial sediments is the potential for incomplete bleaching of the luminescence signal before burial, which can result in significant age overestimation [5,8,13]. This issue is particularly pronounced when dating young Holocene deposits [8,14]. Quartz typically bleaches more rapidly than the infrared-stimulated luminescence signal of K-feldspar [8] and therefore holds greater potential for accurately dating young Holocene sediments.

Research suggests that grain size significantly influences the bleaching efficiency of quartz OSL signals, but findings remain inconclusive [8]. Some studies propose that coarse-grained quartz exhibits superior bleaching characteristics [7,15,16], primarily because these grains are transported via traction or saltation over longer distances and at slower velocities, thereby receiving greater exposure to sunlight prior to deposition [7,17]. In contrast, fine grains may be encapsulated within aggregates, which limit their sunlight exposure and consequently lead to age overestimation in OSL dating [7,17]. Even when coarse grains are not fully bleached, the use of small aliquots or single-grain analysis in combination with statistical models can effectively reduce the impact of incomplete bleaching [8,9,17]. Conversely, other studies argue that fine-grained quartz may bleach more effectively [18,19], largely because these grains can remain suspended in the water column for extended durations, increasing their exposure to sunlight. Meanwhile, coarse grains tend to settle more rapidly on the riverbed, and turbidity in the water column can attenuate light penetration, limiting bleaching during transport [5,20]. Additional findings indicate that the bleaching efficiency of fine-grained quartz varies according to depositional processes. For example, [9] reported that fine-grained quartz in sediments deposited in the Yangtze River Delta between 9~2 ka (1 ka = 1000 years before the present) was well bleached and yielded reliable ages. However, younger deposits from 0.6~0.4 ka exhibited clear signs of incomplete bleaching, likely due to differences in sediment sources and transport mechanisms. Similarly, [16] measured the equivalent doses (De) of quartz in modern Yellow River sediments and found that fine grains within sand layers had De values ranging from 5 to 20 Gy (Gy: gray, the SI unit of radiation dose, where 1 Gy = 1 J/kg, used to measure the total ionizing radiation received by sediments), whereas suspended fine quartz grains—more thoroughly bleached—showed residual De values of less than 3 Gy. These results underscore the significant role of depositional mode in influencing the bleaching efficiency of fine-grained quartz OSL signals. In summary, the complex nature of fluvial deposition leads to spatial and temporal variability in the bleaching efficiency of quartz grains of different sizes. Further comparative studies examining OSL ages across multiple grain sizes, and validated against independent dating methods, are needed to better understand bleaching behavior and its controlling factors under diverse depositional environments.

The North China Plain is the second largest plain in China. Due to frequent river avulsions and flooding, numerous depressions and paleolakes have formed, resulting in the accumulation of abundant fluvial and lacustrine sediments. OSL ages have been obtained using quartz of various grain sizes from both river and lake deposits. For fine-grained quartz, [21] reported good bleaching efficiency in fluvial sediments of Baiyangdian since the late Pleistocene, based on agreement between fine-grained quartz and mixed-mineral OSL ages. However, [22] found that fine quartz in fluvial flood deposits from the past 1 ka in the paleolake area was poorly bleached, leading to age overestimations ranging from 1 to 10 ka. Regarding medium- or coarse-grained quartz, [23] dated Holocene fluvio-lacustrine sediments from a site near the shoreline and concluded that in clay-rich fluvial flood deposits, OSL ages results were overestimated by 3~6 ka. However, they found the method suitable for OSL dating in silt-dominated deposits. In contrast, ref. [22] observed that coarse-grained quartz in silt-dominated flood layers of <1 ka deposits also yielded age overestimations of 0.5~3 ka. In conclusion, for fluvial sediments from diverse environmental settings across the region, it remains uncertain which quartz grain size exhibits better bleaching behavior and how to reliably determine accurate OSL ages.

This study focuses on a Holocene river–lake sedimentary section located in the North China Plain. Based on sedimentary characteristics, fluvial and lacustrine units were identified. OSL dating was conducted on fine-, medium-, and coarse-grained quartz extracted from fluvial deposits, and results were cross-validated using OSL and radiocarbon (¹⁴C) ages from adjacent lacustrine layers to evaluate the reliability of the fluvial OSL ages. We assessed the bleaching efficiency of different quartz grain sizes and investigated the causes of variation in bleaching behavior based on grain size analysis. The aim of this study is to determine how reliable OSL ages can be obtained for fluvial sediments in the North China Plain.

2. Materials and Methods

2.1. Study Area

The North China Plain is an alluvial plain shaped by the Yellow, Huai, Hai, and Luan Rivers. During historical periods, numerous depressions and lakes existed across this region. Daluze was once among the largest lakes in the North China Plain [24]. It has shrunk into two small lakes, which are Ningjinbo in the north and Daluze in the south [25]. At present, the area is occasionally inundated by floodwaters. The Daluze area is located in the central part of the North China Plain, within a structural depression between the piedmont fans of the Taihang Mountains and abandoned courses of the Yellow River. It represents a typical fan-front and inter-fan depression formed by differential accumulation. The Laozhang River and the Hutuo River lie to its south and north, respectively (Figure 1). Historically, Long-source rivers such as the Yellow River, Laozhang River, Hutuo River, as well as short-source rivers originating from the Taihang Mountain such as Fuyang River, Sha River, and many other rivers, have all flowed into or through Daluze/Ningjinbo lake [26,27].

The sedimentary strata of Daluze consist of alluvial-type lacustrine deposits, which are representative of those found across the North China Plain. Previous studies have shown that the Holocene strata in this region are characterized by lacustrine deposits interbedded with multiple fluvial flood layers [22,24,26]. Numerous scholars have examined Holocene lake evolution and climate change based on these sediments [28,29,30,31,32,33,34,35] and have analyzed the fluvial components using indicators such as grain size, magnetic susceptibility, organic carbon content, pollen, and elemental composition. It has been suggested that, in addition to materials transported by piedmont rivers, some sediments may have originated from the Yellow River [24,26].

2.2. Sedimentary Characteristics of the Section and Sample Collection

The YC (Yanchi) section (37°16′17.17″ N, 114°53′48.48″ E) is located in the central part of Daluze (Figure 1) and has a total depth of 1085 cm. Lithologic descriptions are provided in

Table 1. Lithology and sedimentary unit of YC section.

Sedimentary Unit	Layer No.	Depth (cm)	Lithological Description
Modern soil layer (S)	23	0~30	Light brown soil, with plant roots and root holes
Fluvial Flood Layer 4 (F4)	22	30~60	Reddish-brown silty clay with shell fragments
	21	60~115	Gray–yellow silt; 95~100 cm is brown–yellow clayey silt
	20	115~150	Light brown–yellow clayey silt; reddish-brown at 115~125 cm
	19	150~330	Gray–yellow silt
	18	330~340	Interbedded reddish-brown silty clay and gray–yellow clayey silt
	17	340~380	Interbedded brown clayey silt and brown–yellow silt
	16	380~420	Brown silty clay with greasy luster; bluish-gray at ~395 cm
Mountain flood Layer (M)	15	420~510	Dark brown silty clay, blocky and structureless
Fluvial Flood Layer 3 (F3)	14	510~525	Bluish-gray clayey silt with shell fragments
	13	525~535	Gray–yellow silt
	12	535~555	Brown–yellow clayey silt
	11	555~620	Gray–yellow silt
	10	620~630	Bluish-gray silt
	9	630~660	Brown silty clay
Lacustrine Layer 2 (L2)	8	660~710	Gray–black silty clay with shell fragments; carbonaceous debris at 660 cm
Fluvial Flood Layer 2 (F2)	7	710~745	Brown–yellow clayey silt
	6	745~920	Gray–yellow silt
	5	920~926	Brownish clayey silt
Lacustrine Layer 1 (L1)	4	926~970	Gray–black silty clay
Fluvial Flood Layer 1 (F1)	3	970~1020	Gray–yellow silt; coarsens downward, transitional contact
	2	1020~1060	Brown clayey silt with calcareous nodules
	1	1060~1085	Brown–yellow fine sand, well sorted; large shells at base

, with the lithologic column and field photographs shown in Figure 2. Overall, the YC section contains two distinct gray–black silty clay layers (Figure 2), situated at depths of 660~710 cm and 926~970 cm, respectively, and separated by an approximately 210 cm thick layer of gray–yellow silt. Above 660 cm, the deposits mainly consist of interbedded gray–yellow silt and brown to reddish-brown silty clay. Between 970 and 1085 cm, the lithology exhibits marked changes, transitioning downward through gray–yellow silt, brown clayey silt, brown–yellow fine sand, and bluish-gray clayey silt. The thickness of individual layers ranges from 20 to 50 cm.

In the study area, fluvial flood deposits and lacustrine sediments differ markedly in color, grain size, and sedimentary structure. In addition, abrupt contacts between adjacent strata provide clear evidence for the identification of flood events. Accordingly, this study distinguishes fluvial flood deposits and delineates sedimentary units based on field characteristics such as color, lithology, sedimentary structure, fabric, and contact relationships, in combination with regional sedimentological insights. The main sedimentary units identified in the YC section include the modern soil layer, mountain flood layer, fluvial flood layer, and lacustrine layer. Detailed stratigraphic divisions and their characteristics are presented in Table 1.

Modern soil layer: This unit, composed of modern soil, exhibits clear signs of pedogenesis and aggregate structures, including plant roots and root holes (Figure 2a).

Fluvial flood deposits: These deposits typically consist of interbedded gray–yellow silt and reddish-brown silty clay (Figure 2b), occasionally containing brown/yellow fine sand or clayey silt, consistent with characteristics of fluvial flood deposits in the North China Plain [36,37]. Variations in grain size reflect hydrodynamic fluctuations during flood events with stronger flows depositing coarser material. Silt layers are generally thicker (50~180 cm) than silty clay layers (20~40 cm). Flood Layers 2 and 4 each contain approximately 1.7 m thick gray–yellow silt beds, indicating high-magnitude floods and abundant sediment supply. Contacts with underlying mountain flood layer or lacustrine layers are sharp (Figure 2c,d).

Mountain flood layer: This unit consists of thick brown silty clay that is blocky, dense, and relatively homogeneous, containing aggregates with a diameter of mainly 0.5~10 mm, exhibiting characteristics similar to those formed through pedogenesis (Figure 2c). Its structural characteristics resemble those of paleosols, leading to the inference that this layer was deposited by mountain flood events transporting slope-derived soils to the site. In addition to distinct sedimentary structures compared to fluvial flood deposits, the mountain flood layer is notably thicker (nearly 100 cm), whereas silty clay layers within fluvial flood deposits are generally thinner (20–40 cm).

Lacustrine deposits: These deposits consist of dense gray–black silty clay (Figure 2d) that is clearly distinguishable from other sediment types by its distinct color and sharp stratigraphic boundaries.

Grain size analysis samples, as well as OSL and ¹⁴C dating samples, were collected from the YC section. A total of 377 grain size analysis samples were obtained at 2~5 cm intervals from top to bottom except for the modern soil layer. To constrain the chronology of the stratigraphic units, 27 OSL samples were collected from various depths. Additionally, 3 ¹⁴C dating samples were taken from the tops of the two lacustrine layers (lacustrine Layers 1 and 2) for age comparison. Previous studies have addressed the OSL dating results of quartz grains of different particle sizes from strata above 660 cm [22]. This study focuses on the OSL dating results of sediments below 660 cm, alongside a comprehensive analysis of the bleaching efficiency of quartz grains of varying sizes and the factors influencing it.

2.3. Methods

2.3.1. Grain Size Analysis

Each sample (3~5 g) was placed in a 100 mL beaker, to which 10 mL of a 2:1 hydrogen peroxide solution (by volume) was added. The mixture was heated while distilled water was continuously added to prevent drying. After bubbling ceased, 10 mL of a 2:1 hydrochloric acid solution was added, with distilled water again added as needed. Following boiling, the beaker was removed from heat and allowed to cool. Approximately 50 mL of distilled water was then added, and the sample was left to stand for 12 h. The supernatant was decanted, leaving about 20 mL of sediment. Subsequently, 10 mL of 36 g/L sodium hexametaphosphate dispersant was added and stirred thoroughly. The sample was placed in an ultrasonic bath and sonicated for approximately 10 min at 50% power. Grain size analysis was conducted using a Mastersizer 2000 Laser Particle Size Analyzer, produced by Malvern Instruments, Malvern, UK. Sediment grain sizes were classified as follows: D > 62.5 μm; silt: 3.9~62.5 μm; clay: D < 3.9 μm.

2.3.2. OSL Dating

OSL dating samples were prepared under red light conditions (wavelength 640 ± 10 nm) in the laboratory. Approximately 2 cm was removed from each end of the stainless steel tube, and around 20 g of sediment was extracted for determining the concentrations of U, Th, and K and for measuring water content. The remaining central portion of each sample was treated with 30% hydrogen peroxide (H₂O₂) and 10% hydrochloric acid (HCl) to remove organic matter and carbonates. Subsequently, samples were etched with 30% hydrofluorosilicic acid for approximately 5 days and repeatedly washed with distilled water until neutral pH was achieved. For fine-grained or medium-grained samples, fine grains (4~11 μm, abbreviated as “f”) were separated using the sedimentation method, while medium grains (41~65 μm, abbreviated as “m”) were separated by wet sieving. These fractions were further etched with 30% hydrofluorosilicic acid for about 5 days and washed to neutrality. Coarse-grained samples (65~90 μm or 90~125 μm, abbreviated as “c”) were separated by sieving and etched with 40% hydrofluoric acid (HF) for 40~60 min, followed by washing with distilled water, low-temperature drying, and removal of magnetic minerals.

The purity of each sample was confirmed by infrared signal measurement, verifying that the quartz was sufficiently pure. De values were measured by using the Lexsyg Smart TL/OSL Reader (Freiberg Instruments GmbH, Dresden, Germany) following the single-aliquot regenerative dose (SAR) protocol [38]. Details are provided in

Table 2. Protocol for SAR method.

Step	Procedure
1	Give regenerative dose Di, for the natural sample Di = 0
2	Preheat for 10 s at 220 °C (determined by preheat plateau test)
3	Stimulate for 70 s at 125 °C to measure signal Li (Ln for natural sample)
4	Give test dose (DT)
5	Cutheat to 160 °C, 0 s hold
6	Stimulate for 70 s at 125 °C to measure signal Ti (Tn for natural sample)
7	Return to 1

.

Annual dose rates were calculated based on the concentrations of uranium, thorium, and potassium measured using inductively coupled plasma mass spectrometry (ICP-MS). Conversion factors followed the values recommended by [39]. Cosmic dose rates were estimated following the method of [40]. Since water attenuates α, β, and γ radiation, water content during burial has a direct impact on the annual dose rate. After collection, samples were sealed and water content measured as soon as possible. The estimated average burial water content was derived from measured values and depositional environment conditions. Dose rate corrections were applied according to the method of [41]. Environmental dose rates and age calculations were performed using DRAC (Dose Rate and Age Calculator) [42].

2.3.3. ¹⁴C Dating

¹⁴C sample analysis involved sample pretreatment and accelerator mass spectrometry (AMS). Sample preparation included both raw sample pretreatment and graphite target preparation. Pretreatment aimed to remove inorganic carbon (particularly minerals such as calcite, marble, and dolomite) as well as exogenous organic material attached to sample surfaces. A standard acid–alkali–acid (ABA) procedure (2 mol/L HCl, 0.5% NaOH) was employed to eliminate exogenous carbon contamination and purify the target component for dating [43]. The purified carbon was then converted into graphite targets within a vacuum system [44] and analyzed using AMS. Background correction was performed using blank samples from the same batch. Isotopic fractionation correction was applied to the measured ¹⁴C/¹²C ratios using online δ¹³C values obtained from AMS, based on the isotope fractionation correction model. All ¹⁴C samples were analyzed at Lanzhou University. Calibrated ages were obtained using the OxCal 4.4 (https://c14.arch.ox.ac.uk/oxcal/OxCal.html, accessed on 20 June 2025) with the latest IntCal20 calibration curve [45], converting conventional radiocarbon years (δ¹³C-corrected BP ages) into calendar years (cal BP).

3. Results

3.1. Grain Size Analysis Results

This study analyzed the sand, silt, and clay content of each sample, along with parameters such as median grain size and D90 (the grain size at which cumulative volume exceeds 90%). Variations in grain size characteristics with depth are shown in Figure 3, and parameters for different sedimentary units are summarized in

Table 3. Statistical parameters of grain size characteristics for fluvial flood layer, mountain flood layer, and lacustrine layer in YC section.

Sediment Type	Value Type	Median Grain Size (μm)	D90	Content (%)
				Clay	Silt	Sand
Fluvial Flood Layer	Maximum	123.08	259.61	53.39	99.88	80.01
	Minimum	3.71	9.22	0.00	17.06	0.00
	Average	24.66	58.36	16.40	71.73	11.87
Mountain Flood Layer	Maximum	5.48	25.06	66.47	65.42	4.38
	Minimum	2.98	6.59	34.58	33.53	0.00
	Average	3.88	11.08	52.35	47.04	0.61
Lacustrine Layer	Maximum	41.56	130.80	41.85	84.15	23.39
	Minimum	5.08	23.28	5.02	57.55	0.60
	Average	14.60	49.84	20.90	72.84	6.26

.

It can be seen that although sediments from all three depositional units (mountain flood deposits, fluvial flood deposits, and lacustrine deposits) are predominantly composed of silt, notable differences exist among them. On one hand, the range of variation in grain size parameters differs significantly between sediment types. Compared to the other two units, the fluvial flood deposits exhibit the greatest variability in parameters such as median grain size, D90, and clay/silt/sand content (Table 3, Figure 3). This suggests that both the sediment sources and hydrodynamic conditions of the fluvial deposits likely varied considerably across different events or stages within the same flooding episode. In contrast, the mountain flood deposits show the smallest range of variation in these parameters, while the lacustrine deposits generally fall in between, with silt content exhibiting the least variability among all.

On the other hand, the mean values of grain size characteristics and related parameters differ among the various depositional units (Table 3, Figure 3). Fluvial flood deposits are characterized by larger median grain sizes and higher D90 values, accompanied by lower clay content and higher silt and sand content. In contrast, mountain flood deposits exhibit smaller median grain sizes and lower D90 values, with higher clay content and reduced silt and sand proportions. Lacustrine deposits generally display median grain sizes and D90 values intermediate between those of the fluvial and mountain flood deposits.

Notably, grain size characteristics also vary between flood and lake layers from different periods (Figure 3). For lacustrine deposits, Layer L2 exhibits a smaller median grain size and higher clay and sand contents compared to Layer L1. Among the fluvial flood layers, F1 has the lowest median grain size; F2 shows the highest silt content, lowest clay and sand contents, and a moderate median grain size; F3 contains the lowest silt but the highest sand content and median grain size; while grain composition of F4 is similar to F1 but with a larger median grain size.

3.2. Dating Results

3.2.1. OSL Signal Characteristics

Figure 4 shows the decay and growth curves for representative sample YC-OSL-Z1. The OSL signal decays rapidly to background levels within 1 s, indicating that fast components dominate the signal. This confirms the appropriateness of the SAR protocol [38]. The growth curve fits an exponential saturation function well, showing no evidence of saturation.

Preheating is a crucial step in determining the De values in OSL dating, as it helps remove unstable OSL signal components within quartz minerals. Before applying the single-aliquot regenerative dose (SAR) protocol to measure De values, a preheat plateau test is performed to identify the optimal preheat temperature that minimizes the influence of unstable signals. The procedure involves testing natural aliquots of each sample using the SAR protocol (Table 2), with three discs per group, each group subjected to a different preheat temperature. Preheat temperatures ranged from 160 °C to 280 °C, at 20 °C intervals. For each temperature, the average De values of the three aliquots was calculated to assess dose stability across the preheat range. For sample YC-OSL-Z1, a De plateau was observed between 200 °C and 260 °C (Figure 5), with recycling ratios between 0.9 and 1.1 and recuperation below 5%. However, the 200 °C group showed greater variability, indicating that the range of 220~260 °C is more appropriate. Consequently, 220 °C was selected as the standard preheat temperature for this study.

To evaluate the accuracy of the measurement conditions, dose recovery tests were performed on samples YC-OSL-Z1 and YC-OSL-Z10. These samples were first bleached under solar simulation for 1 h to reset their luminescence signals, then given known radiation doses and subsequently measured. The recovered De values matched the administered doses within ±10% (

Table 4. Results of dose recovery experiments.

Sample No.	Given Dose (Gy)	Measured De (Gy)
YC-OSL-Z1	3	3.03 ± 0.24
YC-OSL-Z1	6	6.04 ± 0.36
YC-OSL-Z10	20	19.94 ± 0.81
YC-OSL-Z10	40	41.61 ± 1.69

), demonstrating that the testing conditions were appropriate.

3.2.2. OSL Age Distribution

For each sample from the YC section, only aliquots with a recycling ratio between 0.9 and 1.1 and recuperation rates below 5% were used to calculate the De values. This selection criterion ensures accurate sensitivity correction and minimizes the influence of thermal transfer during measurements, thereby enhancing the reliability of the resulting OSL ages. The distribution of OSL ages for quartz grains of different size fractions is shown in Figure 6. All of the fine-grained quartz samples and the coarse-grained sample YC-OSL-Z9c exhibit relatively narrow and concentrated age ranges, approximating normal distributions. In contrast, coarse-grained samples YC-OSL-Z5c and YC-OSL-Z10c, as well as the medium-grained sample YC-OSL-Z6m, display wider age ranges and more dispersed distributions, each featuring a distinct secondary peak beyond the primary age peak.

3.2.3. OSL Age Results

The OSL age distribution of fine-grained quartz samples from the fluvio-lacustrine sequence at the YC section is relatively concentrated and nearly symmetrical (Figure 6), so the Common Age Model (CAM) [46] was applied for age calculation. In contrast, most medium- and coarse-grained samples exhibit more dispersed De distributions with noticeable asymmetry and distinct secondary peaks, suggesting the presence of incompletely bleached grains. To avoid influence, the Minimum Age Model (MAM) [46] was also applied to the medium- and coarse-grained samples alongside the Common Age Model. The De values, dose rate, and resulting OSL ages for all samples are presented in

Table 5. Equivalent doses (De), annual doses, and OSL ages of OSL samples.

Sample No.	Depth (m)	U (ppm)	Th (ppm)	K (%)	De (Gy)		Dose Rate (Gy/ka)	Water Content (%)	Age(ka)		OD	Grain Size (μm)
					Common Age Model	Minimum Age Model			Common Age Model	Minimum Age Model	(%)
YC-OSL-Z1	6.7	4.56	16.8	2.26	5.03 ± 0.04		4.45 ± 0.28	23 ± 5	1.1 ± 0.1		0.69	4~11
YC-OSL-Z2	7.0	2.26	15.4	2.40	12.23 ± 0.19		3.84 ± 0.22	23 ± 5	3.2 ± 0.2		2.04	4~11
YC-OSL-Z3	7.4	2.09	12.09	2.02	15.32 ± 0.25		3.26 ± 0.18	22 ± 5	4.7 ± 0.3		3.27	4~11
YC-OSL-Z5f	8.3	1.82	8.50	1.73	13.94 ± 0.17		2.66 ± 0.15	22 ± 5	5.3 ± 0.3		3.76	4~11
YC-OSL-Z5c	8.3	1.82	8.50	1.73	17.17 ± 0.27	12.84 ± 1.74 (σ = 14)	2.29 ± 0.10	22 ± 5	7.5 ± 0.3	5.6 ± 0.8 (σ = 14)	31.31	65~90
						11.19 ± 1.49 (σ = 10)				4.9 ± 0.7 (σ = 10)
						10.43 ± 1.00 (σ = 5)				4.6 ± 0.5 (σ = 5)
YC-OSL-Z6f	9.14	2.32	10.70	1.57	14.28 ± 0.16		2.84 ± 0.17	22 ± 5	5.0 ± 0.3		0	4~11
YC-OSL-Z6m	9.14	2.32	10.70	1.57	13.36 ± 0.14	12.45 ± 0.99 (σ = 14)	2.55 ± 0.11	22 ± 5	5.2 ± 0.2	4.9 ± 0.4 (σ = 14)	16.65	41~65
						11.68 ± 0.82 (σ = 10)				4.6 ± 0.4 (σ = 10)
						11.06 ± 0.57 (σ = 5)				4.3 ± 0.3 (σ = 5)
YC-OSL-Z7	9.4	2.21	13.8	1.70	21.22 ± 0.21		2.92 ± 0.18	30 ± 5	7.3 ± 0.4		0	4~11
YC-OSL-Z8	9.6	2.30	12.0	1.86	27.63 ± 0.34		3.16 ± 0.19	22 ± 5	8.8 ± 0.5		0	4~11
YC-OSL-Z9f	9.8	1.82	11.91	1.77	30.20 ± 0.31		2.96 ± 0.17	22 ± 5	10.2 ± 0.6		0	4~11
YC-OSL-Z9c	9.8	1.82	11.91	1.77	26.58 ± 0.21	26.54 ± 1.23 (σ = 14)	2.49 ± 0.10	22 ± 5	10.7 ± 0.5	10.7 ± 0.7 (σ = 14)	3.80	90~125
						26.54 ± 0.68 (σ = 10)				10.7 ± 0.5 (σ = 10)		90~125
						26.54 ± 0.66 (σ = 5)				10.7 ± 0.5 (σ = 5)		90~125
YC-OSL-Z10f	10.6	1.92	10.2	1.72	30.67 ± 0.31		2.85 ± 0.17	20 ± 5	10.8 ± 0.6		0	4~11
YC-OSL-Z10c	10.6	1.92	10.2	1.72	29.77 ± 0.30	29.77 ± 0.83 (σ = 14)	2.41 ± 0.10	20 ± 5	12.4 ± 0.6	12.4 ± 0.9 (σ = 14)	12.69	90~125
						27.17 ± 1.80 (σ = 10)				11.3 ± 0.9 (σ = 10)
						25.33 ± 1.23 (σ = 5)				10.5 ± 0.7 (σ = 5)

.

When applying the MAM, it is necessary to specify a parameter, σ_b, which represents the unexplained overdispersion in the De values beyond known measurement uncertainties. This parameter is typically set to the overdispersion (OD) value observed in well-bleached samples from a single sediment source. Previous studies report average and minimum OD values of 14% and 10%, respectively, for well-bleached fluvial sediments measured in small or medium aliquots [47]. In this study, the well-bleached coarse-grained sample YC-OSL-Z9c exhibited a notably low OD of 3.80 ± 1.10%. To assess the sensitivity of MAM age calculations to σ_b, we therefore applied values of 14%, 10%, and 5% when calculating MAM ages for the medium- and coarse-grained samples.

The results (Table 5) indicate that when the selected σ_b is smaller than the sample’s OD value, the MAM ages are lower than the CAM ages. Conversely, when σ_b exceeds the sample OD, the MAM ages approach the CAM ages but exhibit larger uncertainties. For the coarse-grained samples YC-OSL-Z5c and YC-OSL-Z10c, as well as the medium-grained sample YC-OSL-Z6m, decreasing σ_b consistently results in lower De values. In contrast, for the coarse-grained sample YC-OSL-Z9c, reducing σ_b has minimal effect on the De value, though it does reduce its uncertainty. This difference arises because, for samples YC-OSL-Z5c, YC-OSL-Z10c, and YC-OSL-Z6m, at least two of the tested σ_b values are smaller than the actual OD, while for sample YC-OSL-Z9c, all tested σ_b values exceed the sample’s OD, thus having little influence on the modeled De.

In summary, the CAM and MAM age results for sample YC-OSL-Z9c are consistent, whereas the CAM ages for samples YC-OSL-Z5c, YC-OSL-Z10c, and YC-OSL-Z6c are significantly larger than their corresponding MAM ages. The OSL ages of fine-grained samples increase with burial depth (Figure 6), aligning with stratigraphic expectations within error margins. When applying the MAM to medium- and coarse-grained samples using σ_b = 10%, the resulting ages generally agree well with those of the fine-grained samples, also within the margin of error.

3.2.4. ¹⁴C Age Results

¹⁴C ages were obtained for samples collected from the top of the lacustrine layers using AMS. The results are presented in Figure 6 and

Table 6. AMS ¹⁴C dating of lacustrine sediments of the YC section in the Daluze area.

Lab No.	Sample ID	Depth (cm)	Material	14C Age (BP)	2σ Calibrated Age (cal BP)	Mean Cal Age (cal BP)
LZU14171	YC-14C-1	926~929	Silty clay	4350 ± 20	4855~4966	4910.5
LZU14170	YC-14C-2	660	Organic matter	1070 ± 20	927~997	962
LZU14170-1	YC-14C-2-2	660	Silty clay	1145 ± 25	963~1080	1021.5

. The ¹⁴C ages of the organic matter sample (YC-14C-2) and the silty clay sample (YC-14C-2-2) from the same depth are consistent, and the ¹⁴C dating data from different layer match the stratigraphy, indicating that the ¹⁴C age results of each sample in this section are relatively reliable.

4. Discussion

4.1. Comparison of ¹⁴C and OSL Ages of Lacustrine Sediments

Both organic matter and sediments from lacustrine deposits were dated by using the ¹⁴C dating method. A potential issue with ¹⁴C dating for organic matter is whether the samples can represent the age of the stratum they are in [2]. In this paper, sample YC-14C-2 was taken from a thin organic matter layer at the top of Lacustrine Layer 2, which primarily consists of mainly of gray–black silty clay. This layer is well-preserved, undisturbed, and exhibits distinct depositional contact with the overlying fluvial layer (Figure 2d). It is inferred that a large number of organisms were rapidly buried due to external events (e.g., river flooding), leading to the formation of an in situ organic matter layer likely reflecting the local depositional age. The main challenge for ¹⁴C dating of sediments is the influence of the carbon reservoir effect, potentially causing age overestimation [3]. In this paper, the consistency between the ¹⁴C ages of the organic matter sample (YC-14C-2) and the clay sample (YC-14C-2-2) from the same depth suggests that the reservoir effect can be ignored. It should be noted that all ¹⁴C samples in this paper were collected from the top of lacustrine sediment layers rapidly covered by fluvial sediments, so the ¹⁴C ages can represent the age of ancient lake extinction well.

The OSL dating of lacustrine sediments in this study primarily utilized fine-grained quartz samples. The OSL age distributions for these samples are relatively concentrated, exhibiting low OD values, with the highest being only 2.04%, indicating a relatively uniform degree of bleaching. Previous studies have demonstrated that fine quartz grains suspended in water tend to have lower residual doses [16], suggesting that fine particles suspended and slowly deposited in lacustrine environments are generally well bleached.

OSL and ¹⁴C ages of lacustrine sediments have be compared. The OSL age of sample YC-OSL-Z1 from Lacustrine Layer 2 is 1.1 ± 0.1 ka, which agrees well with ¹⁴C ages from both organic matter (YC-14C-2) and silty clay (YC-14C-2-2) from the same stratigraphic horizon. The OSL age of the sample YC-OSL-Z7 from lacustrine layer 1 is 7.3 ± 0.4 ka, while the ¹⁴C age of the sample YC-14C-1 from the top of the layer is 4855~4966 cal a BP. The OSL ages of YC-OSL-Z7 and YC-OSL-Z8 suggest a slow sedimentation rate of approximately 0.13 m/ka, and using this rate, the OSL ages are consistent with the ¹⁴C results.

4.2. OSL Signal Bleaching Efficiency of Quartz in Different Grain Sizes of River Sediments

Studies have shown that well-bleached samples typically exhibit concentrated De distributions, whereas partially bleached samples display broader and more dispersed De distributions [8]. Thus, the spread of De values can be used to assess whether the luminescence signal was fully reset prior to burial.

The OD value serves as a useful indicator of the spread of De values [46] and, to some extent, reflects the uniformity of bleaching [8]. The OD values for fine-grained samples are all less than 3.76%, whereas those for medium- and coarse-grained samples range from 3.80% to 31.31%, significantly higher than the fine-grained counterparts. Although small aliquots used for medium and coarse grains contain fewer quartz grains than the large aliquots used for fine grains—resulting in more dispersed De values due to reduced averaging [47]—the observed variation in OD among samples still indicates incomplete bleaching of medium- and coarse-grained grains. Specifically, coarse-grained samples YC-OSL-Z5c and YC-OSL-Z10c, along with medium-grained sample YC-OSL-Z6m, exhibit OD values ranging from 12% to 32%, exceeding OD values typical for well-bleached fluvial sediments [47], suggesting partial bleaching. In contrast, the De distribution for coarse-grained sample YC-OSL-Z9c is normally distributed with an OD of only 3.8%, indicating effective bleaching.

While De distribution and OD values are useful for assessing the degree of bleaching, they can also be influenced by other factors such as microdosimetry effects, variations in luminescence sensitivity, and experimental uncertainties [48,49]. Therefore, additional indicators are necessary for a comprehensive assessment of bleaching completeness. Previous studies have demonstrated that, for well-bleached grains, the natural OSL signal intensity normalized by the test dose (L_n/T_n) should theoretically show no correlation with the De value. In contrast, incompletely bleached grains retain residual OSL signals at deposition, resulting in elevated L_n/T_n values that positively correlate with De values [48,50]. Consequently, examining the relationship between De and L_n/T_n values offers an alternative and complementary method for evaluating bleaching completeness.

Figure 7 illustrates the relationship between De and L_n/T_n value for the medium- and coarse-grained samples from the YC section. Samples YC-OSL-Z5c, YC-OSL-Z10c, and YC-OSL-Z6m exhibit a clear positive correlation between De and L_n/T_n value, indicating incomplete bleaching. In contrast, YC-OSL-Z9c shows no such correlation, suggesting that its grains were well bleached. Overall, the samples with dispersed OSL age distributions and high OD values (Figure 6) also display positive De–L_n/T_n correlations (Figure 7), supporting the conclusion that the broad De distributions are primarily due to incomplete bleaching.

The results of this study indicate that the coarse-grained sample YC-OSL-Z9c experienced sufficient bleaching. Grain size analysis shows that this sample has the lowest median grain size among all fluvial OSL samples (Figure 3), suggesting weak hydrodynamic conditions during deposition [51]. Previous research has demonstrated that under such low-energy conditions, quartz grains are more likely to be well bleached due to reduced turbidity, which enhances sunlight exposure [8,52]. Therefore, the effective bleaching observed in YC-OSL-Z9c is likely a result of these low-energy flow conditions and clearer water, which facilitated greater solar exposure during transport and deposition.

In this study, the OSL signals of fine-grained quartz from the fluvial sediments were found to be relatively well bleached, whereas most medium- and coarse-grained quartz samples showed signs of partial bleaching prior to burial. Several factors may explain this difference: (1) Fine-grained sediments likely originate from more distant sources, resulting in longer transport times that allow greater exposure to sunlight and more complete bleaching. In contrast, medium- and coarse-grained particles tend to settle more rapidly in the water column, shortening their transport duration and reducing their opportunity for sufficient solar exposure, thus leading to incomplete bleaching during fluvial transport [53,54]. (2) Fine particles can remain suspended in the water column for extended periods, increasing their sunlight exposure and promoting more effective bleaching. Conversely, coarser grains settle more quickly on the riverbed. Under low-energy hydrodynamic conditions, water turbidity is lower, allowing more sunlight to penetrate to the riverbed and enabling even coarser grains to bleach reasonably well. However, during stronger hydrodynamic events, increased suspended sediment load raises turbidity, reducing sunlight penetration and limiting the bleaching of coarser particles transported along or near the bed [5,20].

4.3. Differences in Quartz OSL Bleaching Efficiency of Fluvial Sediments Across Different Periods

According to the OSL dating results (Figure 6 and Figure 8), fluvial sediments between 660 and 1085 cm depth (before 1 ka) in the YC section exhibit well-bleached fine-grained quartz. Although medium- and coarse-grained quartz often show signs of partial bleaching, applying the MAM to small aliquots produces ages consistent with those from fine-grained samples, indicating that well-bleached grains can still be isolated. However, in the upper part of the section (above 660 cm, corresponding to sediments deposited after 1 ka, Figure 8), despite coarse grains exhibiting better bleaching than fine grains, MAM results from small aliquots still show significant overestimation—some samples even yield ages older than those from ~5 ka layers [22]. This suggests incomplete bleaching of the OSL signal in sediments deposited after 1 ka.

The temporal difference in bleaching efficiency can be largely attributed to changes in sediment sources. During the early to middle Holocene (e.g., around 4~5 ka), the Yellow River flowed through the Daluze region via the Shan-Jing and Yugong Rivers [55], as supported by sedimentary evidence from the HSC-01 section (Figure 1) [26]. Consequently, fluvial sediments deposited during this period were likely sourced from the Yellow River. Fine grains experienced long-distance transport, allowing for thorough bleaching, while coarse grains may have included more locally derived sediments, with shorter transport and therefore poorer bleaching. After 1 ka, the Yellow River ceased to flow through this area. Instead, frequent flooding from rivers in the Haihe Basin [56,57] supplied abundant proximal sediments. Moreover, extensive deforestation in the Taihang Mountains since the Tang and especially the Northern Song Dynasties led to severe soil erosion, as indicated by increased tree pollen in Daluze sediments [31] and the occurrence of mountain flood layers in the section (Figure 2c). The presence of silty clay layers overlying fluvial flood deposits further supports this interpretation. These sediments likely originated from eroded hillslope soils, where fine quartz particles were aggregated, reducing their sunlight exposure and bleaching efficiency. Overall, sediments deposited after 1 ka were predominantly short-transport, proximal deposits from the Haihe Basin and mountain soils, resulting in limited bleaching for both coarse and fine grains.

On the other hand, differences in bleaching efficiency across periods are also closely linked to hydrodynamic conditions. Grain size characteristics (Figure 3) show that in the YC section, fluvial sediment layers below 660 cm (units F1 and F2) have lower median grain size, sand content, and D90 values, indicating weaker hydrodynamic energy. Under these conditions, lower water turbidity would allow both coarse and fine particles to receive sufficient sunlight exposure, thereby enhancing bleaching efficiency. In contrast, fluvial layers above 660 cm (units F3 and F4) exhibit higher median grain size, sand content, and D90 values, reflecting stronger hydrodynamic forces. These conditions are generally accompanied by increased turbidity, which limits sunlight penetration and reduces the bleaching effectiveness for both fine-grained and coarse-grained quartz.

Beyond the current study, similar observations were reported by [9] for Holocene sediments in the Yangtze River Delta. It was concluded that sediments deposited before approximately 2 ka experienced more stable environmental conditions, weaker hydrodynamic forces, and better bleaching of OSL signals. However, since around 2 ka, stronger hydrodynamics and increased water turbidity have led to poorer bleaching conditions. These findings suggest that fluvial sedimentation processes in the eastern Chinese plains underwent significant changes around 1~2 ka. On one hand, Holocene climatic records indicate an increasingly arid climate since about 1~2 ka, with most lakes in the plains shrinking or disappearing during the late Holocene [58]. The loss of these lacustrine buffers diminished their regulatory effect on surface runoff, thus increasing river flow energy. On the other hand, archaeological and historical evidence shows that human activity intensified markedly during this period [59], especially through extensive deforestation, which led to enhanced soil erosion and the reworking of older sedimentary materials by river systems.

4.4. Implication of OSL Dating for Fluvial Sediments in Plain Areas

Based on the results from the fluvio-lacustrine sediments in the Daluze area of Hebei, this study recommends that OSL dating of fluvial sediments should preferably be carried out on sections where fluvial layers are bounded above and below by stable sedimentary units, such as paleosols or lacustrine deposits. Sampling multiple stratigraphic levels for dating is advised. Fluvial deposits enclosed between stable sedimentary units are easier to identify [60,61], and their age results are less likely to be influenced by later disturbances such as bioturbation or hydraulic reworking. Additionally, these layers, which are rich in organic matter, can be dated using radiocarbon (¹⁴C) methods, allowing cross-verification with OSL ages [62] and providing stronger chronological control for the fluvial units [10].

Under weaker hydrodynamic conditions and with more distal sediment sources, fine-grained quartz is often well bleached. Although some medium- and coarse- grained quartz may be incompletely bleached, reliable age estimates can still be obtained using the MAM. In this study, applying a σ_b value of 10% for small aliquots of partially bleached medium- and coarse-grained quartz produced OSL ages generally consistent with those from fine-grained samples. Considering uncertainties, the typical OD for medium- and coarse-grained quartz is also around 10% [63].

It is important to note that previous studies in both the lower Yellow River Daluze region [22] and the Yangtze River Delta [9] have found that incomplete bleaching is common across all quartz grain sizes in fluvial sediments deposited since 1~2 ka, often resulting in overestimated OSL ages. Therefore, when dating late Holocene fluvial sediments in the eastern plains in China, the potential for age overestimation due to incomplete bleaching must be carefully considered. Specifically, fine-grained quartz sometimes yields older OSL ages than coarse grains because of the erosion and re-deposition of older soil layers from surrounding mountains during strong water flow events, leading to significant age overestimation, as much as 10 ka [22]. In contrast, coarse-grained quartz analyzed with small aliquots or single-grain using MAM tends to produce more reliable ages, with typical overestimates after 1 ka [9,22].

Overall, the bleaching behavior of quartz grains in plain-area fluvial sediments is highly complex. Spatial and temporal variations in sediment provenance, hydrodynamic conditions, and depositional processes result in significant differences in bleaching completeness across different grain sizes. Therefore, it is essential to use different-grain-size quartz to obtain dating results and discuss bleaching completeness based on both the De distribution and the relationship between De values and L_n/T_n values.

5. Conclusions

(1)

Modern soil layer, mountain flood layer, fluvial flood layer, and lacustrine layer in the YC section has been identified based on field observations of lithological characteristics and sedimentary structures, as well as laboratory grain size analysis.

(2)

For the fluvial sediments deposited in the early and middle Holocene, fine-grained quartz exhibits concentrated De distributions, whereas most medium-grained and coarse-grained quartz shows more dispersed De values with significant positive correlations with L_n/T_n, indicating incomplete bleaching. By applying the MAM to small aliquots, the OSL ages of medium- and coarse-grained samples align well with those of fine grains.

(3)

The OSL ages of fluvial sediments deposited before 1 ka in the YC section are relatively reliable, attributed to better bleaching of finer grains. This contrasts with fluvial sediments deposited after 1 ka in the region, which commonly show age overestimation due to incomplete bleaching, despite better bleaching efficiency in coarser grains. The variation in bleaching efficiency suggests differences in sediment source distances and depositional processes around 1 ka ago. Based on the results, for fluvial sediments in the eastern plain of China, it is recommended to strengthen the comparison of OSL dating results for different grain-size quartz, combined with appropriate analytical models.

(4)

Transitions between lacustrine and fluvial sediments have occurred in the Daluze region during the Holocene. The main periods of fluvial deposition were concentrated in three intervals: 10.8~10.2 ka, 5.3~4.7 ka, and after 1 ka.