Seventeen OSL samples were treated with 10% HCl and 30% H₂O₂ successively to remove carbonates and organic materials. Wet sieving was utilized to obtain coarse-grained fractions (90–125 µm) according to availability. 90–125 µm fractions were treated with 40% HF for ∼40 min to remove feldspar and washed with 10% HCl for about 30 min to remove the fluoride precipitation generated during etching. The purity of quartz fractions was tested by the infrared stimulated luminescence (IRSL) signals.

90–125 µm quartz fractions were mounted on the center (5 mm diameter) of 9.7-mm diameter stainless-steel discs using silicone oil for equivalent dose (D_e) measurements. Irradiation, preheating, and OSL measurements were conducted on a Risø TL/OSL-DA-20 reader equipped with a⁹⁰Sr/⁹⁰Y beta source and blue LEDs (λ = 470 ± 20 nm) (Bøtter-Jensen et al., 1999). All quartz signals were stimulated at 130 °C for 40 s and recorded by an EMI 9235QA photomultiplier tube fitted with a 7.5-mm Hoya U-340 filter. The D_e in this study was measured by the SAR-SGC method (Lai and Ou, 2013), a combination of single aliquot regenerative dose (SAR) protocol (Murray and Wintle, 2000) and standard growth curve (SGC) method (Roberts and Duller, 2004; Lai, 2006). The preheat for natural and regenerative dose signals was 260 °C for 10 s (Wintle and Murray, 2006), and the preheat for test doses response was 220 °C for 10 s. Preheat plateau test and dose recovery test results from this study and neighboring Pearl River Delta showed that preheat temperature at 260°C for 10 s was appropriate for dating (Xu et al., 2020; Lin et al., 2022; Xu et al., 2022; Lin et al., 2023). For samples SY3-2-G01, SY3-2-G02, SY3-2-G03, and those from core SY2-2, 6 aliquots were measured using the SAR protocol, and 12 aliquots were measured for the natural L_N/T_N measurement. Given the saturation of the OSL signal in samples from cores SY3-2 and SY2-2, the D_e value of samples SY3-2-G04, SY3-2-G05, and samples from cores SY3-1, HS02 were determined based solely on the natural L_N/T_N measurement to expedite the measurement process. After eliminating any obvious statistical outliers, the final D_e value for a sample was calculated.

Tests including the preheat plateau, dose recovery, recycling ratio, and recuperation were conducted on sample SY3-2-G01 to examine the suitability of luminescence properties for the SAR protocol (Wintle and Murray, 2006).

The preheat plateau test was conducted with a preheat temperature ranging from 200°C to 300°C with an interval of 20°C for 10 s and cut-heat temperatures kept at 220°C for 10 s, using a heating rate of 5 C/s. Twelve aliquots (two aliquots per preheat temperature) were measured at each temperature point. The results indicate a preheat plateau between 260°C to 280°C (Figure 2A).

The dose recovery test is to examine whether the D_e measurement protocol can recover a known laboratory dose (Wintle and Murray, 2006). Ideally, the measured dose is in agreement with the given laboratory dose (Wintle and Murray, 2006). The dose recovery test was applied to twelve natural aliquots of sample SY3-2-G01. The given laboratory dose is 20.83 Gy. The measured average D_e value at the 260°C preheat temperature was 21.3 Gy, resulting in the ratio of dose recovery of 1.02 (Figure 2B). The results are within 10% of the natural dose, indicating that the SAR protocol can recover a laboratory dose.

Recycling ratio and recuperation tests are mainly examining whether no obvious thermal transfer was present and whether sensitivity changes could be well corrected in the measurement (Wintle and Murray, 2006). Research showed that a reliable D_e value can preferably meet two requirements including recuperation < 5% and the recycling ratio within 0.9–1.1 (Wintle and Murray, 2006). The average recycling ratio for sample SY3-2-G01 at the 260°C preheat temperature was 1.01 (Figure 2C), indicating that the sensitivity changes were well corrected. The recuperation for sample SY3-2-G01 at the 260°C preheat temperature was 1.7% (Figure 2D), suggesting that no obvious thermal transfer was present. Besides, the accepted D_e aliquots for each sample in this study matched the criteria of recycling ratios between 0.9 and 1.1 and recuperation ratios <5%.

Inductively coupled plasma mass spectrometry (ICP-MS) was used for measuring uranium (U) and thorium (Th). Inductively coupled plasma/optical emission spectrometry (ICP/OES) was utilized for determining potassium (K). The cosmic ray dose was calculated depending on the depth, altitude, and geomagnetic latitude of each sample. The moisture content was estimated to be 25% ± 5% for all OSL samples, considering the variation of moisture content within the burial period in the study region. The dose rates and final ages were calculated on the website program DRAC (Durcan et al., 2015).

Representative decay and growth curves of samples SY3-2-G01 and SY2-2-G15 are shown in Figure 3. The decay curves show that OSL intensity decreasing rapidly during the first second of stimulation toward background levels, indicating that the OSL signals are dominated by the fast component in these samples (Wintle and Murray, 2006). The well-fitted growth curves show that the combined SAR–SGC protocol is appropriate for all the samples in this study. Our quartz OSL results from the four cores are listed in Table 1 and can be shown in Figure 4, with ages ranging from 4.9 ± 0.3 ka to 139 ± 28 ka. Quartz OSL signal saturation could be observed in sixteen samples, with D_e exceeding 190 Gy, indicating that the obtained dating results are regarded as minimum ages (Lai, 2010; Murray et al., 2021; Long et al., 2022; Xu et al., 2022).

The D_e is usually saturates at ∼150 Gy, resulting in age underestimation for sediments over 50 ka (Buylaert et al., 2007; Lai, 2010; Timar-Gabor et al., 2011; Lai and Fan, 2014; Chapot et al., 2016). Underestimation of quartz OSL age is common in coastal deposits. OSL and thermally transferred OSL (TT-OSL) signals of quartz samples from the coastal plain of Israel demonstrated that the upper limit of quartz D_e from Nilotic origin is close to 140 Gy (Faershtein et al., 2019). In the western Bohai Sea (China), the D_e values of quartz OSL samples are >200 Gy, resulting in ages saturation at >80 ka (Long et al., 2022). Core HPQK01 in the Pearl River Delta also showed that quartz OSL ages ranging from 125 ± 18 ka to 58 ± 6 ka are considered as minimum ages due to the OSL signal saturation >150 Gy (Xu et al., 2022). Age underestimation of fine quartz (11–44 μm) from the Lianjiang River plain occurred on samples older than ∼130 ka (Tang et al., 2018). Our results suggest that the quartz OSL ages older than ∼60 ka are underestimated as a result of signal saturation (∼180 Gy).

The OSL dating results obtained from core sediments in the neritic area of the Lianjiang River plain suggest that the timing of deposition is between 4.9 ± 0.3 ka and >83.6 ± 6.5 ka, with sedimentation thickness of less than 6.4 m since at least MIS 5 (Figure 4). The sediment thickness in the inner Lianjiang River plain is at least 80 m since MIS 5, which gradually decreases to less than 20 m thick in the outer Lianjiang River plain (Song et al., 2012; Tang et al., 2018) (Figure 5). In the neritic area of the Lianjiang River plain, the Holocene sediments are around the same thickness as those in the outer Lianjiang River plain, less than 4.2 m (Tang et al., 2018) (Figure 5). However, core sediments at depths of 12.9 m, 18 m, and 19.4 m were dated to 1,442 ± 65 cal a BP, 11,712 ± 508 cal a BP, and 9,321 ± 221 cal a BP, respectively, indicating that nearly 20 m thick of the Holocene sedimentation formed in the neritic area between the Chaoshan plain and Nanao Island (Sun et al., 2007). The Bohai Sea and the China Sea contain the Holocene sediments that are more than 12 m thick (Liu et al., 2017; Long et al., 2022), while they are thinner in the SCS, around 10 m (Wei et al., 2015; Wang et al., 2020). The neritic area of the Lianjiang River plain has comparatively poor conditions for sedimentary preservation. The preservation of the late Quaternary sediments in the Lianjiang River plain is still debated. Some studies attributed it to tectonic activities, while others suggested the corresponding of the sea-level variations (Song et al., 2012; Tang et al., 2018).

The South China Block comprises four NE-trending faults, namely, the Littoral, Changle-Nan’ao, Zhenghe-Dabu, and Shaowu-Heyuan-Yangjiang faults (Figure 6) (Sun et al., 2014; Wang et al., 2014). Of these faults, the Littoral and Changle-Nan’ao faults are active since the Quaternary, and have a significant impact on the Liangjiang River plain and its neritic area (Xu et al., 2010; Wang et al., 2014). These faults have influenced sedimentary evolution and paleo-depositional environments, resulting in accumulation and transportation of the Quaternary along the faults (Sun et al., 2007).

However, the southeastern neritic area of the Lianjiang River plain is under intense interaction of the SCS and Lianjiang River. Global sea level declined since the end of the last interglacial period and reached the maximum at ca. 130 m bmsl during the last glacial maximum (LGM; ca. 30–20 ka) (Hanebuth et al., 2006; Hodgson et al., 2006; Lambeck et al., 2014; Spratt and Lisiecki, 2016). Low-stand sea level led to the incision of the Lianjiang River and the exposure of the continental shelf in the northern SCS (Wei et al., 2015; Xu et al., 2019). A series of buried paleochannels extending from the southeastern neritic area of the Lianjiang River plain to approximately 25 km offshore were discovered and considered as the paleo-Lianjiang River (Figure 7) (Liu et al., 2005). The cores in this study are located at the southern part of the paleo-channels of the Lianjiang River, where they likely experienced fluvial incision during the last glacial period. Lan et al. (1991) demonstrated that the medium coarse sands in the Taiwan Shoal were mainly transported by the currents from the coastal region of southeastern China during 10–20 ka BP based on the ¹⁴C dating. Sediments from the core sites were largely transported to the Taiwan Shoal by the Hanjiang diluted water and Guangdong Coastal Currents (Figure 1A), resulting in limited sediments in the neritic area of the Lianjiang River plain (Lan et al., 1991; Lian and Li, 2011). Studies have shown that terrigenous sediments can be transported and spread by oceanic currents once they enter continental margins (Liu et al., 2008; Liu Z. et al., 2016; Zhang et al., 2022). For instance, the sediment from the Pearl River is predominantly transported southwestward via coastal currents, while Taiwan-derived sediments are among the principal contributors in the NESCS due to the influence of deep-water currents and surface currents that vary seasonally (Liu Z. et al., 2016; Zhang et al., 2022). Therefore, fluvial incision during the last glacial period and strong transportation by coastal currents may result in poor preservation of sediments since at least 83.6 ka in the study area.

The Chaoshan plain neritic area experienced poor sediment preservation since MIS 5, while other regions experienced transgressions and regressions resulting in more than 30 m thick of deposition. Studies from the Bohai Sea, the western South Yellow Sea, and the East China Sea have revealed evidence of sea-level changes and their impact on sedimentary processes (Liu J. et al., 2016; Liu et al., 2017; Wang et al., 2019; Wang et al., 2020; Long et al., 2022). Further research indicated that sea-level changes were identified as the primary control for sedimentation on the northeastern SCS, with Kuroshio intrusion being responsible for the transport of Taiwan-derived sediment during the late Quaternary (Zhang et al., 2022). In summary, the sedimentary processes in the eastern continental shelf of China were mainly influenced by sea-level changes.