A total of 464 samples were selected at ∼10–20 cm intervals for analyses of grain size, which was measured using a Malvern Mastersizer 2000 laser particle size analyser (Malvern Instruments, United Kingdom) at the Qingdao Institute of Marine Geology, China Geological Survey, Qingdao, China. The organic matter and biogenic carbonate fractions in the samples were removed prior to analysis. Grain-size parameters were determined following Folk and Ward (1957), and the sediment types were described with reference to the classification scheme of Shepard (1954).

Foraminifera identifications were conducted to assess palaeoenvironmental conditions. In this study, 326 samples were collected from core YRD-1401 at an interval of ∼20 cm. These samples were dried at 60°C and then sized through a 63 μm sieve after decanting the organic debris. After sieving, the samples were dried again at 60°C, and the >63 μm fraction of benthic foraminifera was examined under a stereoscopic microscope. The abundances and simple diversity of the foraminifers were calculated with respect to a dry sample size of 50 g. The benthic foraminiferal data of the core have been published elsewhere (Qian et al., 2021).

Eight samples were selected for AMS ¹⁴C dating using suitable materials (including benthic foraminifera, organic carbon, bivalve mollusc shells, and gastropods). Dating was performed at the Beta Analytics Laboratory, Miami, Florida, United States. Radiocarbon ages were corrected for the regional marine reservoir effect (△R = −178 ± 50 yr; Southon et al., 2002) and calibrated using Calib Rev. 7.0.4 (Stuiver and Reimer, 1986). The radiocarbon dates are expressed as calibrated calendar years relative to 1950 CE (cal yr BP; or using “ka” in place of “cal kyr BP”) with one standard deviation (1σ) uncertainty. In addition, uncalibrated ages are expressed as ¹⁴C years BP (¹⁴C yr BP). The measured AMS ¹⁴C ages are listed in Table 1.

Sediments collected for luminescence dating were sampled using opaque plastic cylinders. All OSL samples were prepared under red light in the luminescence dating laboratory of Qinghai Institute of Salt Lakes, Chinese Academy of Sciences, Xining, China.

The middle parts of the sampling cylinders were unexposed and can be used to extract quartz and potassium feldspar (K-feldspar) for determinations of equivalent dose (D_e). To remove carbonates and organic material, the samples were treated with 10% HCl and 30% H₂O₂, respectively. Then, samples were wet-sieved to retrieve 38–63 and 90–125 μm fractions. Heavy liquids with densities of 2.58, 2.62, and 2.75 g/cm³ were used to separate quartz and K-feldspar grains in the 90–125 μm fraction. K-feldspar grains were treated with 10% HF for 10 min to remove the outer layer irradiated by alpha particles. To remove the feldspar and/or alpha-irradiated layer from quartz grains, the 38–63 and 90–125 μm fractions were treated using H₂SiF₆ (35%) and HF (40%), respectively (Lai and Wintle, 2006). Any fluorides created during the HF etching were removed using 10% HCl, and the purity of the quartz extracts was checked by IR stimulation (Duller, 2003). Quartz grains with obvious IR signals were re-etched with H₂SiF₆ or HF to avoid age underestimation (Lai and Bruckner, 2008).

Water content was measured by weighing the sample before and after drying, assuming an uncertainty of ±5% for all of the samples. Contents of U and Th were determined using inductively coupled plasma–mass spectrometry (ICP–MS), whereas K contents were determined using ICP–optical emission spectroscopy (ICP–OES). The a-value for the 38–63 μm fraction was taken as 0.035 ± 0.003 (Lai et al., 2008). In addition, contents of K and Rb of 12.5 ± 0.5% and 400 ± 100 ppm were assumed for K-feldspar dose rates, respectively (Huntley and Baril, 1997).

The quartz and K-feldspar extracts were mounted on the central part (∼0.7 cm diameter) of stainless-steel discs (∼0.97 cm diameter) using silicone oil as the fixing agent. All measurements employed a Risø TL/OSL -DA-20 reader equipped with blue diodes (λ = 470 ± 20 nm) and IR diodes (λ = 830 nm). Laboratory irradiation was performed using ⁹⁰Sr/⁹⁰Y sources mounted within the reader, with a dose rate of 0.1222 Gy/s. Quartz OSL was detected through a U-340 filter and feldspar IRSL through a combination of BG-39 and coring-759 filters.

Quartz OSL dating in this study employed the SAR-SGC method (Lai and Ou, 2013), which combines Single Aliquot Regeneration (SAR) with Standard Growth Curve (SGC) protocols (Murray and Wintle, 2000; Roberts and Duller, 2004; Lai, 2006). The suitability of the SAR procedure for determinations of D_e was checked using preheat plateau and dose recovery tests. The preheat plateau test was performed on sample OSL-3. Preheat temperatures ranged from 220 to 300°C at a 20°C interval for 10 s, with a heating rate of 5°C/s. The cut-heat temperature was kept constant at 220°C for all measurements. Every four aliquots for each temperature were measured and calculated as the mean values of D_e. A plateau was observed for temperatures from 260 to 300°C (Figure 2). In addition, the recycling ratios for different temperatures of sample OSL-3 ranged from 0.95 to 1.03 (lying within the range of 0.9–1.1) (Figure 2). According to these results, a preheat temperature of 260°C and a cut-heat of 220°C were employed to measure the D_e of the quartz fraction.

For each sample, an SGC was built using the SAR procedure from 6 aliquots, and the D_e values of 12 additional aliquots were then obtained using this SGC. The D_e value was calculated using the initial 0.64 s integral of the OSL decay curve minus the last 8 s integral. The OSL decay curves and reconstructed dose-response curves for sample YRD-1401-OSL-3 are presented in Figure 3. The quartz OSL signals decrease quickly during the first second of stimulation, indicating that the signal is dominated by the fast component (Singarayer and Bailey, 2003). The values of D_e and ages for each sample were calculated and are listed in Table 2.

K-feldspar pIRIR dating utilizing the pIRIR₁₇₀ and pIRIR₂₉₀ signals was used on pre-treated luminescence samples for age estimations. The integrated signal for both pIRIR and IRSL₅₀ was calculated from the first 2 s minus the background from the last 40 s. The stimulation time for pIRIR signals used in the current pIRIR SAR protocol was 200 s. The test dose was constant at ∼30% of the total measured dose. The pIRIR₁₇₀ decay curve and reconstructed dose-response curves for sample YRD-1401-OSL-4 are presented in Figure 4. The IRSL 50°C signal decreases more rapidly than that of pIRIR₁₇₀, which suggests that the latter bleaches more slowly.

Vertical variations in grain size, foraminiferal abundance, simple diversity, and relative abundance of the main foraminiferal species are shown in Figures 5, 6. On the basis of lithofacies characteristics and benthic foraminiferal assemblages, the sedimentary succession in core YRD-1401 can be divided into six stratigraphic units, DU 6 to DU 1 from bottom to top (Figure 7).

We established the chronostratigraphic framework of the depositional units in core YRD-1401 using the dating results of quartz OSL, feldspar pIRIR, and AMS ¹⁴C. As mentioned above, radiocarbon and OSL dating have their particular advantages and limitations. In this study, ¹⁴C is regarded as the preferred method for young dates and yields reliable and accurate ages for Holocene sediments. OSL performs well and outperforms ¹⁴C dating for the old (>25 ka) sediments. Considering the saturation levels of quartz signals, ages obtained by quartz OSL dating can be regarded as reliable for pre-Holocene sediments when the D_e value is less than ca. 230 Gy (Lai, 2010). However, only minimum ages can be provided when the quartz signals are greater than ca. 230 Gy. Therefore, for samples that exceed the upper limit of quartz OSL dating, the results derived from pIRIR dating may be more accurate.

For DU 6, eight samples yielded the quartz OSL ages between 159 ± 13 and 118 ± 8 ka, which lie beyond the dating limit (with respect to the above-mentioned D_e threshold value) and are, therefore, regarded as minimum ages of the sediments (Wintle and Murray, 2006; Lai, 2010). To further constrain the timing of DU 6, the pIRIR protocol with an elevated temperature stimulation of 290°C was used. One infinite pIRIR₂₉₀ age [YRD-1401-OSL-15 (f)] of >378 ka (minimum age: 378 ± 18 ka) at 80.50 m was determined from the basal part of DU 6, for which the value of D_e is greater than 1000 Gy. Another pIRIR₂₉₀ age [YRD-1401-OSL-10 (f)] of 236.1 ± 14.9 ka obtained at ∼65.00 m corresponds to MIS 7. Therefore, we suggest that DU 6 was formed during MIS 6 or even older.

Three AMS ¹⁴C dates for DU 5 (47.07–39.87 m) are beyond the age limit of radiocarbon dating (>43.5 kyr BP). Two OSL dates of 105.0 ± 8.0 and 81.0 ± 5.5 ka were obtained from this unit, indicating that it was deposited during MIS 5.

In DU 4, an AMS¹⁴C date for the upper part of DU 4 (at 30.75 m) is beyond the limit of the radiocarbon method, and only an infinite age of 43.5 ka can be given. In addition, the stratigraphic interval of DU 4 was dated between 64.1 ± 5.1 and 63.0 ± 5.2 ka using quartz OSL dating, which corresponds to early MIS 3 considering the age error. These ages were cross-checked using the pIRIR protocol with an elevated temperature stimulation of 170°C. The aliquots of sample YRD-1401-OSL-4 were bleached for 56 h under sunlight (8 h per day for 7 days) to determine the residual dose, which has an important influence on pIRIR dating. After subtraction of the residual dose (46.1 Gy), the resultant pIRIR₁₇₀ age (YRD-1401-OSL-4 (f)) of 73.8 ± 4.9 ka is older than the quartz OSL age (63.0 ± 5.2 ka) at the same depth of 37.30 m. This age inconsistency for the same sample may have been caused by a residual signal that is unbleached or hard to bleach by pIRIR dating, as the natural bleaching process in coastal areas is uncertain (Li et al., 2014; Li et al., 2018; Long et al., 2019). Within age uncertainties, the timing of deposition of DU 4 corresponds to early MIS 3.

An OSL age for YRD-1401-OSL-2 of 47.5 ± 5.9 ka was obtained at 29.20 m, and an AMS ¹⁴C date (>43.5 kyr BP) for the lower part of DU 3 is beyond the age limit of radiocarbon dating, indicating that DU 3 was deposited during late MIS 3.

In DU 2, three samples were selected and dated at 2006, 4228, and 8888 cal yr BP using AMS ¹⁴C dating. An OSL age of 6.3 ± 0.5 ka obtained at 19.4 m, together with ¹⁴C results, constrains this unit to the Holocene. According to the progradation history of the Yellow River delta (Xue, 1993; Saito et al., 2000) and the lithology and stratigraphic position of DU 1, DU 1 is interpreted as deltaic deposits of the modern Yellow River.

The sedimentary sequences recorded in the coastal areas of eastern China vary with respect to sediment supply, stratigraphic completeness, and core locations (Liu et al., 2016). Therefore, the sedimentary evolution of the Bohai Sea is best established by using sedimentary records of multiple cores. Here, we use the results from core YRD-1401, together with results from other cores examined by previous studies, to make a regional comparison of sedimentary strata deposited since ca. 200 ka in the coastal areas of Bohai Sea.

A comparison of the stratigraphic sequence of core YRD-1401 with that of other cores, including LZK06 (Li et al., 2019), BT113 (Chen et al., 2012), YRD-1101 (Liu et al., 2016), Lz908 (Yi et al., 2012), and YRD-1402 (Zhang et al., 2016), are shown in Figure 9. The sedimentary sequences can be divided into four depositional units, designated L4 to L1 in ascending order.

Sedimentary sequences in L4 were deposited prior to 130 ka. Fluvial and tidal-river sedimentation prevailed in Liaodong Bay (LZK06 core) and the western coast of the Bohai Sea (e.g., cores BT113, YRD-1101, and YRD-1401). L4 in Laizhou Bay is dominated by alternations of sediments deposited in tidal-flat and fluvial environments (core Lz908). All of the quartz OSL ages for this unit are beyond the upper limit of this dating method and are presented as minimum ages of the sediments. Most of the pIRIR ages vary between 260 and 130 ka, with IRSL signals in some ages being saturated. Considering one infinite pIRIR₂₉₀ age of >378 ka and another pIRIR₂₉₀ age of 236.1 ± 14.9 ka for core YRD-1401 (Figure 7), it is speculated that fluvial and tidal-river environments prevailed in the Bohai Sea during MIS 6 or earlier.

L3 corresponds to sedimentary sequences that formed between 130 and 60 ka and show wide differences between the compared cores. In cores BT113 and YRD-1101, this unit consists mainly of neritic and fluvial deposits, indicating sea-level fluctuation in the Bohai Sea during MIS 5, and these deposits are consistent with those in the Pearl River delta (Fu et al., 2020). In other cores, this unit is dominated by neritic and tidal-flat deposits, with the OSL ages falling in the 100–60 ka range. The chronology and lithology demonstrate that alternating marine and terrestrial environments on the northern and western coasts of the Bohai Sea and tidal-flat environments on the south coast of the Bohai Sea occurred during early MIS 3–MIS 5.

Unit L2 is dominated by fluvial-facies sediments and is estimated to have formed from ca. 60 to 10 ka. OSL ages constrained the fluvial sediments to mid-MIS 3–MIS 2, including cores BT113, YRD-1101, YRD-1402, and YRD-1401. However, some marine-influenced sediments in cores LZK06 and Lz908 are also observed in cores LZK06 and Lz908 and have OSL ages of ca. 40 ka.

The sedimentary sequences of L1 were deposited during the Holocene and are characterized by neritic and tidal-flat deposits associated with sea-level rise. Dating results of ¹⁴C and OSL methods suggest that this unit was affected by the Holocene transgression in the Bohai Sea (e.g., Yi et al., 2012; Liu et al., 2016). L1 also contains Yellow River sediments deposited since AD 11 in the modern Yellow River delta.

Reconstruction of sea-level change and transgression history over the past 200 ka for the Bohai Sea is challenging because of discrepancies among dating results, and it is still unclear whether the seawater covered the present coastal areas of the Bohai Sea during MIS 3. Previous studies of stratigraphic imprints and chronology have elucidated transgression histories in the coastal to shelf areas of eastern China (Qin and Zhao, 1985; Xu et al., 2011), where three marine units (i.e., M-3, M-2, and M-1) have been identified. It is well established that M-1 was deposited during the Holocene (e.g., Zhang et al., 2016; Li et al., 2019), as constrained by numerous dating results. However, it remains unclear whether the second transgression (M-2) occurred during MIS 3 or MIS 5, despite considerable research carried out on this problem (Li et al., 2019; Gao et al., 2021). Put another way, it is uncertain whether the MIS 3 transgression occurred in the coastal areas of eastern China.

Accurate and consistent chronology is the key to solving this problem. Previous studies have determined age for M-2 of ca. 40 ka using ¹⁴C dating of shells, wood, and benthic foraminifera (e.g., Hanebuth et al., 2006; Shang et al., 2018), corresponding to middle to late MIS 3. However, these dating materials are vulnerable to contamination by new carbon, which would lead to an underestimation of radioactive ages (Pigati et al., 2007). For example, the apparent age of one old sample (beyond the upper limit of ¹⁴C dating) mixed with 1% new carbon will give an age of 38 ka (Aitken, 1990). ¹⁴C ages older than 35 ka should be treated with caution, and additional dating methods are needed to verify (Gao et al., 2019). The M-2 deposits, originally dated as MIS 3 using ¹⁴C dating (e.g., Zhao et al., 1978; Hanebuth et al., 2006; Liu et al., 2009; Wang et al., 2014; Shang et al., 2018), have been re-evaluated to MIS 5 or even older based on U-series and OSL dating results (Wang et al., 2021).

After re-examination of the available geological and geochronological data, Chen et al. (2012) and Gao et al. (2019) concluded that there was no reliable evidence for the deposition of marine sediments during MIS 3 in the Yangtze River delta and the western Bohai Sea, And it was presumed that all OSL ages for M-2 would belong to MIS 5. However, Yao et al. (2010) identified a transgression layer with abundant benthic foraminifera in core Lz908. In a subsequent study, radiocarbon dates and OSL ages constrained this transgression event to MIS 5–3 (Yi et al., 2012). Moreover, a more recent study showed that M-2 corresponds to MIS 3 and did not detect MIS 5 marine deposits in the northern coast of the Bohai Sea (Li et al., 2019). In core YRD-1401, tidal-flat deposits are revealed in DU 4, as inferred from low-abundance and sporadic benthic foraminifera, tidal bedding, and rusty-brown stains. OSL ages of 65.5 ± 3.1 and 63.0 ± 5.2 ka were obtained from this unit in the present study, which correspond to early MIS 3. Considering the saturated dose of 230–150 Gy (i.e., a quartz OSL age of ca. 70 ka) (Lai et al., 2010), it is necessary to further constrain the timing of formation of these tidal-flat deposits using pIRIR dating. A pIRIR₁₇₀ age of 73.8 ± 4.9 ka was obtained for sample YRD-1401-OSL-4, which is older than its quartz OSL age. Given the complex and uncertain bleaching process of IRSL signals in coastal areas, the residual doses of samples from such areas may be large, and therefore the pIRIR₁₇₀ age of 73.8 ± 4.9 ka of sample YRD-1401-OSL-4 may have been slightly overestimated. Therefore, we argue that the quartz OSL age and pIRIR₁₇₀ age of sample YRD-1401-OSL-4 are consisting within error ranges and correspond to early MIS 3 (ca. 60 ka).

It has been suggested that marine transgressions in the Bohai Sea are influenced by multiple factors, including sea-level change, sediment input, and palaeotopography (Liu et al., 2009). The Bohai Sea is separated from the Yellow Sea by the Miaodao Islands Uplift (Figure 1), which has blocked the passage of seawater from the Yellow Sea into the Bohai Sea since its uplift during the late Mesozoic (Yi et al., 2016). However, as a result of subsequent subsidence of the Miaodao Island Uplift, regional seawater has entered the Bohai Sea since the late Pleistocene (Liu et al., 2016). Hence, it is necessary to reconstruct the paleo sea level and establish whether seawater could have flooded the Miaodao Islands Uplift during early MIS 3. Palaeo sea level is commonly represented by the elevation of associated tidal-flat deposits. The Bohai Sea has been in a state of continuous subsidence over the past 6 Ma, at a rate of 0.12–0.50 m/ka (Guo et al., 2007; Liu et al., 2016). Considering that the rate of subsidence has decreased substantially during the late Pleistocene, it is estimated that the amount of tectonic subsidence in the Bohai Sea since early MIS 3 is ∼7.2 m, using the minimum subsidence rate (0.12 m/ka). In addition, the reconstruction of the palaeo sea level could be affected by sediment compaction (Qin et al., 2020). In this study, the sediments in YRD-1the 401 core are composed of silty sand, which has a limited effect on compaction. Therefore, the thickness of compaction is assessed at ∼1 m. The tidal-flat deposits at 37.64–30.71 m in core YRD-1401 correspond to early MIS 3. The sea level during early MIS 3 is estimated to have ranged from 26.8 to 19.9 m below the present sea level in the Bohai Sea when the elevation of the core (+2.64 m), tectonic subsidence (7.2 m), and post-deposition compaction (∼1 m) are taken into account. An analogous high stand of sea level has been reported in other areas, including the Yangtze River and Red River deltas (Hanebuth et al., 2006; Pico et al., 2016; Shang et al., 2018). According to the rate of tectonic subsidence (0.12–0.50 m/ka), the subsidence of the Miaodao Islands Uplift since early MIS 3 (ca. 60 ka) is between 7.2 and ∼30 m. The water depth at the bottom of the southern trough of the Miaodao Islands Uplift and the maximum depth of the Laotieshan channel would have exceeded 20 and 50 m, respectively, which would have enabled the seawater to enter the Bohai Sea via the Miaodao Islands Uplift during early MIS 3.

Furthermore, the southern Bohai Sea has been in a stage of continuous subsidence since the late Quaternary, which would have provided sedimentary accommodation space for marine transgressions. However, the MIS 5 transgression was weak and/or absent in the northern Bohai Sea because the depositional elevation in the northern Bohai Sea may be higher than the sea level (Li et al., 2019; Wang et al., 2020). Moreover, some researchers have ascribed the unexpectedly strong transgression during MIS 3 in the marginal seas off eastern China to continuous tectonic subsidence since MIS 5 (e.g., Yan et al., 2006; Wang et al., 2013). However, our results show that the MIS 3 transgression along the southwestern coast of the Bohai Sea occurred only in the early part of this MIS and that a tidal-flat environment prevailed. The foraminiferal assemblages and abundances (Figure 6) and lithologic characteristics of rusty-brown stains and tidal bedding (Figure 8) show that the marine transgression during MIS 3 was weaker than that during MIS 5, despite the continuing subsidence. This contradiction between the intensity of transgression and sea-level change is likely to have been caused by the lack of a clear and reliable chronological framework rather than uncertainty regarding the amount of tectonic subsidence.

Finally, the basis of the record of marine transgressions is stratigraphic completeness (Yi et al., 2012; Liu et al., 2016). During the late Quaternary, erosion occurred with increasing frequency in the complex sedimentary environments of coastal and nearshore areas. The transportation of sediments in channels and periods of erosions are marked by unconformities in stratigraphic records (Straub et al., 2020). In fact, the high degree of instability of the climate and sea-level fluctuations during the Last Glacial, produced extreme erosion events, including the channelization of rivers into the underlying sediments. During the Last Glacial Maximum, a large number of palaeo river channels developed and incised the coastal to shelf areas, with a sea level of approximately 130 m below the present sea level (Liu et al., 2010). These palaeo channels were filled and/or erased by subsequent transgression, with deposits being only partially preserved in the geological record (Fagherazzi et al., 2008). As a result, the study of sedimentary environments and the associated marine transgression during MIS 3 requires detailed analyses of successive sedimentary facies and a reliable chronological framework.

The sedimentary characteristics and chronostratigraphic framework of core YRD-1401, together with other cores examined by previous studies, enable the sedimentary evolution of the Bohai Sea to be understood from the perspective of sea-level change, sediment supply, and geomorphology over the last 200 ka.

The Miaodao Islands Uplift has existed since the late Mesozoic and has played an important role in the sedimentary evolution in the Bohai Sea (Yi et al., 2016). Several weak transgressions are recorded in the Bohai Sea at ca. 0.83 Ma and ascribed to the tectonic subsidence of the Miaodao Islands Uplift (Liu et al., 2016). The climate was relatively warm and wet during MIS 7, although it was interrupted by a cold substage (MIS 7d) and did not reach the peak of that of MIS 5 and the Holocene (Rowe et al., 2014; Columbu et al., 2019). Subsequently, the global sea level fell by 120–130 m during MIS 6, which was characterized by a cold and dry climate (Antonioli et al., 2004). The deposits of DU 6 suggest that in the study area, a tidal-river environment prevailed from MIS 7 to MIS 6. The rusty-brown stains and carbonaceous spots observed in DU 6 are interpreted as representing subaerial exposure. The sand beds with basal erosional surfaces and cross-bedding are indicative of tidal-channel deposition, as indicated by the relatively low abundance and simple diversity of foraminifera as well as by the foraminiferal assemblages, which are dominated by low-salinity species. In addition, high abundances of H.germanica in the benthic foraminiferal assemblages were observed in core YRD-1401 (57.80–51.35 m), which suggests that strong tidal currents occurred during MIS 7 to MIS 6 (Hayward and Hollis, 1994).

Following MIS 6, the global sea level rose and reached its peak during MIS 5e (Shackleton, 1987), and marine deposits were ubiquitous in the coastal to shelf areas of eastern China. This large-scale transgression in the South Yellow Sea is recorded by Asterorotalia spp. and involved a cold-water mass in the central South Yellow Sea (Wang et al., 1981; Yang et al., 1998). During MIS 5, neritic deposits were common in the Bohai Sea, with OSL ages of 100–90 ka (Liu et al., 2016). This view is further confirmed by the deposits of DU 5. The abundance and simple diversity of foraminifera both reach their highest values in this unit in core YRD-1401, and the assemblages are dominated by shallow-water marine/euryhaline species, which suggests that a strong transgression occurred in the Bohai Sea during MIS 5. However, sedimentary evidence from the Pearl River delta and Yangtze River delta shows that sea level fluctuated and these areas were not covered by seawater continuously during MIS 5 (Zhao et al., 2008; Fu et al., 2020). These sea-level fluctuations are also revealed on western coast of the Bohai Sea, where the sedimentary sequences contain alternations of tidal-flat and neritic deposits and are intercalated with fluvial and tidal-river deposits (Figure 9). It is worth noting that the completeness of sedimentary sequences in some cores is poor, which could be a result of subsequent river erosion and/or scouring by transgressive processes. In addition, some factors including high elevation in palaeotopography and sediment discharge during MIS 5 seem to have been unfavorable for the development of marine deposits (Wang and Tian, 1999; Wang et al., 2008). For example, the marine sedimentary interval of MIS 5 is not observed in Liaodong Bay (Li et al., 2019) due to its high elevation in palaeotopography (Wang et al., 2020).

During MIS 4, the sea level fell to ∼80 to ∼90 m below the present sea level, and rivers developed in coastal areas (Chappell et al., 1996; Liu et al., 2010). With the development of rivers, fluvial processes further eroded late Quaternary strata in coastal to shelf areas (Fagherazzi et al., 2008). The MIS 3 transgression along the southwestern coast of the Bohai Sea occurred in early MIS 3, when a tidal-flat environment developed on the southwestern coast of the Bohai Sea. During middle MIS 3 to MIS 2, the lowered sea level exposed the coastal to shelf areas, and a terrestrial environment prevailed (Zhou et al., 2014; Liu et al., 2016).

A peat layer underlying DU 2 at a depth of 22.62 m yielded an age of 8888 cal yr BP in this study, indicating a prevailing coastal marsh sedimentary environment during 9000–8000 yr BP in the Bohai Sea (Qin et al., 1990; Saito et al., 2000). Later, coastal marshes were quickly inundated by seawater, and the sedimentary environment changed from terrestrial (fluvial or lacustrine) to coastal marine (Qiao et al., 2011). During the maximum Holocene transgression, the western coastline of the Bohai Sea lay 50–90 km west of the present coastline (Liu et al., 2016). The ages (including OSL ages and AMS¹⁴C ages) and foraminiferal assemblages of DU 2 suggest that a marine transgression (i.e., M-1) occurred along the southern Bohai Sea during the Holocene.