The Bohai Sea is a semi–enclosed interior continental shelf of the East China Sea (Figure 1), with an average water depth of 18 m, and it is connected to the Huanghai Sea by the narrow Bohai Strait. Core Lz908 is located onshore near the southern coast of the Bohai Sea (37°09′N, 118°58′E, elevation 6 m). The core was drilled to a depth of 101.3 m with an average recovery rate of 75% in the summer of 2007, by the First Institute of Oceanography, Ministry of Natural Resources of China. The drill site was submerged until the middle of the 20th century.

The upper 54.3 m of the core contains marine and coastal sediments (Figure 2), and has been chosen for paleoenvironmental studies (Yi et al., 2012a; Yi et al., 2012b). Integrated the radiocarbon and luminescence–based age model, the geochronology of core Lz908 was established by synchronously tuning sediment grain–size index to the July insolation at 65°N (Yi et al., 2012a), and the tuning bottom (260 ka) has been well constrained by magnetostratigraphy (Yi et al., 2015). Based on this age model (Figure 2), the core was sampled in 10–cm interval to investigate mineral properties of coastal sediments in the south Bohai Sea.

The CIELAB color space, in which any sediment’s color can be expressed by L*a*b* values, can provide useful stratigraphic information (Ortiz, 2011; Sun et al., 2011; Yi et al., 2016). Measurements of diffuse spectral reflectance (DSR) were collected at 10–cm resolution (373 samples in total) in the State Key Laboratory of Marine Geology, Tongji University of China. To reduce uncertainties of water losses, the samples of core Lz908 were dried, ground, and compacted before the measurement. The instrument is the Minolta CM–700d spectrophotometer (400–700 nm wavelength range; 10 nm resolution; 3 mm spot size).

In order to reduce uncertainties of paleoenvironmental inference using original color data (L*a*b*) in such a dynamic region (Yi et al., 2016), we employ principle component analysis (PCA) to identify the contribution of different mineral assemblages to the downcore color variations. The PCA was calculated using the correlation matrix of the center–weighted derivatives of the DSR data (Ortiz et al., 2004; Ortiz, 2011) from core Lz908.

As suggested in previous studies (Ortiz, 2011; Sun et al., 2011; Yi et al., 2016), the results of X–Ray Diffraction (XRD) of the sediments can provide an independent check for the DSR–based mineral identification. Thus, four representative samples from core Lz908 (15, 25, 35, and 45 m in depth) were measured by the XRD, according to the methods described by Liu et al. (2004) as follows.

The XRD clay mineral study was carried out on the <2 μm fraction, which was separated by conventional Stokes’ settling after the removal of carbonate and organic matter by acetic acid (15%) and hydrogen peroxide (10%), respectively. Clay minerals were then identified by using an X’Pert PRO, PANalytical XRD instrument (40 kV and 40 mA) at the First Institute of Oceanography, Ministry of Natural Resources of China. Identification of clay minerals was made according to the position of the (001) series of basal reflections observed on the XRD diagrams (Moore and Reynolds, 1997).

The sediment grain–size record from the Bohai Sea, namely the GS record of core Lz908, can be used to indicate orbital changes of the EASM in the late Quaternary (Yi et al., 2012a), because the sediment grain–size variation is controlled by river discharge and resuspension process in the tidal zone (Chen et al., 2013; Su et al., 2016). Although the influence of regional sea–level changes on the GS record of core Lz908 cannot be excluded (Yi et al., 2012b), the EASM is the predominant factor controlling sedimentary dynamics in the south Bohai Sea, since ∼80% of total variance of the GS record can be explained by solar insolation (Yi et al., 2018).

For the CF–1 record, mineral properties of coastal sediments in the south Bohai Sea are highlighted. Similar because these clay minerals were mainly from the Luzhong Mountains (Figure 1), changes in the CF–1 may be linked to weathering processes and river discharge, rather than regional sea–level changes. For example, minerals illite and goethite in the Chinese loess plateau and the Huanghe River delta usually indicate a cold and dry climate in the past (e.g., Ji et al., 2007; Niu et al., 2015). In this study, the low–frequency changes of the CF–1 are in–phase correlated with orbital eccentricity (Figure 4), inferring that more illite and goethite were produced during glacial intervals in the Luzhong Mountains. Moreover, regional river inputs usually decrease to a much lower level in winters (Yi et al., 2016), and most of detrital materials are transported to the south Bohai Sea in summers (Yi et al., 2012a), via local rivers, such as the Mihe, Xiaoqinghe and Weihe Rivers. Since that the Asian monsoon is the major factor controlling clay minerals in the monsoon domain, such as in the Chinese loess plateau and the Huanghe River delta (e.g., Sun et al., 2011; Zhao et al., 2017; Chen et al., 2021), the in–phase relation between the CF–1 and solar insolation is expected (Figure 4). Thus, the CF–1 can be linked to the EASM by mineralogical responses. When the EASM strengthens, a warmer and wetter climate would produce less minerals illite and goethite in the study area; but when the EASM weakens, the CF–1 value increase, indicating more minerals illite and goethite in the sediments.

In summary, there are two monsoonal records derived from core Lz908, and the GS and the CF–1 are linked to the EASM by sedimentary dynamics and mineral properties, respectively (Figure 6). In specific, the GS record is correlated with the EASM in the late Quaternary by the processes of sedimentary dynamics and regional rainfall, dominated in precession bands (Figure 5C). By regional weathering and river inputs, the CF–1 record is linked to the EASM through both rainfall and temperature changes, highlighted in eccentricity and precession bands (Figure 5A).

Based on stalagmite records from Hulu, Dongge and Sanbao Caves of China, a clean and dominant precession cycle was observed (e.g., Cheng et al., 2016). In these records, speleothem δ¹⁸O changes can indicate δ¹⁸O variation of local meteoric precipitation and thus characterize the EASM intensity (Yuan et al., 2004; Dykoski et al., 2005; Cheng et al., 2016), which might be largely controlled by the large-scale monsoon circulation and concomitant latitudinal shifts of the monsoon rain belt (Zhang et al., 2021). This relationship leads to the idea that the EASM can be directly driven by solar insolation (Kutzbach, 1981; Zhang et al., 2021), explaining a high similarity in spectral characteristics between the EASM and solar insolation (Jiang and Lang, 2010; Shi et al., 2012; Shi et al., 2019). The linkage involves the shift of the ITCZ location (Wang et al., 2005; Cheng et al., 2009; Jo et al., 2014), which modulates atmospheric circulations (Chao and Chen, 1999) and triggers the onset of Asian monsoon (Chao, 2000). Hence, although the physical significance of stalagmite δ¹⁸O are still hotly debated (Zhang et al., 2019), it is reasonable to speculate that monsoonal moisture is the major factor, since that tropical hydrological cycle varies mainly in precession bands (e.g., Huang et al., 2020) and that summer rainfall can be well constrained by stalagmite δ¹⁸O in the Chinese loess plateau (e.g., Tan et al., 2020).

Loess magnetic susceptibility as a proxy of the EASM is based on pedogenesis and mineral transformation (e.g., Yang et al., 2015), via changing rainfall and temperature in the Chinese loess plateau between glacial and interglacial intervals. The monsoon circulation on glacial–interglacial timescales was likely associated to the continental–scale land–sea thermal contrast (Wallace and Hobb, 1977; Webster et al., 1998; Wang and Ding, 2008). Since not all of the rainfall could be used in pedogenesis and because of the loss of evaporation and surface runoff, changes in magnetic susceptibility may reflect effective precipitation (Yi et al., 2018). Changes in global ice volume are proposed to modulate the thermodynamics of land–sea contrast (An et al., 1990), and it is reasonable to speculate that temperature changes were relatively important, considering that pedogenesis and chemical weathering in the Chinese loess plateau are likely not evident when winter temperature is below 0°C (Hao and Guo, 2001). In addition, solar insolation could also modulate the sea-land thermal contrast and this influence was likely evident during interglacial intervals (e.g., Ma et al., 2017).

For the GS record of core Lz908, regional rainfall is the dominant factor (Yi et al., 2012a), while the influence of tropical cyclones (Yi et al., 2018) and sea levels (Yi et al., 2012b) can be excluded, and the former were likely paced by global temperature changes (e.g., Zhou et al., 2019). Because of this, the GS record is dominated by precession cycles with a significant contribution in eccentricity and obliquity bands (Table 1). For the CF–1, the intensified monsoon (rainfall) increases regional river discharge (Chen et al., 2021), thus introducing precessional variabilities into the sediments. On the other hand, the intensified monsoon (temperature) may also strengthen regional pedogenesis in the Luzhong Mountains (Figure 1), likely resulting in a transit of variabilities of eccentricity and obliquity into the sediments. Because of this, the CF–1 record is dominated by eccentricity cycles with an equivalent component in precession bands (Table 1). Hence, monsoonal signals in the Bohai records may be not singly linked to temperature or rainfall changes like the loess–and the speleothem–based proxies, but demonstrate an integrated forcing of ice–sheet evolution and solar insolation.

Overall, although the uncertainties in monsoonal proxies cannot be excluded in this study, the Bohai records likely fill the gap between the new hypothesis of Cheng et al. (2021) and the diverse periodicities of monsoon proxies, supporting that different archives preferentially record a certain aspect of the EASM. Moreover, the difference between the two Bohai records responding to the EASM might highlight a distinct role of precession and eccentricity cycles, which are worth of further investigation in future.