The Huanghe originates from the Qinghai-Tibet Plateau, passes through the Loess Plateau and North China Plain before emptying into the Bohai Sea ( Figure 1B ), with a total length of 5460 km (Xue, 1993; Yu et al., 2013). The high sediment discharge (an average annual sediment discharge of 6.88 × 10⁸ t) of the Huanghe has caused frequent avulsions, leading to continuous shifts in its lower reaches over the past 6000 years, forming 10 deltaic superlobes (Bi et al., 2021). Except one located in the Yellow Sea (1128–1855 CE), all of the superlobes are located in the Bohai Sea, a semi-enclosed marginal sea located off the northeast coast of China with an average water depth of ca. 18 m.

After avulsion at Tongwaxiang, the Huanghe shifted northward into the Bohai Sea, and the MHD began to prograde in 1855 CE, covering an area of ca. 6000 km² (Pang and Si, 1980; Wang and Ye 1990; Yu et al., 2013) ( Figures 1A, B ). Eight delta lobes of the latest present-day active superlobe (the MHD) have formed along the southwestern coast of the Bohai Sea due to frequent avulsions and lobe-switching events (Pang, 1979; Xue, 1994, 2009; Liu et al., 2019) ( Figure 1C ). Among these, the DKL was activated from 1964 to 1976 CE ( Figures 1A, C ); while active accumulation of the QSGL (the present-day active delta lobe) and abandonment/degradation of the DKL were as a result of the latest major avulsion of Huanghe at southeast of DKL in 1976 CE (Chu et al., 2006) ( Figures 1A, C ).

The MHD is located in a temperate monsoon region. Under the control of the East Asian monsoon, the wave direction shifts with the wind direction, primarily southeast in the summer and northeast (strong waves of over 3 m wave height) in the winter (Luo and Liu, 2015; Ren et al., 2022). In addition, the water and sediment discharge of the Huanghe also exhibit pronounced seasonal variations, with the average water and sediment discharge during the flood season (July to November) accounting for 80% of the annual total (Cheng and Cheng, 2000; Fan et al., 2006; Yang et al., 2008).

The coastal zone off the Huanghe mouth is micro-tidal with semidiurnal tidal patterns (Xiong, 2012). The tidal range is small, averaging between 0.22 m and 1.0 m, with velocities of 0.5–1.0 m/s (Zhao et al., 1995; Ye et al., 2023). The southeast flood tidal currents are stronger (velocity up to 2.2 m/s) than the northwest ebb tidal currents (Zhang et al., 1998; Fan et al., 2006). After filtering out tidal and other periodic currents, residual currents persist (Bi et al., 2014), from northwest to southeast, with velocities between 0.1–0.3 m/s, and is the main driving force for sediment transport in the Huanghe (Liu, 1985; Wang, 1989; Wang et al., 2014). Tides, waves, and residual currents together form the longshore transport and/or erosional regimes along the southwest coast of the Bohai Sea (Chen, 2009; Fu et al., 2021) ( Figure 1A ).

There is also a tidal shear front roughly parallel to the coastline near the present-day active river mouth (east of the QSGL) ( Figure 1C ), characterized by low flow velocities and converging flow directions on either side (Wang et al., 2024). Due to the significant differences in hydrodynamic characteristics between the frontal zones (Bi et al., 2011), a substantial amount of sediment (ca. 50%) is trapped by the tidal shear front, with most sediments depositing within 15 m of water depth (Li et al., 2001; Wang et al., 2007). This leads to the formation of a “high platform” (ca. 15 m thick) in the upstream region of the delta front ( Figure 1A ), connected to the steep delta front slope at its seaward end (the middle and lower parts of the delta front), creating geomorphological conditions conducive to the formation of sediments gravity flows (Liu et al., 2022).

To our knowledge, the earliest documentation of cut-and-fill structures, described as moderately disturbed areas just below the seafloor near the QSGL, was provided by Prior et al. (1986) using high-resolution geophysical surveys. Liu et al. (2022) identified both buried and active (exposed on the seafloor) SCs that are widespread off the and the regions between the DKL and QSGL. These features have been attributed to hyperpycnal flows, mass movements, and resuspension-related SGFs. Other studies have also linked the formation of these disturbed areas to processes such as liquefaction, submarine landslides, scouring, and deformation sliding (Xu et al., 2009; Li, 2015; Wang et al., 2020; Yu et al., 2022).

From a geotechnical perspective, SCs are widely recognized as an unfavorable geological condition and a significant concern (Peng et al., 2014). They are considered potential triggers for various subaqueous geohazards, such as submarine slumps, regional landslides, and differential erosion, posing serious threats to marine engineering facilities (Gao, 2011). As a result, regions with SCs are generally avoided when selecting sites for marine engineering structures (Xu et al., 2008; Wen et al., 2018). However, comprehensive analyses of SCs (whether buried or exposed) across the entire Huanghe subaqueous delta front are still limited. Most studies have focused on relatively small areas and lacked high-resolution geophysical data covering a wide extent of the delta front, which has resulted in a gap in understanding the distribution and evolution of SCs in relation to the overall development of the MHD.

The high-resolution seismic data used in this study were obtained from three seismic surveys conducted between 2008 and 2009 CE in the subaqueous areas of the MHD ( Figure 1A ). In 2008 CE, the boomer data were collected using a CSP2200 sub-bottom profiling system (Applied Acoustic Engineering Co., UK) with a source power of 300–500 J; in 2009, the chirp data (first reported) were collected using a Geoacoustics Chirp II system (GeoAcoustics Ltd., UK) with a source frequency range of 2–7 kHz and an Edgetech 3100P system equipped with a 512i towfish (Edgetech Co., USA) with a source frequency range of 0.7–12 kHz. In total, approximately 1500 km of seismic profiles were obtained. The seismic survey tracks covered the southwestern part of the Bohai Sea, with water depths ranging from 5 to 25 m, covering multiple subaqueous delta lobes ( Figures 1B, C ). The Boomer data and Edgetech chirp seismic data used in this study have been partly reported (Liu et al., 2014, 2016, 2022). Focusing on the formation mechanism of SCs, this study further interprets these data in combination with additional data.

All seismic data were processed using Madagascar 2.0 software (http://www.ahay.org/wiki/Main_Page), including swell filtering, automatic gain control, and water column muting. Subsequently, the processed seismic data were interpreted using DecisionSpace™ software (Halliburton Co, US), including the identification and tracking of stratigraphic surfaces and seismic units, and the recognition of SGF-related accumulations and SCs based on the configuration of the reflection and superposition. Following previous studies, the two-way travel time (TWTT) was converted to depth using an average acoustic velocity of 1650 m/s (Liu et al., 2014, 2016).

To understand the lithology characteristics of the SCs, a published core (YD02, 119.35°E, 37.71°N) collected offshore of the QSGL in 2018 CE (Chen et al., 2021) was reviewed in this study. Part of the data was reprocessed for analysis with special emphasis on the lithology of SCs, including raw grain size data (sampled at 3 cm intervals) and core photographs.

Our seismic stratigraphy interpretation, particularly the identification of SGF-related seismic units, is consistent with Liu et al. (2022), with a primary focus on the SCs. We identified a pronounced, sub-horizontal unconformity (the MHD base) at depths 15–25 m below the sea level. This unconformity was formed under a dominant erosional regime after the Huanghe River’s rechanneling to discharge into the southern Yellow Sea between 1128 and 1855 CE (Liu et al., 2016, 2022). The sediments overlaying this surface correspond to the MHD and are divided into units A, B, and C ( Figure 2 ). Unit C, the lowest, features internal reflections that converge seaward, forming a representative Gilbert-type clinoform ( Figure 2 ), corresponding to the older MHD clinoform described by Liu et al. (2022). Together, Units B and C form an apron-like (or wedge-like) accumulation body superimposed on the bottomset of the older MHD clinoform, with their internal reflections downlapping onto the MHD base. These two units correspond to the previously mentioned SGF-related accumulations, with their boundary surface, R2, formed in association with the lobe-switching event in 1976 CE ( Figure 2 ). As with Liu et al. (2022), SCs are characterized by both cut-and-fill structures buried within units A, B, and C, and gullied/jagged seafloor features ( Figure 2 ). The reflective characteristics, morphology, and spatial distribution of these SCs, which are the focus of this study, are described in detail later.

The seismic lines are oblique with the SCs in most cases ( Figure 3 ), and to make the profiles provided in this study resemble along strike or dip direction of the SCs for further discussion, some representative profiles from the dense seismic data were selected to present ( Figures 1 , 3 ). Based on observation of numerous sub-bottom profiles, most SCs are rugged, and have jagged basement, with deeper incisions in the center, as depicted in Figures 4 , 5 , 6 . The longitudinal axes of the SCs appear to be shore-normal given the widths of the SCs are generally greater in the profiles in shore-normal direction than in the that in shore-parallel direction ( Figures 4 – 6 ). These SCs vary considerably in size, with widths ranging from 200 m to 6 km, and depths from 1 to 12 m below the sea level, and extend up to ca. 15 km in the shore-normal direction ( Figures 4 – 6 ). Based on whether or not buried, SCs are divided into two main types:

Exposed/active SCs (jagged/gullied seafloor): Such SCs appear as jagged/gullied seafloor, distinguished from flat surfaces, and are not filled or not fully filled with sediments ( Figure 4 ). They are present in both the abandoned DKL and the active QSGL, with more occur near the DKL.

Buried SCs (subsurface cut-and-fill structures): These SCs are observed in both the abandoned DKL and the active QSGL ( Figure 4 ). They are buried below the seafloor in seismic units A, B, and C, and incising several meters (exceed ca. 10 m) into the underlying seismic units ( Figure 4 ). The parallel/sub-parallel reflections within the units A, B, and C are truncated by such buried SCs ( Figure 4 ). Two types of acoustic reflections of the filling deposits of buried SCs are observed, e.g., acoustically transparent and chaotic reflection ( Figure 5 ). The acoustically transparent fills lack clear stratification and discernible structures, indicating a homogeneous and structureless character, while acoustically chaotic fills suggest heterogeneous and poorly sorted fill sediments (Liu et al., 2013) ( Figure 5 ). Notably, the relief of exposed SCs is lower compared with the jagged basement of buried SCs ( Figure 4 ).

Buried SCs exhibit varying cross-sectional characteristics in the shore-parallel direction, which can be categorized into three types: symmetrical, asymmetrical, and composite SCs ( Figure 6 ).

Symmetrical SCs: Such SCs typically display a U-shaped or V-shaped cross-sectional shape with nearly same slope gradients on both sides ( Figure 6 ). This kind of SCs is less frequently observed and smaller in size, with the maximum width not exceeding 1 km ( Figure 6 ).

Asymmetrical SCs: These SCs feature uneven slopes on both sides, with a tilted surface or irregular depressions at the bottom ( Figure 6 ). Their sizes are comparable to that of the symmetric SCs, with widths up to 1 km and a maximum depth of ca. 6 m below the sea level ( Figure 6 ). Asymmetrical types are less frequently observed among the identified SCs ( Figure 6 ).

Composite SCs: These SCs are primarily the merging of two or more of the previously described types (symmetrical and asymmetrical SCs) of SCs ( Figure 6 ), which account for a large proportion of the observed SCs. They are larger in scale, with a maximum width of ca. 8 km and a maximum depth of ca. 6 m ( Figure 6 ).

Following a detailed analysis of high-resolution seismic profiles, numerous SCs have been identified ( Figure 3 ). These SCs are generally distributed at/near the landward flank of the SGF-related accumulations (upper delta front) and along the delta front slope, which lies between the high platform and the toe of the delta front, up to ca. 10 km away from the shore ( Figures 2 , 3 ).

The SCs distributed area exhibit an elongated, shore-parallel shape, spanning almost the entire subaqueous delta of the DKL and the northern subaqueous delta of the QSGL ( Figure 3A ). It has a length of ca. 70 km in the shore-parallel direction and a width of 2–10 km in the shore-normal direction ( Figure 3A ). The whole range of the SCs distribution in the Huanghe subaqueous delta remains not fully uncovered due to the limited coverage of the existing seismic data. Nevertheless, based on the identified spatial distribution of SCs, it is reasonable to infer that their distribution may extend beyond the seismic data coverage area, potentially including the western part of DKL and the eastern part of QSGL (the eastern part of the present-day active river mouth).

In cross-sections perpendicular to the shoreline, exposed SCs are distributed throughout the study area. Buried SCs are more concentrated, forming clusters, with most located at depths between 5 and 25 m below the sea level ( Figures 4 – 6 ). From the DKL to the QSGL, buried SCs display varying burial depth characteristics. Seismic profiles near the DKL area show that SCs are embedded into nearly the same depth below the seafloor, within unit A (SC1 in Figure 2A ). In contrast, SCs near the QSGL area are embedded below the seafloor into several distinct depths, concentrated within units A, B, and C (SC1, 2, and 3 in Figure 2B ).

Core YD02, which penetrated the buried SCs ( Figure 7 ), provides a valuable opportunity for analyzing the sedimentary characteristics of the fill of the buried SCs.

Poorly sorted, gray-brown silt/fine sand and interlayered organic material characterize the interval of 0.3–2 m in core YD02, producing an upward-fining succession, with sand layers in the uppermost 0.2 m and wavy thin clay/silt bedding in 0.80–0.85 m ( Figure 7B ). Between 2 and 4 m is mostly made up of yellow-brown silt and bioturbation, with many interbedded sand/mud layers. The lower section (4.0–9.23 m) is a comparatively homogeneous, structureless layer of yellow-brown/gray-brown silt ( Figure 7B ). The blue-gray fine sand in this core is between 9.23 and 10 m ( Figure 7B ).

The correlation between the seismic stratigraphy and the core YD02 is illustrated in Figure 7A . The core was retrieved 9 years later that the acquisition of the seismic profile in 2009 CE, during which a significant difference in water depth is produced due to sediment accumulation. Therefore, there is a certain discrepancy between the seafloor and the top of the core YD02 ( Figure 7A ).

The most prominent correlation is at depths of 6.2–6.5 m and 9.10–9.24 m (homogeneous, structureless gray-brown silt) in core YD02, which corresponds to the internal of buried SCs ( Figure 7 ). This composition is consistent with the lithological characteristics observed in numerous SCs worldwide (e.g., Mulder et al., 2003; Correggiar et al., 2005a). The interval of poorly sorted, gray-brown silt/fine sand from 0.3 to 2 m, which is consistent with the lithology of many SGF-dominated sediments at the toes of the delta front (Mulder and Alexander, 2001), suggests that they were deposited by SGFs ( Figure 7B ). The 0.3-2 m section is missing from the seismic data, suggesting that it was deposited between 2008 and 2018 CE. The observed upward-fining trend could potentially be linked to the annual water and sediment regulation scheme implemented since 2002 CE (Chen et al., 2021). During the early stages of water and sediment regulation scheme, the downstream riverbed underwent significant erosion, causing the grain size entering the sea to become coarser. However, as the riverbed’s resistance to erosion strengthened over time, the supply of coarse-grained sediments from the downstream area diminished, leading to a finer sediment deposition into the sea (Wu et al., 2017; Jiang et al., 2018). Additionally, the core YD02 contains a significant amount of conventional deltaic mud accumulations (e.g., Zhou et al., 2016; Hu et al., 2024), and lithology characteristics indicative of the longshore transport and/or erosional regimes reworking (Fan et al., 2013; Chen et al., 2021) ( Figure 7B ), e.g., interbedded sand/mud layers and bioturbation.

We utilizing more extensive and denser seismic data, conclude that the buried SCs found in the Huanghe subaqueous delta are formed by the incision of the seafloor during the downslope transport of SGFs, eventually rapid deposit on the upper delta front/downlapping SGF-related accumulations ( Figure 2 ). Our findings are compared with previous hypotheses that attributed buried SCs (cut-and-fill structures) to submarine geohazards such as sediment liquefaction, slumping, and submarine landslides (e.g., Feng et al., 2004; Wang and Liu, 2016). This section focuses on discussing the seismic data evidence supporting this conclusion:

(1) According to previous studies, the origin of “cut-and-fill structures” is often associated with the collapse depression feature formed after hydrodynamic disturbances, where sediments gradually move downslope and spill over the initial arched failure area; Finally, erosive silt flow tongues slide downslope, forming cut-and-fill structures (Wang et al., 2020). Our seismic data indicate that the buried SCs basement is irregularly jagged ( Figure 4 ), which contrasts with the smooth basement created by slides (Yu et al., 2022). The preserved steep incision of the SCs basement suggests that the internal sediments were rapidly infilled after the incision (Correggiar et al., 2005b) ( Figure 8C ). Based on this, we propose that the buried SCs observed in the MHD are more likely formed by erosion.

(2) The SCs are generally located at/near the upper delta front/the landward flank of the SGF-related accumulations (seismic units A and B) ( Figure 2 ), which is consistent with the position where SGF incise the seafloor (e.g., Zavala et al., 2011). The longitudinal axes of the SCs also appear to be consistent with the direction of downslope transport of the SGF sediment (e.g., Liu et al., 2022) ( Figures 4 – 6 ). Both features indicate a strong correlation between SCs and SGFs.

(3) The SCs driven by SGFs can also explain why the relief at the exposed SCs’ basement is generally smaller than that of buried SCs ( Figure 4 ). A possible explanation is that SCs may still be filled with sediments after the seafloor is incised by SGFs, ultimately reducing the reliefs. Alternatively, the erosive intensity of the SGFs forming exposed SCs may has decreased compared to earlier periods. Since the lobe switched to the QSGL in 1976 CE, collapse-related turbidity flows have frequently occurred in the study area, transporting large masses from the upper delta front to the middle and lower delta front, thereby significantly reducing the delta front slope (Prior et al., 1986; Bi et al., 2021; Liu et al., 2022) ( Figure 9B ), which in turn has reduced the frequency and erosive intensity of these flows. This will be discussed in detail in subsequent sections.

(4) The cut-and-fill characteristics of buried SCs are consistent with buried SCs driven by SGFs in several delta fronts worldwide, including the Yangtze Delta (Feng et al., 2024), the Mississippi Delta (Malonely et al., 2020), the modern Po Delta (Correggiar et al., 2005b), and many other ancient and modern deltas (Dasgupta and Buatois, 2015). Numerous lithological characteristics of cores in previous studies also indicate that the interiors of SCs in the Huanghe subaqueous delta are predominantly composed of homogeneous (acoustically transparent material), high water content clay or silt (e.g., Liu et al., 2014; Zhou et al., 2016) ( Figure 7B ), which is inconsistent with the loose, lower water content sediment characteristics following liquefaction and slumping in previous studies (e.g., Chang and Jia, 2010).

Based on the above evidence and previous studies (Zavala et al., 2006; Xu, 2014), this paper further explained the complete evolution process of the cut-and-fill structures of SCs ( Figure 8 ): Initially, SGFs are generated at the upper slope of the subaqueous delta, with their flow velocity increasing continuously as they move downslope ( Figure 8A ). When the velocity exceeds the erosion threshold, SGFs will incise the seafloor, forming gullied SCs ( Figure 8B ). As these flows continue downslope, they persist in their erosive action, causing the SCs to enlarge along the slope ( Figure 8B ). Eventually, the SGFs attenuate in a pulsating manner, and the sediments they carry gradually deposit, ultimately forming the cut-fill structures of SCs (Xu, 2014) ( Figure 8C ).