The YS is geographically separated into the Northern Yellow Sea and the SYS by a line joining the Shandong peninsula with Baengnyeong Island in South Korea (Zhang et al., 2018) ( Figure 1 ). The SYS shelf covers an area of 300 × 10³ km² with an average water depth of 44 m (Zhao et al., 2021). Water circulation in the SYS consists of the Yellow Sea Warm Current (YSWC), the Korea Coastal Current (KCC), and the China Coastal Current (CCC) ( Figure 1 ). The northward-flowing YSWC, a branch of the Kuroshio Current (KC) characterized by warm and saline water, feeds into the YS ( Figure 1 ). The YSCWM forms in the bottom waters of the central SYS from late spring to autumn as a consequence of cooling of the YSWC during winter (Song, 2011). South-flowing currents include the CCC along the Chinese coast and the KCC along the Korean coast ( Figure 1 ).

The sediment types in the SYS are influenced by a combination of hydrodynamic features and the locations of river inputs. Muddy sediments have developed in the central SYS and the western part of the Northern Yellow Sea (Shi et al., 2004). Cyclonic eddies and YSCWM have contributed to the formation of mud areas in the central SYS (i.e., at sample sites N3513, N3517, N3609, and N3613); the sources of these sediments have been traced back to the Yellow River and old Yellow River (Zhao et al., 2018) ( Figure 1 ). A mud patch in the Eastern YS (northwest of Jeju Island; N3503) has been attributed to deposition of fine-grained materials from the Kuem River, facilitated by anticyclonic eddies (Shi et al., 2004; Song, 2011) ( Figure 1 ). In the Southern Yellow Sea Mud Area (SYSMA), silt and clay together constitute more than 80% of the total sediment (Zhao et al., 2021). In contrast, high-energy sedimentary environments in the SYS, driven by coastal and tidal currents, exhibit continuous erosion and transportation of fine sediment (Shi et al., 2004). Consequently, the outer regions of the SYSMA (N3509, N3707, N3711) are characterized by a higher proportion of sand-size sediment than of fine sediment ( Figure 1 ).

An SYS sampling survey was conducted aboard R/V Isabu from April 11 to 14, 2022 ( Figure 1 ). The vertical distribution of water temperature, salinity, and dissolved O₂ were measured with a CTD system (Seabird, SBE911plus). Undisturbed sediment samples were collected using intact acrylic cores (length = 25 cm, inner diameter = 6.5 cm) for analysis geochemistry, and incubation cores (length = 14 cm, inner diameter = 10 cm) for TOU measurement. The sediment was subsampled using a box corer (0.25 m²).

The total sediment O₂ uptake rate was measured using ex situ incubation (Lee et al., 2019). The incubation chamber core was carefully inserted into the sediment without disturbing the surface. The sediment core was immediately closed with a gas-tight lid, leaving an internal water height of ~7 cm. The chamber core was submerged in a circulated water bath (EYELA, NTT-2100, Tokyo, Japan), with a constant in situ bottom water temperature, using an immersion cooler (EYELA, ECS-0, Tokyo, Japan), for a limited time (< 2 h). The O₂ concentration in the chamber was measured at 10-s intervals with an inline O₂ optode (Pyroscience GmbH, OXFTC2, Aachen, Germany). The optode sensors were calibrated using two-point calibration (i.e., zero concentration and air saturation of in situ bottom water). After the TOU measurement, the vertical O₂ profile across the SWI was measured using an O₂ microsensor (Glud, 2008). The O₂ microsensor (Unisense, OX-25, Aarhus, Denmark) (tip diameter ~ 25 μm and low stirring sensitivity <1%) provided readings at 100 μm vertical intervals, with stirring of the overlying water. The output current of the microsensor was amplified by a picoammeter (Unisense, PA-3000, Aarhus, Denmark) and stored on a laptop computer via an analog-to-digital converter. Two-point calibration of the O₂ microsensor was performed between 100% air saturation and zero concentration in situ bottom water, respectively. Water samples for BNFs estimation were collected at the start and end times of the TOU measurements. The collected BNFs samples were immediately filtered (0.45 μm, ADVENTEC, Tokyo, Japan) and stored in a deep freezer (< −60°C).

To determine grain size, the sediment samples from 0–2 cm depth were oxidized and disassembled with 10% H₂O₂ and 0.1 N HCl. The gravel, sand, and silt in the sample were separated using the wet sieving method, and the weights of the fractions were measured. The fine fraction of the samples (< 62.5 μm) was determined using a Mastersizer 2000 laser diffraction analyzer (Malvern Instruments Ltd, Malvern, UK). The OC and TN content of the surface sediments were determined using a CHN analyzer (CE instrument, EA 1110, Milan, Italy) after removing CaCO₃ with 10% HCl. The δ¹³C in the surface sediment was measured using an isotope ratio mass spectrometer (Isoprime, GV Instruments, Manchester, UK) interfaced with an elemental analyzer. The analytical precision for δ¹³C was ± 0.09‰, estimated by IAEA standards (CH-3 for carbon). The chlorophyll-a (Chl-a) in the surface sediment (0–2 cm) was extracted using 90% acetone for 24 hours and measured using a spectrophotometer (Optizen POP, KLAB, Seoul, Korea) (Parsons et al., 1984). The water samples in the incubation core for the sum of nitrate (NO₂ ^– + NO₃ ^–), NH₄ ⁺, PO₄ ^3–, and Si(OH)₄ were analyzed using a nutrient autoanalyzer (QUAATRO, Seal Analytical, Wrexham, UK). Certified reference material (Lot. CP; Kanso Technos Co., Ltd, Osaka, Japan) was included in each batch of nutrient samples and was used to ensure the accuracy of the samples. Reproducibility was generally within ± 5%.

The TOU and nutrient flux, via the SWI, were estimated from a least squares linear regression using the concentration gradients, with incubation time and height of the benthic chamber as follows:

where F is the TOU or the BNF (mmol m^–2 d^–1), dC/dt is the slope of the linear regression of time and O₂ concentration, or the slope between the initial and final nutrient concentrations, V is the chamber volume (m³), and A is the chamber area (m²).

Based on the O₂ vertical profiles, the upper diffusive boundary layer and the surface sediment position were determined and used to estimate the DOU (Jørgensen and Revsbech, 1985). DOU was calculated from the linear O₂ slope (decreasing) relative to the upper diffusive boundary layer (Glud, 2008):

where DOU is the diffusive O₂ flux (mmol m⁻² d⁻¹), D₀ is the O₂ molecular diffusion coefficient in seawater at a given salinity and temperature (Ramsing and Gundersen, from the table for seawater and gases, Unisense A/S) (cm⁻² s⁻¹), and dC/dZ is the linear O₂ slope in the diffusive boundary layer (mmol cm⁻⁴).

The sampling sites had water depths ranging from 30 to 85 m, with N3503 being the shallowest ( Figure 2 ; Table 1 ). Apart from site N3517, salinity remained consistent with depth, and surface water temperatures were generally higher than bottom temperatures ( Figure 2 ). The site nearest to the old Yellow River, N3517, has the lowest surface layer salinity among the sites ( Table 1 ). The surface salinities and water temperatures at N3503 and N3707, which are near the Han River and Yeongsan River, were relatively lower than at other stations, likely as a result of terrestrial water inflow ( Figure 2 ; Table 1 ). Overall, salinities and temperatures in both the surface and bottom waters increased from the Korean coastal waters toward the SYS, indicating freshwater discharge from rivers ( Figure 2 ; Table 1 ). Dissolved O₂ concentrations ranged from 241 to 317 μmol L^–1, with surface waters having higher O₂ concentrations than bottom waters ( Table 1 ). Dissimilar to the patterns observed for salinity and temperature, O₂ concentrations decreased from east to west ( Table 1 ).

The grain sizes in the surface sediments varied from clay to gravel. Gravel content was 3% and was only found at site N3707 ( Table 2 ). The sand content ranged from 2 to 86%, with the highest fraction observed at N3707 ( Table 2 ). The silt content ranged from 8 to 76%, and the clay content ranged from 3 to 40%; N3503 had the highest silt content and N3613 had the highest clay content ( Table 2 ). The median grain size ranged from 2.5 to 7.8 Ø, with finer-grained sediment at N3613 and coarser-grained sediment at N3707 ( Table 2 ).

The OC and TN contents in the surface sediment (0–2 cm) ranged from 0.28 to 1.58% and from 0.03 to 0.22%, respectively. The highest values were in the fine sediment at N3613 and the lowest in coarse sediment at N3707 ( Table 2 ). The carbon to nitrogen (C/N) molar ratios in the surface sediments ranged from 7 to 11, with the highest at N3707 ( Table 2 ). The Chl-a inventory in the surface sediment showed values from 30 to 123 mg m^–2, with the highest at N3517 and the lowest at N3711 ( Table 2 ). The stable isotope δ¹³C values were all between –23.81 and –22.23‰, showing insignificant spatial variation ( Table 2 ).

The O₂ concentration in the incubation chamber gradually decreased over time ( Figure 3 ). The TOU ranged from 11.9 ± 0.02 to 20.5 ± 0.03 mmol m^–2 d^–1, with N3613 exhibiting approximately 2 times higher TOU than N3707 ( Table 3 ). Vertical profiles of O₂ (n = 3) in the pore water are shown in Figure 4 . Unfortunately, we could not measure the vertical O₂ distribution at N3711 because the microsensor failed after collisions with coarse particles or shell debris. The O₂ penetration depth (OPD) varied from 0.27 ± 0.02 cm at N3613 to 0.93 ± 0.11 cm at N3509 ( Table 3 ). The DOU ranged from 9.7 ± 2.47 to 19.4 ± 2.15 mmol m^–2 d^–1, with the highest value at N3613 and the lowest at N3509 ( Table 3 ).

The BNF results are also in Table 3 . During the incubation, only silicate was released from the sediment ( Figure 5 ). The BNFs ranged from –0.56 to 0.87 mmol m^–2 d^–1 for nitrate, –0.12 to 0.68 mmol m^–2 d^–1 for NH₄ ⁺, –0.05 to 0.07 mmol m^–2 d^–1 for PO₄ ^3–, and 0.26 to 6.32 mmol m^–2 d^–1 for Si(OH)₄ ( Table 3 ).

TOU values were significantly correlated with the surface sediment’s median grain size (r = 0.73, p < 0.05) and OC contents (r = 0.78, p < 0.05) ( Figure 6 ). Our findings indicate that OC content may act as a control factor, in combination with grain size, for benthic OC oxidation in the YS, which is consistent with previous studies (Upton et al., 1993; Giles et al., 2007). In the SYS, sediment grain size can vary depending on local hydrodynamic characteristics (Shi et al., 2004). Zhao et al. (2021) suggested that the weak current energy and cyclonic cold eddies in the central SYS contribute to the formation of the SYSMA. In contrast, the muddy patch in the Eastern part of the SYS is formed by gradual deposition of suspended particles driven by anticyclonic circulation (Shi et al., 2004). Given that OC is preferentially adsorbed onto clay surfaces, the OC content in YS sediment could be influenced by the presence of fine sediments (Bock and Mayer, 2000; Ma et al., 2018; Guo et al., 2021). Consequently, the highest TOU was observed in fine sediment with high OC content (N3613), while coarse sediment exhibited the lowest TOU and OC content (N3707). Therefore, spatial variations in sediment OC oxidation in the SYS may be attributed to the distribution of OC across sediment types.

Although TOU, in this study, varied with grain size and OC content, the spatial differences in the SYS were not significant ( Tables 2 , 3 ). The TOU values in SYS are comparable to those measured in temperate shelf areas with similar water depth but lower than those in temperate shelf areas with higher OC contents (1.45–3.45%). Conversely, TOUs in this study were relatively higher than those in polar regions with refractory OC composition (C/N: 7.1–17.3) and lower temperatures (–1.7 to 4.0°C) ( Table 4 ). The sedimentary OC oxidation was estimated using the Redfield ratio of 138O₂:106C, and the total sediment OC oxidation rate (TC_ox) ranged from 9.1 to 15.8 mmol C m^–2 d^–1, which is higher than the values reported in previous studies in the YS, which ranged from 1.7 to 8.8 mmol C m^–2 d^–1 (Song et al., 2016; Zhao et al., 2018).

In this study, PP in the euphotic zone of the SYS ranged from 11.8 to 199.3 mmol C m^–2 d^–1 (Noh et al., 2024, unpublished). Assuming that all pelagic PP was remineralized through sediment OC oxidation, the TC_ox accounts for from 5 to 87% of PP in the water column in the SYS ( Table 5 ). Except for N3503, TC_ox/PP ranged from 5 to 26%, which is comparable with values reported for the YS (8%; Song et al., 2016) and East China Sea (17%; Kim et al., 2020). In contrast, the higher TC_ox/PP at N3503 (87%) may be due to the low PP, likely caused by high turbidity (Song, 2011). Indeed, at N3503, Brightness Transmission (64.5–64.8%) and the Beam Attenuation Coefficient (1.73–1.75m^–1) in the surface water (0–5 m) were 1.3 times lower and 2.6 times higher, respectively, compared to the average values of the other stations (84.6% and 0.67 m^–1).

The δ¹³C and C/N ratios in sediments are commonly used as proxies for determining the origin and/or quality of OC ( Figures 1 , 7 ; Table 4 ). For example, marine origin OC shows a higher δ¹³C ratio (–21 to –19‰) and lower C/N ratio (~ 7) than terrestrial OC (δ¹³C: –26 to –24‰; C/N: > 10). Any mixture of marine and terrestrial-derived OC can have values between these two endmembers (Yao et al., 2015; Chen et al., 2021). Furthermore, the normalized TOU relative to the C/N ratio can be used to distinguish the remineralization states of OC. Higher values may indicate oxidation of fresh OC, whereas lower values may suggest the oxidation of refractory OC ( Figure 7 ).

Previous studies have suggested that the properties of OC in the surface sediment of the SYS are influenced by contributions from both marine and terrestrial sources (Zhao et al., 2018; Liu X. et al., 2020; Chen et al., 2021). The lowest value of TOU/(C/N) was observed at station N3707, which had a relatively high contribution of terrestrial OC. In contrast, other stations, with C/N ratios from 7 to 8 and δ¹³C values ranging from –22.96 to –22.23‰, indicated a greater fraction of marine OC, resulting in higher TOU/(C/N) ratios, approximately twice those at N3707 (Chen et al., 2021).

Terrestrial OC undergoes pre-aging before reaching an aquatic system, making it more refractory than marine OC (Arndt et al., 2013; Zhao et al., 2018). Burdige (2007) reported that terrestrial OC is more likely to be preferentially buried, while marine OC tends to be remineralized in marine sediment. Indeed, there is a distinct difference between the burial efficiencies of terrestrial OC (68%) and marine OC (7%) in YS sediments (Zhao et al., 2021). Consequently, increased terrestrial OC in sediments can reduce the efficiency of OC oxidation. These factors, which influence the fate of sediment OC, are based on its quality and are particularly significant in the SYS, where the complex distribution of OC is influenced by interactions between various OC sources and hydrodynamic characteristics.

The nutrients accumulated in pore water through benthic OC remineralization are exchanged across the SWI, potentially serving as a nutrient source for primary producers in the water column (Mortimer et al., 1999; Hopkinson et al., 2001; An et al., 2019; Guo et al., 2020; Yang et al., 2021). Therefore, estimating nutrient flux is essential for evaluating benthic-pelagic coupling. The BNFs measured in this study were within the ranges reported in most of the related literature ( Table 4 ). The fluxes of nitrate, NH₄ ⁺, and PO₄ ^3– showed transport in either direction across the SWI, at different stations, indicating that these nutrients were either released from the sediment or absorbed into it.

During sediment OC oxidation, dissimilatory nitrate reduction to ammonia can increase NH₄ ⁺ while decreasing nitrate (Nunnally et al., 2013; Yang et al., 2021). However, our results did not show simultaneous efflux of NH₄ ⁺ and nitrate influx ( Figure 5 ). Additionally, NH₄ ⁺ and nitrate fluxes were not correlated with TOU (Spearman, r = 0.35, p = 0.40), and the average OPD was 0.61 ± 0.21 cm ( Table 3 ). These findings indicate that during spring the SYS sediment was oxic, favoring nitrification (Song et al., 2021). Janssen et al. (2005) demonstrated that in highly permeable, coarse-grained sediment, greater oxygen penetration depth promotes ammonium nitrification, leading to relatively lower ammonium release compared to fine-grained sediments. Under oxic conditions, nitrification can decrease NH₄ ⁺ and increase nitrate in the upper sediment layer, promoting NH₄ ⁺ influx and nitrate efflux (Rowe and Phoel, 1992; Denis and Grenz, 2003; Rysgaard et al., 2004; Louis et al., 2021). Zhou et al. (2022) suggested that a combination of nitrification and denitrification processes may control the NH₄ ⁺ and nitrate fluxes in the SYS. Therefore, the relative contributions of nitrification and denitrification processes, for OC oxidation, may affect NH₄ ⁺ and nitrate fluxes at the SWI.

The benthic PO₄ ^3– flux was not correlated with TOU (Spearman, r = 0.53, p = 0.18). Therefore, the low efflux and uptake of PO₄ ^3– might be attributed to removal processes exceeding regeneration of inorganic nutrients via benthic OC oxidation. The decrease in PO₄ ^3– flux is likely a consequence of adsorption onto Fe-oxides in the oxidative environment of the samples (Liu J. et al., 2020; Yang et al., 2021). An et al. (2019) reported that the release of PO₄ ^3– from iron-bound inorganic P (Fe-P) results from the biotic and/or abiotic reduction of Fe(III) under anaerobic conditions. In addition, Louis et al. (2021) reported that low porosity can reduce the thickness of the oxic layer in sediment, which can enhance the solubilization of Fe(III)-bound phosphorus, thereby increasing phosphorus flux across the sediment-water interface. Unlike dissolved inorganic nitrogen (DIN) and phosphate (DIP), Si(OH)₄ was released from sediment at all stations ( Table 3 ). Previous studies have shown that silicate efflux is related to the biogenic silica concentration in the sediment and benthic remineralization (Conley et al., 1989; Denis and Grenz, 2003; Miatta and Snelgrove, 2021; Yang et al., 2021). Although dissolved inorganic silicate (DSi) flux was not correlated with Chl-a (Spearman, r = 0.10, p = 0.82), it was significantly correlated with TOU (Spearman, r = 0.78, p < 0.05). These results imply that SYS sediments could potentially serve as major sources of Si(OH)₄ with sediment OC degradation (Liu et al., 2016).

Using the Redfield ratio (C:N:P = 106:16:1, Redfield et al., 1963; Conley et al., 1989), we estimated the contribution of BNF to the nutrient requirements for PP in the water column. The levels of DIN and DIP were, respectively, 1–33% and 2–14% of the nutrients required for PP ( Table 5 ). These findings suggest that the significance of BNF in supplying N and P from sedimentary OC remineralization in the SYS during spring, may vary among the sediment types. In addition, the BNFs in this study were comparable with, or less than, those in other shelf seas with similar water depths ( Table 4 ). Moreover, the benthic-pelagic coupling in the SYS during spring appears to be less significant than in other shelf areas (Denis and Grenz, 2003; Kim et al., 2020; Yang et al., 2021).