Two acoustic Doppler current profilers (ADCPs) were moored in the inner (M1: 35.1881°N, 128.5861°E) and outer (M2: 35.1657°N, 128.5947°E) parts of the main channel of Masan Bay ( Figure 1A ). The mean water depths at M1 and M2 were 13.7 and 17.1 m, respectively, and the distance between them was 2.9 km. Up-looking broadband ADCPs at M1 (1200 kHz RDI Workhorse Sentinel) and M2 (1000 kHz RDI Sentinel V20) installed in a trawl-resistant bottom mount were configured to measure the profiles of the three-dimensional current velocities, water level, temperature, and echo intensity every minute and 10 minutes, respectively, from September 22 to October 13, 2017. The data for M1 was adjusted to the same time resolution as M2 in post processing. The ADCP transducers were positioned 0.4 m above the bed (mab), and the vertical bin size was set to 0.2 m (M1) and 0.25 m (M2). In post-processing, the coordinates of the current velocity at M1 and M2 were rotated +1°C and −31°C, respectively, to obtain the along- and across-channel components (u and v in Figure 1 ).

An along-channel axial survey was conducted with a research vessel on October 12, 2017. The castings of a CTD (SBE-19plus V2, SeaBird) equipped with an optical backscatter sensor (OBS, Campbell) were conducted to measure the turbidity profile (in NTU). Water samples were taken using a Niskin water sampler at each CTD station to convert the turbidity to the suspended sediment concentration (SSC, mg l⁻¹). All water samples were vacuum filtered through pre-weighed glass fiber filters (GF/F, pore size: 0.7 μm; Whatman). The residues on filters were oven-dried at 105°CC for 24 h. The masses of the filtered sediments were determined by the difference in the re-weight of the filters divided by the filtered water volume. The OBS calibration showed good linear regression between turbidity and SSC (r ² = 0.80).

The echo intensities (E, counts) recorded from the ADCPs were calibrated to the SSC_ADCP (mg l⁻¹) through a comparison with simultaneous field-collected SSC. Calibration was performed using the following sonar Equations 1, 2, and 3 suggested by Deines (1999):

where S _v is the volume scattering strength (dB), SSC_ADCP (hereafter referred to as SSC) and SSC_OBS are ADCP- and OBS-derived SSC (mg l⁻¹), respectively. R is the range to scatterers along the beam (m), α is the sound attenuation coefficient by the water parcel. L_DBM is the 10·log₁₀ of the transmit pulse length (0.02), and P _DBW is the 10·log₁₀ of the transmit power (23.8). E _r is the reference echo intensity (40 counts). K _c (0.55) and C (−64.78), the signal calibration coefficient, were determined by linear regression between the echo intensity (E−E _r) and SSC_OBS (Deines, 1999; Kim et al., 2004; Ha et al., 2011; Seo et al., 2022).

Wind data were obtained at the Masan Tide Station (35.2100°N, 128.5889°E) maintained by the Korea Hydrographic and Oceanographic Agency ( Figure 1 ).

The instantaneous flux (F, kg m⁻² s⁻¹) of suspended sediments at the ADCP mooring stations were calculated as follows (Equation 4):

where u is the along-channel current velocity, and z is the vertical coordinate for each bin measured by ADCPs. F was decomposed into u = u _res + u _tide and SSC = SSC_res+ SSC_tide. A subscripts res and tide denote a tidal average and fluctuation of a variable, respectively, resulting from the application of 36-h low- and high-pass Butter-worth filters (Geyer et al., 2001; Sommerfield and Wong, 2011; Morgan-King and Schoellhamer, 2013; Gong et al., 2023; Seo et al., 2023). The coastal seiches velocity (u _seiche) and seiche-induced SSC (SSC_seiche) were calculated by band-pass filtering the u and SSC with a frequency of 1 h (±30 min).

The tidal and residual mechanisms of sediment flux were identified by decomposing the tidally averaged F ( F ¯ ) as follows (Equation 5):

where u r e s · S S C r e s ¯ is the component of mean-flow flux (F _mean) that is longer than the tidal cycle, and u t i d e · S S C t i d e ¯ is the component of correlation flux (F _corr) that is shorter than the tidal cycle (Geyer et al., 2001; Scully and Friedrichs, 2007; Sommerfield and Wong, 2011; Kim et al., 2024). The F _mean is composed of several processes, such as river discharge, gravitational circulation, tidal straining, and Stokes drift, while F _corr comprises tidal pumping, tidal straining, and coastal seiches (F _seiche = u s e i c h e · S S C s e i c h e ¯ ) processes (Schulz et al., 2018; Figueroa et al., 2020).

The total along-channel sediment flux was calculated by depth-integrated F ¯ , as follows (Equation 6):

where q _total is the depth-integrated sediment flux (kg m⁻¹ s⁻¹) from the bottom to the water surface of the water column, indicating total sediment transport per unit width. The depth-integrated F _mean and F _corr are referred to as q _mean and q _corr, respectively. Here, q _corr includes q _seiche (depth-integrated F _seiche).

The periods of natural harbor resonance within a semi-enclosed bay were calculated using Equation 7 to understand the fluctuation of coastal seiches (Park et al., 2020; Kwon et al., 2021):

where T _n is the resonant harbor oscillation period of the n^th mode for a semi-enclosed bay, h is the averaged depth (ca. 15 m) from head to mouth, l is the length (17 km) of the bay, and g is the gravitational acceleration.

The SSC anomalies caused by coastal seiches were calculated by subtracting 30 min (three samples) moving-averaged SSC from the SSC over the entire measurement period.

The PSD of depth-averaged current velocity for each mooring station was calculated through a Hamming window of four-day segments with half the window overlapping ( Figure 2 ). The estimated PSDs were classified into two main frequency bands: (1) the low frequencies (residual current) in the range of 0−0.027 cph (period T > 36 h) and (2) the tidal frequencies (tidal current) in the range of 0.027−3 cph (T< 36 h). Generally, the PSD at M2 was higher than that at M1 over the entire frequencies. In the tidal frequency band, the semi-diurnal (12 h and 25 min) frequency dominated the diurnal (24 h and 50 min) frequency ( Figure 2A ). In particular, the power spectra were unstable in the range of 0.5−1.5 cph and had selective enhancement. Such broad-banded spectra, presumably defined by the topographic characteristics (i.e., width, depth, and curvature) of the whole area and free long waves propagating, were indicative of the form of coastal seiches. The signal of coastal seiches with a high correlation (~0.8) continued to be detected at M1 and M2 throughout the measurement period ( Figure 2B ). For the total time span, the coastal seiches were dependent on each other in terms of the major oscillation frequencies. Their energies at M1 and M2 were 19.32% and 19.99% of the total (0.01−3 cph), which is not negligible in the microtidal bay. A comparison of the coastal seiches between two mooring stations showed that the PSD at M2 (3.71×10⁻⁴ cm² s⁻² cph⁻¹) was three times higher than that at M1 (1.14×10⁻⁴ cm² s⁻² cph⁻¹) on average ( Figure 2A ).

The longitudinal wave speed where the phase of a wave propagated could be estimated using the 1-h (±30 min) band-pass filtered current data at M1 and M2 ( Figure 3 ). The time lag of the waves propagating between the two stations, approximately 2.9 km apart, was 304.7 s on average. At a given separation distance and time lag, the wave propagated toward the land at a speed of 9.52 m s⁻¹. This is slower than the theoretical shallow water wave speed, 11.59–12.95 m s⁻¹, in this region. It was reasonable to believe that the wavelength of the coastal seiches was 34.27 km (half-wavelength: 17.14 km) ( Figure 1 ). These properties of coastal seiches were also similar to those calculated using Equation 7. The resonance periods for open-end bay in this geometric condition were 94, 31, and 19 min for the first three modes. The resonance period of the first mode (n = 1) was close to the wave periods of the incoming long waves from the open sea. This suggests that the wave oscillations (0.5−1.5 cph) measured at the two mooring stations originated from amplitude perturbations of waves with a node near the mouth of Jinhae Bay.

The current patterns in the total velocity fields had two different vertical structures depending on the relative importance between the tidal (u _tide) and residual (u _res) currents ( Figure 4 ). The u in the spring tides (S _I and S _II) with a tidal range of 1.5–2 m were vertically uniform as a barotropic structure, showing landward or seaward flow with the tidal phase ( Figures 4A, D ). The u _tide fluctuated in the range of −0.025–0.022 m s⁻¹ (−0.053–0.047 m s⁻¹) at M1 (M2) on average and dominated the u _res of 0.012 m s⁻¹ (0.018 m s⁻¹) in the bottom layer. In neap tides (N _I and N _II) with a tidal range of 0.4–1 m, u had two-layered current flows as a baroclinic structure, showing seaward flow in the surface layer and landward flow in the bottom layer ( Figures 4A, D ). The u _tide at M1 and M2 stagnated in the range of −0.012–0.012 m s⁻¹ and −0.026–0.023 m s⁻¹, respectively, while their u _res reached 0.022 m s⁻¹ and 0.044 m s⁻¹ in the bottom layer, on average. As the tidal forcing weakened, two-layered residual circulations (also known as exchange flow) were developed ( Figures 4C, F ), intensifying the vertical shear (s ², du/dz) at approximately 8 mab of M1 and 5 mab of M2 ( Figures 5C, E ).

Unlike u _tide, the signal of coastal seiches with 1-h period appeared persistently over the entire measurement periods ( Figures 3 , 4A, D ). Coastal seiches affected the water column through repetitive back-and-forth horizontal motions along the main channel but had different characteristics depending on the distance from the node. Coastal seiches measured at M1 and M2 had a motion close to the standing wave with phase lags of 66.44°C and 68.56°C between current velocity and water elevation, respectively ( Figures 3B, D ). As they propagated toward shallow onshore from M2 to M1, their velocity weakened while the amplitude of the sea level variation was amplified ( Figures 3A, C ). In practice, the coastal seiches at M2 close to the entrance of the bay had a high velocity (u _seiche) fluctuations in the range of −0.08–0.08 m s⁻¹ with relatively small fluctuations of sea level (−0.06–0.06 m) ( Figure 6 ). At M1 near the head of the estuary, the u _seiche decreased by 38% compared to M2 and was only in the range of −0.05–0.05 m s⁻¹, while the range of sea level variation increased by 160% to −0.1–0.1 m. Over the entire measurement period, the mean magnitudes of u _seiche were 0.010 m s⁻¹ at M1 and 0.017 m s⁻¹ at M2. The u _seiche were relatively small in magnitude but contributed 20.75% and 31.62% to the total u at M1 and M2, respectively, and 40.72% and 39.42% to the u′ on average ( Supplementary Figure S1 ).

The mobile sediment pool distributed along the main channel was disturbed by near-bed currents, forming the turbidity maxima ( Figure 5F ). In view of variability in SSC, the sediment behaviors at M2 were more dynamic than at M1. For the M1 between ETM and STM, the sediment resuspension occurred mainly in S _I and N _I, with near-bed SSC of 7–51 mg l⁻¹ ( Figure 5B ). However, these resuspension events occurred episodically, and a high SSC (> 51 mg l⁻¹) was frequently detected in the surface rather than the bottom because of the surface dispersion of the turbid plume. The near-bed SSC at M2 was influenced by the STM that developed approximately 1.5 km wide in the central part of the main channel, which was 7–80 mg l⁻¹ with the instantaneous SSC > 100 mg l⁻¹ ( Figures 5D, F ). The SSC had a diurnal pattern in S _I and N _I depending on the sea breeze ( Figures 5A, D ). The bed sediments resuspended during northerly winds generated the STM, and the suspended sediments then temporarily settled down during southerly winds. The near-bed SSC increased gradually to 80 mg l⁻¹ as the sea breeze was repeated in S _I (September 20–24, 2017). The height, where sediments were resuspended in the water column, also increased to reach the water surface. This height was suppressed to < 5 mab by increased s ² when two-layered residual circulation was intensified in N _I ( Figures 5D, E ). The position of STM was moved gradually toward the land because the residual circulation strengthened in N _I and S _II. Sediment resuspension barely occurred, even though the current velocity applied to the bed in S _II was similar to that in S _I. Thus, the SSC in the water column returned to the background SSC (~7 mg l⁻¹) ( Figure 5D ).

A series of resuspension at both M1 and M2 had a marked effect on an increase in the instantaneous flux F ( Figures 7A, D ). During the spring tides (S _I and S _II), F at both stations exhibited landward and seaward periodicity during the flood and ebb tides, respectively. Compared to F (flood: 1.12×10⁻³ kg m⁻² s⁻¹; ebb: −0.59×10⁻³ kg m⁻² s⁻¹) at M1 with low sediment supply, the F at M2 influenced by STM reached up to 6.74×10⁻³ kg m⁻² s⁻¹ in the bottom layer during the flood tide and −2.52×10⁻³ kg m⁻² s⁻¹ during the ebb tide. When the STM moved toward the land in S _II, the F at M2 decreased sharply regardless of the tidal cycle and had a low F of 0.30×10⁻³ kg m⁻² s⁻¹ for the wash load. In addition, during the neap tides (N _I and N _II), regardless of the tidal cycle, the F at M1 and M2 were 0.41×10⁻³ kg m⁻² s⁻¹ and 1.23×10⁻³ kg m⁻² s⁻¹ in the bottom layer, respectively, on average. Over the measurement period, the suspended sediments at M1 and M2 were generally transported toward the land with positive q _total under the balance between tidal and residual currents ( Figures 7B, E ). The suspended sediments were controlled mainly by the residual currents owing to the high buoyancy of STM (or ETM), resulting in a q _mean at M1 and M2 of 92% and 94% of q _total, respectively. The cumulative q _corr at M1 and M2 reached up to 150 kg m⁻¹ and 429 kg m⁻¹, respectively, and tended to be slightly transported toward the land when the events of sediment resuspension occurred ( Figures 7C, F ).

The coastal seiches occurred consistently with 1-h (±30 min) interval as a barotropic structure ( Figures 8A, B, G, H ). In accordance with the period of coastal seiches, the sediments concentrated near the bed, and turbid plumes in the bay fluctuated ( Figures 8E, F, K, L ). As shown in Figure 9 , the SSC anomalies were symmetrical in the range of −35–50 mg l⁻¹ (skewness: 7.37) at M1 and −43–46 mg l⁻¹ (skewness: 10.72) at M2, along with a symmetrical coastal seiches velocity. The results suggest a strong correlation between u _seiche and SSC anomaly because of similar histogram features ( Supplementary Figure S2 ), but this did not mean that a larger u _seiche indicated a larger SSC anomaly. For example, when the sediment resuspension (SSC > 40 mg l⁻¹) occurred at M2 on September 22, 2017, the SSC anomaly fluctuated consistently in the range of −30–37 mg l⁻¹ from 04:00 AM to 7:00 PM along with two peaks of resuspension on 06:00 AM and 6:00 PM ( Figure 8F ). In contrast, u _seiche, which was minimal (< 0.01 m s⁻¹) at 07:00 AM, increased rapidly to 0.07 m s⁻¹ at 12:00 AM and remained almost constant without an apparent decrease for 10 h. This result shows that the magnitudes between u _seiche and SSC anomaly were not highly correlated despite varying the same frequency. To clarify the relationship between them, the SSC anomaly did not increase proportionally to the u _seiche when u _seiche over the entire period was rearranged in ascending order ( Figures 9E, F ). At M1 and M2, even weak u _seiche of less than 0.01 m s⁻¹ caused the SSC anomaly of up to 50 mg l⁻¹, whereas even strong u _seiche of 0.08 m s⁻¹ caused an SSC anomaly of only 5 mg l⁻¹. Although the coastal seiches influenced the fluctuations of suspended sediments, they did not contribute predominantly to enhancing the SSC anomaly. The SSC anomaly in the water column depended on whether the bed sediments are eroded or the turbid plumes is spread, by other forcings such as winds and tides.

The sediments resuspended from the bed were influenced repeatedly by the horizontal motions of coastal seiches. At each pulse of coastal seiches at M2 (M1), they were transported toward the land or sea, with F _seiche of 0.040×10⁻³ kg m⁻² s⁻¹ (0.027×10⁻³ kg m⁻² s⁻¹) near the bed, on average. When the STM was developed, F _seiche increased to > 0.3×10⁻³ kg m⁻² s⁻¹ at M1 and to > 0.4×10⁻³ kg m⁻² s⁻¹ at M2. However, the bidirectional F _seiche canceled each other out owing to the characteristics of back-and-forth horizontal motions of coastal seiches, so thus the q _seiche was not heavily biased to either side. Although the cumulative q _seiche at M2 (M1) remained only −36.21 kg m⁻¹ (−17.04 kg m⁻¹), the q _seiche was equivalent to q _mean and q _corr when the resuspension rarely occurred (i.e., S _II). Over the entire period, these q _seiche at both stations accounted for 10% and 1% of the cumulative q _corr and q _total ( Figures 7C, F ).

The coastal seiches in shallow regions are usually accompanied by horizontal currents in the bottom layer and enhance near-bed current velocity (Rabinovich, 2009). The bed sediment particles can be resuspended if the seiche-enhanced current velocity exceeds an erosion threshold (Gloor et al., 1994; Basterretxea et al., 2011). Depending on the local properties, coastal seiches may be one of the causes of sediment resuspension when they are enhanced comparable to tidal and residual currents. As listed in Table 1 , the characteristics of coastal seiches in Jiaozhou Bay (China), Port Kembla Harbor (Australia), and Mediterranean Harbor (Italy), which have a period of several tens of minutes and current velocity of around 0.1 m s⁻¹, were similar to those in Masan Bay except for the energy source (Jordi et al., 2008; Luick and Hinwood, 2008; Yuan et al., 2008). Most played as much of a significant role in sediment resuspension as tide and wind forcing. They kept the bed sediment unconsolidated and made it readily resuspendable in subsequent events, forming a high SSC of several tens of mg l⁻¹ despite the coarser silt and sand composition ( Table 1 ) (Yuan et al., 2008; Grifoll et al., 2019). The coastal seiches in Masan Bay also contributed to enhancing the near-bed current velocity through back-and-forth horizontal motions (M1: −0.05−0.05 m s⁻¹; M2: −0.08−0.08 m s⁻¹), but it was considered not to cause sufficient bed disturbances to resuspend the bed sediments.

The mobile sediment pool with a thick layer of loosely aggregated cohesive sediments was distributed along the main channel of Masan Bay (Geyer, 1993; Lin and Kuo, 2001; Seo et al., 2022). In S _I, it was easily disturbed and resuspended, even under the slow currents (< 0.1 m s⁻¹). Fine-grained sediments on the mobile sediment pool could be supplied to the upper water column by even northerly winds of < 2.8 m s⁻¹ during slack tides, resulting in STM with > 40 mg l⁻¹. In S _II, the consolidated bed exposed by the STM being moved toward the land required a current velocity of at least > 0.18 m s⁻¹ to be resuspended (Seo et al., 2022). Therefore, the sediment bed with different erosion thresholds was distributed depending on the STM position, even though the coastal seiches consistently caused a disturbance of the sediment bed over the entire period. If the coastal seiches caused a sufficient disturbance of the bed, the erodible sediments with erosion threshold less than 0.1 m s⁻¹, such as the mobile sediment pool (Seo et al., 2022), would have been thickly distributed regardless of the STM position. Even if the coastal seiches could not cause sediment resuspension directly, the SSC anomaly by coastal seiches would also have been detected without significant differences between spring tides (S _I and S _II). Considering these conditions, the contribution of coastal seiches to the sediment behavior depended on the position of the STM and was premised on resuspension by external forcings.

The effects of coastal seiches on the resuspension of bed sediments were insignificant in Masan Bay without external forcings, such as tides and winds. As long as the bed sediments were resuspended by them, however, the coastal seiches had a marked effects on SSC anomaly. Once the bed sediments were entrained by the external forcings, they could be immediately diffused toward the direction of decreasing concentrations and maintained in suspension (Yuan et al., 2008) ( Supplementary Figure S3 ). As the diffusion occurred, the spatial distributions of SSCs were expanded from the bed to the surface layers ( Figure 10B ). Consequently, in Masan Bay, the vertical motion of coastal seiches helped enhance the sediment horizontal flux in the surface layer. A weak vertical motion in the range of −0.01–0.01 m s⁻¹ coincided with the back-and-forth horizontal motions at the anti-node at each pulse of coastal seiches. They would have reacted with the tidal currents to partially produce a flow reversal and provide a vertical mixing (Talke and Stacey, 2008; Yuan et al., 2008). Despite relatively low contribution of u _seiche on the vertical mixing compared to u _tide ( Supplementary Figure S4 ), the suspended sediments were lifted to the upper layer where they would not have risen if not for the vertical motion of coastal seiches. As shown in Figures 8 and 10 , in practice, the suspended particles within the STM at M2 fluctuated in accordance with the period of coastal seiches. Repetitive coastal seiches vertically pumped the position of diffusing suspended sediment particles, generating the SSC fluctuations in the sawtooth form at 1-h period. At high SSC of 40 mg l^-1, the amplitudes of its isoline were only within a few tens of centimeter ( Figure 8B ). The lower the SSC, the higher amplitudes of its isoline due to vertical mixing, which had a sufficient diffusion range to cover the entire water column ( Figure 10B ). This lifting effect applied equally to the turbid plume (M1) and near-bed load regardless of the behavior types of suspended sediments ( Figure 10A ). For all behaviors of the suspended sediments over the entire period, the horizontal fluxes for those uplifted above the position of suspended sediments assumed to be free of fluctuations by coastal seiches reached 51×10⁻³ kg m⁻¹ s⁻¹ (standard SSC: 20 mg l⁻¹) at M1 and 83×10⁻³ kg m⁻¹ s⁻¹ (standard SSC: 20 mg l⁻¹) at M2 ( Figure 10 ). Although the sediment diffusion was suppressed by enhanced s ² in the neap tides (N _I and N _II) (Teng et al., 2022; Zhang et al., 2022), the cumulative uplifted fluxes at M1 and M2 over the entire periods were up to 3.05 kg m⁻¹ and 2.77 kg m⁻¹, respectively, regardless of the direction. The upper layer of the water column could be supplied continuously with other sediment particles from the lower layer by the coastal seiches, equivalent to approximately 1% of the cumulative q _seiche at both M1 and M2.

As a feature of the back-and-forth coastal seiches, the cumulative q _seiche accounted for only 1% of the total, which does not appear to contribute much to sediment transport. On the other hand, the tendency, in which suspended sediments were lifted and advected as much as 0.4×10⁻³ kg m⁻² s⁻¹ by a pulse of coastal seiches, suggested that they significantly affected the sediment behaviors. The contribution of coastal seiches to sediment transports in the bay can be evaluated differently for each sedimentary condition because this effect could be enhanced or reduced drastically depending on whether the STM was developed or not. To determine the distance traveled by a water parcel (or sediment particle) during a half period of coastal seiches, regardless of the STM, the excursion length (El _seiche in m) was calculated using following Equation 8:

where v ° and T are the velocity amplitude (m s⁻¹) and periods (s) of coastal seiches, respectively (Johansson, 2010; Cerralbo et al., 2015; Björk et al., 2017; Lafta, 2021). The sediment particles at M2 could travel toward the land or sea as much as El _seiche (< 57 m) for every pulse of coastal seiches unless the u _seiche does not strengthen ( Figure 11 ). In particular, the El _seiche at M1 decreased to < 32 m because the u _seiche gradually slowed down as it moved away from the node, further shortening the sediment transport trajectory. Unlike the tidal (El _tide ~420 m) or residual current that caused even STM movements (Seo et al., 2022), the El _seiche was too short compared to their transportable amounts of suspended sediments, indicating a redistribution process at the immediate vicinity ( Figure 11 ). Given the long wavelength (~35 km) of coastal seiches and the shallow water depth (~15 m), this process was driven primarily by the seiche-induced lifting effects on the suspended sediments than the advection. As more coastal seiches were repeated, the bidirectional coastal seiches could partially offset more current flows, which induced sediment stagnation rather than circulation in the bay. Applying the redistribution processes to the STM, the El _seiche was only approximately 3% of the width (1.5 km) of the STM located in the central part of Masan Bay. Even if the sediment particles within the STM were advected (0.4×10⁻³ kg m⁻² s⁻¹ at M2) by coastal seiches, they could not leave its region. Under the dominant tidal and residual currents, the coastal seiches continued to trap the suspended sediments partially and pump vertically. Considering the sluggish water circulation in Masan Bay (Park et al., 2011; Seo et al., 2022) and the short El _seiche at M1, coastal seiches would cause a redistribution in a narrower range as the STM moves toward the land in N _I and S _II. Unless the coastal seiches lift the suspended sediments to sea level and discharge them out to sea by estuarine circulation, more sediment stagnation will occur in the inner part of the bay.

The effect of persistent coastal seiches on the vertical redistribution of sediments can be maximized when sediment resuspension events occur or turbidity maxima develop. Eventually, this can promote the potential to cause secondary problems, such as pollutant release, an increase in residence time, and a change in inorganic nutrients (Heyes et al., 2004; Basterretxea et al., 2011; Spieckermann et al., 2022). The water exchange with the open sea by the coastal seiches that can induce circulation in the bay must be accompanied to prevent these problems. However, as above, the short El _seiche was only approximately 3% of an almost 2 km long entrance. This allowed a part of the water only to be pumped back and forth within the channel instead of renewing the bay water volume. In other words, the impact of coastal seiches on the water column is limited only within the bay. They would not be sufficient to address the possible problems caused by redistribution despite repetitive fluctuation and be dependent on the tidal and residual currents.