Journal Search Engine

ISSN : 1229-3431(Print)
ISSN : 2287-3341(Online)

Journal of the Korean Society of Marine Environment and Safety Vol.29 No.6 pp.562-575
DOI : https://doi.org/10.7837/kosomes.2023.29.6.562

Salinity Changes and Bottom Water Particle Exchange Simulations in Response to Sluice Gate Operations at Saemangeum Lake

Seonghwa Park^*, Jonggu Kim^**†, Minsun Kwon^***

^*PhD Candidate, Dept. of Civil & Environmental Engineering, Kunsan National University, Kunsan 54150, Korea
^**Professor, Dept. of Environmental Engineering, Kunsan National University, Kunsan 54150, Korea
^***PhD, Ocean Physics Dept., Land & Ocean Environmental Eng., Suwon 16690, Korea

* First Author : andriack@kunsan.ac.kr, 063-469-1871

^† Corresponding Author : kjg466@kunsan.ac.kr, 063-469-1874

Received July 25, 2023 Review September 13, 2023 Accepted October 27, 2023

Abstract

In an effort to improve water quality, the South Korean government has implemented measures to increase seawater circulation in Saemangeum Lake. We analyzed the effect of increasing the frequency of seawater circulation based on salinity levels and bottom water exchange in the lake, using an environmental fluid dynamics code model. When the sluice gate opening and shutting frequency increased from once to twice per day, the internal water level of Saemangeum Lake increased by up to ~0.7 m. The salinity increased by 2.12 psu near the western breakwater and decreased by 1.18 psu near the freshwater inlet. We analyzed the extent of bottom water exchange using a particle tracing method and observed that the residual rate of particles shallower than 5 m in water depth decreased by 2.52% in Case 2 (opening and shutting twice per day) compared to Case 1 (opening and shutting once per day). This indicates that increasing the frequency of sluice gate opening and shutting would promote enhanced bottom water exchange. Consequently, the increased salinity and bottom water exchange associated with increased seawater circulation are expected to improve water quality in Saemangeum Lake.

Key Words : Saemangeum Lake , Stratification , Water quality , Particle tracing , Salinity change

새만금 배수갑문 운영에 따른 염분 변화와 저층수의 입자교환 모의

박 성화^*, 김 종구^**†, 권 민선^***

^*군산대학교 토목환경공학부 박사과정
^**군산대학교 환경공학과 교수
^***국토해양환경기술단 연구원

초록

새만금 호의 수질 개선을 위하여 국가에서 해수 유통을 증가시킴에 따라 해수 유통 빈도 증가로 인한 새만금 호 내 염분과 저층수 교환 변화를 알아보기 위하여, EFDC(Environmental Fluid Dynamics Code) 모델을 이용하였다. 갑문 개폐 횟수를 하루 1회에서 2회로 증가했을 때, 새만금 호 내부 수위는 최대 약 0.7 m 상승하였다. 염분은 서측 방조제 근방에서 2.12 psu 증가하였으며, 담수 유입 부근에서는 1.18 psu 감소하였다. 입자추적을 이용하여 저층수 교환 정도 분석한 결과, 수심 5m 이하 입자 잔류율은 Case 2(1일 2회 개방)에서 Case 1(1일 1회 개방)에 비해 2.52% 감소한 것으로 나타났다. 이는 수문 개폐 횟수를 증가시켰을 때, 저층수 교환이 더 활발해 질 수 있다는 것을 알 수 있다. 따라서 해수 유통 증가에 따른 염분 및 저층수 교환 증가로 새만금 호의 수질 개선이 될 수 있다고 판단된다.

키워드 : 새만금 호 , 성층 , 수질 , 입자추적 , 염분변화

This article has been cited by 0 article in crossref

Cited-By

Funding:

1. Introduction

Saemangeum Lake is an artificial lake established within the Saemangeum breakwater construction in 2006. Saemangeum is located on the west coast of South Korea, in Jeollabuk-do. Efforts toward desalination have led to a degradation in the water quality of the dam. Currently, seawater is circulated through the lake as part of a water quality improvement policy; as a pilot project, sluice gate operation was increased from once to twice per day since 2021. The Ministry of Environment Republic of Korea (2021) indicated that the increased inflow of seawater could promote the occurrence of a hypoxic water mass within Saemangeum, as observed in various environments in the United States, due to stratification derived mainly from differences in salinity, water temperature, and nutrient salt supply due to freshwater inflow. With the formation of an influent channel for seawater into Yeong Rang Lake, intensive salinity stratification occurred (Kim et al., 2008). This creates a concern that the river channel (≥40 m above sea level) near the sluice gate of Saemangeum could cause intensified stratification.

Stratification analysis is complex for lakes where sea and freshwater enter at a boundary, such as for Saemangeum. In Mobile Bay, AL, USA, where rivers are broad and water depth is low, the formation and dissipation of stratification are considerably impacted by stream flow and wind velocity rather than tidal flux; transitions of a few days to months have been observed from complete mixing to intense stratification depending on the relative scale of stream flow and wind velocity (Schroeder et al., 1990). In contrast, in the Rhine region of freshwater influence (ROFI), the variation in stratification intensity presented a semi-diurnal pattern due to tidal effects (Souza and Simpson, 1997). Wind is an important external force impacting stratification. For example, while wind blowing upstream from the mouth of the York River in the United States weakened stratification, wind blowing downstream intensified stratification (Scully et al., 2005). Density current, which contributes greatly to evaporation in Shark Bay, Australia, was found to play an important role in the formation of tide lines (Nahas et al., 2005).

Current evidence indicates that the occurrence and dissipation of stratification in Saemangeum Lake are governed by intricate interactions among various factors. Pae et al. (2009) reported that the salt concentration and quantitative change in seawater and freshwater inflow will strongly impact stratification. Similarly, Oh et al. (2013) suggested that the most important factors affecting the salinity distribution in Saemangeum Lake were the amounts of seawater circulation and upstream freshwater inflow along with the topographical conditions of the Mangyeong and Dongjin waters. The increased water volume in Saemangeum Lake because of dredging, increased time of bottom layer residence, and considerable increase in water depth could facilitate upstream seawater inflow. Indeed, the Ministry of Environment Republic of Korea (2021) reported pronounced vertical salinity stratification in downstream Saemangeum. This stratification was more expanded in the upstream Dongjin waters than those in the Mangyeong waters, implying that both seawater and freshwater inflow are important drivers.

The formation and dissipation of stratification within Saemangeum, along with changes in salinity and water quality, occur through complex mechanisms, while the long-term effect of seawater circulation in Saemangeum remains unclear. Park et al. (2023) suggested that the change in salinity could impact the euryhaline phytoplankton distribution, while Jeong et al. (2018) reported that the efficient management of water levels could improve water quality. However, no study has quantitatively assessed the impact of increasing seawater inflow on water and salinity levels or the mixture of bottom water in Saemangeum Lake. Therefore, this study quantitatively analyzes the extent of changes in bottom water exchange and salinity with differential seawater inflow, that is, varying frequencies of sluice gate operation. We used a comprehensive numerical model to consider all possible factors rather than analyzing the contribution/superiority of individual factors on the formation of stratification.

2. Material and Method

2.1 Theoretical background

2.1.1 Model outline

We selected the widely used environmental fluid dynamics code (EFDC_DS) model (Dynamic Solutions LLC [DSLLC], US). It includes hydrodynamics, water quality, sediment transport, and toxicity modules that apply to various aquatic environments, such as coasts, estuaries, lakes and marshes, and wetlands. This model is also suitable for three-dimensional intertidal zones in shallow sea areas.

2.1.2 Basic equations

The EFDC incorporates the Navier–Stokes equation for fluid flow and the advection–diffusion equations for salinity and temperature (Hamrick, 1992). Transforming the vertically hydrostatic boundary layer form of the turbulent equations of motion and utilizing the Boussinesq approximation for variable density generates the momentum and continuity equations and transport equations for salinity and temperature [Eqs. (1–7)].

A. Motion

\begin{array}{l} \partial_{t} (m_{x} m_{y} H u) + \partial_{x} (m_{y} H u u) + \partial_{y} (m_{x} H υ u) \\ + \partial_{z} (m_{x} m_{y} w u) - (m_{x} m_{y} f + υ \partial_{x} m_{y} - u \partial_{y} m_{x}) H υ \\ = - m_{y} H \partial_{x} (g ζ + p) - m_{y} (\partial_{x} h - z \partial_{x} H) \partial_{z} P \\ + \partial_{z} \frac{m_{x} m_{y} A_{υ} \partial_{z} u}{H} + Q_{u} \end{array}

(1)

\begin{array}{l} \partial_{t} (m_{x} m_{y} H υ) + \partial_{x} (m_{y} H u υ) + \partial_{y} (m_{x} H υ υ) \\ + \partial_{z} (m_{x} m_{y} w υ) - (m_{x} m_{y} f + υ \partial_{x} m_{y} - u \partial_{y} m_{x}) H u \\ = - m_{y} H \partial_{y} (g ζ + p) - m_{x} (\partial_{y} h - z \partial_{y} H) \partial_{z} P \\ + \partial_{z} \frac{m_{x} m_{y} A_{υ} \partial_{z} u}{H} + Q_{υ} \end{array}

(2)

B. Continuity

\begin{array}{l} \partial_{t} (m_{x} m_{y} ζ) + \partial_{x} (m_{y} H u) + \partial_{y} (m_{x} H υ) \\ + \partial_{z} (m_{x} m_{y} w) = 0 \end{array}

(3)

C. Water depth integral continuity (applied in the vertical boundary condition)

\partial_{t} (m ζ) + \partial_{x} (m_{y} H \int_{0}^{1} u d z) + \partial_{y} (m_{x} H \int_{0}^{1} υ d z) = 0

(4)

D. Statics

\partial_{z} p = - \frac{g H (ρ - ρ_{0})}{ρ_{0}} = - g H b

(5)

Here,

u, υ : Horizontal velocity component in x, y directions
h, ζ : Water depth, elevation
H : Total water depth (= h + ς)
m_x , m_y : Square root of the metric-tensor diagonal component that satisfies the arbitrary distance in the curvilinear coordinate system
w : Vertical velocity component in the converted dimensionless vertical coordinate system z

Bottom friction was determined based on ocean floor shear stress as follows:

\frac{A_{υ} \partial_{z} {(u, υ)}_{z = 0}}{H} = (τ_{b x}, τ_{b y}) = c_{b} \sqrt{u_{1}^{2} + υ_{1}^{2}} (u, υ_{1})

(6)

c_{b} = {(κ^{- 1} ln (Δ_{1} / 2 z_{0}))}^{- 2}

(7)

Here,

c_b : Drag coefficient
κ : von Karman constant (0.4)
Δ₁ : Dimensionless bottom surface thickness
z₀ : = $z_{0}^{*} / H$ , dimensionless roughness height (0.0125)

2.1.3 Heat transfer equation

The sigma-coordinate basic heat transfer equation was calculated as follows (Ji et al., 2008):

\begin{array}{l} \frac{\partial}{\partial t} (m H T) + \frac{\partial}{\partial x} (P T) + \frac{\partial}{\partial y} (Q T) + \frac{\partial}{\partial z} (m w T) \\ = \frac{\partial}{\partial z} (\frac{m}{H} A_{b} \frac{\partial T}{\partial z}) = \frac{\partial I}{\partial z} + S_{T} \end{array}

(8)

Here,

x, y : Orthogonal curvilinear coordinate in horizontal direction (m)
z : Sigma-coordinate (dimensionless)
t : Time (s)
m_xm_y : Square root (m) of diagonal component of metric tensor
m : Jacobian matrix, (m2)
T : Temperature (°C)
H : Total water depth (m)
P , Q : Mass flux per unit time in x, y coordinate system
u, υ : Horizontal velocity component in curvilinear coordinate system (m s-1)
w : Vertical velocity component (m s-1)
A_υ : Vertical warm current eddy viscosity coefficient (m2 s-1)
I : Short wave solar radiation energy (W m-2)
S_t : (Source of heat exchange) / (Sink) (J s-1)

I (D) = I_{s} exp (- K_{e} D)

(9)

Here,

I (D) : Solar radiation energy (W m-2) delivered to depth D from surface
I_s : Solar radiation energy (W m-2) at surface
D = H (1 - z) : Depth (m) from surface
K_e : Light attenuation coefficient (1 m-1)

Solar radiation is absorbed by water as it passes through the surface and proceeds deep into the water column according to the light extinction coefficient (i.e., light attenuation coefficient), which measures the intensity of light absorbed.

2.1.4 Wind stress

We calculated wind shear stress, an important factor acting on the surface of the water, based on Eqs. (10) and (11). Wind stress at the water surface was determined using the composition factors x and y.

\frac{1}{ρ_{w}} [\begin{matrix} τ_{s x} \\ τ_{s y} \end{matrix}] = C_{D} \frac{ρ_{a}}{ρ_{w}} W_{S} [\begin{matrix} W_{s x} \\ W_{s y} \end{matrix}]

(10)

W_{s} = \sqrt{W_{s x}^{2} + W_{s y}^{2}}

(11)

Here,

W_s : Wind velocity at 10 m above surface
W_sx : composition factor of wind velocity at 10 m above surface
W_sy : composition factor of wind velocity at 10 m above surface
τ_sx : Surface drag due to wind in direction
τ_sy : Surface drag due to wind in -direction
C_D : Wind direction coefficient
ρ_a: Air density
ρ_w : Water density

2.1.5 Lagrangian particle tracking module

The Lagrangian particle tracking module in EFDC is an effective tool for solving many problems in fluid dynamics related to the simulation and prediction of the trajectory of objects traveling in rivers, lakes, and marine systems. Research on the trajectories of solid particle movement in a fluid environment appeared very early in mechanical sciences and was considered a Lagrangian approach. The advantage of this method is that it can track the movement of each specific particle with more detail and accuracy compared with the averaging of grid-cell concentrations.

The position equation for a Lagrangian particle assuming a random walk can be calculated as follows (DSI LLC., 2023):

d x = (u + \frac{\partial A_{H}}{\partial x}) d t + (2 p - 1) \sqrt{2 A_{H} d t}

(12)

d y = (υ + \frac{\partial A_{H}}{\partial y}) d t + (2 p - 1) \sqrt{2 A_{H} d t}

(13)

d z = (w + \frac{\partial A_{b}}{\partial z}) d t + (2 p - 1) \sqrt{2 A_{b} d t}

(14)

where dt is the time step (s ), A_H is the horizontal momentum and mass diffusivity (m^s/s ), A_b is the vertical turbulent eddy diffusivity (m^s/s ), and p is a random number from a uniformly distributed random variable generator with a mean of 0.5. Eqs. (12) –(14) follow the 3D random walk approach used by Dunsbergen and Stelling (1993).

2.2 Data collection

We verified the model using data for conductivity, temperature, and depth (CTD) for May 2021. The survey point of the water temperature salinity data is shown in Fig. 1. Note that we could not obtain data for the exact same study period; therefore, the model simulation was performed from May 15, 2021, to June 14, 2021, according to the water temperature and salinity observation period. Tide level observation data (T-2) were obtained from the Biando Observatory (KHOA, 2023) for the simulated period. As the current observation data (C-1) relates to measurements from April to May 2012 measured in 12JB05 (KHOA, 2023), model verification was performed using the current curve predicted from the harmonic constant.

2.3 Model construction

Surface and bottom water exchanges in Saemangeum were simulated in a sufficiently broad study area, extending from where the inner stream flows to the outer sea. The horizontal range of the model was approximately 40 km from south to north and east to west. A tunable grating system with grid spacing from 240 m to 300 m was applied considering the width of the Sinsi and Garyeok sluice gates. A total of 141 grids were placed each in the x- and y-direction, and a total of 9,224 active cells were generated. Five vertical Sigma layers were included, and a digital chart was used for water depth input (Fig. 2). Breakwaters and topography within the range of the model were treated using the masking technique.

The boundary condition of the model and initial conditions of water temperature and salinity were applied as shown in Tables 1 –3. We used the equilibrium temperature module for the surface water temperature calculation.

The boundary condition of the atmospheric variables were applied as shown in Table 4. We used meteorological data from Gunsan (140), including sea-level (atmospheric) pressure, temperature, relative humidity, rainfall, wind velocity, and wind direction. Solar radiation and total cloud cover data were sourced from Jeonju (146) (Fig. 3).

2.4 Experimental conditions

The EFDC model is limited in simulating the opening and closing of the sluice gate in Saemangeum. Thus, we developed a subroutine to reproduce the natural opening and closing of the sluice gate (Fig. 4). The subroutine is called from the flow calculation code to read the gate opening and closing flags recorded in the GATE.inp file and applies gate control to the model through the masking method. The GATECONTROL subroutine reproduced sluice gate opening and closing based on actual operational records of the Saemangeum Project Office of Korea Rural Community Corporation.

In the actual data, 0 (closed) and 1 (open) were used as sluice gate opening and closing flags; however, they are indicated as 0 and –1 on the floor plan for readability.

The Case 2 model simulation proceeded with the sluice gate operation data, excluding nighttime opening (Case 1, Table 5).

Changes in salinity were analyzed over three periods, including Day 10.5 (1 d after sluice gate opening), Day 16.208 (immediately after closing for the spring tide), and Day 24.75 (immediately before opening for the spring tide; Fig. 5).

2.5 Particle exchange experiment

Particle tracing was performed using the verified model. Particles were dropped shallower than 5 m in water depth, and the movement of bottom water was observed according to the opening of the sluice gate.

Oh and Choi (2015) reported that the vertical stratification of salinity within Saemangeum was most pronounced within 5–10 m in water depth, and found that the salt concentration rapidly changed within the upper 5 m during summer. Therefore, we dropped particles below the reference depth for bottom water exchange (i.e., 5 m) and analyzed the degree of exchange based on the number of particles remaining under 5 m.

R (%) = [1 - \frac{P_{t}}{P_{0}}] \times 100

(15)

Here, P₀ is the initial number of particles and P_t is the number of particles remaining after 30 day.

3. Results

3.1 Verification of results

3.1.1 Verifying the operation of GATECONTROL

To verify that the GATECONTROL subroutine works properly, we compared the water level inside Saemangeum observed since January 28, 2019, with the model-calculated water level. Although the timeline of the observed data does not overlap with the simulation period, it is considered sufficient to verify the operation of GATECONTROL. We confirmed that the model effectively reproduced the observed water level (Fig. 6). The validation point (T-1) of the model is shown in Fig. 1.

3.1.2 Verification of tide and current

The validation points (T-2, C-1) of the model are shown in Fig. 1. We verified the model using the method of Kim and Yoon (2011). We calculated the Absolute Average Error (AAE) and Relative Absolute Average Error (RAAE) as quantitative indicators, and the Correlation Coefficient (CC), Index of Agreement (IA), and Cost Function (CF) as qualitative indicators (Tables 6 and 7; Figs. 7 and 8). Tide level verification for Case 1 showed that the AAE was 0.13 m and CC was 0.9950. For Case 2, the AAE was 0.14 m and CC was 0.9949. Tidal current verification for Case 1 showed that the AAE and CC were 0.03 m s-1 and 0.98 for the u component and 0.04 m s-1 and 0.99 for the v component, respectively. For Case 2, the AAE and CC were 0.1 m s-1 and 0.77 for the u component and 0.03 m s-1 and 0.99 for the v component, respectively. These results indicate that the model effectively reproduced the tidal characteristics near the inner and outer boundaries of Saemangeum Lake.

3.1.3 Water temperature and salinity verification

To verify the water temperature and salinity results, we compared the calculated values at the same point and time with the CTD survey data in 2021 (Figs. 9 and 10). The validation points of the model are shown in Fig. 1. The vertical profile of water temperature and salinity matched the calculated and observed values for all points except ME2 and DE2. The difference in salinity at these points was attributed to vertical resolution and differences in water depth.

Regarding salinity simulation, the amount of inflow from the Mangyeong and Dongjin Rivers, which was incorporated into the model, was estimated from existing literature rather than precise measurements, which may have constrained the reproduction of actual salinity levels. Nonetheless, the model provided a relatively sound reproduction of the results.

3.2 Flow reproduction

We simulated the water flow in Saemangeum using the flow model (Fig. 11). In the outer seaside of Saemangeum, the rising tide progresses northeast and shows different characteristics from the Gogunsan Islands. The inflowing tide branches off due to Biando Island and the Saemangeum port breakwater, a part of the flow enters inner Saemangeum through the Sinsi and Garyeok sluice gates, and, at the north end of the Gogunsan Islands, it flows north along the northern breakwater. The ebb tide current showed reverse flow characteristics to the rising tide.

The inner flow of Saemangeum was considerably impacted by the sluice gate. During neap tide when the sluice gates are not open, the flow rate is weaker and irregular compared to that in spring tide. The flow rate was stronger when Sinsi sluice gate was open than when Garyeok sluice gate was open.

3.3 Inner water level changes according to frequency of sluice gate opening

3.3.1 Inner water level change comparison

We examined the changes in water level according to the frequency of sluice gate opening for Cases 1 and 2 (Fig. 12). Due to the additional opening and closing of the model sluice gate, the water level in Case 2 was 0.7 m lower than that for Case 1.

3.4 Salinity change experiment according to the frequency of sluice gate opening

3.4.1 Salinity change simulation

We compared the changes in salinity according to the frequency of sluice gate opening between Days 10.5, 16.208, and 24.75 (Figs. 13–15).

The level of salinity was higher in the Dongjin River–Garyeok sluice gate region than that in the upper Mangyeong River–Sinsi sluice gate region on the east–west road. However, bottom water salinity was higher at the front of Garyeok sluice gate than that at the front of Sinsi sluice gate, which was attributed to the 40-m-deep trough at Sinsi sluice gate.

On Day 10.5 during the initial sluice gate opening, the salt concentration within the lake increased due to the inflow of seawater. Salt concentrations increased further by Day 16.208, before closing the spring tide sluice gate, and decreased at Day 24.75 after neap tide with a closed sluice gate, due to the inflow of freshwater.

Opening the sluice gate allows the inflow of seawater and the outflow of lake water during ebb tide. Consequently, salt concentrations decreased in the nearby sea area. This impact persisted even after neap tide when the sluice gate had closed.

We performed a more detailed analysis of the changes in salinity according to the frequency of sluice gate opening (Figs. 16 –18). With an increasing frequency of sluice gate opening, the salt concentration increased in inner Saemangeum from the breakwaters. In particular, the rising tide caused an inflow of seawater, thereby increasing the salt concentration in downstream Dongjin by a considerable amount. Additionally, during spring tide with an open sluice gate, salt concentration varied between the surface and bottom layers with an increase in the frequency of sluice gate opening.

The salinity at the front outer part of the sluice gate decreased in Case 2 due to an increase in the outflow of freshwater to the sea through the Sinsi and Garyeok sluice gates.

After spring tide, the difference in salinity between Cases 1 and 2 decreased due to the inflow of freshwater during neap tide with no sluice gate opening. With increased seawater inflow in Case 2, freshwater inflow to the inner part of the lake stagnated, leading to a reduction in salinity where freshwater entered.

The changes in salinity according to the opening of the sluice gate were quantitatively analyzed and compared at seven fixed points (Tables 8 and 9). At the surface layer on Day 10.5 (immediately after opening the sluice gate during spring tide) the level of salinity was higher at ME2, DE2, and DL2 and lower at MK7 and ML3 near the breakwater for Case 2 compared to that for Case 1. On Day 16.208, the level of salinity was higher at MK7, ML3, and DL2 near the breakwater and lower at ME2, DE1, and DE2 at the inflow of freshwater for Case 2 compared to that for Case 1. However, on Day 24.75, the level of salinity was higher at MK7, ML3, DE1, DE2, and DL2 within the lake and only lower at ME2 (where freshwater flows in from the Mangyeong basin) in Case 2 compared to that in Case 1. With increasing sluice gate opening, the maximum increase in salinity at the surface layer was 2.12 psu at DE2 on Day 10.5, and the maximum decrease was –1.18 psu at ME2 on Day 24.75.

In the bottom layer at Day 10.5, the level of salinity was higher at ME2, DE2, and DL2, while it was lower at MK7 and ML3 for Case 2 compared to that for Case 1. On Day 16.208, at the freshwater inflow, elevated salinity levels were observed at MK7, ML3 DE2, and DL2 while decreased levels occurred at ME2 and DE1 for Case 2 compared to that for Case 1. On Day 24.75, higher salinity levels were observed at MK7, ML3, DE1, DE2, and DL2 and lower levels at ME2 for Case 2 compared to that for Case 1, which is similar to the findings for the surface layer. With increased opening of sluice gate, the maximum increase in salinity at the bottom layer was 1.25 psu at DE2 on Day 16.208, and the maximum decrease was –1.94 psu at ME2 on Day 24.75.

These findings indicated that the maximum increase in salinity occurred on Day 10.5 for the surface layer and Day 16.208 for the bottom layer. This was attributed to the rapid change in salinity at the surface layer due to the inflow of seawater after opening the sluice gate and halite water sinking to the bottom layer over time. In contrast, the maximum decrease in salinity at the surface and bottom layers was observed at ME2 on Day 24.75, ~8 days after the opening and closing of the sluice gate, which we attributed to the continuous inflow of freshwater being stagnated by the impact of the seawater.

3.5 Particle exchange experiment

3.5.1 Bottom water particle exchange simulation

To quantitively assess the extent of bottom water exchange according to the operation of the sluice gate, particle tracing was analyzed at below 5 m in water depth (Figs. 19 and 20). The particles escaped outwards during spring tide sluice gate opening, while the outside particles flowed back into the lake. During neap tide, the particles moved only within the lake, and only freshwater inflow and wind acted on them, causing minor particle movement.

After 720 hour, the number of residual particles under 5 m in water depth was considerably reduced, especially during spring tide. This was attributed to the stable impact of freshwater inflow on the exchange of the upper- and lower-layers during spring tide.

Figs. 21–22 illustrate the changes in the number of residual particles within Saemangeum Lake over time above and below 5 m in water depth. The residual rate of particles under 5 m in water depth was lower for Case 2 than that for Case 1.

Cases 1 and 2 showed rapid surfacing of the particles dropped to 5 m or below due to the inflow of seawater. Subsequently, the particles tend to rapidly escape the lake through the sluice gate during spring tide opening.

The number of particles that escaped was considerably higher in Case 2 than that in Case 1. Therefore, on Day 30, the residence of particle residues at the surface layer was longer in Case 2 than that in Case 1 due to seawater inflow. On Day 27 during the third spring tide, a greater quantity of particles resided at 0–5 m in water depth in Case 2, despite having the same sluice-gate-opening schedule.

3.5.2 Comparison of residual rate of particles in the bottom layer

Table 10 presents a comparison of the residual rate of particles at or below 5 m in water depth using the simulated results for 5 day of spring and neap tide.

The residual rate of particles was 2.52% lower in Case 2 than that in Case 1, suggesting that an increased frequency of sluice gate opening would promote bottom water exchange.

4. Conclusion

This study quantitatively analyzed the differences in bottom water exchange and salinity level according to the frequency of sluice gate opening in Saemangeum Lake using a numerical model. Seawater inflow through the sluice gate during ebb tide was greater at the lower Garyeok sluice gate than that at the upper Sinsi sluice gate, and salinity was higher at the inner part of Garyeok sluice gate than that at the inner part of Sinsi sluice gate. After 10.5 day (initial period of sluice gate opening during spring tide), the salt concentration within the lake increased due to seawater inflow; the salt concentration at the bottom layer increased after 16.208 day (end of sluice gate opening during spring tide) compared to that on Day 10.5. However, after 24.75 day (before opening the sluice gate after neap tide), the salt concentration decreased within the lake due to freshwater inflow.

To quantitatively assess the extent of bottom water exchange according to the operation of the sluice gates, we performed particle tracing and found that the particles initially moved within the lake due to wind and eventually escaped the lake during the rising spring tide because of the opening of the sluice gate. During ebb tide, seawater flowed in and filled the deep layers within the lake, causing particles at the bottom layer to float. After 30 day, the rate of residual particles at or below 5 m in water depth decreased due to an increase in bottom water exchange associated with increased sluice gate opening.

Our results indicated that seawater flows in during ebb tide with open sluice gates despite stratification in Saemangeum Lake, and the exchange rate of bottom water increased with an increasing frequency of sluice gate opening because the bottom water was substituted with seawater by the outflow of surface water and mixture during the rising tide. Overall, water quality is expected to improve because of increased bottom water exchange and salinity associated with increased seawater circulation in Saemangeum Lake.

On the other hand, dredging and reclamation work is currently being conducted at Saemangeum Lake. As a result, the model was unable to perfectly reflect the exact topography at the time of observation. Therefore, large discrepancies arose between the observed and simulated values during the vertical verification of water temperature and salinity. Although the model has limitations and does not accurately reflect the terrain, it is clear that increasing the number of gate openings improves water quality. Future simulations will require new observations and accurate terrain information after all construction work in Saemangeum is completed.

Figure

Fig. 1.

Fixed survey point and water depth grid map. (ME: Mangyeong Estuary, MK: Mangyeong-Kunsan, ML: Mangyeong Lake, DE: Dongjin Estuary, DL: Dongjin Lake)

Fig. 2.

Input air temperature time-series data (KMA, 2023).

Fig. 3.

Input wind speed and direction time-series data (KMA, 2023). Spd.: speed; Dir.: direction.

Fig. 4.

Summary of code parameters for the subroutine GATECONTROL.

Fig. 5.

Input flags for gate control.

Fig. 6.

Comparison between observed and modeled water surface elevation in Saemangeum.

Fig. 7.

Comparison of time series for observed and simulated tides.

Fig. 8.

Comparison of time series for observed and simulated tidal current.

Fig. 9.

Comparison of observed and calculated values of vertical water temperature in Saemangeum.

Fig. 10.

Comparison of observed and calculated values of vertical salinity in Saemangeum.

Fig. 11.

Current vector map: (a) spring flood tide, (b) spring ebb tide, (c) neap flood tide, and (d) neap ebb tide.

Fig. 12.

Comparison of water surface elevation between Case 1 and Case 2 near Saemangeum gates.

Fig. 13.

Salinity simulation after 10.5 day: (a) Case 1—surface layer, (b) Case 1—bottom layer, (c) Case 2—surface layer, and (d) Case 2—bottom layer.

Fig. 14.

Salinity simulation after 16.208 day: (a) Case 1—surface layer, (b) Case 1—bottom layer, (c) Case 2—surface layer, and (d) Case 2—bottom layer.

Fig. 15.

Salinity simulation after 24.75 day: (a) Case 1—surface layer, (b) Case 1—bottom layer, (c) Case 2—surface layer, and (d) Case 2—bottom layer.

Fig. 16.

Difference in salinity simulation results between Case 1 and Case 2 (Case 2 – Case 1) after 10.5 day: (a) surface layer, (b) bottom layer.

Fig. 17.

Difference in salinity simulation results between Case 1 and Case 2 (Case 2 – Case 1) after 16.208 day: (a) surface layer, (b) bottom layer.

Fig. 18.

Difference in salinity simulation results between Case 1 and Case 2 (Case 2 – Case 1) after 24.75 day: (a) surface layer, (b) bottom layer.

Fig. 19.

Initial period of particle tracing.

Fig. 20.

Particle tracing after 720 hour: (a) Case 1, (b) Case 2.

Fig. 21.

Particle numbers according to water depth with sluice gate opening once.

Fig. 22.

Particle numbers according to water depth with sluice gate opening twice.

Table

Table 1.

Tidal harmonics (KHOA, 2023) in open boundaries

Classification	Gyeokpo Port	Bieung Port
M2	203.4	75.2	210.1	81.5
S2	78.2	127.3	82	134
K1	33.6	275.6	33.7	278.2
O1	25.1	237.8	25.9	240.3

Table 2.

River inflows into Saemangeum Lake (Ministry of Environment Republic of Korea, 2020)

Classification	Mangyeong	Dongjin
Flow rate	7.86 m3 s-1	8.68 m3 s-1

Table 3.

Initial and boundary conditions of water temperature and salinity in EFDC model

Classification	Salinity (psu)	Water temperature (°C)
Boundary condition	Quayside	Surface layer	30.6	18.2
Bottom layer	31.1	18.4
Dongjin	Average	0	24.4
Mangyeong	Average	0	24.9
Initial condition	Quayside	Average	30.9	16.8
Inner side	Average	15.6	19.7

Table 4.

Boundary condition time series of atmospheric variables

Classification	Average	Unit
Atmospheric pressure	1009.15	mb
Air temperature	19.51	°C
Humidity	0.78
Rainfall	0.0002	m/d
Evaporation	0.0014	m/d
Solar radiation	211.31	W/m2/d
Cloud cover	0.58

Table 5.

Model scenarios

Experiment	Sluice gate opening
Case 1	Open once a day
Case 2	Open twice a day

Table 6.

Verification of tidal elevations

Variable	Target value	Reach
AAE (m)	0.00	0.13
RAAE (%)	0.00	2.42
CC	1.0000	0.9950
IA	1.0000	0.9975
CF	0.0000	0.0863

Table 7.

Verification of tidal currents

12JB05
AAE (m/s)	0.03	0.03
RAAE (%)	6.09	7.47
CC	0.9877	0.9825
IA	0.9937	0.9909
CF	0.1309	0.1561

Table 8.

Surface layer salinity (psu) at seven points for Days 10.500, 16.208, and 24.750 compared between Case 1 and 2

Day	Case	ME1	ME2	MK7	ML3	DE1	DE2	DL2
10.5	Case 1	0.00	5.64	13.26	15.25	0.00	11.45	18.32
Case 2	0.00	5.97	13.25	15.09	0.00	13.57	19.41
Delta	0.00	0.33	–0.01	–0.16	0.00	2.12	1.09
16.208	Case 1	0.00	9.02	15.29	17.46	0.08	17.13	22.44
Case 2	0.00	7.97	15.38	18.42	0.00	16.63	24.14
Delta	0.00	–1.05	0.09	0.96	–0.07	–0.50	1.70
24.75	Case 1	0.00	8.09	17.91	18.51	0.01	13.24	20.84
Case 2	0.00	6.91	18.45	19.24	0.04	14.61	22.13
Delta	0.00	–1.18	0.54	0.73	0.04	1.37	1.29

Table 9.

Bottom layer salinity (psu) at seven points for Day 10.500, 16.208, and 24.750 compared between Case 1 and 2

Day	Case	ME1	ME2	MK7	ML3	DE1	DE2	DL2
10.5	Case 1	0.00	6.49	19.09	23.37	0.00	20.31	27.41
Case 2	0.00	7.12	18.39	23.00	0.00	20.71	27.7
Delta	0.00	0.63	–0.69	–0.38	0.00	0.40	0.29
16.208	Case 1	0.00	12.05	21.78	25.86	0.15	24.97	28.40
Case 2	0.00	11.28	22.14	26.47	0.00	26.22	28.55
Delta	0.00	–0.77	0.36	0.61	–0.14	1.25	0.15
24.75	Case 1	0.00	16.19	21.48	24.30	0.00	23.45	27.81
Case 2	0.00	14.25	22.00	24.97	0.32	24.10	28.28
Delta	0.00	–1.94	0.52	0.67	0.32	0.65	0.47

Table 10.

Residual rate of bottom-layer particles

	Frequency of sluice gate opening	Residual rate	Note
Particles at or below 5 m	Once (Case 1)	10.22%
Twice (Case 2)	7.70%

Reference

DSI LLC (2023), EFDC+ Theory, Version 11. Published by DSI LLC, Edmonds WA. Available at https://www.eemodelingsystem.com/wp-content/Download/Documentation/EFDC Theory Document Ver 11.pdf
Dunsbergen, D. W. and G. Stelling (1993), A 3d particle model for transport problems in transformed coordinates. Communications on hydraulic and geotechnical engineering, No. 1993-07
Hamrick, J. M. (1992), A three-dimensional environmental fluid dynamics computer code: Theoretical and computational aspects. Technical report, Virginia Institute of Marine Science, College of William and Mary.
Jeong, S. K. , K. H. Ryu, Y. H. Jung, and S. G. Lee (2018), Optimal management of water level for water quality security in Saemangeum Basin, Journal of the Korean Society of Hazard Mitigation, Vol. 18, Issue 5, pp. 301-309.
Ji, H. , C. J. Adkins, B. R. Cartwright, and K. L. Friedman (2008), Yeast est2p affects telomere length by influencing association of rap1p with telomeric chromatin. Molecular and Cellular Biology, Vol. 28, pp. 2380-2390.
KHOA(Korea Hydrographic and Oceanographic Agency) (2023), Ocean Data in Grid Framework. Available at http://www.khoa.go.kr/oceangrid/gis/category/reference/distribution.do.
Kim, B. C. , J. Y. Lee, J. S. Choi, W. M. Heo, and S. B. Park (2008) Fish kill caused by eutrophication and stratification in a brackish coastal Lagoon, Korean Society of Water & Wastewater Joint Fall Forum Paper, pp. 23-24.
Kim, T. Y. and H. S. Yoon (2011), Skill assessments for evaluating the performance of the hydrodynamic model, Journal of the Korean Society for Marine Environment & Energy, Vol. 14, Issue 2, pp. 107-113.
KMA National Climate Data Center (2023) Open MET Data Portal. Available at https://data.kma.go.kr/data/grnd/selectAsosRltmList.do?pgmNo=36.
Ministry of Environment Republic of Korea (2020), Saemangeum Lake sediment monitoring (Ⅷ).
Ministry of Environment Republic of Korea (2021), Water quality impact analysis and reduction measure according to Saemangeum salinity/water temperature stratification.
Nahas, E. L. , C. B. Pattiaratchi, and G. N. Ivey (2005), Processes controlling the position of frontal systems in Shark Bay, Western Australia, Estuarine, Coastal and Shelf Science, Vol. 65, Issue 3, pp. 463-474.
Oh, C. S. and J. H. Choi (2015), Consideration on changes of density stratification in Saemangeum reservoir, Journal of the Korean Society for Marine Environment and Energy, Vol. 18, Issue 2, pp. 81-93.
Oh, C. S. , J. H. Choi, and Y. K. Cho (2013), Numerical simulation on hydrodynamic characterization changes associated with the construction of dikes and dredging operations in Saemangeum Lake, Journal of Environmental Science International, Vol. 22, Issue 9, pp. 1115-1129.
Pae, Y. S. , S. H. Choi, E. T. Kang, S. H. Kim, and C. H. Kang (2009), Growth characteristic of stratification in Saemangeum Lake, Proceedings of the Korean Society of Agricultural Engineers Conference, pp. 180-180.
Park, S. H. , J. G. Kim, G. O. Myung, and M. S. Kwon (2023), Relationship between phytoplankton distribution and salinity in the giant artificial brackish Lake Saemangeum, Ecohydrology & Hydrobiology, Vol. 23, Issue 2, pp. 251-260.
Schroeder, W. A. , L. M. Kay, and R. S. Mills (1990), Spectrophotometric analysis of amino acids using ninhydrin chemical reaction, Analytical Chemistry, Vol. 22, pp. 760-760.
Scully, M. E. , C. Friedrichs, and J. Brubaker (2005), Control of estuarine stratification and mixing by wind-induced straining of the estuarine density field, Estuaries, Vol. 28, pp. 321-326.
Souza, A. J. and J. H. Simpson (1997), Controls on stratification in the Rhine ROFI system, Journal of Marine Systems, Vol. 12, Issues 1-4, pp. 311-323.

Classification	Gyeokpo Port		Bieung Port
Classification	Amplitude (cm)	Phase (°)	Amplitude (cm)	Phase (°)
M2	203.4	75.2	210.1	81.5
S2	78.2	127.3	82	134
K1	33.6	275.6	33.7	278.2
O1	25.1	237.8	25.9	240.3