Influence of Salinity Gradient Changes on Phytoplankton Growth Caused by Sluice Construction in Yongjiang River Estuary Area

Though the number of sluices and dams in coastal areas has increased rapidly in recent years, the influence of their construction on phytoplankton in estuary areas is hardly known. This paper aims to provide a reference for quantitative research on the ecological influence of sluice construction and give ecological justifications for the setting of environmental standards in the estuary areas. The survey data gained at the lower reach of the Yongjiang River and its estuarine areas in June 2015 were used in MIKE21 software (Danish Hydraulic Institute (DHI), Denmark)) for establishing a two-dimensional numerical model to simulate the salinity field distribution after sluice construction. Based on the simulation results, the salinity gradient changes caused by the construction were analyzed. The one-dimensional Gaussian model was applied to calculated the phytoplankton’s ecological threshold interval over the salinity changes, which helped predict the influence of salinity changes on phytoplankton cell density. The study shows that salinity in the Yongjiang estuary increases obviously, beyond the phytoplankton ecological threshold, after sluice construction without water discharge. Salinity will become a restriction factor to phytoplankton growth after sluice construction in the study area, which may cause a sharp decrease of certain phytoplankton species.


Introduction
The increasing demand for oceanic resources and services from mankind has brought great pressure to the aquatic ecosystem [1,2]. Phytoplankton, as the most important primary producer in the aquatic ecosystem and a great participant in the global carbon cycle [1,3,4], is widely spread on the earth and plays a key role in energy flow, material circulation and information transmission [5]. Rojo C et al. studied phytoplankton species richness related to drought in the semiarid wetland of Las Tablas de Daimiel National Park in Central Spain and found out that phytoplankton could be used to effectively track the environmental changes in the wetland [6]. Rath A R et al. investigated the relationship between seasonal-spatial distribution of phytoplankton and environmental factors in the New Mangalore Port along the western coast in India. The authors found that chlorella and dinoflagellate's high richness matched their high nutrition in monsoon seasons, which emphasized some diatom species' potential as hydrology and water quality indicators in coastal ecosystems [7]. Several studies have analyzed the reaction of phytoplankton to ecological factors [8][9][10][11], but little analysis has been done from the perspective of an ecological threshold. Phytoplankton is very suitable for analyzing the ecological threshold of environmental gradients on large spatial scales because of its unique performance as an ecological indicator of environmental changes in different ecosystems [12]. area is shown in Figure 1.
The Yongjiang basin has a subtropical humid monsoon climate with an obvious change between winter and summer monsoons. The wind direction is southeast in summer and northwest in winter. The Yongjiang tide is an irregular semi-diurnal tide, with an annual maximum tide level of 4.86 m, a minimum of 0.13 m, an annual average high tide level of 3.14 m, a low level of 1.46 m and an average tide difference of 1.82 m. Figure 1. The study area: The lower reach below the Sanjiangkou in Ningbo is the Yongjiang River. The Yongjiang estuary is north to Huibie Sea in Hangzhou Bay and east to Jintang Island.

Data Collection
Data of hydrology, water quality and biology of the study area involved in this paper were provided by Bureau of Hydrology and Water Resources Investigation at Lower Reaches of Yangtze River, Hydrology Bureau of Yangtze River Water Conservancy Commission, and came from the measurements carried out in June 2015 at sampling locations presented in Figure 2.
In Figure 2, there were 45 sampling points in total, 12 in the lower reach of the Yongjiang River (with numbers 1 through 12). From these 12, 9 were taken from three cross-sections, sampled evenly along the section three times (e.g., 01~03, marked with "⊕"). The other 33 points (with numbers 13 through 45) were taken in 20 locations in the estuary, with 13 locations sampled in both upper and lower water (e.g., 20/21, marked with "⊕") thus representing 26 points. The other single sampling locations were marked with "◯" in Figure 2. The Yongjiang basin has a subtropical humid monsoon climate with an obvious change between winter and summer monsoons. The wind direction is southeast in summer and northwest in winter. The Yongjiang tide is an irregular semi-diurnal tide, with an annual maximum tide level of 4.86 m, a minimum of 0.13 m, an annual average high tide level of 3.14 m, a low level of 1.46 m and an average tide difference of 1.82 m.

Data Collection
Data of hydrology, water quality and biology of the study area involved in this paper were provided by Bureau of Hydrology and Water Resources Investigation at Lower Reaches of Yangtze River, Hydrology Bureau of Yangtze River Water Conservancy Commission, and came from the measurements carried out in June 2015 at sampling locations presented in Figure 2.
In Figure 2, there were 45 sampling points in total, 12 in the lower reach of the Yongjiang River (with numbers 1 through 12). From these 12, 9 were taken from three cross-sections, sampled evenly along the section three times (e.g., 01~03, marked with "⊕"). The other 33 points (with numbers 13 through 45) were taken in 20 locations in the estuary, with 13 locations sampled in both upper and lower water (e.g., 20/21, marked with "⊕") thus representing 26 points. The other single sampling locations were marked with " " in Figure 2.
After the phytoplankton samples were collected, they were immediately fixed with Luge's solution (usually 10 mL per liter of water). Then each sample was concentrated to 30~50 mL and poured into the calibrated precipitator. After 24~48 h, the supernatant was siphoned off. The sample could be stored in the sample bottle until it was concentrated to 30~50 mL. The types of samples were identified directly under the microscope and recorded. Their numbers were counted and calculated by microscope counting method. locations account for 45 points, 12 in the lower reach of the river and 33 in the estuary. Of the 12 points, nine were taken from three cross-sections, sampled evenly along the section three times. The other 33 points (with numbers 13 through 45) were taken in 20 locations in the estuary, with 13 locations sampled in both upper and lower water. Single sampling points were marked with "◯" in their locations, while the multiple ones were marked with "⊕".
After the phytoplankton samples were collected, they were immediately fixed with Luge's solution (usually 10 mL per liter of water). Then each sample was concentrated to 30~50 mL and poured into the calibrated precipitator. After 24~48 h, the supernatant was siphoned off. The sample could be stored in the sample bottle until it was concentrated to 30~50 mL. The types of samples were identified directly under the microscope and recorded. Their numbers were counted and calculated by microscope counting method.

ET and Gaussian Curve Theory
Most ET studies focus on the known single environmental factor that can directly affect aquatic communities [15]. The determination of the ET zone is inseparable from Shelford's Law of tolerance [30]. In Shelford's thought, an organism's adaptation to various environmental factors has its ecological minimum and maximum, between which the organism can survive. The organism's function has its best performance at or near the optimum point, fading towards both endpoints and finally suppressed at those endpoints.
The relationship between organism and environmental factors is usually described by the belltolerance curve, matching the concentration, symmetry and even variability of the one-dimension Gaussian curve. Therefore, this paper simulated the above relationship and presented an ET zone range by using a Gaussian curve.
The one-dimension Gaussian model Equation is: In the Equation, refers to one index that represents phytoplankton growth condition; is the maximum value of corresponding index; is the corresponding environmental factor value when its biological index reaches the maximum value; means the tolerance of the index, which describes

ET and Gaussian Curve Theory
Most ET studies focus on the known single environmental factor that can directly affect aquatic communities [15]. The determination of the ET zone is inseparable from Shelford's Law of tolerance [30]. In Shelford's thought, an organism's adaptation to various environmental factors has its ecological minimum and maximum, between which the organism can survive. The organism's function has its best performance at or near the optimum point, fading towards both endpoints and finally suppressed at those endpoints.
The relationship between organism and environmental factors is usually described by the bell-tolerance curve, matching the concentration, symmetry and even variability of the one-dimension Gaussian curve. Therefore, this paper simulated the above relationship and presented an ET zone range by using a Gaussian curve.
The one-dimension Gaussian model Equation is: In the Equation, y refers to one index that represents phytoplankton growth condition; c is the maximum value of corresponding index; u is the corresponding environmental factor value when its biological index reaches the maximum value; t means the tolerance of the index, which describes biological species' ecological threshold. Generally speaking, one biological specie's ET zone is (u − 2t, u + 2t), and its optimum ET zone is (u − t, u + t). Figure 3 shows the Gaussian curve.

Ecological Factor Selection
Ecological factor is the general term of various environmental factors that influence organisms. A large number of studies indicate that phytoplankton growth, its community structure changes and regional distribution are affected by many environmental factors such as dissolved oxygen, salinity, nutrient salts, water temperature, and PH value [31][32][33][34][35]. A Pearson Correlation Analysis was made on the main environmental factors in the Yongjiang estuary. The result (see Table S1 in the Supplementary Materials section) indicates that there is no obvious correlation between them despite their respective influence on phytoplankton growth. In other words, their influence on phytoplankton can be regarded as independent. Therefore, a preliminary study of phytoplankton ET under salinity changes is undertaken in this paper. In most cases, salinity is a significant factor in the estuarine system controlling phytoplankton's activity and diversity [36]. Sluice and dam construction will become a physical barrier to nutrient transport and will inevitably change the original salinity field in the estuary [10,36].

Numerical Simulation by MIKE 21 Software (DHI, Denmark)
Two-dimensional and three-dimensional numerical simulations have been widely applied in hydrology and water quality simulations of rivers, estuaries, and oceans [37][38][39][40][41][42][43]. MIKE 21(DHI, Denmark) is a professional engineering software to simulate hydrodynamic force, water quality, sediment and wave, which has advanced pre-processing and post-processing function and a friendly customer page and enjoys popularity in simulating water current, wave, sediment, and other environmental variables of river, lake, estuary, bay, coast, and ocean.
The following equations, the conservation of mass and momentum integrated over the vertical, describe the flow and water level variations: (1) Continuity Equation of water flow:

Ecological Factor Selection
Ecological factor is the general term of various environmental factors that influence organisms. A large number of studies indicate that phytoplankton growth, its community structure changes and regional distribution are affected by many environmental factors such as dissolved oxygen, salinity, nutrient salts, water temperature, and PH value [31][32][33][34][35]. A Pearson Correlation Analysis was made on the main environmental factors in the Yongjiang estuary. The result (see Table S1 in the Supplementary Materials section) indicates that there is no obvious correlation between them despite their respective influence on phytoplankton growth. In other words, their influence on phytoplankton can be regarded as independent. Therefore, a preliminary study of phytoplankton ET under salinity changes is undertaken in this paper. In most cases, salinity is a significant factor in the estuarine system controlling phytoplankton's activity and diversity [36]. Sluice and dam construction will become a physical barrier to nutrient transport and will inevitably change the original salinity field in the estuary [10,36].

Numerical Simulation by MIKE 21 Software (DHI, Denmark)
Two-dimensional and three-dimensional numerical simulations have been widely applied in hydrology and water quality simulations of rivers, estuaries, and oceans [37][38][39][40][41][42][43]. MIKE 21(DHI, Denmark) is a professional engineering software to simulate hydrodynamic force, water quality, sediment and wave, which has advanced pre-processing and post-processing function and a friendly customer page and enjoys popularity in simulating water current, wave, sediment, and other environmental variables of river, lake, estuary, bay, coast, and ocean.
The following equations, the conservation of mass and momentum integrated over the vertical, describe the flow and water level variations: (1) Continuity Equation of water flow: Water 2020, 12, 2492 6 of 15 (2) Momentum Equations pf water flow: In the Equation: x, y, and t are the space and time coordinates respectively; z is water level; u and v are the components of vertical mean velocity in the direction of x and y respectively; M and N are the components of discharge per unit width in the directions of x and y; M = hu, = hv; n is the Manning roughness coefficient; ; v t is the turbulent viscosity coefficient; g is the gravitational acceleration.   The hydrodynamic module was calculated by low order of accuracy, in which calculation time was from 08:00:00 16 June 2015 to 08:00:00 26 June 2015; time step was set to 3600s, CFL numbers of the shallow water Equation and the transport Equation were all set to 0.8 to control equation convergence. The upstream boundary was measured by the water level at Qingshuipu Bridge in the same period, and the downstream boundary was the deep sea. The Global Tide Prediction Model programmed in MIKE 21 (DHI, Denmark) was applied to create a tidal water level. After calibration, the value of eddy viscosity coefficient was set to 0.7 and manning coefficient was set to 60 m 1/3 /s. The Transport Module predefined in the software was selected to simulate salinity.

Tide Level Verification
The measured tide level values of the spring tide on 18 June 2015 and neap tide on 24 June 2015 were selected to verify those of the Zhenhai, Zhenhaikou, and Dong'ao observation stations (Figure 4).

Current Velocity Verification
The measured current velocity of the neap tide on 24 June 2015 was selected to verify the current velocity and flow direction at three observation points in the offshore area.
The offshore sea current compound is composed of the tidal current, wind current, and geostrophic current. It can be divided into periodic tidal current and non-periodic residual current. Therefore, the measured current velocity values could not be used for verification directly. A tidal current quasi-harmonic analysis according to short-term measured values was needed to predict tidal current values. The quasi-harmonic analysis uses a linear equation to analyze and calculate the northward and eastward components of tidal currents, solving tidal harmonic constants and residual current of partial tide by the least square method to present the north component u and the east component v at a designated time [44]. The harmonic constants of the tides selected by the quasi-harmonic analysis were used to forecast the flow corresponding to the simulation time for verification.

Sluice Construction Simulation
There are studies related to sluice construction practicability and analysis of sluice site selection in Yongjiang already in China [45]. Based on them, this paper selected the ESTUARY GATE scheme. The upper boundary hydrological condition adopted the monthly lowest average flow of 90% assurance rate, 4.12 m 3 /s, which was calculated by monthly average flow from 1985 to 2014 at the nearby Hongjiata station. The gate was set as a full water column considering the tide-resisting performance at the sluice site.

Results of Sample Analysis
After preliminary identification, 53 kinds of phytoplankton were found in the study areas, in which there were 23 green algae covering 43.4% of total kinds, 12 blue-green algae covering 22.6%, and 11 diatoms covering 20.8%. The algae belonging to Bacillariophyta, Cyanophyta and Chlorophyta (e.g., Oscillaria sp, Melosira granulata, Tribonema Derbes et Solier, Microcystis aeruginisa, Pediastrum semplex) have the highest percentage in cell density.

Results of Water Level Verification
After calculation, the result shows that the average relative errors between measured tide level and the simulated value at the three tide stations are within the range of 3.3-4.4% during the spring tide and between 2.4-3.8% during the neap tide, meeting the simulation requirements.
Comparison between simulated water level and the measured water level is shown in Figure 5. The simulated water level values are basically consistent with the measured water level values at each station. and the simulated value at the three tide stations are within the range of 3.3-4.4% during the spring tide and between 2.4-3.8% during the neap tide, meeting the simulation requirements.
Comparison between simulated water level and the measured water level is shown in Figure 5. The simulated water level values are basically consistent with the measured water level values at each station.

Results of Velocity Verification
Based on the quasi-harmonic analysis, tidal current matching the simulation time is predicted by harmonic constants of partial tide K1, M2, M3, M4, 2MK5, M6, 3MK7, M8.
After calculation, the average relative errors between the simulated values and the measured values of flow velocity and direction at the three observation stations are within the range of 1.6-3.8% during the neap tide, which meet the simulation requirements.

Results of Salinity Simulation
The average values of salinity during the simulation period in the estuary are introduced to Surfer15, interpolated by the Kriging method to draw contour plots. The two contour plots of salinity gradient changes before and after sluice construction are shown in Figure 6. Contour map (a) was drawn with the measured data and contour map (b) was generated with the simulated data when a gate was added at the entrance to the estuary. The sampling points in the estuary (with number 13 through 45) were marked in both maps to analyze the salinity changes in these locations.
The average values of salinity during the simulation period in the estuary are introduced to Surfer15, interpolated by the Kriging method to draw contour plots. The two contour plots of salinity gradient changes before and after sluice construction are shown in Figure 6. Contour map (a) was drawn with the measured data and contour map (b) was generated with the simulated data when a gate was added at the entrance to the estuary. The sampling points in the estuary (with number 13 through 45) were marked in both maps to analyze the salinity changes in these locations.

ET Calculation
Natural logarithms are taken on both sides of the Equation (1) to get the quadratic Equation (5), The result of a quadratic fitting of Equation (5) and the measured salinity values is: y = −0.432x 2 + 17.92x − 162 (R 2 = 0.839). The fitting curve is shown in Figure 7.

Descriptive Analysis of Salinity Values
Descriptive analysis of salinity values in 2422 computational meshes in the study area before and after sluice construction was conducted. The results are shown in Table 1.

Descriptive Analysis of Salinity Values
Descriptive analysis of salinity values in 2422 computational meshes in the study area before and after sluice construction was conducted. The results are shown in Table 1. Before sluice construction, the salinity range in Yongjiang offshore sea water varied roughly from 21.7 to 23.6% , as shown in Figure 6a, while the low salinity values were concentrated in the river channel area with a minimum salinity of 13.776% , as shown in Table 1.

Checking of Gaussian Curve
Based on the fitted one-dimension Gaussian Equation curve, a comparison can be drawn between measured phytoplankton cell density and Gaussian model theoretical values. As shown in Figure 8, the changing trend of the Gaussian model is consistent with that of the measured values, which suggests that the relationship between phytoplankton and salinity can be well described by the Gaussian Equation as calculated. Thus, the threshold zone (18.75, 23.07) (% ) of salinity calculated in this paper is adequately representative of the appropriate salinity condition to which phytoplankton growth is subject in the Yongjiang estuary. obtained in this work [47]. Yi Li et al. analyzed the correlation between Labidocera euchaeta and the salinity of sea water in the same area as Yang, finding that Labidocera euchaeta ET zone under salinity gradient ranged (18.99,23.73) (‰) and optimum salinity ET zone was (20.17,22.55) (‰) [48]. Thus, it can be argued that applying the Gaussian curve for ET studies is effective to quantify biological tolerance.

Prediction of Phytoplankton Density
After sluice construction, salinity in the estuary showed an increase since the gate acted as a physical barrier to hinder the exchange between fresh water and sea water to some degree [10,36]. According to the simulation, the maximum salinity reached 24.021‰, up from 23.643‰, while the average value surged from 22.137‰ to 23.146‰. On average, the level of salinity increased by about 1% in the entire estuary after sluice construction. At the sample point 12#, there was a notable increase in salinity, by around 0.4‰, which was close to where the planned sluice was located. For the sample points 13-15# and 17-19#, which were near the Zhoushan Bridge, their salinity showed no notable increase, as it rose by merely 0.2‰. However, a high salinity region, covering sample points 16#, 28-33#, 38-45#, was concentrated in the south of the Jintang Island, where the level of salinity increased by 0.4-0.8‰, as shown in Figure 6.
According to the salinity the ET of phytoplankton solved by the Gaussian curve, when salinity in the Yongjiang estuary falls within the range of 18.75~23.07‰, phytoplankton survives. When salinity ranges between 19.83‰ and 21.99‰, phytoplankton is left in the optimum condition with a higher phytoplankton cell density. After the sluice construction, salinity in the estuary is much higher than the threshold range suitable for phytoplankton growth. According to the Gaussian curve, the salinity range in the western sea area of the Jintang Island is 23.0 ± 0.6‰, which is beyond the optimum range for phytoplankton growth and is close to the lethal point. Moreover, the salinity value in the southern sea area of the Jintang Island reached 24.021‰, which is also beyond the lethal point, thus making it difficult for phytoplankton to grow in this area after sluice construction. As diatom, green algae and blue-green algae are the dominant groups in the phytoplankton community before The Gaussian Equation also shows its reasonability in the ET studies of species in other regions. Baoshan Cui et al. adopted Suaeda salsa (L.) Pall. biomass as an index to study the ET of the Suaeda salsa (L.) Pall. community under the scope of changes in water depth and soil salinity in the Yellow River Delta. The authors applied a Gaussian model, which revealed that the water depth in the ET zone of Suaeda salsa (L.) Pall. ranged (−0.92, 0.08) (m), and that salinity in the ET zone ranged (5.17, 20.25) (g/kg). Furthering this, a discussion was conducted around the potential impact of water-salt interaction on Suaeda salsa (L.) Pall. growth [46]. By taking the Gaussian model approach, Yiyi Yang et al. conducted a study on the relationship between phytoplankton cell density and various environmental factors at the Yueqing Bay located in Zhejiang Province, China, which is close to the Yongjiang River. It was found that the salinity in the ET zone fell within the range of (24.36,27.88) (% ), with an optimum ET range of (25.24,27.00) (% ), which is closely matched by the results obtained in this work [47]. Yi Li et al. analyzed the correlation between Labidocera euchaeta and the salinity of sea water in the same area as Yang, finding that Labidocera euchaeta ET zone under salinity gradient ranged (18.99,23.73) (% ) and optimum salinity ET zone was (20.17,22.55) (% ) [48]. Thus, it can be argued that applying the Gaussian curve for ET studies is effective to quantify biological tolerance.

Prediction of Phytoplankton Density
After sluice construction, salinity in the estuary showed an increase since the gate acted as a physical barrier to hinder the exchange between fresh water and sea water to some degree [10,36]. According to the simulation, the maximum salinity reached 24.021% , up from 23.643% , while the average value surged from 22.137% to 23.146% . On average, the level of salinity increased by about 1% in the entire estuary after sluice construction. At the sample point 12#, there was a notable increase in salinity, by around 0.4% , which was close to where the planned sluice was located. For the sample points 13-15# and 17-19#, which were near the Zhoushan Bridge, their salinity showed no notable increase, as it rose by merely 0.2% . However, a high salinity region, covering sample points 16#, 28-33#, 38-45#, was concentrated in the south of the Jintang Island, where the level of salinity increased by 0.4-0.8% , as shown in Figure 6.
According to the salinity the ET of phytoplankton solved by the Gaussian curve, when salinity in the Yongjiang estuary falls within the range of 18.75~23.07% , phytoplankton survives. When salinity ranges between 19.83% and 21.99% , phytoplankton is left in the optimum condition with a higher phytoplankton cell density. After the sluice construction, salinity in the estuary is much higher than the threshold range suitable for phytoplankton growth. According to the Gaussian curve, the salinity range in the western sea area of the Jintang Island is 23.0 ± 0.6% , which is beyond the optimum range for phytoplankton growth and is close to the lethal point. Moreover, the salinity value in the southern sea area of the Jintang Island reached 24.021% , which is also beyond the lethal point, thus making it difficult for phytoplankton to grow in this area after sluice construction. As diatom, green algae and blue-green algae are the dominant groups in the phytoplankton community before sluice construction, it is expected that the abundance of freshwater green algae, which ranks as top in the phytoplankton community in the Yongjiang estuary, will be reduced significantly, while the brackish water algae (e.g., the diatom) may be subjected to only limited impact. Since the quantity of phytoplankton is an essential index for the evaluation of primary productivity, it is foreseeable that it will show a decreasing trend in the estuary after sluice construction.

Limitation and Prospect
The ET of phytoplankton solved in this paper is limited to targeting the environmental gradient of salinity of the study area. There is the possibility that the sluice construction may affect other hydrological and nutritional factors to varying degrees, as a result of which the accuracy of the ET zone may be compromised to some extent. Besides, multiple sampling in a location is also a limitation imposed on this study. There may be autocorrelation biases of the sampling points on the same cross-section, as they are much closer to each other than the overall sampling points. The points sampled at different depths in the estuary are also the potential influencing factors for the results.
The estuarine ecosystem is highly complex in structures, which makes it necessary to conduct more systematic and comprehensive studies on ET. Based on the results obtained in this work, the dominant phytoplankton species on different nutritional levels in the estuarine ecosystems should be monitored in future research. Besides, the number of samples should be increased and the sampling locations ought to be distributed in a more sensible way. More accurate indoor breeding experiments can be conducted to study the micro physiological response of phytoplankton to the water-salt gradient and its ET under other environmental gradients. Only in this way can the state of the ecosystem be better understood.
Author Contributions: M.Y. finished the first draft of the paper and completed the calculation and verification of numerical models; C.J. edited and reviewed the manuscript; X.W. and M.Z. helped deal with the data. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.