Hou, L. J.; Li, H. L.; Zheng, C. M.; Ma, Q.; Wang, C. Y.; Wang, X. J., and Qu, W. J., 2016. Seawater-groundwater exchange in a silty tidal flat in the south coast of Laizhou Bay, China.
There were few studies about seawater-groundwater exchange in silty, low-permeability tidal flats with very gentle slopes. This paper reports the monitoring data and preliminary analytical results on a typical transect in a silty tidal flat with large-scale seepage faces at the south coast of Laizhou Bay, China. The “pair-wells method”, which was an improvement of the single-well method used by Ma et al. (2015) was used to estimate the seawater-groundwater exchange rate. We selected 14 locations along a typical transect in the intertidal zone, with a slope of 0.4‰ and cross-shore length of 3514 m, to install pair-wells for monitoring the groundwater head, salinity and temperature at two different depths simultaneously once per hour from 20:00 August 10th to 23:00 September 12th, 2014. The vertical hydraulic conductivity measured in situ ranges from 5.4 ×10−7 to 1.1×10−5 m s−1. The salinity of the underground brine ranges from 25 to 56 g l−1. Based on the observed data and the generalized Darcy's law, the submarine groundwater discharge (SGD) and inflow along the entire transect were estimated to be 164.0 and 6.5 m2 d−1, respectively. It is found that 80.4% of the total SGD occurred between W7 and W10, where the hydraulic conductivity is one or two orders of magnitude greater than that at other wells; and 64.6% of the total inflow occurred between W2 and W4. The single-well method, on the other hand, yields a SGD value of 223.0 m2 d−1 and inflow value of 13.3 m2 d−1. Neglecting the density effect may lead to an overestimation of the SGD by 25.6% and underestimation of the inflow by 27.4%.
With the new technologies and methods developed, the study of seawater-groundwater exchange has been a hot spot of hydrogeology and oceanography. Submarine groundwater discharge (SGD) has significant impacts and implications on ecological environment due to its important role in transport of nutrients, contaminations and other chemicals (Li and Jiao, 2013; Li et al., 1999; Slomp and Cappellen, 2004). An increasing number of scholars have paid attention to SGD estimation.
Many different methods have been developed to estimate SGD. Since 1996, natural radioactive isotopes Ra and Rn were widely used as tracers to estimate SGD in large scale areas (Gu et al., 2012; Lee et al., 2012; Moore, 1996, 2007; Moore and Oliveira, 2008; Wang et al., 2015; Xu et al., 2014). Seepage meters were also used to estimated SGD at local area, such as manual seepage instrument developed by Cable et al. (1997), heat pulse seepage meter developed by Taniguchi and Fukuo (1993) and ultrasonic flow meter developed by Paulsen et al. (2001). Numerical simulations usually focused on limited coastal transect, such as Guo et al. (2010), Robinson et al. (2007), Xia and Li (2012), Xia et al. (2010).
In comparison with many existing studies in gravel or sand beaches, there were few studies of seawater-groundwater exchange in silty beaches. A series of experiments conducted in gravel beaches of Prince William Sound, Alaska, indicated that the two-layered structure of a high-permeability surface layer underlain by a low-permeability layer had significant impacts on the persistence of spilled oil (Guo et al., 2010; Li and Boufadel, 2010; Xia et al., 2010). Xia and Li (2012) made a comparative study between a bald beach and a mangrove beach with similar topography and tidal actions in the intertidal zones of Dongzhaigang in Haikou City of Hainan Province, indicating that the inland freshwater recharge and the mud-sand two-layered structure played key roles in the development of mangrove and their spatial distribution.
Ma et al. (2015) developed a simple method (“single-well method”) for estimating seawater-groundwater exchange rate in a silty beach, taking into account the effects of seepage face and density and based on field measurements of groundwater hydraulic head, temperature and salinity. The exchange rate can be calculated by the generalized Darcy's law, the observed tidal level and the hydraulic head in the observation well, Their case study in a silty beach at Haimiao of Laizhou Bay indicated that the SGD and inflow were 8.8 and 15.3 m3 d−1 m−1, respectively. A freshwater discharge tube was also identified near the low-tide line. Riedel et al. (2010) quantified SGD in a tidal sand flat margin with a seepage face in the German Wadden Sea using the numerical code SUTRA, and found that the magnitude of SGD was controlled by the maximum hydraulic gradient during low tide.
In this paper, “pair-wells method” is used to improve the single-well method used by Ma et al. (2015) to investigate the seawater-groundwater exchange rate in a typical cross-shore silty-beach transect in the intertidal zone of Qingxiang Town (Liutuan Town), Laizhou Bay. Fourteen pair-wells were used to monitor the pore water pressure, salinity and temperature at two different depths simultaneously along the entire transect. Based on these observed data, we quantitatively estimated the seawater-groundwater exchange rate. The impacts of density and groundwater pumping for salt production on SGD and inflow were discussed quantitatively. The difference between the results of this study and those of Ma et al. (2015) were analysed. Comparisons were made between the single-well and pair-wells methods. The advantages of the pair-wells method over the single-well one are discussed.
FIELD DESCRIPTION AND METHODS
Laizhou Bay, an important fishing and salt production base of China (Figure 1), is located in the south of Bohai Sea, north of Shandong Peninsula. The beach in the southern coast of Laizhou Bay (from Hutouya to Yangjiaogoukou) is very flat, and the width of the intertidal zone in the cross-shore direction is about 4000 m. The tidal flat in the study area consists of silty-muddy sediments, and the aquifers are mainly composed of late Pleistocene to Holocene sediments. Brine in Quaternary aquifers is mainly distributed within 10 km of the shoreline (Han et al., 2014; Zhang and Peng, 1998). There are three layers containing brine in the sediments (Xue et al., 2000), which has resulted from a combination of past and recent processes, including a number of phases of regression, transgression, brine formation and mixing (Han et al., 2014; Li et al., 2000; Xue et al., 2000).
The weather in Laizhou Bay belongs to sub-humid monsoon climate, with an annual average temperature of 11.9°C (Han et al., 2014). The annual mean precipitation is 660 mm, and 60–70% of the precipitation generally concentrates in June–August. The annual average potential evapotranspiration, approximately 1774 mm, is much greater than the precipitation (Bi et al., 2012; Han et al., 2014). Under natural conditions, the unconfined aquifer is recharged by precipitation and groundwater from south (Han, Wu, and Zhang, 2011; Han et al., 2014). The area is exposed to semi-diurnal tides, and the average durations for both rising and falling tides are about 6 hours.
The typical transect that we selected in the intertidal zone is located in Qingxiang Town (Figure 1). Along the entire transect, we dug 14 pits to install pair-wells, which are numbered as W1, W2, …, and W14 sequentially seaward (Figure 2). The beach surface elevations of the pair-wells were surveyed by an electronic total station (TS–800, Johanna). These data and the elevations of each pressure transducer installed at the 14 observation pair-wells were given in Table 1. There were several times of raining during the monitoring period (from 20:00 PM August 11th to 23:00 PM September 12th, 2014, almost 33 days), with the precipitation ranging from 28.3 mm to 51.6 mm.
Elevations of beach surface and each LTC-Diver at the 14 observation wells
Each of pair-wells consists of two LTC-Divers (Level, Temperature and Conductivity Diver; Levelogger Junior, Solinst, Canada), inner and outer PVC pipes, and a plexiglass rod linking the two LTC-Divers together (Figure 3a). The pipes were slotted uniformly with holes about 5.0 mm in diameter to keep a good hydraulic connection around the pair-wells. Furthermore, to prevent holes from being blocked by silty-muddy sediments generally occurring in the silty beach, fine stainless-steel meshes were screened outside the two pipes, and the room between the two pipes was filled with coarse sands (diameter of 0.5 – 2.0 mm). Then, a LTC-Diver was placed in each inner pipe (Figure 3b). In each pit, a pair of wells were installed to record the pore water pressure, temperature and conductivity hourly at two different depths (Table 1). At the same time, a LTC-Diver was installed in the sea to monitor the tidal fluctuation; and a Baro-Diver (Schlumberger) was placed above the ground surface to obtain the air pressure and temperature once an hour.
Based on the definition of salinity given by UNESCO (1966), the electrical conductivity measured by a LTC-Diver can be transformed into salinity. More details can be seen in Cox, Culkin, and Riley (1967); Ma et al. (2015); and, Wooster, Lee, and Dietrich (1969). Then, the observed pressures of air and ground-water, and elevations of each LTC-Diver, were transformed into freshwater-equivalent hydraulic head (Figure 4).
The in-situ multiple diameter falling-head method (Wang et al., 2014) was used to measure the vertical hydraulic conductivity Kv(Table 2) of intertidal sediments at the locations of some pair-wells (W2, W4, W6, W8 and W12). More details about the apparatus and method can be found in Li et al. (2010), Ma et al. (2015), and Wang et al. (2014).
Kv–values obtained by the method in Wang et al. (2014)
Seepage Face Development Condition
The presence/absence of seepage face development is defined by the characteristics of tidal actions and tidal flat face in the study area (Turner, 1995).
In our study area, RT = 0.25–1.6 m, T = 12 h, KH = 1.1 × 10−5 m s−1(Table 2), β = 0.4‰, n = 0.03 – 0.35. Considering that the horizontal hydraulic conductivity is greater than or equal to the vertical one (KH ⩾ KV) generally, a maximum value of the vertical hydraulic conductivity (1.1×10−5 m s−1) was used for conservative estimation. No measurements were conducted for the effective porosity of the silty beach sediments. According to Fetter (1994), the porosity of silt, silt sands, fine sands and sandy clay ranges from 0.03 to 0.35, which is used here for conservative estimation. According to these parameters and Equation (1), the seepage face parameter ranged from 108 to 1010, much greater than 1. Therefore, seepage faces will develop during low tides in the gentle silty tidal flat.
Direct Estimation of Water Exchange Rate
The observed data and preliminary analysis (Figure 4) indicated that the beach surface between W2 and W14 was saturated during the entire monitoring period. The generalized Darcy's law considering the density effect can be used to estimate the vertical flow rate by the following equation (Ma et al., 2015):
The parameter ϵ is a constant, equaling 7.143×10−4 m3 kg−1, which is used to describe the linear relationship between density and salinity, which is defined by
Therefore, the generalized hydraulic gradient J is defined as:
The generalized hydraulic gradients J of the 12 pair-wells (W2–W8, W10–W14) were shown in Figure 5.
If J >0 (i. e., qn >0), the groundwater flows upward (SGD); otherwise, it flows downward (inflow). Normally, the density effect leads to downward flow due to the positive parameter ϵ.
As shown in Figure 2, the beach surface of W1 is located in the high intertidal zone, and the low LTC-Diver was inundated only during 7.92–9.67 d, 17.25–19.54 d, and 21.92–23.67 d (Figure 4a). In other words, the LTC – Divers installed at W1 were above the water table and did not able to record the ground-water data during most of the observation time. Besides, the salinities at up and low LTC-Divers were 0 since no groundwater entered the inner pipe when the LTC-Diver were above the water table. Thus, the seawater-groundwater exchange estimation later did not include W1.
As shown in Figure 4, the observed hydraulic heads at the up and low LTC-Divers of observation wells W2–W14 decreased with the falling tide when the tidal level was higher than the beach surface, and became almost constant when the tidal level fell near the beach surface until the next rising tide coming. Note that during the exposure of beach surface, the hydraulic heads of the two divers were higher than the beach surface. Furthermore, the hydraulic head of the low LTC-Diver was always higher than that of up LTC-Diver, especially at W2, W4 – W8, and W13. The generalized hydraulic gradient between the up and low LTC-Divers was positive (except W11) in most time which is about 0.2 at W7 and W8 (Figure 5b), indicating considerable groundwater discharge.
The salinity difference between the up and low LTC-Divers was very small for W2-W14, and was less than 2 g l−1 at W10 and W11 (Figures 4j and 4k). Note that as shown in Figure 4, the salinities at low LTC-Diver at W2, W4, W7 and W8 were 0, probably because the pipe was blocked by fine sediments, which exist extensively in the silty tidal flat, or there were bubbles around the LTC-Diver preventing water from entering. The sudden decrease of hydraulic heads at W3 – W5 (Figures 4c, 4d, and 4e) after 18 d may be caused by the pumping of groundwater for salt production nearby (Feng, Tan, and Liu, 2013; Yu, 2014).
At W9, the low LTC-Diver did not worked (Figure 4i), thus the seawater-groundwater exchange estimation at W9 was not included. As shown in Figure 4i, the hydraulic head at up LTC-Diver was about 0.15 m higher than the beach surface, indicating that the groundwater discharge from the seepage face developed here and might be significant.
Evaluation of SGD and Inflow
The exchange rate qn given by Equation (2) was averaged over the entire monitoring period for the outflow qout and inflow qin at each well using the following equations:Table 3), then the total SGD and inflow could be evaluated by the integral along the x–axis, namely:
Exchange rate qin, qout of every well (except W1 and W9), and corresponding inflow and SGD between the adjacent wells estimated by the pair-wells method along the entire transect; and those (Inflow-s and SGD-s) obtained by the single-well method
The values of qout (or qin) between Xi and Xi+1 are determined by linear interpolations of their values at Xi and Xi+1(i = 2, …, 8, 10,…, 12, 13).
The estimated results of the seawater-groundwater exchange rate at each well are shown in Table 3. The calculated SGD along the entire transect was 164.00 m2 d−1 and inflow was 6.5 m2 d−1. As shown in Figure 6, the inflow concentrating between W2 and W4, accounted for about 64.6% of the total inflow. The SGD concentrated between W7 and W10 (Figure 6), accounting for about 80.4% of the total SGD, which is probably due to the high-permeability (10−5 m s−1 there, one or two orders of magnitudes greater than those of other pair-wells locations) nearby (Table 2). Under the rainfall condition, the SGD always increases with more inland freshwater recharge. The SGD and inflow are season-dependent. It would be different in the wet and dry season with varying inland freshwater recharge, and it still needs further research.
When removing the observation periods of W3 to W5 influenced by groundwater pumping (i. e., when the water pressure equaled essentially zero), the SGD and inflow along the entire transect were 170.7 and 5.6 m2 d−1, i. e., the total SGD increased by 6.7 m2 d−1 (about 4.1%), and the total inflow decreased by 0.9 m2 d−1 (about 14.1%), respectively. Though groundwater pumping had little effect on the water exchange of the entire transect, it did have great impact on the zone between W3 and W5 (details will be given later).
The Impacts of Groundwater Pumping for Salt Production
As shown in Figures 4c–4e, the hydraulic heads and salinities of W3–W5 showed a sudden drop between 18.5 d ~ 20 d, even close to the elevations of LTC-Divers sometimes. This was most likely caused by groundwater pumping for salt production nearby. With pumping effect, the generalized hydraulic gradient changed from positive to negative suddenly, and gradually recovered to normal state after t = 20 d (Figure 5a). Meanwhile, the hydraulic heads and salinities recovered gradually within two days. This phenomenon was not observed at W2 and W6, indicating that pumping only had an impact on local area.
The SGD and inflow between W3 and W5 were 6.9 m2 d−1 and 46.7 m2 d−1 during the pumping period, respectively. The SGD and inflow were 12.3 and 0.6 m2 d−1 when ignoring the pumping effects. The SGD decreased by 5.4 m2 d−1, and the inflow increased by 46.1 m2 d−1 with the pumping effects. During the pumping period, the maximum inflow between W3 and W5 increases to 217.3 m2 d−1. These results indicated that short-term high-rate pumping had great influence on the interactions of seawater-groundwater between W3 and W5.
Comparison with the Single-well Method
We also estimated the SGD and inflow using the data from low LTC-Divers and tidal level. The difference between the two estimation methods is how to calculate the vertical groundwater flow rate qn. The vertical groundwater flow rate is calculated by the single-well method as following:Equation (2); ΔL is the depth of the low LTC-Diver from the beach surface [ L] (Figure 2 and Table 1). The parameter hsurface is defined as following: Ma et al. (2015).
The calculated seawater-groundwater exchange rates using the single-well method are shown in Figure 6. As shown in Table 3, the estimated SGD–s and inflow–s (SGD–s and inflow–s, the SGD and inflow calculated using the single-well method) are 223.0 and 13.3 m2 d−1, much greater than those estimated by the pair-wells method. This might be related to the following aspects:
(1) The salinity difference between seawater and underground brine. As shown in Figure7, the underground brine salinities are between 30 and 40 g l−1, up to about 56 g l−1, which was much higher than the averaged seawater salinity about 23 g l−1. When the averaged salinity between seawater and groundwater was used to estimate the vertical groundwater flow rate qn, which will be less than the averaged salinity of pair-wells.
(2) The different tidal levels at different locations. The tidal level will decrease with the spread of tide landward (Lian and Zhao, 1995; Wang et al., 1995), due to the influence of friction of beach surface (Kamphuis, 1978), and the coefficient of friction increases with decreasing depth (Yu and Li, 2002; Zhang, Lv, and Sun, 2012). But the differences of the tidal level at different locations are neglected in the single-well method. What's more, the SGD and inflow estimation in the pair-wells method used the hydraulic head at the up LTC-Diver, which is more close to the real hydraulic head at the observation well locations than the tidal level or beach surface here used in the single-well method.
Based on these reasons, the estimated results and errors of the single-well method would be greater than those by the pair-wells method.
Density Effect and Salinity Distribution
The generalized Darcy's law used in both the pair-wells method and the single-well method above considers the density effect. The salinity along the entire transect is great, which will have significant impacts on the seawater-groundwater exchange.
Ignoring the density effect, the vertical groundwater flow rate qn can be written as:
The estimation of SGD by Equation (12) increased by 25.6% to 206.0 m2 d−1, and inflow decreased by 27.4% to 4.7 m2 d−1, compared to those estimated by pair-wells method. Due to the high salinity of underground brine, the influence of density effect can change the exchange rate by nearly 25%, indicating that much downward groundwater water flow driven by high density brine.
Considering that the salinity at every LTC–Diver was relatively stable (Figure 4), the salinity observed at each LTC–Diver was averaged and the results were shown in Figure 7. One can see that along the entire transect the salinity decreased from 56.2 to 25.7 g l−1 seaward gradually, but all of them were higher than the seawater salinity, averaged 22.6 g l−1(Figure 7). The reasons for this might include the following aspects: (1) evaporation (the exposure time of the tidal flat increases landward, leading to a landward increasing-evaporation and in turn a landward increasing of salinity), (2) human activities such as exploitation of the underground brine and construction of many salt pans, as shown in Figure 1c (such activities often discharge waste salt brines into the coastal areas, leading to the high salinity there), and (3) the inland fresh groundwater discharge near the low tide zone (Heiss and Micheal, 2014; Li, Boufadel, and Weaver, 2008; Robinson et al., 2007), which may lead to the decrease of salinity of the groundwater in the lower intertidal zone.
Comparison with Ma et al. (2015)
The estimated SGD by the pair-wells method in this study was 164.0 m2 d−1, much greater than the SGD about 8.8 m2 d−1 calculated in Ma et al. (2015); while the inflow along this transect was 6.5 m2 d−1, less than that nearly 15.3 m2 d−1 in Ma et al. (2015) (Table 4). The differences in SGD and inflow between the two results are significant since there are many differences between the two transects.
The comparison of the two typical transects in this paper and Ma et al. (2015)
The length of the cross-shore transect in this paper is 3514 m, which is about nine times long of the transect (392 m) in Ma et al. (2015) (Table 4). There were six single-well along the entire transect in Ma et al. (2015), while 14 pair-wells were set along our transect in this paper (Figures 1 and 2). The selected cross-shore transect in Ma et al. (2015) has two distinct slopes, with the landward part of a relatively steep slope about 2%, and the seaward part of a gentle slope about 0.1%. The slope of the selected transect in this paper is about 0.4 ‰, which is much smaller than the slopes in Ma et al. (2015) (Table 4).
Field measurements were conducted from 11:00 September 9th to 15:00 September 14th, 2012, about 125 hours in Ma et al. (2015) (Table 4). The observation time from 20:00 August 10th to 23:00 September 12th, 2014, almost 800 hours in this paper, is very long compared with most previous studies (Table 4).
The selected transect in this paper is heterogeneous along the cross-shore direction. The permeability ranges from 10−7–10−5 m s−1, which is one or two orders of magnitude less than that about 10−5 m s−1 in Ma et al. (2015) (Table 4). There are little water flowing downward due to the low-permeability surface sediments and the higher hydraulic head of the groundwater when surface was submerged.
The salinity of groundwater along the transect in Ma et al. (2015) ranged from 16 to 27 g l−1, which is less than the salinity of seawater. The transect in this study area is located in brine zones, and the salinity of groundwater at observation locations ranged from 25 to 56 g l−1, which is greater than the seawater salinity averaged about 23 g l−1(Figure 7). The details of the salinity distribution have been discussed in the “Density effect and salinity distribution” section.
This paper investigated the seawater-groundwater exchange in a silty tidal flat with a large-scale seepage face in the south coastal of Laizhou Bay. We selected 14 locations along a typical transect in the intertidal zone with the slope of 0.4‰ and the cross-shore width of 3514 m, to install observation pair-wells monitoring the groundwater head, salinity and temperature at two different depths simultaneously. The vertical hydraulic conductivity ranges from 5.4×10−7 to 1.1×10−5 m s−1, and the salinity of the underground brine is from 25 to 56 g l−1 along the selected transect. Basing on the observed data from 20:00 August 10th to 23:00 September 12th, 2014 and generalized Darcy's law, the submarine groundwater discharge (SGD) and inflow along the studied transect were estimated to be 164.0 and 6.5 m2 d−1, respectively; within 80.4% of the total SGD concentrated between W7 and W10 due to the high-permeability here, and 64.6% of the total inflow between W2 and W4.
The comparative study between the pair-wells method and single-well method indicated that the pair-wells method is more suitable for the long transect with a very gentle slope than the single-well method, due to the pair-wells can monitor the groundwater head, salinity and electrical conductivity at two different elevations simultaneously. The fixed distance of the pair-wells can avoid the uncertain errors of the single-well method.
The observed data indicated that pumping of groundwater has significant impacts on the exchange between seawater and ground-water. During the groundwater pumping period, the outflow decreased by 5.4 m2 d−1 and the inflow increased by 46.1 m2 d−1, even up to 217.3 m2 d−1 within the tidal zone from W3 to W5.
The density effect had great impacts on the seawater-groundwater exchange in the brine zones due to its high salinity ranging from 25 to 56g l−1 along the entire transect. Leaving out the density effect, the exchange rate can change 25%, indicating that much downward groundwater flow driven by the density effect.
The salinity decreased seaward gradually along the entire transect, probably because the evaporation, human activities such as exploitation of the underground brine and construction of many salt pans, and the inland fresh groundwater discharge near the low tide zone.
This work was supported by the Key Program of National Natural Science Foundation of China (Grant No. 41430641), the National Basic Research Program of China (“973” Program, Grant No. 2015CB452902). We would like to thank Kai Xiao, Zongzhong Song, Shuaishuai Liu, Yan Zhang and Shaohong Li for their field work.