- Email: [email protected]
Determination of hydraulic flow patterns in constructed wetlands using hydrogen and oxygen isotopes
Journal of Molecular Liquids 223 (2016) 775–780
Contents lists available at ScienceDirect
Journal of Molecular Liquids journal homepage: www.elsevier.com/locate/molliq
Determination of hydraulic flow patterns in constructed wetlands using hydrogen and oxygen isotopes Haimeng Sun a, Zhen Hu a, Jian Zhang a,⁎, Weizhong Wu b, Shuang Liang a, Shaoyong Lu c, Huaqing Liu a a b c
Shandong Key Laboratory of Water Pollution Control and Resource Reuse, School of Environmental Science & Engineering, Shandong University, Jinan 250100, PR China Department of Environmental Science, College of Environmental Sciences and Engineering, Peking University, Beijing 100871, PR China Research Centre of Lake Environment, Chinese Research Academy of Environmental Sciences, Beijing 100012, PR China
a r t i c l e
i n f o
Article history: Received 6 April 2016 Received in revised form 13 July 2016 Accepted 10 August 2016 Available online 05 September 2016 Keywords: Constructed wetlands Stable isotopes Hydraulics Flow patterns
a b s t r a c t The treatment efficiency of constructed wetlands (CWs) is highly dependent on the stability of the hydraulic flow patterns. To date, general technologies used to study hydraulic flow patterns of CWs mainly include tracer method, model simulation and velocity measurement, which are either expensive, empirical, or having secondary pollution. In this study, a new technology, which was based on the isotopic composition variation in CWs, was applied to detect the hydraulic flow patterns of two different CWs. Results showed that the hydraulic flow patterns of the two studied wetlands could be detected effectively by using hydrogen and oxygen isotopes. Furthermore, the locations of stagnant areas (SAs) and preferential flow areas (PFAs) were also determined. Significant regional difference in isotopic composition existed inside each CW, and two wetland design suggestions are proposed after hydraulic analysis. One is that the influent of CWs is supposed to be distributed uniformly, and another piece of advice is that the vegetation in the direction perpendicular to water flow should be maintained at the same types and density. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Constructed wetlands (CWs), which are designed on the basis of all kinds of reaction process in natural wetlands but make the process occur in more controlled systems, have a rapid momentum of development especially in last three decades [1]. These systems, which are robust, low power consumption, easy to maintain and operate, are suitable for advanced treatment of municipal sewage plants' effluent or decentralized wastewater [2,3]. The pollutants such as suspended material, nitrogen, phosphorus, organic matters, and metals are removed by the complex and abundant reactions relying on physical, chemical and biological processes [4]. Pollutants removal mechanisms are mainly organic matters decomposition by microorganism, nitrogen removal through nitrification – denitrification, and nutrient elements absorption by the plants and substrate. The implementation of these reactions depends on the joint function of plants, matrix, and microorganisms. To a certain extent, nutrients removal efficiencies are dependent on the time that waste water touched with vegetation, matrix, and microorganisms in the CWs. As a result, hydraulic retention time (HRT) has significant effect on water pollutants removal efficiency [5–7]. Furthermore, CWs' water flow is rather complex because of its large areas of plants and artificial matrix. Hydraulic performance of CWs is influenced by several factors, such as plant ⁎ Corresponding author. E-mail address: [email protected] (J. Zhang).
http://dx.doi.org/10.1016/j.molliq.2016.08.115 0167-7322/© 2016 Elsevier B.V. All rights reserved.
species richness and density, the CW's aspect ratio, the inlet and outlet's form [8,9]. A combination of these factors leads to different hydraulic conditions inside the CWs, and the hydraulic flow patterns are the comprehensive performance of all hydraulic conditions which also are the external presentation of the residence time distribution (RTD) [10]. Since RTD is a value of process quantity, it can't reflect the hydraulic conditions inside the CWs in situ. It is important to investigate the hydraulic variation inside CWs, which can help us find the unfavorable hydraulic phenomenon such as stagnant areas (SAs) or preferential flow areas (PFAs). At present, many researchers focus on the application of tracer, hydraulic models and velocity measurement to simulate hydraulic flow patterns in CWs [11–15]. However, in large-scale CWs, adding tracers is not a good choice because of the high cost, potential of secondary pollution caused by adding external material into the CWs, and the intensive continuous monitoring time [16]. As to hydraulic model, its results can not reflect the hydraulic flow patterns synchronously as the parameter is always empirical [17]. Finally direct velocity measurement inside the CWs needs expensive apparatus, not to mention that field measurement is considered a labor-consuming job [18,19]. Stable isotopes of oxygen and hydrogen are a novel technology which has been used in hydrology for investigating water composition and circulation in rivers, lakes, oceans, ground water and atmosphere [20–22]. In liquid water, isotopes molecules' vapor pressure is inversely proportional with its molecular weight. Compared to 18O and 2H, 16O and 1H have higher vapor pressure, indicating that they could separate
776
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
from the liquid phase more easily, which lead to the enrichment of the heavy water isotopes 18O and 2H [23]. In CWs, the cumulative amount of evaporation in different regions has strong relationship with regional water residence time [24]. Moreover, regional water residence time determined the flow patterns. If an accurate survey was done on the distribution of water isotopic composition, the hydraulic flow patterns would be known inside CWs. The isotopes technology utilizes the internal elements in CWs to detect the hydraulic flow patterns real-timely and has a low cost without secondary pollution. The aims of the present study were to characterize hydraulic flow patterns using isotope technology, further, to determine the locations of the SAs and PFAs in CWs. After hydraulic analysis, some engineering suggestions are proposed. The ammonia nitrogen's distribution characteristics in different hydraulic conditions were studied as well. 2. Materials and methods Fig. 1. The schematic diagrams of the W-CW (a) and Q-CW (b). Circles represent sampling points (21 points in W-CW while 31 points in Q-CW), and squares stand for the inlets. The bold lines instead for the outlet and the dashed lines instead for the overflow weir.
2.1. Sites description Two sites were identified based on differences in shapes, plants species and inlet patterns: Wu river constructed wetland (W-CW) (N: 34° 52′6.9″; E: 118°20′42.7″) and Qihe ecological constructed wetland (QCW) (N: 36°56′26.9″; E: 116°47′24.7″). Both sites belong to monsoon climate of medium latitudes, with four distinctive seasons. Both the CWs accept treated sewage water from local sewage treatment plants, and the water quality met Chinese National Class I (Grade A) Sewage Discharge Standard. Detailed information of the two CWs are given in Table 1. The surface flow constructed wetlands (SFCWs) in the two CWs were selected as our experimental subjects to study the hydraulic flow patterns using isotopic technology and velocity distribution. In WCW, the experimental subject was the SFCW between second and third over flow weir, which was about 16.1 ha and straight type without bend, a single culvert set as the inlet pattern (Fig. 1a). In Q-CW, the experimental subject was the five surface flow treated units, which was about 3.5 ha and with two bends in shape. Inlet pattern was several culverts distributing in the bank, an overflow weir set between every two unit for oxygen supplement (Fig. 1b). Vegetation coverage area was another obviously different factor between the two studied wetlands. According to the method proposed by Jiang et al. [25], vegetation coverage ratios were 26% and 83% for W-CW and Q-CW, respectively. W-CW's flow channel was constructed on the river watercourse (named watercourse channel type), mainly water areas. Yet the Q-CW was built artificially (named man-made channel type), which had abundant vegetation species and high density. Vegetation types were mainly emergent plants and submerged plants. The emergent plants were mainly Typha orientalis and Phragmites australis while the submerged plants were mainly Vallisneria natans and Nymphaea tetragona. It would be specially mentioned that a middle inlet was set in the second level wetland, through which a small part of outflow of sub-surface CWs flew into the SFCW directly when high water inflow rate occur. Coincidentally, the middle inlet was open when the measurement was conducted, which increased hydraulic changes. Sampling points in this study were set in the areas that had different hydraulic conditions such as open water areas, various vegetation species coverage areas, inlet and outlet regions (Fig. 1). Isotopes were all collected at about 10 cm below the water surface. The sampling job was conducted after consecutive ten more days without rain. The
collected water was put into 250 mL plastic bottles then preserved into insulation box with ice bags in it. A GPS (72H) was used to determine the sampling points' location. Water sampling job was conducted in August 2013 for W-CW, and June 2015 for Q-CW. There were 21 and 33 sampling points designed in W-CW and Q-CW, respectively (Fig. 1). 2.2. Water isotopes analysis Isotopic composition is expressed δ in ‰ as: δX E ¼
Rsample −1 1000 Rstandard
ð1Þ
where X is the atomic mass of the heavy isotope of element E and R is the ratio of the heavy to light isotope (18O/16O). Rsample is the ratio of heavy (e.g. 2H and 18O) to light (e.g., 1H and 16O) isotope of water samples; Rstandard is the ratio of heavy (e.g., 2H and 18O) to light (e.g., 1H and 16 O) isotope related to VSMOW standard. The isotopic composition of oxygen and hydrogen was analyzed by wavelength-scanned cavityring-downs pectroscopy (WS-CRDS) [26], using Picarro L2140-i δ18O/δ17O/δD/17O-excess high-precision isotopic water analyzer in Beikerui Detection Technology Co., Ltd. The relationship between surface water signatures of δD and δ18O relative to the local meteoric water line (LMWL) showed the degree of evaporation intuitively [27]. Since the experimental subjects located in eastern China, the eastern China meteoric water line (ECMWL) proposed by Yu et al. [28] was selected as LMWL. In order to show the distribution of the isotopes clearly, Surfer 10.0 was used to give a δ18O distribution contour map of the wetland using an ordinary kriging interpolation [29]. If a statistically significant difference in the isotopic distribution was found inside two studied wetlands, it indicated that the hydraulic conditions heterogeneity existed. In the areas of contour values were high, indicating that high evaporation leading to heavy isotopes enrichment, where a long HRT and SAs might exist. On the contrary, in the PFA, the short HRT lead to relative low evaporation enrichment, so small numerical isotopic values line would be distributed in these areas. In the δ18O distribution maps, the range of measured δ18O values would be
Table 1 Properties of the studied wetlands. Parameters Wetlands
Area (ha)
Influent type
Flow rate (m3/d)
Experimental subject (ES)
ES's area (ha)
ES's shape
Plant coverage ratio
W-CW Q-CW
533.33 68.53
Treated sewage water Treated sewage water
3.0 × 105 4.0 × 104
Second surface flow treated unit 5 surface flow treated units
16.1 3.5
Straight type Turning type
26% 83%
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
777
divided into twelve value scales, and the areas of the top two value levels distributed were considered as the SAs, while the areas of the minimum two value levels distributed were considered as PFAs. 2.3. Flow field measurement To verify the stability of water isotopes technology, three-dimensional flow velocities were measured using acoustic Doppler velocimeters (ADVs) at every sampling point. ADV, one of a common technology of Doppler methods [30], has facilitated technologies for flow field measurement, such as current meters technology applied in velocity and turbulence measurements in a two- or three-dimensional flow [31]. At each testing point, triplicate measurements were conducted, and the final results were the average of the three measurements. The results were showed as three-dimensional vector velocity, but in this study, the absolute values of the velocity were more concerned, so three-dimensional vector velocity was converted into absolute velocity. Velocity results were used to get a continuous distribution of the velocity by kriging with a trend, and then the velocity distribution was obtained, from which the hydraulic flow patterns would be showed clearly. On the velocity distribution maps, the range of measured velocity values would be divided into twelve value scales as well, and the areas of the top two value levels distributed were considered as the PFAs, while the areas of the minimum two value levels distributed were considered as SAs. 2.4. NH+ 4 -N concentration analysis Water sample of every point in the two studied wetlands was taken to laboratory and analyzed for NH+ 4 -N concentration according to standard methods [32]. In general, the treatment processes such as mass sedimentation, volatilization, sorption, and transport are based on first-order kinetic models. This model is given as follows: C−C ¼ expð−K v T Þ C 0 −C
ð2Þ
+ where, C is the NH+ 4 -N concentration (mg/L), C0 is the influent NH4 -N ⁎ concentration (mg/L), C is the background concentration (mg/L), T is the retention time (d) and Kv is the first-order volumetric rate constant (d−1). As shown in the formula, ammonia nitrogen removal efficiency is related with residence time T and Kv. In SAs and PFAs, many factors can affect the Kv and residence time T. The NH+ 4 -N concentration would present different characteristics in SAs and PFAs. In order to link the + NH+ 4 -N distribution with the hydraulic flow patterns, NH4 -N concentration contour distribution maps were given by using Surfer 10.0 for the studied wetlands.
3. Results
Fig. 2. The statistics results of the δ18O (a), velocity (b) and ammonia nitrogen (c) of the two CWs.
3.1. Isotopic distribution The δ18O values ranged from − 6.05‰ to −5.40‰ for W-CW, and ranged from −7.85‰ to −7.14‰ for Q-CW (Fig. 2a). Linear correlation between δ18O and δ2H in both wetlands were δ2H = 1.99δ18O-34.85 (Rsquare was 0.80) for W-CW and δ2H = 5.14δ18O-17.71 (Rsquare was 0.97) for Q-CW (Fig. 3). The linear relation in both wetlands were deviated from ECWML δ2H = 7.8δ18O + 6.6 [28], which proved that the evaporation of water had obvious influence on the isotopic composition of both wetlands. Because of the difference in the sampling time, δ18O values were higher in W-CW than that in Q-CW (Fig. 3). The measurement was conducted in August for W-CW, and the temperature was higher than that in June, when the measurement in Q-CW was conducted. The evaporative enrichment in water would be enhanced with temperature increased, which resulted in higher δ18O composition in WCW. In this study, the variation trend of δ18O inside experiment subjects
was the factor to be concerned rather than the different δ18O composition between the two subjects. Interpolated maps of δ18O distributions of the two wetlands are given in Figs. 4a and 5a, respectively, where the variation shows clearly inside the two studied wetlands. In accordance with the provisions of Section 2.2, the marginal values of the SAs and PFAs were −5.50‰ and −5.94‰ for W-CW, while the marginal values of the SAs and PFAs were − 7.22‰ and − 7.72‰ for Q-CW. Two SAs and one PFA were determined in W-CW (Fig. 4a) and four SAs and one PFA were found in W-CW (Fig. 5a). 3.2. Velocity distribution As could be seen from the flow velocity measurement results (Fig. 2b), mixed flow existed in each CW, which indicates a spread of residence time. Compared with Q-CW, the amplitude of velocity
778
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
concentration in SAs was 1.84 mg/L which was close to the lowest concentration. In the areas outside of the SAs of the studied wetlands, the NH+ 4 -N concentration variation was relative stable. 4. Discussion 4.1. Hydraulic flow patterns determined by isotopic distribution
Fig. 3. Relationship between surface water signatures of δ2H and δ18O for W-CW and QCW relative to the eastern China meteoric water line (ECMWL).
fluctuation in W-CW (0.002 m/s to 0.017 m/s) was more stable than that in Q-CW (0.002 m/s to 0.038 m/s), which presumably due to the existence of middle inlet set in the second treatment unit in Q-CW. The average velocity was 0.0139 m/s and 0.0207 m/s before and after middle inlet, respectively. The simulated flow velocity distributions in the W-CW and Q-CW are presented in Figs. 4b and 5b, and both showed regional changes in the CWs clearly. Velocity distribution showed similar hydraulic flow patterns with δ18O distribution. In Fig. 4b, two SAs and one PFA were determined in W-CW, and their locations were similar with that in Fig. 4a. The same experiment results were found in Fig. 5a and b. 3.3. NH+ 4 -N distribution NH+ 4 -N concentration values showed regional differences as well (Fig. 2c). In both studied wetlands, NH+ 4 -N concentration declined gradually from inlet to outlet (Figs. 4c and 5c). The highest NH+ 4 -N concentration was 2.83 mg/L detected around inlet in W-CW, while the highest NH+ 4 -N concentration was 2.90 mg/L observed in the area around the middle inlet (Figs. 4c and 5c). The NH+ 4 -N distribution was highly different in SAs between the studied wetlands. In W-CW, the mean NH+ 4 -N concentration in SAs was 2.57 mg/L which was close to the highest concentration. On the contrary, in Q-CW, the mean NH+ 4 -N
Air temperature was 26.4 °C when the measurement was conducted in W-CW, and improved to 33.8 °C when the measurement was conducted in Q-CW. Evaporation rate in the Q-CW was higher than that in the W-CW due to the higher temperature. In this case, the enrichment of heavy isotopes would be intensified in Q-CW. As a result, isotopic composition in Q-CW was significant higher than that in W-CW. Immediately after influent mixed with residual water in the wetland, water budget began to change due to factors including precipitation, evaporation, absorption and interaction with groundwater [33], which might cause δ18O variation in the treated water. As 2–5 days without precipitation was enough to obtain the required isotope enrichment [34], the precipitation could be neglected when the sampling was conducted after consecutive ten more days without rain. In the studied wetlands, the well-defined facility of substrate was built in the bottom to prevent water infiltration, which led to negligible interaction with groundwater, so the groundwater's influence on the isotopic composition in the surface wetland could be neglected. Furthermore, absorption by vegetation's roots had little impact on the isotopic composition around the roots [35]. In such circumstances, evaporative enrichment was the decisive factor towards isotopic distribution. Due to the climatic condition was same inside both studied wetlands, the evaporation resulting isotopes variation was caused by the HRT. In the studied wetlands, an obvious difference was found in isotopic distribution, which indicated a different HRT distribution in different regions, and the hydraulic flow patterns were showed clearly (Figs. 4a and 5a). Flow exchange was weak between the stagnant areas and smooth flow areas, so the flow velocity was faster in the smooth flow areas. Compared to water flow in stagnant areas, cumulative amount of evaporation was lacking in the smooth flow areas and flow to the outlet rapidly, so the isotopic composition in some regions close to the outlet were lower than that in some regions close to the inlet. Small δ18O values and gradient indicated preferential flow in the front of the inlet in wetlands (Fig. 4a), the same flow condition was determined around the middle inlet in Q-CW (Fig. 5a). This was supported by velocity distribution (Figs. 4b and 5b). δ18O distribution results showed two and four SAs
Fig. 4. The isotopes δ18O distribution (a), velocity distribution (b) and ammonia nitrogen distribution (c) in W-CW.
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
779
Fig. 5. The isotopes δ18O distribution (a), velocity distribution (b) and ammonia nitrogen distribution (c) in Q-CW.
in W-CW and Q-CW, respectively. Isotopic δ18O distribution and velocity distribution correspond very well. Hydraulic flow patterns were relative stable outside the SAs and PFAs. 4.2. Hydraulic analysis In W-CW, the isotope results showed two SAs and one PFA were detected by using water isotopes technology. A strong stagnant flow was observed in the northeast corner. The configuration of the inlet led to the low hydraulic efficiency. The inlet pattern was a single culvert inlet in the northwest corner and waste water flew into wetland in shares. Since higher velocity relative to the flow inside the wetland, the influent strengthen water exchange with the water in the front and weaken the exchange with the water on the side [36]. With time to accumulate, a SA formed on the side, while the preferential flow occurred in the front region of inlet, which also explained the formation of the PFA in the front area of the inlet. Vegetation is a significant factor in wetland systems, but it increases flow resistance and influences hydraulic flow patterns [37]. Another SA in W-CW was vegetated area, where the flow resistance was higher than that in open areas around. The flow in this part had been separated into vegetation coverage condition and without vegetation condition. Compared with the condition without vegetation, high flow resistance led to poor fluidity in vegetation coverage condition, so the SA formed over time. In the second part, the wetland was mainly open area, scarcely covered by vegetation, lacking factors for hydraulic change, so flow patterns were stable. In QCW, the dispersive inlets were in the west bank of the first treated unit, so the flow velocity near the eastern bank was low and the SA was close to the eastern bank (Fig. 5a). In the area around the overflow weir between the first treated unit and the second treated unit, another SA formed. In this SA, flexible submerged vegetation was planted, and on the other side, rigid emergent vegetation was planted. The roughness coefficient was different for the two plant types [38]. For such flow condition, the total drag force of submerged vegetation was higher than emergent vegetation [39]. In this circumstance, the flow path was in the emergent vegetation coverage areas, and SA formed in the submerged vegetation areas. The middle inlet configuration was similar with W-CW's, and the hydraulic flow patterns were also resembled. After waste water flew into the wetland from middle inlet, a PFA formed in front of the inlet and a SA existed on the side near the east bank
(Fig. 5a). In the third treated unit, the SA's plant density (about 130– 150 plants/m2) was highest in the wetland, and also the flow resistance was higher than that in other areas, which was the factor for the formation of the SA. From the hydraulic flow patterns simulation results, proper configuration of inlet and vegetation management should be paid more attention when designing a wetland. The uniform inlet pattern should be selected rather than a single culvert and dispersed inlets. Types and densities should be consistent in the direction perpendicular to water flow, in order to avoid the different flow resistance leading to poor hydraulic efficiency. 4.3. NH+ 4 -N distribution characteristic The significant difference between SA and PFA was residence time, which is the important factor that influences NH+ 4 -N removal efficiency (Eq. 2). Water residence time in SA is plenty to oxidize NH+ 4 -N. On the contrary, in PFA, an inadequate residence time is adverse to the removal of NH+ 4 -N. However, except for T, Kv is also a determinant factor for NH+ 4 -N removal. Kv varies with different types and densities of vegetation in wetlands, which are also the factors that influence hydraulic flow patterns. The combination of T and Kv determined the NH+ 4 -N distribution in different hydraulic flow patterns, which explained the different NH+ 4 -N distribution characteristic between the studied wetlands. W-CW had a deep water depth, and was constructed on the river watercourse, mainly composed by open areas. In W-CW, due to the lack of plants, the effect of microorganisms and plant uptake decreased [40] and the Kv was relative low [17]. But in good fluidity area, dissolve oxygen was high [41], which was helpful for the degradation + of NH+ 4 -N. In conclusion, the NH4 -N accumulated in SAs (Fig. 4c). In Q-CW, there was an opposite circumstance. The wetland had abundant plant species and densities, which enhanced microbial quantities and plants absorption and a high Kv value [17]. In SAs, long residence time T and high Kv value enhanced the removal of NH+ 4 -N, so the low concentration of NH+ 4 -N distributed in SAs (Fig. 5c). 5. Conclusions Water isotopes technology was used to investigate the hydraulic flow patterns in two studied wetlands, and results showed that it could accurately and intuitively determine SAs and PFAs in CWs. The
780
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
ammonia nitrogen concentration presented different distribution characteristics in different types of wetlands. In watercourse channel type CWs, ammonia nitrogen concentration was high in SAs but low in PFAs, and opposite phenomenon was observed in the man-made channel type constructed wetlands. The inlet patterns and vegetation have significant impact on the hydraulic flow patterns in CWs. Also, it is recommended to use uniform inlet pattern and maintain the vegetation at the same types and densities in the direction perpendicular to water flow. Acknowledgement This work received supports from National Natural Science Foundation of China (No. 21307076 & No. 41305124) and Fundamental Research Funds of Shandong University (No. 2014TB003 & No. 2015JC056). References [1] C. Calheiros, R.B. Mesquita, H. Brix, A.O. Rangel, P.M. Castro, Constructed wetland implemented in a tourism unit for wastewater treatment, 2014. [2] J. Vymazal, The use constructed wetlands with horizontal sub-surface flow for various types of wastewater, Ecol. Eng. 35 (2009) 1–17. [3] S. Wu, P. Kuschk, H. Brix, J. Vymazal, R. Dong, Development of constructed wetlands in performance intensifications for wastewater treatment: a nitrogen and organic matter targeted review, Water Res. 57 (2014) 40–55. [4] H. Wu, J. Zhang, H.H. Ngo, W. Guo, Z. Hu, S. Liang, J. Fan, H. Liu, A review on the sustainability of constructed wetlands for wastewater treatment: design and operation, Bioresour. Technol. 175 (2015) 594–601. [5] H. Bodin, J. Persson, J.-E. Englund, P. Milberg, Influence of residence time analyses on estimates of wetland hydraulics and pollutant removal, J. Hydrol. 501 (2013) 1–12. [6] G. Jahid Hasan, A. Kurniawan, S. Ooi, M. Hekstra, Y. Broekema, S. Bayen, Pollutants in mangrove ecosystems: a conceptual model for evaluating residence time, ICWFM 2015: 5th International Conference on Water & Flood Management, ICWFM, 2015. [7] Y. Lin, G. Xiaoshuang, Z. Yue, A study on the effect of along purification in hybrid flow constructed wetlands under different residence time, Acta Agric. Univ. Jiangxiensis (2013). [8] C.-B. Zhang, W.-L. Liu, J. Wang, Y. Ge, B.-H. Gu, J. Chang, Effects of plant diversity and hydraulic retention time on pollutant removals in vertical flow constructed wetland mesocosms, Ecol. Eng. 49 (2012) 244–248. [9] Y. Wang, X. Song, W. Liao, R. Niu, W. Wang, Y. Ding, Y. Wang, D. Yan, Impacts of inlet–outlet configuration, flow rate and filter size on hydraulic behavior of quasi2-dimensional horizontal constructed wetland: NaCl and dye tracer test, Ecol. Eng. 69 (2014) 177–185. [10] C. de la Mora Orozco, K.D. Jones, Evaluation of hydraulic residence time distribution (RTD) characterization and monitoring in a constructed channel wetland in South Texas, Environmental Sustainability Issues in the South Texas–Mexico Border Region, Springer 2014, pp. 159–177. [11] N.-B. Chang, Z. Xuan, M.P. Wanielista, A tracer study for assessing the interactions between hydraulic retention time and transport processes in a wetland system for nutrient removal, Bioprocess Biosyst. Eng. 35 (2012) 399–406. [12] J.-M. Chyan, F.J. Tan, I.-M. Chen, C.-J. Lin, D.B. Senoro, M.P.C. Luna, Effects of porosity on flow of free water surface constructed wetland in a physical model, Desalin. Water Treat. 52 (2014) 1077–1085. [13] H.E. Golden, C.R. Lane, D.M. Amatya, K.W. Bandilla, H.R. Kiperwas, C.D. Knightes, H. Ssegane, Hydrologic connectivity between geographically isolated wetlands and surface water systems: a review of select modeling methods, Environ. Model Softw. 53 (2014) 190–206. [14] J. Laurent, P. Bois, M. Nuel, A. Wanko, Systemic models of full-scale Surface Flow Treatment Wetlands: determination by application of fluorescent tracers, Chem. Eng. J. 264 (2015) 389–398. [15] E. Horstman, T. Balke, T. Bouma, M. Dohmen-Janssen, S. Hulscher, Optimizing methods to measure hydrodynamics in coastal wetlands: evaluating the use and positioning of ADV, ADCP AND HR-ADCP, Coast. Eng. Proc. 1 (2011) 51.
[16] B.H. Schmid, M.A. Hengl, U. Stephan, Salt tracer experiments in constructed wetland ponds with emergent vegetation: laboratory study on the formation of density layers and its influence on breakthrough curve analysis, Water Res. 38 (2004) 2095–2102. [17] J. Kumar, Y. Zhao, A review on numerous modeling approaches for effective, economical and ecological treatment wetlands, J. Environ. Manag. 92 (2011) 400–406. [18] C.M. García, M.I. Cantero, Y. Niño, M.H. García, Turbulence measurements with acoustic Doppler velocimeters, J. Hydraul. Eng. 131 (2005) 1062–1073. [19] A. Lohrmann, R. Cabrera, N.C. Kraus, Acoustic-Doppler velocimeter (ADV) for laboratory use, Fundamentals and Advancements in Hydraulic Measurements and Experimentation, ASCE 1994, pp. 351–365. [20] J.R. Gat, Oxygen and hydrogen isotopes in the hydrologic cycle, Annu. Rev. Earth Planet. Sci. 24 (1996) 225–262. [21] R.M. Price, G. Skrzypek, P.F. Grierson, P.K. Swart, J.W. Fourqurean, The use of stable isotopes of oxygen and hydrogen to identify water sources in two hypersaline estuaries with different hydrologic regimes, Mar. Freshw. Res. 63 (2012) 952–966. [22] L. Wassenaar, P. Athanasopoulos, M. Hendry, Isotope hydrology of precipitation, surface and ground waters in the Okanagan Valley, British Columbia, Canada, J. Hydrol. 411 (2011) 37–48. [23] B. Luz, E. Barkan, R. Yam, A. Shemesh, Fractionation of oxygen and hydrogen isotopes in evaporating water, Geochim. Cosmochim. Acta 73 (2009) 6697–6703. [24] X. Xiao, R. Horton, T. Sauer, J. Heitman, T. Ren, Cumulative soil water evaporation as a function of depth and time, Vadose Zone J. 10 (2011) 1016–1022. [25] Y. Jiang, T. Shi, X. Gu, Healthy urban streams: the ecological continuity study of the Suzhou creek corridor in Shanghai, Cities 59 (2016) 80–94. [26] P. Gupta, D. Noone, J. Galewsky, C. Sweeney, B. Vaughn, A new laser-based, field-deployable analyzer for laboratory-class stable isotope measurements in water, Geochim. Cosmochim. Acta 73 (2009) A480. [27] C.D. Cappa, M.B. Hendricks, D.J. DePaolo, R.C. Cohen, Isotopic fractionation of water during evaporation, J. Geophys. Res. Atmos. 108 (2003). [28] J. Yu, F. Yu, D. Liu, Hydrogen and oxygen isotopic compositions of meteoric waters in the eastern part of China, Chin. J. Geochem. 6 (1987) 367–371. [29] W.C. Van Beers, J.P. Kleijnen, Kriging interpolation in simulation: a survey, Simulation Conference, 2004. Proceedings of the 2004 Winter, IEEE, 2004. [30] A.C. Wilcox, E.E. Wohl, Field measurements of three-dimensional hydraulics in a step-pool channel, Geomorphology 83 (2007) 215–231. [31] I.J. Walker, A.G. Roy, Fluid flow and sediment transport processes in geomorphology: innovations, insights, and advances in measurement, Geomorphology 68 (2005) 1–2. [32] W.E. Federation, A.P.H. Association, Standard Methods for the Examination of Water and Wastewater, American Public Health Association (APHA), Washington, DC, USA, 2005. [33] Y. Mohamed, W. Bastiaanssen, H. Savenije, B. van den Hurk, C. Finlayson, Wetland versus open water evaporation: an analysis and literature review, Phys. Chem. Earth 47 (2012) 114–121. [34] A.-K. Ronkanen, B. Kløve, Use of stabile isotopes and tracers to detect preferential flow patterns in a peatland treating municipal wastewater, J. Hydrol. 347 (2007) 418–429. [35] J.D. Marshall, J.R. Brooks, K. Lajtha, Sources of variation in the stable isotopic composition of plants, Stable Isotopes in Ecology and Environmental Science, Vol. 2, 2007, pp. 22–60. [36] H. Gvirtzman, N. Paldor, M. Magaritz, Y. Bachmat, Mass exchange between mobile freshwater and immobile saline water in the unsaturated zone, Water Resour. Res. 24 (1988) 1638–1644. [37] M. Mahmoudi, M. Nalesso, R. Garcia, F. Miralles-Wilhelm, The Effect of Manning's Roughness Calibration on Flow and Sediment Transport in Wetlands: Vegetation Drag ApproachAGU Spring Meeting Abstracts 2013 07. [38] C. James, A. Birkhead, A. Jordanova, J. O'sullivan, Flow resistance of emergent vegetation, J. Hydraul. Res. 42 (2004) 390–398. [39] J. Aberle, J. Järvelä, Flow resistance of emergent rigid and flexible floodplain vegetation, J. Hydraul. Res. 51 (2013) 33–45. [40] A.M. Ibekwe, S. Lyon, M. Leddy, M. Jacobson-Meyers, Impact of plant density and microbial composition on water quality from a free water surface constructed wetland, J. Appl. Microbiol. 102 (2007) 921–936. [41] H. Haider, W. Ali, S. Haydar, Evaluation of various relationships of reaeration rate coefficient for modeling dissolved oxygen in a river with extreme flow variations in Pakistan, Hydrol. Process. 27 (2013) 3949–3963.
Contents lists available at ScienceDirect
Journal of Molecular Liquids journal homepage: www.elsevier.com/locate/molliq
Determination of hydraulic flow patterns in constructed wetlands using hydrogen and oxygen isotopes Haimeng Sun a, Zhen Hu a, Jian Zhang a,⁎, Weizhong Wu b, Shuang Liang a, Shaoyong Lu c, Huaqing Liu a a b c
Shandong Key Laboratory of Water Pollution Control and Resource Reuse, School of Environmental Science & Engineering, Shandong University, Jinan 250100, PR China Department of Environmental Science, College of Environmental Sciences and Engineering, Peking University, Beijing 100871, PR China Research Centre of Lake Environment, Chinese Research Academy of Environmental Sciences, Beijing 100012, PR China
a r t i c l e
i n f o
Article history: Received 6 April 2016 Received in revised form 13 July 2016 Accepted 10 August 2016 Available online 05 September 2016 Keywords: Constructed wetlands Stable isotopes Hydraulics Flow patterns
a b s t r a c t The treatment efficiency of constructed wetlands (CWs) is highly dependent on the stability of the hydraulic flow patterns. To date, general technologies used to study hydraulic flow patterns of CWs mainly include tracer method, model simulation and velocity measurement, which are either expensive, empirical, or having secondary pollution. In this study, a new technology, which was based on the isotopic composition variation in CWs, was applied to detect the hydraulic flow patterns of two different CWs. Results showed that the hydraulic flow patterns of the two studied wetlands could be detected effectively by using hydrogen and oxygen isotopes. Furthermore, the locations of stagnant areas (SAs) and preferential flow areas (PFAs) were also determined. Significant regional difference in isotopic composition existed inside each CW, and two wetland design suggestions are proposed after hydraulic analysis. One is that the influent of CWs is supposed to be distributed uniformly, and another piece of advice is that the vegetation in the direction perpendicular to water flow should be maintained at the same types and density. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Constructed wetlands (CWs), which are designed on the basis of all kinds of reaction process in natural wetlands but make the process occur in more controlled systems, have a rapid momentum of development especially in last three decades [1]. These systems, which are robust, low power consumption, easy to maintain and operate, are suitable for advanced treatment of municipal sewage plants' effluent or decentralized wastewater [2,3]. The pollutants such as suspended material, nitrogen, phosphorus, organic matters, and metals are removed by the complex and abundant reactions relying on physical, chemical and biological processes [4]. Pollutants removal mechanisms are mainly organic matters decomposition by microorganism, nitrogen removal through nitrification – denitrification, and nutrient elements absorption by the plants and substrate. The implementation of these reactions depends on the joint function of plants, matrix, and microorganisms. To a certain extent, nutrients removal efficiencies are dependent on the time that waste water touched with vegetation, matrix, and microorganisms in the CWs. As a result, hydraulic retention time (HRT) has significant effect on water pollutants removal efficiency [5–7]. Furthermore, CWs' water flow is rather complex because of its large areas of plants and artificial matrix. Hydraulic performance of CWs is influenced by several factors, such as plant ⁎ Corresponding author. E-mail address: [email protected] (J. Zhang).
http://dx.doi.org/10.1016/j.molliq.2016.08.115 0167-7322/© 2016 Elsevier B.V. All rights reserved.
species richness and density, the CW's aspect ratio, the inlet and outlet's form [8,9]. A combination of these factors leads to different hydraulic conditions inside the CWs, and the hydraulic flow patterns are the comprehensive performance of all hydraulic conditions which also are the external presentation of the residence time distribution (RTD) [10]. Since RTD is a value of process quantity, it can't reflect the hydraulic conditions inside the CWs in situ. It is important to investigate the hydraulic variation inside CWs, which can help us find the unfavorable hydraulic phenomenon such as stagnant areas (SAs) or preferential flow areas (PFAs). At present, many researchers focus on the application of tracer, hydraulic models and velocity measurement to simulate hydraulic flow patterns in CWs [11–15]. However, in large-scale CWs, adding tracers is not a good choice because of the high cost, potential of secondary pollution caused by adding external material into the CWs, and the intensive continuous monitoring time [16]. As to hydraulic model, its results can not reflect the hydraulic flow patterns synchronously as the parameter is always empirical [17]. Finally direct velocity measurement inside the CWs needs expensive apparatus, not to mention that field measurement is considered a labor-consuming job [18,19]. Stable isotopes of oxygen and hydrogen are a novel technology which has been used in hydrology for investigating water composition and circulation in rivers, lakes, oceans, ground water and atmosphere [20–22]. In liquid water, isotopes molecules' vapor pressure is inversely proportional with its molecular weight. Compared to 18O and 2H, 16O and 1H have higher vapor pressure, indicating that they could separate
776
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
from the liquid phase more easily, which lead to the enrichment of the heavy water isotopes 18O and 2H [23]. In CWs, the cumulative amount of evaporation in different regions has strong relationship with regional water residence time [24]. Moreover, regional water residence time determined the flow patterns. If an accurate survey was done on the distribution of water isotopic composition, the hydraulic flow patterns would be known inside CWs. The isotopes technology utilizes the internal elements in CWs to detect the hydraulic flow patterns real-timely and has a low cost without secondary pollution. The aims of the present study were to characterize hydraulic flow patterns using isotope technology, further, to determine the locations of the SAs and PFAs in CWs. After hydraulic analysis, some engineering suggestions are proposed. The ammonia nitrogen's distribution characteristics in different hydraulic conditions were studied as well. 2. Materials and methods Fig. 1. The schematic diagrams of the W-CW (a) and Q-CW (b). Circles represent sampling points (21 points in W-CW while 31 points in Q-CW), and squares stand for the inlets. The bold lines instead for the outlet and the dashed lines instead for the overflow weir.
2.1. Sites description Two sites were identified based on differences in shapes, plants species and inlet patterns: Wu river constructed wetland (W-CW) (N: 34° 52′6.9″; E: 118°20′42.7″) and Qihe ecological constructed wetland (QCW) (N: 36°56′26.9″; E: 116°47′24.7″). Both sites belong to monsoon climate of medium latitudes, with four distinctive seasons. Both the CWs accept treated sewage water from local sewage treatment plants, and the water quality met Chinese National Class I (Grade A) Sewage Discharge Standard. Detailed information of the two CWs are given in Table 1. The surface flow constructed wetlands (SFCWs) in the two CWs were selected as our experimental subjects to study the hydraulic flow patterns using isotopic technology and velocity distribution. In WCW, the experimental subject was the SFCW between second and third over flow weir, which was about 16.1 ha and straight type without bend, a single culvert set as the inlet pattern (Fig. 1a). In Q-CW, the experimental subject was the five surface flow treated units, which was about 3.5 ha and with two bends in shape. Inlet pattern was several culverts distributing in the bank, an overflow weir set between every two unit for oxygen supplement (Fig. 1b). Vegetation coverage area was another obviously different factor between the two studied wetlands. According to the method proposed by Jiang et al. [25], vegetation coverage ratios were 26% and 83% for W-CW and Q-CW, respectively. W-CW's flow channel was constructed on the river watercourse (named watercourse channel type), mainly water areas. Yet the Q-CW was built artificially (named man-made channel type), which had abundant vegetation species and high density. Vegetation types were mainly emergent plants and submerged plants. The emergent plants were mainly Typha orientalis and Phragmites australis while the submerged plants were mainly Vallisneria natans and Nymphaea tetragona. It would be specially mentioned that a middle inlet was set in the second level wetland, through which a small part of outflow of sub-surface CWs flew into the SFCW directly when high water inflow rate occur. Coincidentally, the middle inlet was open when the measurement was conducted, which increased hydraulic changes. Sampling points in this study were set in the areas that had different hydraulic conditions such as open water areas, various vegetation species coverage areas, inlet and outlet regions (Fig. 1). Isotopes were all collected at about 10 cm below the water surface. The sampling job was conducted after consecutive ten more days without rain. The
collected water was put into 250 mL plastic bottles then preserved into insulation box with ice bags in it. A GPS (72H) was used to determine the sampling points' location. Water sampling job was conducted in August 2013 for W-CW, and June 2015 for Q-CW. There were 21 and 33 sampling points designed in W-CW and Q-CW, respectively (Fig. 1). 2.2. Water isotopes analysis Isotopic composition is expressed δ in ‰ as: δX E ¼
Rsample −1 1000 Rstandard
ð1Þ
where X is the atomic mass of the heavy isotope of element E and R is the ratio of the heavy to light isotope (18O/16O). Rsample is the ratio of heavy (e.g. 2H and 18O) to light (e.g., 1H and 16O) isotope of water samples; Rstandard is the ratio of heavy (e.g., 2H and 18O) to light (e.g., 1H and 16 O) isotope related to VSMOW standard. The isotopic composition of oxygen and hydrogen was analyzed by wavelength-scanned cavityring-downs pectroscopy (WS-CRDS) [26], using Picarro L2140-i δ18O/δ17O/δD/17O-excess high-precision isotopic water analyzer in Beikerui Detection Technology Co., Ltd. The relationship between surface water signatures of δD and δ18O relative to the local meteoric water line (LMWL) showed the degree of evaporation intuitively [27]. Since the experimental subjects located in eastern China, the eastern China meteoric water line (ECMWL) proposed by Yu et al. [28] was selected as LMWL. In order to show the distribution of the isotopes clearly, Surfer 10.0 was used to give a δ18O distribution contour map of the wetland using an ordinary kriging interpolation [29]. If a statistically significant difference in the isotopic distribution was found inside two studied wetlands, it indicated that the hydraulic conditions heterogeneity existed. In the areas of contour values were high, indicating that high evaporation leading to heavy isotopes enrichment, where a long HRT and SAs might exist. On the contrary, in the PFA, the short HRT lead to relative low evaporation enrichment, so small numerical isotopic values line would be distributed in these areas. In the δ18O distribution maps, the range of measured δ18O values would be
Table 1 Properties of the studied wetlands. Parameters Wetlands
Area (ha)
Influent type
Flow rate (m3/d)
Experimental subject (ES)
ES's area (ha)
ES's shape
Plant coverage ratio
W-CW Q-CW
533.33 68.53
Treated sewage water Treated sewage water
3.0 × 105 4.0 × 104
Second surface flow treated unit 5 surface flow treated units
16.1 3.5
Straight type Turning type
26% 83%
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
777
divided into twelve value scales, and the areas of the top two value levels distributed were considered as the SAs, while the areas of the minimum two value levels distributed were considered as PFAs. 2.3. Flow field measurement To verify the stability of water isotopes technology, three-dimensional flow velocities were measured using acoustic Doppler velocimeters (ADVs) at every sampling point. ADV, one of a common technology of Doppler methods [30], has facilitated technologies for flow field measurement, such as current meters technology applied in velocity and turbulence measurements in a two- or three-dimensional flow [31]. At each testing point, triplicate measurements were conducted, and the final results were the average of the three measurements. The results were showed as three-dimensional vector velocity, but in this study, the absolute values of the velocity were more concerned, so three-dimensional vector velocity was converted into absolute velocity. Velocity results were used to get a continuous distribution of the velocity by kriging with a trend, and then the velocity distribution was obtained, from which the hydraulic flow patterns would be showed clearly. On the velocity distribution maps, the range of measured velocity values would be divided into twelve value scales as well, and the areas of the top two value levels distributed were considered as the PFAs, while the areas of the minimum two value levels distributed were considered as SAs. 2.4. NH+ 4 -N concentration analysis Water sample of every point in the two studied wetlands was taken to laboratory and analyzed for NH+ 4 -N concentration according to standard methods [32]. In general, the treatment processes such as mass sedimentation, volatilization, sorption, and transport are based on first-order kinetic models. This model is given as follows: C−C ¼ expð−K v T Þ C 0 −C
ð2Þ
+ where, C is the NH+ 4 -N concentration (mg/L), C0 is the influent NH4 -N ⁎ concentration (mg/L), C is the background concentration (mg/L), T is the retention time (d) and Kv is the first-order volumetric rate constant (d−1). As shown in the formula, ammonia nitrogen removal efficiency is related with residence time T and Kv. In SAs and PFAs, many factors can affect the Kv and residence time T. The NH+ 4 -N concentration would present different characteristics in SAs and PFAs. In order to link the + NH+ 4 -N distribution with the hydraulic flow patterns, NH4 -N concentration contour distribution maps were given by using Surfer 10.0 for the studied wetlands.
3. Results
Fig. 2. The statistics results of the δ18O (a), velocity (b) and ammonia nitrogen (c) of the two CWs.
3.1. Isotopic distribution The δ18O values ranged from − 6.05‰ to −5.40‰ for W-CW, and ranged from −7.85‰ to −7.14‰ for Q-CW (Fig. 2a). Linear correlation between δ18O and δ2H in both wetlands were δ2H = 1.99δ18O-34.85 (Rsquare was 0.80) for W-CW and δ2H = 5.14δ18O-17.71 (Rsquare was 0.97) for Q-CW (Fig. 3). The linear relation in both wetlands were deviated from ECWML δ2H = 7.8δ18O + 6.6 [28], which proved that the evaporation of water had obvious influence on the isotopic composition of both wetlands. Because of the difference in the sampling time, δ18O values were higher in W-CW than that in Q-CW (Fig. 3). The measurement was conducted in August for W-CW, and the temperature was higher than that in June, when the measurement in Q-CW was conducted. The evaporative enrichment in water would be enhanced with temperature increased, which resulted in higher δ18O composition in WCW. In this study, the variation trend of δ18O inside experiment subjects
was the factor to be concerned rather than the different δ18O composition between the two subjects. Interpolated maps of δ18O distributions of the two wetlands are given in Figs. 4a and 5a, respectively, where the variation shows clearly inside the two studied wetlands. In accordance with the provisions of Section 2.2, the marginal values of the SAs and PFAs were −5.50‰ and −5.94‰ for W-CW, while the marginal values of the SAs and PFAs were − 7.22‰ and − 7.72‰ for Q-CW. Two SAs and one PFA were determined in W-CW (Fig. 4a) and four SAs and one PFA were found in W-CW (Fig. 5a). 3.2. Velocity distribution As could be seen from the flow velocity measurement results (Fig. 2b), mixed flow existed in each CW, which indicates a spread of residence time. Compared with Q-CW, the amplitude of velocity
778
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
concentration in SAs was 1.84 mg/L which was close to the lowest concentration. In the areas outside of the SAs of the studied wetlands, the NH+ 4 -N concentration variation was relative stable. 4. Discussion 4.1. Hydraulic flow patterns determined by isotopic distribution
Fig. 3. Relationship between surface water signatures of δ2H and δ18O for W-CW and QCW relative to the eastern China meteoric water line (ECMWL).
fluctuation in W-CW (0.002 m/s to 0.017 m/s) was more stable than that in Q-CW (0.002 m/s to 0.038 m/s), which presumably due to the existence of middle inlet set in the second treatment unit in Q-CW. The average velocity was 0.0139 m/s and 0.0207 m/s before and after middle inlet, respectively. The simulated flow velocity distributions in the W-CW and Q-CW are presented in Figs. 4b and 5b, and both showed regional changes in the CWs clearly. Velocity distribution showed similar hydraulic flow patterns with δ18O distribution. In Fig. 4b, two SAs and one PFA were determined in W-CW, and their locations were similar with that in Fig. 4a. The same experiment results were found in Fig. 5a and b. 3.3. NH+ 4 -N distribution NH+ 4 -N concentration values showed regional differences as well (Fig. 2c). In both studied wetlands, NH+ 4 -N concentration declined gradually from inlet to outlet (Figs. 4c and 5c). The highest NH+ 4 -N concentration was 2.83 mg/L detected around inlet in W-CW, while the highest NH+ 4 -N concentration was 2.90 mg/L observed in the area around the middle inlet (Figs. 4c and 5c). The NH+ 4 -N distribution was highly different in SAs between the studied wetlands. In W-CW, the mean NH+ 4 -N concentration in SAs was 2.57 mg/L which was close to the highest concentration. On the contrary, in Q-CW, the mean NH+ 4 -N
Air temperature was 26.4 °C when the measurement was conducted in W-CW, and improved to 33.8 °C when the measurement was conducted in Q-CW. Evaporation rate in the Q-CW was higher than that in the W-CW due to the higher temperature. In this case, the enrichment of heavy isotopes would be intensified in Q-CW. As a result, isotopic composition in Q-CW was significant higher than that in W-CW. Immediately after influent mixed with residual water in the wetland, water budget began to change due to factors including precipitation, evaporation, absorption and interaction with groundwater [33], which might cause δ18O variation in the treated water. As 2–5 days without precipitation was enough to obtain the required isotope enrichment [34], the precipitation could be neglected when the sampling was conducted after consecutive ten more days without rain. In the studied wetlands, the well-defined facility of substrate was built in the bottom to prevent water infiltration, which led to negligible interaction with groundwater, so the groundwater's influence on the isotopic composition in the surface wetland could be neglected. Furthermore, absorption by vegetation's roots had little impact on the isotopic composition around the roots [35]. In such circumstances, evaporative enrichment was the decisive factor towards isotopic distribution. Due to the climatic condition was same inside both studied wetlands, the evaporation resulting isotopes variation was caused by the HRT. In the studied wetlands, an obvious difference was found in isotopic distribution, which indicated a different HRT distribution in different regions, and the hydraulic flow patterns were showed clearly (Figs. 4a and 5a). Flow exchange was weak between the stagnant areas and smooth flow areas, so the flow velocity was faster in the smooth flow areas. Compared to water flow in stagnant areas, cumulative amount of evaporation was lacking in the smooth flow areas and flow to the outlet rapidly, so the isotopic composition in some regions close to the outlet were lower than that in some regions close to the inlet. Small δ18O values and gradient indicated preferential flow in the front of the inlet in wetlands (Fig. 4a), the same flow condition was determined around the middle inlet in Q-CW (Fig. 5a). This was supported by velocity distribution (Figs. 4b and 5b). δ18O distribution results showed two and four SAs
Fig. 4. The isotopes δ18O distribution (a), velocity distribution (b) and ammonia nitrogen distribution (c) in W-CW.
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
779
Fig. 5. The isotopes δ18O distribution (a), velocity distribution (b) and ammonia nitrogen distribution (c) in Q-CW.
in W-CW and Q-CW, respectively. Isotopic δ18O distribution and velocity distribution correspond very well. Hydraulic flow patterns were relative stable outside the SAs and PFAs. 4.2. Hydraulic analysis In W-CW, the isotope results showed two SAs and one PFA were detected by using water isotopes technology. A strong stagnant flow was observed in the northeast corner. The configuration of the inlet led to the low hydraulic efficiency. The inlet pattern was a single culvert inlet in the northwest corner and waste water flew into wetland in shares. Since higher velocity relative to the flow inside the wetland, the influent strengthen water exchange with the water in the front and weaken the exchange with the water on the side [36]. With time to accumulate, a SA formed on the side, while the preferential flow occurred in the front region of inlet, which also explained the formation of the PFA in the front area of the inlet. Vegetation is a significant factor in wetland systems, but it increases flow resistance and influences hydraulic flow patterns [37]. Another SA in W-CW was vegetated area, where the flow resistance was higher than that in open areas around. The flow in this part had been separated into vegetation coverage condition and without vegetation condition. Compared with the condition without vegetation, high flow resistance led to poor fluidity in vegetation coverage condition, so the SA formed over time. In the second part, the wetland was mainly open area, scarcely covered by vegetation, lacking factors for hydraulic change, so flow patterns were stable. In QCW, the dispersive inlets were in the west bank of the first treated unit, so the flow velocity near the eastern bank was low and the SA was close to the eastern bank (Fig. 5a). In the area around the overflow weir between the first treated unit and the second treated unit, another SA formed. In this SA, flexible submerged vegetation was planted, and on the other side, rigid emergent vegetation was planted. The roughness coefficient was different for the two plant types [38]. For such flow condition, the total drag force of submerged vegetation was higher than emergent vegetation [39]. In this circumstance, the flow path was in the emergent vegetation coverage areas, and SA formed in the submerged vegetation areas. The middle inlet configuration was similar with W-CW's, and the hydraulic flow patterns were also resembled. After waste water flew into the wetland from middle inlet, a PFA formed in front of the inlet and a SA existed on the side near the east bank
(Fig. 5a). In the third treated unit, the SA's plant density (about 130– 150 plants/m2) was highest in the wetland, and also the flow resistance was higher than that in other areas, which was the factor for the formation of the SA. From the hydraulic flow patterns simulation results, proper configuration of inlet and vegetation management should be paid more attention when designing a wetland. The uniform inlet pattern should be selected rather than a single culvert and dispersed inlets. Types and densities should be consistent in the direction perpendicular to water flow, in order to avoid the different flow resistance leading to poor hydraulic efficiency. 4.3. NH+ 4 -N distribution characteristic The significant difference between SA and PFA was residence time, which is the important factor that influences NH+ 4 -N removal efficiency (Eq. 2). Water residence time in SA is plenty to oxidize NH+ 4 -N. On the contrary, in PFA, an inadequate residence time is adverse to the removal of NH+ 4 -N. However, except for T, Kv is also a determinant factor for NH+ 4 -N removal. Kv varies with different types and densities of vegetation in wetlands, which are also the factors that influence hydraulic flow patterns. The combination of T and Kv determined the NH+ 4 -N distribution in different hydraulic flow patterns, which explained the different NH+ 4 -N distribution characteristic between the studied wetlands. W-CW had a deep water depth, and was constructed on the river watercourse, mainly composed by open areas. In W-CW, due to the lack of plants, the effect of microorganisms and plant uptake decreased [40] and the Kv was relative low [17]. But in good fluidity area, dissolve oxygen was high [41], which was helpful for the degradation + of NH+ 4 -N. In conclusion, the NH4 -N accumulated in SAs (Fig. 4c). In Q-CW, there was an opposite circumstance. The wetland had abundant plant species and densities, which enhanced microbial quantities and plants absorption and a high Kv value [17]. In SAs, long residence time T and high Kv value enhanced the removal of NH+ 4 -N, so the low concentration of NH+ 4 -N distributed in SAs (Fig. 5c). 5. Conclusions Water isotopes technology was used to investigate the hydraulic flow patterns in two studied wetlands, and results showed that it could accurately and intuitively determine SAs and PFAs in CWs. The
780
H. Sun et al. / Journal of Molecular Liquids 223 (2016) 775–780
ammonia nitrogen concentration presented different distribution characteristics in different types of wetlands. In watercourse channel type CWs, ammonia nitrogen concentration was high in SAs but low in PFAs, and opposite phenomenon was observed in the man-made channel type constructed wetlands. The inlet patterns and vegetation have significant impact on the hydraulic flow patterns in CWs. Also, it is recommended to use uniform inlet pattern and maintain the vegetation at the same types and densities in the direction perpendicular to water flow. Acknowledgement This work received supports from National Natural Science Foundation of China (No. 21307076 & No. 41305124) and Fundamental Research Funds of Shandong University (No. 2014TB003 & No. 2015JC056). References [1] C. Calheiros, R.B. Mesquita, H. Brix, A.O. Rangel, P.M. Castro, Constructed wetland implemented in a tourism unit for wastewater treatment, 2014. [2] J. Vymazal, The use constructed wetlands with horizontal sub-surface flow for various types of wastewater, Ecol. Eng. 35 (2009) 1–17. [3] S. Wu, P. Kuschk, H. Brix, J. Vymazal, R. Dong, Development of constructed wetlands in performance intensifications for wastewater treatment: a nitrogen and organic matter targeted review, Water Res. 57 (2014) 40–55. [4] H. Wu, J. Zhang, H.H. Ngo, W. Guo, Z. Hu, S. Liang, J. Fan, H. Liu, A review on the sustainability of constructed wetlands for wastewater treatment: design and operation, Bioresour. Technol. 175 (2015) 594–601. [5] H. Bodin, J. Persson, J.-E. Englund, P. Milberg, Influence of residence time analyses on estimates of wetland hydraulics and pollutant removal, J. Hydrol. 501 (2013) 1–12. [6] G. Jahid Hasan, A. Kurniawan, S. Ooi, M. Hekstra, Y. Broekema, S. Bayen, Pollutants in mangrove ecosystems: a conceptual model for evaluating residence time, ICWFM 2015: 5th International Conference on Water & Flood Management, ICWFM, 2015. [7] Y. Lin, G. Xiaoshuang, Z. Yue, A study on the effect of along purification in hybrid flow constructed wetlands under different residence time, Acta Agric. Univ. Jiangxiensis (2013). [8] C.-B. Zhang, W.-L. Liu, J. Wang, Y. Ge, B.-H. Gu, J. Chang, Effects of plant diversity and hydraulic retention time on pollutant removals in vertical flow constructed wetland mesocosms, Ecol. Eng. 49 (2012) 244–248. [9] Y. Wang, X. Song, W. Liao, R. Niu, W. Wang, Y. Ding, Y. Wang, D. Yan, Impacts of inlet–outlet configuration, flow rate and filter size on hydraulic behavior of quasi2-dimensional horizontal constructed wetland: NaCl and dye tracer test, Ecol. Eng. 69 (2014) 177–185. [10] C. de la Mora Orozco, K.D. Jones, Evaluation of hydraulic residence time distribution (RTD) characterization and monitoring in a constructed channel wetland in South Texas, Environmental Sustainability Issues in the South Texas–Mexico Border Region, Springer 2014, pp. 159–177. [11] N.-B. Chang, Z. Xuan, M.P. Wanielista, A tracer study for assessing the interactions between hydraulic retention time and transport processes in a wetland system for nutrient removal, Bioprocess Biosyst. Eng. 35 (2012) 399–406. [12] J.-M. Chyan, F.J. Tan, I.-M. Chen, C.-J. Lin, D.B. Senoro, M.P.C. Luna, Effects of porosity on flow of free water surface constructed wetland in a physical model, Desalin. Water Treat. 52 (2014) 1077–1085. [13] H.E. Golden, C.R. Lane, D.M. Amatya, K.W. Bandilla, H.R. Kiperwas, C.D. Knightes, H. Ssegane, Hydrologic connectivity between geographically isolated wetlands and surface water systems: a review of select modeling methods, Environ. Model Softw. 53 (2014) 190–206. [14] J. Laurent, P. Bois, M. Nuel, A. Wanko, Systemic models of full-scale Surface Flow Treatment Wetlands: determination by application of fluorescent tracers, Chem. Eng. J. 264 (2015) 389–398. [15] E. Horstman, T. Balke, T. Bouma, M. Dohmen-Janssen, S. Hulscher, Optimizing methods to measure hydrodynamics in coastal wetlands: evaluating the use and positioning of ADV, ADCP AND HR-ADCP, Coast. Eng. Proc. 1 (2011) 51.
[16] B.H. Schmid, M.A. Hengl, U. Stephan, Salt tracer experiments in constructed wetland ponds with emergent vegetation: laboratory study on the formation of density layers and its influence on breakthrough curve analysis, Water Res. 38 (2004) 2095–2102. [17] J. Kumar, Y. Zhao, A review on numerous modeling approaches for effective, economical and ecological treatment wetlands, J. Environ. Manag. 92 (2011) 400–406. [18] C.M. García, M.I. Cantero, Y. Niño, M.H. García, Turbulence measurements with acoustic Doppler velocimeters, J. Hydraul. Eng. 131 (2005) 1062–1073. [19] A. Lohrmann, R. Cabrera, N.C. Kraus, Acoustic-Doppler velocimeter (ADV) for laboratory use, Fundamentals and Advancements in Hydraulic Measurements and Experimentation, ASCE 1994, pp. 351–365. [20] J.R. Gat, Oxygen and hydrogen isotopes in the hydrologic cycle, Annu. Rev. Earth Planet. Sci. 24 (1996) 225–262. [21] R.M. Price, G. Skrzypek, P.F. Grierson, P.K. Swart, J.W. Fourqurean, The use of stable isotopes of oxygen and hydrogen to identify water sources in two hypersaline estuaries with different hydrologic regimes, Mar. Freshw. Res. 63 (2012) 952–966. [22] L. Wassenaar, P. Athanasopoulos, M. Hendry, Isotope hydrology of precipitation, surface and ground waters in the Okanagan Valley, British Columbia, Canada, J. Hydrol. 411 (2011) 37–48. [23] B. Luz, E. Barkan, R. Yam, A. Shemesh, Fractionation of oxygen and hydrogen isotopes in evaporating water, Geochim. Cosmochim. Acta 73 (2009) 6697–6703. [24] X. Xiao, R. Horton, T. Sauer, J. Heitman, T. Ren, Cumulative soil water evaporation as a function of depth and time, Vadose Zone J. 10 (2011) 1016–1022. [25] Y. Jiang, T. Shi, X. Gu, Healthy urban streams: the ecological continuity study of the Suzhou creek corridor in Shanghai, Cities 59 (2016) 80–94. [26] P. Gupta, D. Noone, J. Galewsky, C. Sweeney, B. Vaughn, A new laser-based, field-deployable analyzer for laboratory-class stable isotope measurements in water, Geochim. Cosmochim. Acta 73 (2009) A480. [27] C.D. Cappa, M.B. Hendricks, D.J. DePaolo, R.C. Cohen, Isotopic fractionation of water during evaporation, J. Geophys. Res. Atmos. 108 (2003). [28] J. Yu, F. Yu, D. Liu, Hydrogen and oxygen isotopic compositions of meteoric waters in the eastern part of China, Chin. J. Geochem. 6 (1987) 367–371. [29] W.C. Van Beers, J.P. Kleijnen, Kriging interpolation in simulation: a survey, Simulation Conference, 2004. Proceedings of the 2004 Winter, IEEE, 2004. [30] A.C. Wilcox, E.E. Wohl, Field measurements of three-dimensional hydraulics in a step-pool channel, Geomorphology 83 (2007) 215–231. [31] I.J. Walker, A.G. Roy, Fluid flow and sediment transport processes in geomorphology: innovations, insights, and advances in measurement, Geomorphology 68 (2005) 1–2. [32] W.E. Federation, A.P.H. Association, Standard Methods for the Examination of Water and Wastewater, American Public Health Association (APHA), Washington, DC, USA, 2005. [33] Y. Mohamed, W. Bastiaanssen, H. Savenije, B. van den Hurk, C. Finlayson, Wetland versus open water evaporation: an analysis and literature review, Phys. Chem. Earth 47 (2012) 114–121. [34] A.-K. Ronkanen, B. Kløve, Use of stabile isotopes and tracers to detect preferential flow patterns in a peatland treating municipal wastewater, J. Hydrol. 347 (2007) 418–429. [35] J.D. Marshall, J.R. Brooks, K. Lajtha, Sources of variation in the stable isotopic composition of plants, Stable Isotopes in Ecology and Environmental Science, Vol. 2, 2007, pp. 22–60. [36] H. Gvirtzman, N. Paldor, M. Magaritz, Y. Bachmat, Mass exchange between mobile freshwater and immobile saline water in the unsaturated zone, Water Resour. Res. 24 (1988) 1638–1644. [37] M. Mahmoudi, M. Nalesso, R. Garcia, F. Miralles-Wilhelm, The Effect of Manning's Roughness Calibration on Flow and Sediment Transport in Wetlands: Vegetation Drag ApproachAGU Spring Meeting Abstracts 2013 07. [38] C. James, A. Birkhead, A. Jordanova, J. O'sullivan, Flow resistance of emergent vegetation, J. Hydraul. Res. 42 (2004) 390–398. [39] J. Aberle, J. Järvelä, Flow resistance of emergent rigid and flexible floodplain vegetation, J. Hydraul. Res. 51 (2013) 33–45. [40] A.M. Ibekwe, S. Lyon, M. Leddy, M. Jacobson-Meyers, Impact of plant density and microbial composition on water quality from a free water surface constructed wetland, J. Appl. Microbiol. 102 (2007) 921–936. [41] H. Haider, W. Ali, S. Haydar, Evaluation of various relationships of reaeration rate coefficient for modeling dissolved oxygen in a river with extreme flow variations in Pakistan, Hydrol. Process. 27 (2013) 3949–3963.