Phosphorus removal from lagoon-pretreated swine ...

Science of the Total Environment 576 (2017) 490–497

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Phosphorus removal from lagoon-pretreated swine wastewater by pilot-scale surface flow constructed wetlands planted with Myriophyllum aquaticum Pei Luo a, Feng Liu a,⁎, Xinliang Liu a, Xiao Wu a,b, Ran Yao a,c, Liang Chen a, Xi Li a, Runlin Xiao a, Jinshui Wu a a Key Laboratory of Agro-ecological Processes in Subtropical Regions, Changsha Research Station for Agricultural & Environmental Monitoring, Institute of Subtropical Agriculture, Chinese Academy of Sciences, Hunan 410125, China b College of Resources and Environment, Huazhong Agricultural University, Wuhan 430070, China c University of Chinese Academy of Sciences, Beijing 100049, China

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• M. aquaticum could tolerate and treat swine wastewate in field condition. • Pilot-scale SFCWs planted with M. aquaticum can remove TP with high efficiency. • The TP removal efficiency and removal rate constant had temporal variations. • TP removal was dominated by sediment adsorption in CW1 but plant uptake in CW2/CW3.

a r t i c l e

i n f o

Article history: Received 31 August 2016 Received in revised form 11 October 2016 Accepted 13 October 2016 Available online 26 October 2016 Editor: Jay Gan Keywords: Constructed wetland Phosphorus removal Swine wastewater Myriophyllum aquaticum Harvest management

⁎ Corresponding author. E-mail address: [email protected] (F. Liu).

http://dx.doi.org/10.1016/j.scitotenv.2016.10.094 0048-9697/© 2016 Elsevier B.V. All rights reserved.

a b s t r a c t Although constructed wetlands (CWs) are used as one relatively low-cost technology for livestock wastewater treatment, the improvement of phosphorus removal in CWs is urgently needed. In this study, a three-stage pilot-scale CW system consisting of three surface flow CWs (SFCWs; CW1, CW2, and CW3) in series from inlet to outlet was constructed to treat swine wastewater (SW) from a lagoon. The CWs were planted with Myriophyllum aquaticum. Considering different inlet loading rates, three strengths of swine wastewater (low: 33% SW, medium: 66% SW, and high: 100% SW) were fed to the CW system to determine total phosphorus (TP) removal efficiency and clarify the important role of plant harvest. Results from the period 2014–2016 indicate that the three-stage CW system had mean TP cumulative removal efficiencies and removal rates of 78.2–89.8% and 0.412–0.779 g m−2 d−1 respectively, under different inlet loading rates. The TP removal efficiency and removal rate constant had temporal variations, which depended on temperature condition and the annual growth pattern of M. aquaticum. The harvested phosphorus mass was 15.1–40.9 g m−2 yr−1 in the CWs except for CW1 with high strength SW, and contributed 22.5–59.6% of TP mass removal rate by the SFCWs. The TP removal was mainly by adsorption and precipitation in the substrate in CW1 but by uptake and multiple harvests of M. aquaticum in CW2 and CW3. The results suggest the three-stage CW system planted with M. aquaticum is suited for removing high TP concentrations from swine wastewater with a high removal efficiency. However, TP removal in high strength SW amounted to 70.1 ± 23.3%, and the outflow concentration of 17.0 ±

P. Luo et al. / Science of the Total Environment 576 (2017) 490–497

491

14.9 mg L−1 was still high. Optimal loading rates for high strength SW still need to be investigated for the CW system presented here. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The release of livestock and poultry breeding wastewater with large amounts of nitrogen, phosphorus, and other pollutants has caused severe surface and underground water pollution worldwide. Constructed wetlands (CWs) are increasingly being recognized as a relatively lowcost, energy-efficient, natural, and ecological technology for treating wastewaters (Wu et al., 2015a), while at the same time offering the potential for multiple benefits. Many studies have demonstrated that CWs can remove phosphorus from dairy and swine wastewaters but with relatively low removal efficiency (lower than 70%) (Cronk, 1996; Lee et al., 2004; Liu et al., 2016; Majer Newman et al., 1999; Poach et al., 2004; Reddy et al., 2001; Stone et al., 2004; Tanner et al., 1995). Therefore, enhancing removal efficiency of CWs and developing a costeffective way to remove nutrients from swine wastewater is urgently needed to improve water quality in agricultural areas. It is well established that mechanisms of phosphorus removal in CWs include sorption on substrates, storage in biomass, and the formation and accretion of new sediments and soils (Kadlec and Wallace, 2009). The selection of suitable substrates and plants in CWs is important for future practical applications (Arias and Brix, 2005; Vohla et al., 2011; Wu et al., 2013; Wu et al., 2015a; Yan and Xu, 2014). The finite sorption capacity of substrates cannot always maintain long-term or sustainable phosphorus removal (Dunne and Reddy, 2005). On the other hand, vegetation often plays a significant role in phosphorus assimilation and storage. Plant harvest has been reported to be an important way to remove phosphorus from CWs, and harvest of aboveground plants could significantly increase P removal compared to non-harvest (Dunne and Reddy, 2005; Vymazal, 2007; Wu et al., 2013). Aquatic plant species have been reported to possess various phosphorus removal capacities in CWs. For example, Buddhavarapu and Hancock (1991) reported that up to 89% of total phosphorus (TP) was removed in the pilot-scale CWs, and duckweed harvest accounted for up to 73% of the removed TP. Wu et al. (2011) found the microcosm CWs planted with Typha orientalis, Phragmites australis, Scirpus validus, and Iris pseudacorus could remove mean 47.8–66.5% of TP from polluted river water, but plant uptake was responsible for only 10.8–34.2% of the removed TP. Ong et al.(2010) observed that the average TP removal efficiency in CWs planted with Manchurian wild rice (52%) was significantly higher than those planted with Phragmites australis (34%). Selecting suitable plants should therefore be the focus of current research on sustainable design of CWs (Vymazal, 2011; Vymazal, 2013). We hypothesize that selection of an ideal plant in CWs can lower the burden and increase the lifetime of the CWs for removal of P in the long-term. It has been found that multiple harvests may increase the total plant standing stock due to high nutrient concentrations in young shoots (Vymazal et al., 2010). Optimized harvest management could increase the removal of phosphorus by 37% and nitrogen by 42% (Březinová and Vymazal, 2015). Březinová and Vymazal (2015) revealed that the amount of phosphorus removed via multiple harvests can increase up to 43% compared to a single harvest. Vymazal et al. (2010) found multiple harvests of aboveground biomass of Phalaris arundinacea could be beneficial for removal of some trace elements from municipal sewage in a horizontal sub-surface flow CW. However, many investigations employing plant harvesting strategies for increasing pollutant removal in CWs have been reported in different regions exclusively for lightly loaded systems (e.g., b 10–20 g P m− 2 yr− 1) (Vymazal, 2007; Wu et al., 2015b). It remains to be determined whether multiple harvests are effective in treating high-loaded wastewater, such as swine wastewater, when the plant can tolerate swine wastewater and have high nutrient uptake capacity.

To fill this knowledge gap, we constructed a three-stage surface flow CW system to treat swine wastewater, and selected Myriophyllum aquaticum as the wetland plant. We have previously demonstrated that M. aquaticum is able to tolerate high-strength swine wastewater and effectively remove nutrients from polluted waters in laboratory scale tests (Liu et al., 2016). Therefore, the objectives of this study were: (1) to analyze TP removal efficiency of surface flow CWs (SFCWs) planted with M. aquaticum in treating swine wastewater in a natural field condition; (2) to investigate the seasonal variation of phosphorus removal; and (3) to assess TP removal mechanisms and the role of multiple harvests on long-term TP removal efficiency in SFCWs. 2. Materials and methods 2.1. Constructed wetlands Eleven sequencing batches of three-stage SFCWs were set up at the Changsha Research Station for Agricultural & Environmental Monitoring, Changsha, Hunan Province, China (28° 30′ N, 113° 18′ E). The area is characterized by a subtropical climate with an average annual air temperature of 17.5 °C and annual precipitation of 1330 mm. Each three-stage SFCW contained three identical units (length 5 m, width 2 m, water depth 0.2 m), named CW1, CW2, and CW3 according to wastewater flow direction from inlet to outlet (Fig. 1). These SFCWs was covered with a paddy soil as substrate. The soil characteristics included pH value of 6.20, bulk density of 1.31 g cm−3, total nitrogen content of 1.20 g kg−1, TP content of 0.42 g kg−1, soil organic carbon of 12.9 g kg−1, and a clayed loam texture with sand, silt and clay content of 31.5%, 36.3%, and 32.2%, respectively. The SFCWs were planted with M. aquaticum, a widespread herb that can grow as a submergent or emergent plant. The pilot-scale SFCWs were exposed to swine wastewater (SW) of three different strengths: high (HS; pristine swine wastewater without dilution, i.e., 100% SW; n = 3), medium (MS; swine wastewater diluted with fresh water at a 2:1 ratio, i.e., 66% SW; n = 3), and low loading rates (LS; swine wastewater diluted with fresh water at a 1:2 ratio, i.e., 33% SW; n = 3). The characteristics of the swine wastewater are listed in Table S1 of the Supplementary material

Fig. 1. Schematic diagram of the surface flow constructed wetlands (SFCWs) planted with M. aquaticum, and the harvest schedule of M. aquaticum in the SFCWs. Harvest schedule of M. aquaticum was as follows: one third of each unit was harvested at approximately 0.5 m intervals. For example, we divided CW2 into nine parts and labeled them as 1, 2, 3, 1, 2, 3, 1, 2, and 3. We harvested the three parts labeled as “1” for the first harvest, those labeled as “2” for the second, and finally the third harvest took place in the sections labeled “3”. Other wetland units were similarly harvested.

492


(“S” indicates figures and tables in the Supplementary material afterwards). Additional SFCWs (n = 2) with fresh-water inflow were set up as the control treatment. Treatment of the four SFCWs was operated in an intermittent flow regime with a total of 0.18 m3 d−1 wastewater. The theoretical hydraulic retention time was 11 d in each unit (CW1, CW2, and CW3). The total hydraulic retention time in the SFCWs of 33 d was comparable to the previous studies of CWs treating swine wastewater, such as 33 d by Ibekwe et al. (2016) and 10–18 d by Poach et al. (2004). The SFCWs could remove most of ammonia nitrogen and organics from the lagoon-pretreated swine wastewater, and the majority of effluents were lower than the Wastewater Discharge Standard for livestock and poultry breeding in China (ammonia nitrogen, 80 mg L−1; chemical oxygen demand, 400 mg L−1; GB 18596-2001). 2.2. Sample collection and analysis The SFCWs were built in 2014, and formally operated from July 2014. To enhance the removal efficiency and prolong the service lifetime of the SFCWs, emergent shoots of M. aquaticum in all SFCW units were harvested above the water level and removed from the wetlands at approximately one month intervals. Each time, one third of each unit was harvested at approximately 0.5 m intervals. More details on the harvesting method are provided in Fig. 1. In each SFCW unit, water samples (inflow and outflow) were collected twice a week for the first three months (July–September 2014) and then once a week from October 2014 to January 2016. Monthly sediment samples were collected from July 2014. A dozen of sediment subsamples (0–5 cm top layer) from different evenly distributed points in the same CW unit were mixed completely to form one homogeneous sample. The samples were transported to the laboratory, and water samples were analyzed immediately using the National Standard Methods of China (GB 11893-89).The sediment samples were freeze-dried, ground into fine powder, and filtered through a 0.15 mm sieve, and then stored for TP analysis. The TP analysis of sediment samples was conducted using the National Standard Methods of China (GB 9837-88). Plant samples were dried at 105 °C for 30 min and oven dried at 80 °C until a constant weight was reached (Liu et al., 2016). The dried plant material was ground, and passed through a 1 mm sieve. The TP content in M. aquaticum was determined using colorimetric analysis after digestion with hydrogen peroxide and perchloric acid. All analyses were performed on duplicate samples. Water temperature, pH, and dissolved oxygen were immediately determined in the field by using a portable Mettler Toledo SG68-SevenGo Duo multi-parameter water quality meter (Mettler Toledo, Switzerland). 2.3. Data analysis The TP removal efficiency (%) by the SFCWs was calculated using the difference in TP concentration (mg L−1) between the inlet and outlet divided by TP concentration (mg L− 1) in the inlet. The TP loading rate (g m−2 d−1) was calculated by multiplying the hydraulic loading rate (0.018 m d−1; which is calculated by inlet flow rate (0.18 m3 d−1) divided by the area of CW unit (10 m2)) by the inlet TP concentration (mg L−1). The TP removal rate (g m−2 d−1) was defined as the hydraulic loading rate (m d−1) multiplied by the difference in TP concentration (mg L−1) between the inlet and outlet. It has been reported that the removal rate constants are usually not affected by the inlet concentrations of pollutants by simulation methods (Kadlec, 2000). To exclude the effect of TP inlet concentration and loading rate and estimate the removal capacity of CWs, the area-based removal rate constant was estimated by the first-order plug flow kinetic model (Kadlec and Wallace, 2009): ln

C o −C k ¼− C i −C q

ð1Þ

where C* is the background concentration (mg L−1; 0 was selected in this study); q is the hydraulic loading rate (m yr−1); and k is a firstorder removal rate constant (m yr−1). For a harvest of M. aquaticum, the harvested phosphorus mass of M. aquaticum in the SFCWs was calculated by multiplying harvest biomass (dry weight) and phosphorus content of the harvested part of M. aquaticum(dry weight). The annual harvested phosphorus mass was sum of the all harvest mass during one year. Pearson correlation analysis was applied to explore the correlations of TP concentrations in the inlet and outlet, and correlations of TP removal rates with loading rates. The independent-samples and paired-samples t-test were used to observe significant difference in the TP concentrations of sediments, and removal rate constants for different loading rates and phosphate adsorption capacity by CW1 sediments, respectively. All statistical analyses were performed with PASW Statistics 18 (Chicago, IL, USA). 3. Results and discussion 3.1. Treatment performance of SFCWs Over the monitoring period, the average phosphorus mass loading rates of the swine wastewaters LS, MS, and HS in the SFCWs were 0.459 ± 0.217, 0.801 ± 0.307, and 0.996 ± 0.408 g m−2 d−1, respectively (Table 1). Fig. 2 shows TP concentrations in the influent and effluent and removal efficiency by the SFCWs over time. The average TP concentrations of LS, MS, and HS were 25.5 ± 12.0, 44.5 ± 17.0, and 55.4 ± 22.6 mg L−1 in the influent, and 3.0 ± 4.4, 8.1 ± 9.4, and 17.0 ± 14.9 mg L−1 in the effluent, respectively. During the period 2014–2016, the mean cumulative TP removal efficiencies for LS, MS, and HS were 78.2%, 83.9%, and 89.8%, respectively. The overall TP removal efficiencies for LS, MS, and HS were 89.4 ± 14.6%, 80.9 ± 19.9%, and 70.1 ± 23.3%, respectively. The TP removal efficiencies in the SFCWs planted with M. aquaticum were greater than those in other CWs used to treat domestic wastewater (removal efficiency: 37–60%) (Vymazal, 2007), dairy farm wastewater (removal efficiency: 37–74%) (Tanner et al., 1995), and swine wastewater (removal efficiency: 13–45%) (Poach et al., 2004; Reddy et al., 2001). The mean areal TP mass removal rates in the SFCWs fed with LS, MS, and HS were 0.412 ± 0.210, 0.672 ± 0.309, and 0.779 ± 0.385 g m−2 d−1, respectively, higher than those reported by other studies for treating livestock wastewaters, such as 0.13–0.32 g m−2 d− 1 by Tanner et al. (1995), mean 0.10–0.40 g m−2 d− 1 by Poach et al. (2004), and 0.029–0.041 g m−2 d−1 by Reddy et al. (2001). This suggests that the pilot-scale SFCWs planted with M. aquaticum were effective in treating swine wastewater with high TP concentrations. Some previous studies also demonstrated that in a laboratory scale environment M. aquaticum could tolerate high-strength swine wastewater and remove nutrients from polluted waters with high removal efficiency (Liu et al., 2016; Souza et al., 2013). We confirm M. aquaticum can tolerate highstrength swine wastewater and effectively remove P from wastewater in long-term cycles in natural environmental field conditions. Nevertheless, TP removal rates in the present study were lower than those in CWs planted with Eichhornia crassipes (water hyacinth) treating pretreated swine effluent (0.84–2.12 g m− 2 d−1) (Lee et al., 2004), but the latter with low removal efficiency (47–59%). The TP loading rates and removal rates by the SFCWs were linearly correlated and fitted the equation TP removal rates = 0.779 × TP loading rates (R2 = 0.713; Fig. S1). The ratio of TP removal rates to loading rates in the SFCWs ranged from 0.5 to 1 with different loading rates (Fig. S2), and only some ratios with high loading rates were lower than 0.5, suggesting that the SFCWs planted with M. aquaticum can remove phosphorus stably and efficiently. The SFCWs were less effective in TP removal when loaded with more than ~ 0.6 g m− 2 d− 1 swine wastewater, which has also been observed in other studies (Kadlec and Wallace, 2009; Lin et al., 2002; Tanner et al., 1995) where TP removal rates or efficiencies were lower when the loading was over a certain


493

Table 1 The mean inlet and outlet TP concentrations, inlet loading rate, removal efficiency, the first-order removal rate constant and removal rate in the three-stage surface flow constructed wetland system with three different strengths of wastewater. CW1, CW2, and CW3 indicate the first, second, and third unit of the constructed wetlands (Fig. 1).

Inlet concentration (mg L−1) Outlet concentration (mg L−1) Inlet loading rate (g m−2 d−1) Removal efficiency (%) First-order removal rate constant (m yr−1) Mass removal rate (g m−2 d−1) CW1 CW2 CW3 Total

Low strength

Medium strength

High strength

25.5 ± 12.0 3.0 ± 4.4 0.459 ± 0.217 89.4 ± 14.6 22.3 ± 10.4

44.5 ± 17.0 8.1 ± 9.4 0.801 ± 0.307 80.9 ± 19.9 17.7 ± 11.8

55.4 ± 22.6 17.0 ± 14.9 0.996 ± 0.408 70.1 ± 23.3 12.1 ± 10.2

0.247 0.117 0.044 0.412

0.310 0.231 0.136 0.672

0.360 0.243 0.178 0.779

± ± ± ±

0.148 0.098 0.057 0.210

threshold. On the other hand, a statistically significant positive correlation between TP inflow and outflow concentrations was observed for LS (p = 0.005) but not found for MS (p = 0.081) or HS (p = 0.101). This suggests there may be different roles of the various removal mechanisms in TP removal in the SFCWs. The area-based first-order TP removal rate constants for the three loading rates were 22.3 ± 10.4, 17.7 ± 11.8, and 12.1 ± 10.2 m yr−1, respectively. The TP removal rate constants were significantly decreased

Inlet Concentration

± ± ± ±

0.223 0.140 0.085 0.309

± ± ± ±

0.296 0.166 0.112 0.385

with increasing loading rates (p b 0.01; paired-samples t-test). A possible explanation was that higher phosphorus loading rates may decrease the plants uptake rates in CWs (Tanner et al., 1995). We observed that the growth rate and status of M. aquaticum under three loading rates were different in appearance, i.e., the largest harvest biomass was found in LS and the smallest biomass in HS (Table 2). The TP removal rate constants in the present study were much higher than those for marsh-pond-marsh wetland systems (1.1–1.7 m yr− 1) (Stone et al.,

Outlet Concentration

80

Removal Efficiency

(a) Low Strength

100 80

60

60 40 40 20

20

-1

(b) Medium Strength

0 100 80

80 60 40 40 20 0 160

(c) High Strength

Removal Efficiency (%)

TP concentration (mg L )

0 120

0 100 80

120

60 80 40 40

20

0 0

100

200

300

400

500

0 600

Sampling Time (d) Fig. 2. Mean TP concentrations in inlet and outlet and removal efficiency of TP in the constructed wetlands planted with M. aquaticum treating swine wastewater during the monitoring period (2014–2016).


Table 2 Annual harvest biomass of M. aquaticum, plant uptake of TP, and contribution of plant harvest to TP removal in the three-stage surface flow constructed wetland system with three different strengths of wastewater. CW1, CW2, and CW3 indicate the first, second, and third unit of the constructed wetlands (Fig. 1). Low strength Plant harvest biomass (kg yr CW1 252 ± 3 CW2 308 ± 19 CW3 243 ± 38 Total 804 ± 56 TP uptake by plant (g yr−1) CW1 302 ± 28 CW2 387 ± 32 CW3 207 ± 56 Total 897 ± 80

Medium strength

High strength

148 244 305 697

± ± ± ±

15 26 21 28

40.2 ± 5.6 180 ± 17 240 ± 7 461 ± 17

151 382 409 942

± ± ± ±

25 52 26 70

27.2 ± 5.6 237 ± 22 394 ± 39 640 ± 37

−1

)

TP mass removal rate by the SFCWs (g yr−1) CW1 902 ± 543 1131 ± 812 CW2 428 ± 356 842 ± 509 CW3 162 ± 209 495 ± 311 Total 1505 ± 766 2454 ± 1128

1313 ± 1079 888 ± 606 649 ± 410 2845 ± 1406

Average percentage of TP plant uptake to mass removal rate (%) CW1 33.5 13.4 CW2 90.4 77.2 CW3 100 82.6 Total 59.6 38.4

2.1 26.7 60.7 22.5

2004) and livestock wastewater treatment wetlands (mean: 8 m yr−1; range: 2–18 m yr−1) (Knight et al., 2000), but were comparable with those reported in Kadlec and Wallace (2009) (median: 10.0 m yr−1; 10th–90th percentile range: 1.4–60 m yr−1). The TP concentration of water in the SFCWs declined with increasing distance from the inlet in an approximately exponential pattern (Fig. 3), which has also been observed in other studies (Knight et al., 2000). This trend is important for constructing wetlands for treating wastewater. For example, when a stable loading rate is given, we can calculate the length and area of the CW needed in order to remove pollutants from wastewater effectively. In the present study, for LS, MS, and HS, we need mean lengths of 10, 15, and 30 m of SFCW respectively to obtain removal of TP, resulting in the outlet TP concentrations being lower than the Wastewater Discharge Standard of TP for livestock and poultry breeding in China (8 mg L−1; GB 18596-2001). Therefore, the trend is important for the design and construction of treatment wetlands for practical application.

3.2. Temporal variation of phosphorus removal in SFCWs TP removal rates during the initial stage were close to 100%, and then decreased in about 100–300 d and continually decreased and fluctuated over time (Fig. 2). It has been reported that over 90% of the total influent TP concentrations were removed in young CWs but that the ratios declined sharply after a period of time, even at low loading rates (lower than 2.7 g m− 2 yr−1) (Dzakpasu et al., 2015; Faulkner and Richardson, 1989). The consistently high retention rate at the beginning was likely due to higher availability of sorption sites, and increased (micro)biological activity and plant uptake of TP during the rapid expansion of plant (Hijosa-Valsero et al., 2012; Kadlec and Wallace, 2009). The high loading of nutrients may influence the rapid growth of M. aquaticum, and shows a sharp increase in biomass at the initial stage (Souza et al., 2013). TP removal rates decreased after approximately three months or one year in all SFCWs and this can be attributed to the gradual saturation of sorption sites for TP (Dzakpasu et al., 2015; Kadlec and Wallace, 2009; Tanner et al., 1995). A strong seasonal pattern was found that TP removal rates in spring and autumn were higher than those in summer and winter (Fig. 4). Barca et al. (2013) observed that TP removal efficiencies decreased

1.5

(a) Low Strength

1.0

.5

0.0 1.5

TP Removal Rates (g m-2 d-1)

494

(b) Medium Strength

1.0

.5

0.0 2.0

(c) High Strength 1.5

1.0

.5

0.0 Spring

Fig. 3. Change of TP concentrations in water in the constructed wetlands planted with M. aquaticum with increasing distance from inlet by exponential fitting. Wastewater Discharge Standard of TP for livestock and poultry breeding in China is 8 mg L−1 (GB 18596-2001).

Summer

Autumn

Winter

Fig. 4. Seasonal variation of TP removal rate in the constructed wetlands planted with M. aquaticum with (a) low, (b) medium, and (c) high loading rates. Black solid and red dashed lines represent the median and mean values, respectively. The box plots represent 25th–75th percentiles, and the whiskers represent the 10th and 90th percentiles.


drastically during winter and then began to gradually increase during spring. Reduced performance of the SFCWs during winter and summer was likely the result of several differences between seasons (Majer Newman et al., 1999). Water temperature can influence the physicochemical removal mechanisms of phosphorus in SFCWs. It has been reported that TP adsorption to the substratum is temperature dependent and more efficient in warmer climates (Barca et al., 2013; Shilton et al., 2006). On the other hand, growth of M. aquaticum would be inhibited to some extent in high temperature conditions (N35 °C) during summer or even stopped in low temperature conditions (b 5 °C) during winter. This decreased TP uptake by M. aquaticum and affected TP removal efficiencies in summer and winter. The mean TP removal rate constants were highest at 30.1 m yr−1 (mean) for all loading rates in the first few months and gradually decreased to stabilize around a value of 5 m yr−1 (Fig. S2). This decrease in TP removal rate constants could be attributed to the decrease in temperature from summer to winter and associated decrease in plant growth capacity (Barca et al., 2013). The decreased rate of TP removal rate constants was

1.0

(a) Low Strength

495

fastest in HS, and slowest in LS. After about 200 and 400 d, the TP removal rate constants seemed to slightly improve in spring and autumn and also varied seasonally. A plausible explanation is the annual patterns of vegetation in temperate climates (Kadlec and Wallace, 2009). Another probable reason was that the solubility of various calcium phosphates decreases with increasing temperature, favoring precipitation (Barca et al., 2013). 3.3. Possible phosphorus removal mechanisms The four phosphorus retention mechanisms in the CWs included chemical and physical adsorption, precipitation, plant uptake, soil accretion and sedimentation (Arias and Brix, 2005; Kadlec and Wallace, 2009; Vymazal, 2007; Wu et al., 2015a). Distribution of phosphorus in the CWs was mainly concentrated in two media, i.e., adsorption and precipitation in substrate, and uptake and removal by plants. Therefore, substrate and plant can play vitally important roles in TP removal in CWs. In the present study, TP concentrations in sediment gradually increased over time in CW1, independent of inlet loading rates (Fig. 5).

CW1

CW2

CW3

.8 .6 .4

TP concentration in sediment (mg/kg)

.2 0.0 1.0

(b) Medium Strength

.8 .6 .4 .2 0.0 1.0

(c) High Strength

.8 .6 .4 .2 0.0

07 08 09 10 11 12 01 02 03 04 05 06 07 08 09 10 11 14 14 14 14 14 14 15 15 15 1 5 15 1 5 15 15 15 15 15 20 20 20 20 20 20 20 20 20 20 20 2 0 20 2 0 2 0 20 20

Sampling time Fig. 5. Concentrations of TP in sediment of the constructed wetlands planted with M. aquaticum over time. CW1, CW2, and CW3 indicate the first, second, and third unit of the constructed wetlands (Fig. 1).

496


This suggests that adsorption and precipitation by sediment was an important removal pathway in CW1. From the beginning, TP concentrations of CW1 sediments did not reach an asymptote, suggesting the SFCW had not reached a maximum TP storage capacity. Meanwhile, we also found phosphate adsorption capacity by CW1 sediments was significant higher (p b 0.01; paired-samples t-test) in September 2015 than that in September 2014 (detailed in the Supplementary material), suggesting that the compositions of sediment may have changed over time. It has been reported that the chemical composition of electric arc furnace steel slag by X-ray fluorescence analyses changed with continuous treatment in CWs (Drizo et al., 2002). On the other hand, our previous study showed that vigorous growth of M. aquaticum could rapidly accumulate organic carbon in ditch soils, which enhanced phosphorus adsorption capacity of the soils (Liu et al., 2013). Furthermore, we observed some black deposits on the surface of sediment in CW1 but not CW2 or CW3, which can be attributed to precipitation of total suspended solids from the swine wastewater in CW1. Dzakpasu et al. (2015) also observed that most suspended solids from domestic wastewater were deposited in the first wetland cell. On the other hand, the black deposits in CW1 may contain a certain amount of sulfites or sulfides. The environmental conditions near the bottom, especially in the inlets, may be locally anaerobic, in favor of sulfate reduction and consequent formation of sulfites or sulfides (Chen et al., 2016). The TP concentrations in the CW1 sediment decreased slightly in the early spring (March and April), which may be due to the quick growth of M. aquaticum and relatively low inlet loading rates during this period. The TP concentrations of sediment barely increased in CW2 and CW3 for LS and MS, and CW3 for HS, while the TP concentrations of water decreased markedly in the inlet and outlet of CW2 and CW3 (Fig. S3). These results are largely attributed to uptake and harvest management of M. aquaticum, besides the substance sorption (Table 2). In CW units, except for HS CW1, the phosphorus content of the harvested part of M. aquaticum ranged from 3.5 to 8.9 g kg−1 dry weight, and the annual harvested phosphorus mass was 15.1–40.9 g m−2 yr−1 (harvest biomass was 13.0–32.8 kg wet weight m−2 yr−1, i.e., 1.95–4.90 kg dry weight m−2 yr−1). These results were higher than those reported for other wetland plants, such as Cyperus papyrus (5.0 g m−2 yr−1), Phragmites australis (12.0 g m− 2 yr−1), Typha latifolia (18.0 g m−2 yr− 1), Pistia stratiotes (4.0 g m− 2 yr− 1), Potamogeton pectinatus (4.0 g m−2 yr− 1), and Ceratophyllum demersum (1.0 g m− 2 yr−1) (Kivaisi, 2001). This indicates that TP uptake by M. aquaticum and its multiple harvests were key to TP removal in our pilot-scale SFCWs for treating swine wastewater. Comparison of TP removal by harvest of M. aquaticum with the total mass removal rates by SFCWs (Table 2) showed that TP uptake by M. aquaticum in CW3 with different strengths was higher than other mechanisms, accounting for 60.7–100% of TP removal, demonstrating the important role of M. aquaticum uptake in TP removal in the SFCWs. The amount of nutrients removed through harvest has been considered to be limited and usually does not represent a substantial part of the inflow loading for heavily loaded systems, while removal via harvesting may be important only for lightly loaded systems (Březinová and Vymazal, 2015). In the present study, however, TP removal through multiple harvests of M. aquaticum was reasonable in treating heavily loaded swine wastewater. An important reason for this is that M. aquaticum can be subjected to multiple harvests and can tolerate high concentrations of swine wastewater. Multiple harvesting of M. aquaticum from the SFCWs can contribute 22.5–59.6% of TP removal mass, suggesting M. aquaticum has great potential for aquaculture wastewater treatment. On the other hand, multiple harvests of M. aquaticum can lower burdens on the substrate and prolong the service time in the long-term, because TP concentrations in sediments in CW2 and CW3 for LS and MS barely increased for more than a year and a half. Nevertheless, TP uptake by M. aquaticum in CW1 is low efficiency, and only contributed 2.1–33.5% TP removal. It is possible that black deposits on the surface

of the sediment adversely affected plant growth. Several studies have suggested that overall nutrient removal would be higher if a multiple harvest scheme were adopted (Jinadasa et al., 2008; Vymazal et al., 2010). The multiple harvest scheme in the present study may be appropriate. Additionally, the burial of plant residuals in CWs is also a potential TP removal pathway. It was reported that about 10–20% of the assimilated phosphorus in plant could be permanently stored in sediment as the residual during the decomposition process (Kadlec and Wallace, 2009). Therefore, the decaying biomass of M. aquaticum could not decompose completely and some phosphorus from the undecomposed biomass would become a part of newly formed sediment layer on the bottom.

3.4. Application potential of M. aquaticum in constructed wetland Based on the findings of this study, long-term nutrient removal by multiple harvests of M. aquaticum shoots can be considered as a better plant management approach in CW systems. The harvested biomass of M. aquaticum could be processed into composts and animal feed, because the aquatic plant possesses many advantages, such as a high production rate, easy harvest, high crude protein content, and low fiber content. Therefore, M. aquaticum may be a better alternative plant in CWs treating high-load wastewater and may harbor a promising prospect for application to sewage treatment. However, frequent aboveground harvesting may slow growth and biomass development (Álvarez and Bécares, 2008). Some plants in CWs cannot be harvested multiple times while preserving a high or normal removal efficiency in wastewater. For example, harvesting Phragmites spp. during the growing season may lead to serious damage of the stand because this plant translocates reserve products only very late in the season (Asaeda et al., 2006). Furthermore, multiple harvests of plants should be based on consideration of economic, climatic and wetland operation conditions. Therefore, further studies are needed (including effects of harvesting on physiological and biochemical characteristics of M. aquaticum) in order to provide more evidence for multiple harvests of M. aquaticum in field applications. Optimization of the harvesting strategy needs further investigations aimed at enhancing the performance of pollutant removal. An appropriate plant harvesting strategy is dependent on plant species, growth stages, times of harvest, the density of harvested biomass, parts of CWs, and climatic conditions. In the present study, the SFCWs were operated for relatively short durations (less than two years). Therefore, the results derived may not be reflective of long-term removal capacity. Furthermore, we did not discuss the role of microbial processes on TP removal. Many studies have reported that polyphosphate-accumulating organisms can take up phosphorus from wastewater (Oehmen et al., 2007; Yuan et al., 2012). Further studies are needed to acquire more data for SFCWs planted with M. aquaticum for treatment of swine wastewater.

4. Conclusions The pilot-scale SFCWs planted with M. aquaticum can remove phosphorus effectively from swine wastewater, with a mean removal efficiency of 70.1–89.4%. Mean TP removal rate constants were 12.1–22.3 m yr−1. The main TP removal mechanisms in the SFCWs varied in each wetland unit from inlet to outlet. Phosphorus in the swine wastewater was removed mainly via adsorption and precipitation by sediment in CW1 and via uptake and harvest of the plant M. aquaticum in CW3. Multiple harvests of M. aquaticum could enhance TP removal efficiency and played an important role in TP removal in the SFCW units with relatively low loading rates (such as CW2 and CW3). M. aquaticum is appropriate for application in constructed wetland systems for treating swine wastewater.


Acknowledgments The present study was financially supported by the National Science and Technology Supporting Project (2014BAD14B01, 2014BAD14B05), the National Natural Science Foundation of China (no. 41601272), the Key CAS Program (KZZD-EW-11-3), and the Key Projects of International Cooperation in Science and Technology Innovation (S2016G0053). Special thanks to the Public Service Technology Center, Institute of Subtropical Agriculture, Chinese Academy of Sciences for instrument support. We thank all participants in sample collection and analysis. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2016.10.094. References Álvarez, J.A., Bécares, E., 2008. The effect of plant harvesting on the performance of a free water surface constructed wetland. Environ. Eng. Sci. 25, 1115–1122. Arias, C.A., Brix, H., 2005. Phosphorus removal in constructed wetlands: can suitable alternative media be identified? Water Sci. Technol. 51, 267–273. Asaeda, T., Rajapakse, L., Manatunge, J., Sahara, N., 2006. The effect of summer harvesting of Phragmites australis on growth characteristics and rhizome resource storage. Hydrobiologia 553, 327–335. Barca, C., Troesch, S., Meyer, D., Drissen, P., Andrès, Y., Chazarenc, F., 2013. Steel slag filters to upgrade phosphorus removal in constructed wetlands: two years of field experiments. Environ. Sci. Technol. 47, 549–556. Březinová, T., Vymazal, J., 2015. Seasonal growth pattern of Phalaris arundinacea in constructed wetlands with horizontal subsurface flow. Ecol. Eng. 80, 62–68. Buddhavarapu, L.R., Hancock, S.J., 1991. Advanced treatment for lagoons using duckweed. Water Environ. Technol. 3, 41–44. Cronk, J.K., 1996. Constructed wetlands to treat wastewater from dairy and swine operations: a review. Agric. Ecosyst. Environ. 58, 97–114. Chen, Y., Wen, Y., Zhou, Q., Huang, J., Vymazal, J., Kuschk, P., 2016. Sulfate removal and sulfur transformation in constructed wetlands: the roles of filling material and plant biomass. Water Res. 102, 572–581. Drizo, A., Comeau, Y., Forget, C., Chapuis, R.P., 2002. Phosphorus saturation potential: a parameter for estimating the longevity of constructed wetland systems. Environ. Sci. Technol. 36, 4642–4648. Dunne, E.J., Reddy, K.R., 2005. Phosphorus biogeochemistry of wetlands in agricultural watersheds. In: Dunne, E.J., Reddy, K.R.R., Carton, O.T. (Eds.), Nutrient Management in Agricultural Watersheds: A Wetlands Solution, pp. 105–119. Dzakpasu, M., Scholz, M., McCarthy, V., Jordan, S.N., 2015. Assessment of long-term phosphorus retention in an integrated constructed wetland treating domestic wastewater. Environ. Sci. Pollut. Res. 22, 305–313. Faulkner, S.P., Richardson, C.J., 1989. Physical and chemical characteristics of freshwater wetland soils. In: Hammer, D.A. (Ed.), Constructed Wetlands for Wastewater Treatment: Municipal, Industrial, and Agricultural. CRC Press LLC, Boca Raton, Florida, pp. 41–72. Hijosa-Valsero, M., Sidrach-Cardona, R., Bécares, E., 2012. Comparison of interannual removal variation of various constructed wetland types. Sci. Total Environ. 430, 174–183. Ibekwe, A.M., Ma, J., Murinda, S., Reddy, G.B., 2016. Bacterial community dynamics in surface flow constructed wetlands for the treatment of swine waste. Sci. Total Environ. 544, 68–76. Jinadasa, K., Tanaka, N., Sasikala, S., Werellagama, D., Mowjood, M., Ng, W., 2008. Impact of harvesting on constructed wetlands performance—a comparison between Scirpus grossus and Typha angustifolia. J. Environ. Sci. Health A 43, 664–671. Kadlec, R.H., 2000. The inadequacy of first-order treatment wetland models. Ecol. Eng. 15, 105–119. Kadlec, R.H., Wallace, S.D., 2009. Treatment Wetlands. CRC Press, Boca Raton, FL. Kivaisi, A.K., 2001. The potential for constructed wetlands for wastewater treatment and reuse in developing countries: a review. Ecol. Eng. 16, 545–560.

497

Knight, R.L., Payne Jr., V.W.E., Borer, R.E., Clarke Jr., R.A., Pries, J.H., 2000. Constructed wetlands for livestock wastewater management. Ecol. Eng. 15, 41–55. Lee, C.-Y., Lee, C.-C., Lee, F.-Y., Tseng, S.-K., Liao, C.-J., 2004. Performance of subsurface flow constructed wetland taking pretreated swine effluent under heavy loads. Bioresour. Technol. 92, 173–179. Lin, Y.-F., Jing, S.-R., Lee, D.-Y., Wang, T.-W., 2002. Nutrient removal from aquaculture wastewater using a constructed wetlands system. Aquaculture 209, 169–184. Liu, F., Xiao, R., Wang, Y., Li, Y., Zhang, S., Luo, Q., et al., 2013. Effect of a novel constructed drainage ditch on the phosphorus sorption capacity of ditch soils in an agricultural headwater catchment in subtropical Central China. Ecol. Eng. 58, 69–76. Liu, F., Zhang, S., Wang, Y., Li, Y., Xiao, R., Li, H., et al., 2016. Nitrogen removal and mass balance in newly-formed Myriophyllum aquaticum mesocosm during a single 28day incubation with swine wastewater treatment. J. Environ. Manag. 166, 596–604. Majer Newman, J., Clausen, J.C., Neafsey, J.A., 1999. Seasonal performance of a wetland constructed to process dairy milkhouse wastewater in Connecticut. Ecol. Eng. 14, 181–198. Oehmen, A., Lemos, P.C., Carvalho, G., Yuan, Z., Keller, J., Blackall, L.L., et al., 2007. Advances in enhanced biological phosphorus removal: from micro to macro scale. Water Res. 41, 2271–2300. Ong, S.A., Uchiyama, K., Inadama, D., Ishida, Y., Yamagiwa, K., 2010. Performance evaluation of laboratory scale up-flow constructed wetlands with different designs and emergent plants. Bioresour. Technol. 101, 7239–7244. Poach, M.E., Hunt, P.G., Reddy, G.B., Stone, K.C., Johnson, M.H., Grubbs, A., 2004. Swine wastewater treatment by marsh-pond-marsh constructed wetlands under varying nitrogen loads. Ecol. Eng. 23, 165–175. Reddy, G.B., Hunt, P.G., Phillips, R., Stone, K., Grubbs, A., 2001. Treatment of swine wastewater in marsh-pond-marsh constructed wetlands. Water Sci. Technol. 44, 545–550. Shilton, A.N., Elmetri, I., Drizo, A., Pratt, S., Haverkamp, R.G., Bilby, S.C., 2006. Phosphorus removal by an ‘active’ slag filter–a decade of full scale experience. Water Res. 40, 113–118. Souza, F.A., Dziedzic, M., Cubas, S.A., Maranho, L.T., 2013. Restoration of polluted waters by phytoremediation using Myriophyllum aquaticum (Vell.) Verdc., Haloragaceae. J. Environ. Manag. 120, 5–9. Stone, K.C., Poach, M.E., Hunt, P.G., Reddy, G.B., 2004. Marsh-pond-marsh constructed wetland design analysis for swine lagoon wastewater treatment. Ecol. Eng. 23, 127–133. Tanner, C.C., Clayton, J.S., Upsdell, M.P., 1995. Effect of loading rate and planting on treatment of dairy farm wastewaters in constructed wetlands—II. Removal of nitrogen and phosphorus. Water Res. 29, 27–34. Vohla, C., Kõiv, M., Bavor, H.J., Chazarenc, F., Mander, Ü., 2011. Filter materials for phosphorus removal from wastewater in treatment wetlands—a review. Ecol. Eng. 37, 70–89. Vymazal, J., 2007. Removal of nutrients in various types of constructed wetlands. Sci. Total Environ. 380, 48–65. Vymazal, J., 2011. Plants used in constructed wetlands with horizontal subsurface flow: a review. Hydrobiologia 674, 133–156. Vymazal, J., 2013. Emergent plants used in free water surface constructed wetlands: a review. Ecol. Eng. 61 (Part B), 582–592. Vymazal, J., Kröpfelová, L., Švehla, J., Štíchová, J., 2010. Can multiple harvest of aboveground biomass enhance removal of trace elements in constructed wetlands receiving municipal sewage? Ecol. Eng. 36, 939–945. Wu, H., Zhang, J., Li, P., Zhang, J., Xie, H., Zhang, B., 2011. Nutrient removal in constructed microcosm wetlands for treating polluted river water in northern China. Ecol. Eng. 37, 560–568. Wu, H., Zhang, J., Ngo, H.H., Guo, W., Hu, Z., Liang, S., et al., 2015a. A review on the sustainability of constructed wetlands for wastewater treatment: design and operation. Bioresour. Technol. 175, 594–601. Wu, H., Fan, J., Zhang, J., Ngo, H.H., Guo, W., Liang, S., et al., 2015b. Strategies and techniques to enhance constructed wetland performance for sustainable wastewater treatment. Environ. Sci. Pollut. Res. 22, 14637–14650. Wu, H., Zhang, J., Li, C., Fan, J., Zou, Y., 2013. Mass balance study on phosphorus removal in constructed wetland microcosms treating polluted river water. Clean: Soil, Air, Water 41, 844–850. Yan, Y., Xu, J., 2014. Improving winter performance of constructed wetlands for wastewater treatment in northern China: a review. Wetlands 34, 243–253. Yuan, Z., Pratt, S., Batstone, D.J., 2012. Phosphorus recovery from wastewater through microbial processes. Curr. Opin. Biotechnol. 23, 878–883.