A pilot study was conducted for 7 months for the City of Oxnard, California, on the use of constructed wetlands to treat concentrate produced by microfiltration and reverse osmosis (RO) of reclaimed wastewater. The treatment performance of a transportable subsurface-flow wetland was investigated by monitoring various forms of nitrogen, orthophosphate, oxygen demand, organic carbon, and selenium. Significant mass removal of constituents was measured under two hydraulic residence times (HRTs) (2.5 and 5 days). Inflow and outflow concentrations of nitrate-N and ammonia-N were significantly different for both HRTs, whereas nitrite-N and total organic carbon (TOC) were significantly different during HRT2. Mass removal by the constructed wetland averaged 61% of nitrate-N, 32% of nitrite-N, 42% of ammonia-N, 43% of biochemical oxygen demand, 19% of orthophosphate as P, 18% of TOC and 61% of selenium. Mass removal exceeded concentration reductions through water volume loss through evapotranspiration. Calibrated first-order area-based removal rates were consistent with literature ranges, and were greater during HRT1 consistent with greater mass loads, higher hydraulic loading and shorter HRTs. The rate constants may provide a basis for sizing a full-scale wetland receiving a similar quality of water. The results indicated that engineered wetlands can be useful in the management of RO membrane concentrate for reclaimed water reuse.

INTRODUCTION

Membrane technologies for water treatment produce significant quantities of concentrate with elevated concentrations of metals, nutrients and inorganic ions often exceeding water quality standards. Common disposal methods for concentrate include ocean discharge to coastal waters and deep-well injection or evaporation ponds inland (Mickley 2009; Xu et al. 2013). Stringent water quality regulations may further prohibit direct disposal and will mandate extensive treatment prior to disposal. Managing reverse osmosis (RO) concentrate or brine or brackish waters produced by membrane treatment is a significant challenge and there is an urgent need for environmentally compatible and cost-effective management options for the concentrate (Perez-Gonzalez et al. 2012).

The use of wetlands to treat and reuse concentrate is receiving increased consideration as a concentrate management technique (WateReuse Foundation 2006; Kepke et al. 2009). Treatment wetlands have a proven capacity for removal of pollutants (Kadlec & Wallace 2009), even in high salinity wastewaters (Webb et al. 2012). A conceptual alternative to ocean disposal could be to use the intrinsic chemical, biological and physical processes of wetlands to remove excess nutrients and contaminants for discharge or for beneficial use, such as brackish wetland creation or restoration.

As part of a program to make efficient use of water resources, the City of Oxnard, California, USA, has constructed the Advanced Water Purification Facility (AWPF) to provide advanced treatment of the secondary effluent from the current wastewater plant, thereby producing a high-quality recycled water that meets the water quality criteria for groundwater recharge (GWR) and unrestricted irrigation. To meet GWR water quality requirements, the treatment options for the AWPF have been designed to include microfiltration (MF), RO, and ultraviolet disinfection including advanced oxidation capabilities, degasification and water stabilization (Lozier & Ortega 2010). A portion of the RO concentrate will be treated through a constructed wetland treatment system to demonstrate performance.

It was anticipated during design of the AWPF that the constructed wetlands will treat RO concentrate with high concentrations of nitrogen, phosphorus, selenium, and dissolved solids. Therefore, prior to full plant construction a pilot wetland study was conducted with the following goals: (1) monitor the survival and growth of brackish marsh plants receiving the RO concentrate similar to full plant operations; (2) assess the pollutant removal performance of wetlands treating the RO concentrate; (3) evaluate if discharge is ecologically safe to wetland biota; and (4) confirm that the aesthetics of the treatment wetland would be acceptable (i.e., no offensive odors or colors would be generated).

METHODS AND MATERIALS

RO concentrate and wastewater effluent characteristics

The inflow to the pilot wetland was concentrate from MF and RO pilot test units treating secondary treated wastewater. The three-stage RO pilot unit was designed to treat 169 L/d of filtrate from the MF and operate at 85% recovery. The pilot system had a total permeate output of approximately 142 L/d based on an average flux of 0.43 m3/(m2·d). Table 1 compares the secondary effluent quality from the Oxnard wastewater treatment plant with RO concentrate produced in this pilot study. The concentration of chemicals in the concentrate is many times greater than secondary treated wastewater.

Table 1

Comparative water quality of Oxnard wastewater treatment plant effluent and pilot RO concentrate

  Secondary effluent RO concentrate 
Constituent Units Mean SD Mean SD 
Total dissolved solids mg/L 1,750 112.8 11,833 592.2 
Total suspended solids mg/L 4.5 2.0 1.9 0.2 
Turbidity NTU 3.3 1.3 0.9 0.66 
Nitrate as N (NO3-N) mg/L 1.2 1.4 14 4.0 
Total nitrogen as N (TN) mg/L 25.9 5.8 170 
Ammonia as N (NH3-N) mg/L 22.2 3.3 121.7 49.1 
Total organic carbon mg/L 16.6 2.2 72.3 3.6 
Alkalinity (as CaCO3mg/L 316 12.6 1,487 424.9 
Boron mg/L 1.22 0.5 3.6 – 
Chloride mg/L 415 38.4 2,773 137.1 
Iron (total) μg/L 284 88.1 451 66.8 
Silica mg/L 26.7 1.9 169.7 17.6 
Sulfate mg/L 480 63.3 3,678 610.4 
pH (standard units) – 7.83 0.10 7.3 0.1 
  Secondary effluent RO concentrate 
Constituent Units Mean SD Mean SD 
Total dissolved solids mg/L 1,750 112.8 11,833 592.2 
Total suspended solids mg/L 4.5 2.0 1.9 0.2 
Turbidity NTU 3.3 1.3 0.9 0.66 
Nitrate as N (NO3-N) mg/L 1.2 1.4 14 4.0 
Total nitrogen as N (TN) mg/L 25.9 5.8 170 
Ammonia as N (NH3-N) mg/L 22.2 3.3 121.7 49.1 
Total organic carbon mg/L 16.6 2.2 72.3 3.6 
Alkalinity (as CaCO3mg/L 316 12.6 1,487 424.9 
Boron mg/L 1.22 0.5 3.6 – 
Chloride mg/L 415 38.4 2,773 137.1 
Iron (total) μg/L 284 88.1 451 66.8 
Silica mg/L 26.7 1.9 169.7 17.6 
Sulfate mg/L 480 63.3 3,678 610.4 
pH (standard units) – 7.83 0.10 7.3 0.1 

SD: standard deviation; N: number of samples.

Study area

The City of Oxnard is situated in Ventura County about 100 km west of Los Angeles, California. The climate is mild, semi-arid with a temperature range from 6 to 24 °C.

Pilot wetland system

A portable subsurface-flow wetland developed by Mobile Environmental Solutions (MES) of Tustin, California, was installed adjacent to the AWPF pilot facility in the grounds of the City's wastewater treatment plant (Figure 1). The area of the MES wetland trailer was 8.9 m2, filled with 11.9 m3 of soil and gravel growing a mature stand (2 m) of bulrush (Schoenoplectus californicus).

Figure 1

The MES wetland system. The white trailer on the left is the AWPF pilot system. The RO concentrate was conveyed to the constructed wetlands using a flow control device and effluent was sampled from the other end of the trailer.

Figure 1

The MES wetland system. The white trailer on the left is the AWPF pilot system. The RO concentrate was conveyed to the constructed wetlands using a flow control device and effluent was sampled from the other end of the trailer.

Start-up and acclimation

The pilot wetland plant and bacterial communities were acclimated gradually to the total dissolved solids (TDS) levels in the MF/RO concentrate. Secondary effluent was recycled through the wetland for the first week. Over the next 3 weeks sea salt (Instant Ocean®) was added to the wetland to attain a TDS of 5 g/L and finally 11 g/L. During this time the water recycled continuously through the wetland at a rate of 1.9 L/min, for an equivalent hydraulic loading rate (HLR) of 31 cm/d.

All components of the study were operational by the end of September 2008. Initially, the MES wetland was loaded at 1 L/min, yielding an HLR of 12.7 cm/d and a nominal hydraulic residence time (nHRT) of 2.5 d (Table 2).

Table 2

Hydraulic data summary and operational sampling periods

Dates Duration (d) Flows (L/min) nHRT (d) HLR (cm/d) Comments 
1–24 Sept. 2008 23 1.9 1.3 24.5 Initial acclimation period – no sampling 
1 Oct. 2008 to 19 Jan. 2009 110 2.5 12.9 First period of sampling (HRT1) 
20 Jan. to 5 Mar. 2009 40 0.5 6.5 Final period of sampling (HRT2) 
Dates Duration (d) Flows (L/min) nHRT (d) HLR (cm/d) Comments 
1–24 Sept. 2008 23 1.9 1.3 24.5 Initial acclimation period – no sampling 
1 Oct. 2008 to 19 Jan. 2009 110 2.5 12.9 First period of sampling (HRT1) 
20 Jan. to 5 Mar. 2009 40 0.5 6.5 Final period of sampling (HRT2) 

HLR = hydraulic loading rate; nHRT = nominal hydraulic residence time.

Sampling

Between 17 September 2008 and 5 March 2009, the inflow and outflow to the MES wetland were sampled weekly for analysis according to Standard Methods (American Public Health Association 2005) of ammonia-N, nitrate-N, total Kjeldahl nitrogen (TKN), orthophosphate as P, biochemical oxygen demand (BOD5), total organic carbon (TOC), and total selenium (Se). Field measurements of temperature, specific conductance, pH, and ammonia-N were taken bi-weekly initially, then weekly during the study.

Water balance

Wetland inflow was measured in this study but direct outflow measurements were not taken. Hydrologic factors affecting the water balance of wetlands include evaporation, evapotransporation, precipitation, and infiltration (Kadlec & Wallace 2009). Since the MES pilot trailer was sealed and no water was lost through infiltration, water was lost only through evapotranspiration (ET0) and the outflow. By estimating ET0, the outflow was calculated as the residual in the water balance.

Evapotranspiration was estimated based on the difference in measured chloride concentrations at the influent and effluent. The mass balance of a conservative constituent such as chloride is commonly used to confirm water budgets (Kadlec & Wallace 2009). Changes in concentration of chloride are used to measure difference in flow, as: 
formula
1
where Qi = inflow rate, Ci = inflow concentration and Co = outflow concentration.

First-order area-based rate constants

First-order area-based rate constants were calibrated for the treatment wetland model developed by Kadlec & Knight (1996), updated in Kadlec & Wallace (2009). The first-order rate constants determined from this study are provided later in this article.

Operational variation

The nHRT was maintained at 2.5 days during sampling period HRT1 for 110 days, followed by a 5-day nHRT for 40 days during sampling period HRT2. Several interruptions in RO concentrate production were caused by membrane changes, anticoagulant trials, cleaning (backwash) of membranes, power outage and computer or remote control malfunctioning. During HRT1, more dynamic (variable) operating conditions resulted in a wider range of influent concentrations to wetlands than in HRT2. Note that the chemicals added for the operation of the RO unit may have impacted on metal complexes or wetland health but those were not measured or quantified specifically in this study.

RESULTS AND DISCUSSION

Operating conditions and aesthetics

The average HLR decreased from 1 L/min (12.9 cm/d) to 0.5 L/min (6.5 cm/d) during the final sampling period (Table 2).

Estimates of evapotranspiration based on Equation (1) averaged 2.7 mm/d. Water loss determined from inflow and outflow rates varied between 10% and 50% and between 10% and 30% during HRT1 and HRT2, respectively. On average, about 21% of the influent was lost through wetland evapotranspiration.

No objectionable odors were detected from the demonstration wetland. At no point during the 7 month study were indications of stress evident, including tip browning, shoot necrosis, shoot pigment loss, or other indications of plant mortality or injury. Normal flowering and fruiting characteristics were observed.

Water quality analysis

Field parameters

Field measurements for temperature, pH, specific conductance and ammonia-N are summarized in Table 3. Influent water temperature ranged from 14 to 22.5 °C. Effluent temperature ranged between 12 and 18 °C, tracking maximum ambient air temperature. Mean pH decreased slightly by 0.3 units through the wetland during HRT2, indicating alkaline conditions.

Table 3

Summary of field parameter measurements

  Influent Effluent 
Parameter HRT Mean SD Mean SD 
Temperature (°C) 25.30 – 17.40 – – 
18.16 2.8 11 14.91 2.2 11 
pH 7.39 0.1 7.11 0.1 
7.37 0.1 11 7.17 0.1 11 
Conductivity (μS/cm) 15,880 – 17,800 – 
15,103 1,410 11 18,195 984.2 11 
NH3-N (mg/L) 110.6 31.9 12 66.4 30.1 12 
150.0 13.4 11 94.1 8.9 11 
  Influent Effluent 
Parameter HRT Mean SD Mean SD 
Temperature (°C) 25.30 – 17.40 – – 
18.16 2.8 11 14.91 2.2 11 
pH 7.39 0.1 7.11 0.1 
7.37 0.1 11 7.17 0.1 11 
Conductivity (μS/cm) 15,880 – 17,800 – 
15,103 1,410 11 18,195 984.2 11 
NH3-N (mg/L) 110.6 31.9 12 66.4 30.1 12 
150.0 13.4 11 94.1 8.9 11 

SD: standard deviation; N: number of samples.

Effluent specific conductivity was consistently greater than influent conductivity by an average of 17% during HRT2, which agrees with the estimated chloride increase of 21% and is attributable to evapoconcentration caused by water loss through plant evapotranspiration.

Nitrogen

Ammonia and TKN

During both sampling periods, average inflow ammonia-N decreased significantly by about 24% from 146.2 to 112.6 mg/L in HRT1 and from 168.6 to 128.4 mg/L in HRT2 (Table 4). Influent and effluent concentrations for ammonia-N are significantly different (P < 0.05). TKN concentrations were not significantly different. Ammonia-N mass decreased on average by 42% from 14.9 to 8.6 g/(m2·d) during the entire sampling period and TKN mass reduction averaged about 6% (Table 5). Mass removal rates were greater than concentration reductions because of evapotranspiration losses. The non-significant reduction in TKN indicated poorly aerobic conditions in the wetland root zone.

Table 4

Average concentration of parameters in wetland influent and effluent

  Average concentration (mg/L*)   
  Influent Effluent    
Parameter HRT Mean SD Mean SD Percent removal (%) 
NO3-N 7.2 5.2 11 3.0 3.3 11 0.001 58 
14.0 2.3 4.1 2.3 0.004 71 
NO2-N 5.8 7.1 2.1 3.1 0.33 63 
13.7 4.4 6.2 5.1 0.04 55 
NH3-N 146.2 10.7 11 112.6 17.2 11 0.0001 23 
168.6 9.6 11 128.4 6.4 11 0.0001 24 
TKN 146.1 46.7 11 136.4 36.9 11 0.26 
116.1 34.0 113.2 36.2 1.00 
Orthophosphate, as P 17.9 8.6 13.8 4.5 0.43 23 
15.7 4.8 13.9 1.4 0.51 11 
BOD5 10.0 7.2 11 4.9 2.0 11 0.14 51 
4.3 0.8 3.0 1.1 0.13 30 
TOC 77.9 8.9 66.2 15.3 0.15 15 
72.3 3.6 56.5 4.3 0.004 22 
Selenium (μg/L) 35.6 10.2 21.5 6.2 0.08 40 
  Average concentration (mg/L*)   
  Influent Effluent    
Parameter HRT Mean SD Mean SD Percent removal (%) 
NO3-N 7.2 5.2 11 3.0 3.3 11 0.001 58 
14.0 2.3 4.1 2.3 0.004 71 
NO2-N 5.8 7.1 2.1 3.1 0.33 63 
13.7 4.4 6.2 5.1 0.04 55 
NH3-N 146.2 10.7 11 112.6 17.2 11 0.0001 23 
168.6 9.6 11 128.4 6.4 11 0.0001 24 
TKN 146.1 46.7 11 136.4 36.9 11 0.26 
116.1 34.0 113.2 36.2 1.00 
Orthophosphate, as P 17.9 8.6 13.8 4.5 0.43 23 
15.7 4.8 13.9 1.4 0.51 11 
BOD5 10.0 7.2 11 4.9 2.0 11 0.14 51 
4.3 0.8 3.0 1.1 0.13 30 
TOC 77.9 8.9 66.2 15.3 0.15 15 
72.3 3.6 56.5 4.3 0.004 22 
Selenium (μg/L) 35.6 10.2 21.5 6.2 0.08 40 

*Concentrations are in mg/L unless otherwise mentioned.

SD: standard deviation; N: number of samples.

Bold values represent influent and effluent concentrations that are significantly different (P < 0.05) as per Mann–Whitney non-parametric U-test at P < 5% level of significance.

Table 5

Average mass loading rates and mass removal summary

 Average mass loading rates (g/(m2·d))*   
  Influent Effluent  
Parameter Mean SD Mean SD Mass removal (%) 
NO3-N 0.99 0.7 17 0.38 0.4 17 61 
NO2-N 0.5 0.9 15 0.4 0.5 15 32 
NH3-N 14.9 4.2 22 8.6 2.6 22 42 
TKN 15.8 8.8 18 14.9 7.4 18 
Orthophosphate as P 1.9 1.7 14 1.6 1.0 14 19 
BOD5 1.0 0.9 16 0.6 0.3 16 43 
TOC 8.4 0.5 14 7.0 3.0 14 18 
Selenium (mg/(m2·d)) 0.002 0.001 0.001 0.001 61 
Calcium 133.7 52.9 120.3 – 10 
 Average mass loading rates (g/(m2·d))*   
  Influent Effluent  
Parameter Mean SD Mean SD Mass removal (%) 
NO3-N 0.99 0.7 17 0.38 0.4 17 61 
NO2-N 0.5 0.9 15 0.4 0.5 15 32 
NH3-N 14.9 4.2 22 8.6 2.6 22 42 
TKN 15.8 8.8 18 14.9 7.4 18 
Orthophosphate as P 1.9 1.7 14 1.6 1.0 14 19 
BOD5 1.0 0.9 16 0.6 0.3 16 43 
TOC 8.4 0.5 14 7.0 3.0 14 18 
Selenium (mg/(m2·d)) 0.002 0.001 0.001 0.001 61 
Calcium 133.7 52.9 120.3 – 10 

*Mass loading rates represent average for HRT1 and HRT2 sampling periods.

Influent and effluent concentrations of ammonia-N ranged between 190.2 and 75.8 mg/L. The inflow TKN mass load varied between 1,499 and 12,459.3 g/(m2·yr) (average 5,770 g/(m2·yr)). The data ranges were consistent with other highly loaded subsurface flow wetlands, indicating that treatment of RO concentrate from reclaimed water is similar to other high nutrient loaded wetlands system.

Oxidized nitrogen

Nitrate-N showed significant reductions of 58% and 71% during HRT1 and HRT2, respectively (Table 4). Influent and effluent nitrate-N concentrations were significantly different (P < 0.05) as per Mann–Whitney non-parametric U-test at P < 5% level of significance. Nitrate-N reduction varied seasonally, decreasing in February consistent with ambient air temperatures ≤10 °C (Figure 2). Influent water temperature measured in the field followed the trend of air temperature (Utah Climate Center 2015). Nitrite-N showed greater than 50% reduction during both HRTs. Although effluent nitrite-N was not significantly different from inflow during HRT1, effluent nitrite-N was significantly different than inflow during HRT2 as per Mann–Whitney non-parametric U test at P < 5% level of significance (Table 4). The lack of significant P-value during HRT1 is attributed to the greater range of inflow nitrite-N (1–18 mg/L) than during HRT2 (10.9–17.7 mg/L), which appears to track changes in air temperature (Figure 2). These relations to temperature are consistent with general expectations of the influence of temperature on wetland microbial denitrification (Kadlec & Wallace 2009). BOD5, the most labile form of organic carbon, is preferentially used by wetlands to reduce oxidized forms of nitrogen (Kadlec & Wallace 2009). BOD5-rich influent water would have supplied enough energy to initiatea rapid denitrification process in the wetlands treatment.

Figure 2

Influent and effluent NOx and nitrate-N concentrations, influent water temperature and air temperature.

Figure 2

Influent and effluent NOx and nitrate-N concentrations, influent water temperature and air temperature.

NOx (sum of nitrite-N and nitrate-N) mass loading averaged 1.57 g/(m2·d) over the entire study with an average removal efficiency of 56% (Table 5). Mass removal efficiency of nitrate-N and nitrite-N by the constructed wetland averaged 61% and 32%, respectively (Table 5).

BOD and TOC

Influent and effluent TOC were not significantly different during HRT1 but TOC concentrations decreased significantly by 22% during HRT2 (Table 4). Greatest TOC reductions were observed when temperature was ≥10 °C. Mass of TOC was reduced by 18% through the wetland (Table 5). These results suggest assimilation of organic carbon by the wetland, consistent with the microbial requirements to sustain the observed nitrate-N removal rates.

Average BOD5 concentrations decreased by 51% during HRT1 and 30% during HRT2 (Table 4). Inlet and outlet BOD5 concentrations were not significantly different. Operation of the pilot system during HRT1 was more variable, leading to more frequent cleaning, contributing to greater variance in inlet BOD5 than in HRT2. For HRT2, inlet BOD5 was well within the background concentrations expected for treatment wetlands (Kadlec & Wallace 2009). The concentration ratio of BOD5:TOC in the wetland influent varied between 0.06 and 0.13 and in the effluent it varied between 0.05 and 0.07. These ratio differences indicate biologically available carbon compounds were utilized in wetland treatment processes and were discharged with relatively fewer labile components.

Phosphorus

Average inlet and outlet orthophosphate (as P) concentrations were not significantly different. The mass reduction was 19% (Table 4). Both surface flow and subsurface flow treatment wetlands have limited abilities to remove phosphorus (Vymazal 2007). The modest removal of orthophosphate is consistent with the possibility of some export of organic matter in the form of bacterial biomass, root exudates and material, compounded by an evaporative increase in parameter concentrations. Adsorption to the gravel substrate and plant root surface was likely limited given that the wetland had previously been established and sorption sites were relatively saturated.

Inorganic parameters

Inorganic parameters of chloride, TDS, sulfate and alkalinity increased in concentration by 21, 19, 15 and 4%, respectively, over the study (Table 6). Most of this change can be attributed to evapoconcentration through evaporation and plant transpiration. Constructed wetlands have little effect (other than dilution or concentration) on the removal of environmentally conservative parameters such as chloride (Kadlec & Wallace 2009). However, because the wetland was completely ‘root-bound’ and all of the media volume occupied by the plant roots, this density of plants exerts a significant draw upon the water flow through the system. The combined reduction of concentrate flow and salt mass could be a potential benefit of this method of concentrate management.

Table 6

Relative changes in inorganic concentration in water

Parameter Chloride TDS Sulfate Alkalinity 
Inflow (mg/L) 2,908 11,588 3,852 1,612 
Outflow (mg/L) 3,524 13,788 4,432 1,676 
Δ Concentration* 21% 19% 15% 4% 
Parameter Chloride TDS Sulfate Alkalinity 
Inflow (mg/L) 2,908 11,588 3,852 1,612 
Outflow (mg/L) 3,524 13,788 4,432 1,676 
Δ Concentration* 21% 19% 15% 4% 

*Δ Concentration: change in concentration between inflow and outflow (%).

The lesser increase by sulfate and alkalinity indicate the potential for sulfate reduction in the anaerobic root zone and the precipitation of calcium-bound carbonate and sulfate. Calcium mass decreased by about 10% (Table 5). Alkalinity was a likely source of carbon and sustained a neutral system pH for denitrification.

Average total selenium decreased 40% from 35.6 to 21.5 μg/L (Table 4). For reference, the total selenium in the secondary effluent water source averaged 11.5 μg/L, ranging from 11.1 to 12.3 μg/L, and comprised 72% organic selenium (9.62 ± 0.68 μg/L), 14% inorganic selenium (1.93 ± 0.71 μg/L), 11% selenite (1.48 ± 0.65 μg/L), and 3% selenate (0.45 ± 0.20 μg/L). Transformation of selenate to selenite and then to reduced elemental selenium, and the volatilization of organic selenium could likely account for the observed selenium reduction. Organic selenium mass is likely to be adsorbed on soil and plant root surface, and then can readily be converted to volatile forms (Lin & Terry 2003). Outflow selenium concentrations were greater than the US Environmental Protection Agency selenium criterion of 5 μg/L for aquatic wildlife protection (NOAA 1999). Additional treatment through larger area or other processes would be required to discharge to natural systems.

Comparative analysis and first-order rates

Table 7 provides a summary of the calibrated first-order rate coefficients measured for HRT1 and HRT2. Values of C* (background constituent concentration, mg/L) and θ (temperature coefficient, dimensionless) reported in Kadlec & Wallace (2009) were assumed for model development. References to percentile ranges refer to the distributions reported in Kadlec & Wallace (2009).

Table 7

Area-based first-order model rate constants (m/yr) during HRT1 and HRT2

Parameter HRT1 HRT2 
NO3-N 54.6 40.0 
NO2-N 63.9 25.1 
NH3-N 15.7 8.2 
TKN 4.1 0.8 
Orthophosphate as P 15.9 5.6 
BOD5 56.9 16.1 
TOC 10.6 7.5 
Selenium – 17.8 
Parameter HRT1 HRT2 
NO3-N 54.6 40.0 
NO2-N 63.9 25.1 
NH3-N 15.7 8.2 
TKN 4.1 0.8 
Orthophosphate as P 15.9 5.6 
BOD5 56.9 16.1 
TOC 10.6 7.5 
Selenium – 17.8 

In general, rate constants determined from this study are comparable to those previously reported (Kadlec & Wallace 2009). The rate coefficients for HRT1 are higher than those for HRT2 indicating greater rate of removal of constituents at higher mass loading rates.

The denitrification rate estimated in this study is between the 55th and 65th percentile of removal rates. Denitrification rates estimated in this pilot study indicate adequate carbon supply from organic matter and CaCO3 (measured as alkalinity). Presence of elevated and labile BOD5 in the influent likely supported rapid denitrification, more so during HRT1 when the system was receiving greater loading and the microbial biofilm was less limited by carbon energy source than during HRT2.

The first-order rates of ammonia oxidation are between the 25th and 35th percentiles, indicating a relatively oxygen-limited environment in the wetlands with relatively lower nitrification and plant uptake rates. Similarly, the first-order rate coefficient of TKN was in the range of 10th and 30th percentile rates.

The relatively greater first-order rate for BOD5 during HRT1 suggests the effect of high mass loading to the wetland system under relatively shorter HRTs than in HRT2. The system was more stable (i.e., less fluctuation in concentrate production) in HRT2 when the change of concentration was not that rapid with less interruption in the RO concentrate production compared to operation during HRT1.

The first-order area-based rate for orthophosphate was greater during HRT1 than in HRT2, possibly due to microbial uptake and media sorption. The rate during HRT2 is about one-third of the rate estimated during HRT1. Sorption by soil and plant surface may provide initial removal, but this partly reversible storage eventually becomes saturated. Uptake by biota, including bacteria, algae, and duckweed, as well as macrophytes, forms an initial removal mechanism (Kadlec & Wallace 2009). Accumulation of phosphorus within the solids in the wetland bed can lead to a loss of pore volume and hydraulic conductivity. Knight et al. (2000) reported an average rate of 8 m/yr for livestock wastewater effluent treated by constructed wetlands, with a range of estimated values from 2 to 18 m/yr. Kadlec & Knight (1996) estimated a rate of 12 m/yr for total phosphorus for wastewater treatment.

The first-order rate constant for selenium was comparable to that calculated for other wetlands treating selenium The constructed free water surface wetland in Imperial, California, was reported to be 12 m/yr (Kadlec & Wallace 2009). The major process responsible for selenium removal seems to be volatilization. The selection of a horizontal subsurface-flow wetland type minimizes potential wildlife contact. Full-scale design of a constructed wetland for organically bound selenium would conceivably be a wetland operated as a coupled aerobic–anaerobic process for oxidation and reduction.

CONCLUSIONS

The Oxnard reclaimed water concentrate pilot study demonstrated that constructed wetlands can provide treatment of a brackish wastewater produced by an MF/RO process with elevated nitrogen, phosphorus, metal complexes and salts. The wetland plants tolerated the brackish concentrations and grew in the ammonium-rich water. This study illustrated that the subsurface-flow wetland treating membrane concentrate would pose no nuisance concerns, as no odor was detectable from the wetland influent and effluent. The aesthetics of the wetland water were acceptable with no odors, which was one of the primary concerns due to the presence of the treatment plant in a residential area.

Mass of water quality parameters decreased significantly through the pilot treatment wetlands in response to water loss through plant evapotranspiration. Significant reductions in inorganic nitrogen concentration and mass were measured throughout the study. Significant reductions were observed in TOC consistent with denitrification demands. Nitrite concentrations were detectable in the wetland effluent, indicating carbon limitation and/or hydraulic short-circuiting in the test wetland. Selenium concentrations decreased through the pilot wetlands. Mass removals were estimated to be on the order of 40%. In contrast to the removals of nutrients and selenium, TDS increased by 19% to an outlet concentration of 13.8 g/L through evapoconcentration.

As a beneficial use of reclaimed water, the brackish wetland effluent could serve as a water source for estuarine wetland restoration projects. Wetland plants survived and showed no indication of adverse effects. The constructed wetland would improve the water quality of the concentrate, and the created wetland habitat could benefit the local environment. Moreover, the study results indicated that the reductions in nitrogen, phosphorus and selenium in the reclaimed water concentrate were found to be consistent with expectations from the wetland treatment literature. The first-order rate constants determined may provide a basis for sizing a full-scale wetland receiving a similar quality of water. To achieve lower effluent selenium, wetland area could be increased.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the cooperation of the wastewater division of the City of Oxnard, CA. Assistance from Jeff Miller, Mark Moise and Anthony Emmert is appreciated. Special thanks to Dr Stephen Lyon for the design and implementation of the wetlands pilot study and for the use of the portable wetland system designed by the Mobile Environmental Solutions (MES). We also thank the staff of CH2M who participated.

REFERENCES

REFERENCES
American Public Health Association
2005
Standard Methods for the Examination of Water and Wastewater
,
21st edn
,
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
,
USA.
Kadlec
R. H.
Knight
R. L.
1996
Treatment Wetlands
.
CRC Press/Lewis Publishers
,
Boca Raton, FL, USA
.
Kadlec
R. H.
Wallace
S.
2009
Treatment Wetlands
,
2nd edn
,
CRC Press
,
Boca Raton, FL, USA
.
Kepke
J.
Bays
J.
Lozier
J.
2009
Concentrate treatment using wetlands
.
Water (J. Australian Water Association)
36
(
7
),
57
63
.
Knight
R. L.
Payne
V. W. E.
Borer
R. E.
Clarke
R. A.
Pries
J. A.
2000
Constructed wetlands for livestock wastewater management
.
Ecol. Eng.
15
,
41
55
.
Mickley
M.
2009
Treatment Wetland, Desalination and Water Purification Research and Development Program Report no. 155. US Department of the Interior, Bureau of Reclamation.
(
accessed 9 April 2015)
.
NOAA
1999
Screening Quick Reference Tables
. (
accessed 16 May 2013).
Utah Climate Center
2015
(
accessed 10 February 2015)
.
WateReuse Foundation
.
2006
Beneficial and Non-traditional Use of Membrane Concentrate. Prepared by CH2M HILL and Mickley and Associates, Inc., for the WateReuse Foundation. Project 02–006b-01. Alexandria, Virginia, USA
.
Webb
J. M.
Quinta
R.
Papadimitriou
S.
Norman
L.
Rigby
M.
Thomas
D. N.
Le Vay
L.
2012
Halophyte filter beds for treatment of saline wastewater from aquaculture
.
Water Res.
46
,
512
5114
.
Xu
P.
Cath
T. Y.
Robertson
A. P.
Reinhard
M.
Leckie
J. O.
Drewes
J. E.
2013
Critical review of desalination concentrate management, treatment and beneficial use
.
Environ. Eng. Sci.
30
(
8
),
502
514
.