*To whom correspondence should be addressed.

Transcription

1 Pilot Study on the Performance of Reverse Osmosis, Microfiltration and Wetlands Concentrate Treatment of Wastewater with High Membrane Fouling Potential R. K. Chakraborti 1,*, J.C. Lozier 2, M. Witwer 3, J. S. Bays 4, U.G. Erdal 5 and K. Ortega 6 1 CH2M HILL; 325 E. Hillcrest Drive, Suite 125, Thousand Oaks, CA CH2M HILL; 2625 S. Plaza Drive Suite 300, Tempe, AZ CH2M HILL; 3011 S.W. Williston Road; Gainesville, FL CH2M HILL; 4350 W. Cypress St.; Tampa, FL CH2M HILL; 6 Hutton Centre Drive (Griffin Towers); Suite 700; Santa Ana, CA City of Oxnard; 305 W. Third Street, Oxnard, CA *To whom correspondence should be addressed. Rajat.Chakraborti@ch2m.com ABSTRACT A proof pilot study consisting of MF, RO and constructed wetlands systems was conducted during the design of the City of Oxnard s 94.6 ML/d (25 mgd) Advanced Water Purification Facility. This study was intended to evaluate unit performance for wastewater with high membrane fouling potential to achieve a finished water quality suitable for groundwater recharge and irrigation. After some modifications the MF and RO units met all the full-scale design and operational criteria. The RO treatment achieved removal of key constituents below the target levels. The RO permeate contained TOC, TN, TDS and chloride concentrations less than target levels of 0.5, 5, 500 and 250 mg/l, respectively. The RO unit reduced NDMA concentration by approximately 50 percent and 1,4-dioxane by approximately 80 percent. The constructed wetland plant Bullrush was found to tolerate the high levels of TDS and provided significant mass removal of nitrogen and other constituents in the RO concentrate. KEYWORDS Water reuse and water recycle, groundwater recharge, microfiltration, reverse osmosis, fouling, treatment wetlands, wastewater treatment INTRODUCTION The City of Oxnard (City) has embarked on the Groundwater Recovery Enhancement and Treatment (GREAT) program, a water resources project that combines wastewater recycling, groundwater injection, and groundwater desalination to more efficiently use existing local water resources. As part of its water resources master planning process, the City determined that additional alternative water supply sources should be developed to continue meeting the City's goal of providing current and future residents and businesses with a reliable and affordable source of high quality water. Limitations on both the City s local groundwater and imported water sources, plus the increased cost of imported water, prompted the City to conduct an

2 advanced planning study of alternative water supply sources. This study resulted in the development of the GREAT Program. According to the plan, the wastewater plant would produce up to 18.9 ML/d (5 million gallons per day [mgd]) of tertiary treated water and up to 23.7 ML/d (6.25 mgd) of advanced treated water (primarily for agricultural irrigation). The recycled water will be used on the southern Oxnard Plain and Pleasant Valley areas where overdraft conditions and the effects of that overdraft are most severe. A portion of the advanced treated wastewater would be used for groundwater injection during times of low agricultural demand (approximately 3 months per year). The GREAT Program could potentially increase production of the existing City of Oxnard s Wastewater Treatment Plant (OWTP) of tertiary treated water up to a total of 94.6 ML/d (25 mgd) for agricultural irrigation and groundwater injection. As part of the GREAT program the City of Oxnard is implementing an Advanced Water Purification Facility (AWPF) that will treat secondary effluent from the OWTP. The AWPF will include a multiple-barrier treatment train consisting of microfiltration (MF), reverse osmosis (RO) and advanced oxidation (AOX) using ultraviolet (UV) light and hydrogen peroxide. A portion of the RO concentrate will be treated through a demonstration-scale wetland treatment system (WTS). A simplified process schematic is shown in Figure 1. The existing OWTP currently treats the wastewater to secondary treatment levels and disinfects before discharging into the Pacific Ocean via an ocean outfall. The objective of the AWPF design is to produce a high-quality recycled water that meets the water quality criteria for groundwater recharge (GWR) and unrestricted irrigation as specified by the State of California, Department of Public Health (CDPH), Title 22 guidelines. As part of the procurement process for the pre-purchased MF and RO equipment systems, pilot testing was required by each equipment supplier following their selection. The intent of the proof-pilot testing was to confirm that the performance of each system met the requirements described in the respective procurement documents. While detailed design of the full-scale AWPF was being completed, a proof pilot test was conducted on the OWTP secondary effluent. In this study, a Pall MF unit and a three-stage RO unit supplied by Membrane Systems Inc. were proof pilot tested and operated over a six-month period to (1) confirm the ability of the combined MF and RO process to produce a product water (permeate) that complies with the CDPH Title 22 Recycled Water criteria for GWR and (2) demonstrate conformance of the MF and RO systems with performance requirements as developed by the City for full-scale MF and RO equipment pre-selection and procurement. Although not evaluated as part of the pilot testing, advanced oxidation using hydrogen peroxide and ultraviolet (UV) light will be used to further treat the RO permeate at full scale. The full-scale AWPF system will also include a demonstration-scale wetlands system that will treat a portion of the RO concentrate prior to discharging the concentrate to the ocean. AWPF treated water will be used in irrigation of edible food crops, landscape irrigation, injection into groundwater basin to form a barrier to seawater intrusion, and other possible industrial uses.

3 Figure 1. Schematic of Full-Scale AWPF Process PROOF PILOT STUDY OBJECTIVES Pilot testing is critical for projects such as the AWPF because the performance of membrane systems is not readily predicted using effluent water quality data alone and certain constituents in the effluent that result in membrane fouling and performance declines cannot be easily quantified. Likewise, removal efficiency of key water quality parameters (e.g., total organic carbon (TOC) and nitrogen) by RO is best determined at pilot scale. Pilot testing was conducted after pre-selecting the MF and RO systems to streamline project delivery by allowing for AWPF detailed design to progress in parallel with pilot testing. The proof pilot plant was operated from July 20th, 2008 until March 13th, Characterization of the OWTP secondary effluent (prior to disinfection) showed relatively high levels of TOC (average of 17 mg/l), total dissolved solids (TDS) (average of 1,700 mg/l), total nitrogen (TN) (average of 27 mg/l), ammonia as N (average of 22 mg/l), chloride (average of 415 mg/l) and n-nitrosodimethylamine (NDMA) (average of 69 ng/l). The major objective of advanced treatment is to produce a recycled water that complies the TOC and TN requirements of the enforced by CPDH draft Title 22 recycled water regulations for GWR of 0.5 mg/l and 5.0 mg/l, respectively. Although NDMA is a non-regulated compound, both USEPA and CDPH have an action level of 10 ng/l for recycled water. The primary proof pilot test objectives are: Determine the removal efficiency of TOC, TN, TDS and NDMA from the secondary effluent of OWTP using MF and RO treatment and ensure the water quality complies with the Title 22 GWR criteria. Demonstrate that the selected MF and RO system can meet the performance criteria Characterize RO concentrate quality and quantity for residual disposal treatment and permitting

4 Conduct limited testing of wetlands for treatment of RO concentrate Verify the proposed cleaning regimes for the MF and RO systems are adequate to meet the performance criteria Refinement of the design criteria for full-scale MF and RO systems based on pilot operating data in conjunction with completion of AWPF final design PROCESS OVERVIEW AND METHODOLOGY In the pilot plant design, secondary effluent was pumped from the OWTP secondary clarifier to the proof pilot system. Sodium hypochlorite (12.5 percent) was added to the secondary effluent upstream of an in-line static mixer to form chloramines upon reaction with the naturally occurring ammonia in the effluent. Continuous chloramination was necessary to control biological growth in the pilot plant and in particular manage biofouling of the RO system. The chloraminated feed water was screened using a 300-μm strainer located on the MF unit to remove larger particulate matter in the secondary effluent. The MF filtrate was discharged to a break tank to accommodate MF system backwashes and enhanced flux maintenance cleanings while maintaining RO system operation. Sulfuric acid and scale inhibitor were added to the RO feed water upstream of the RO feed booster pump. The RO feed pump further pressurized the RO feed water. RO permeate was discharged to waste, with a portion of the permeate directed to the permeate tank for use in RO unit flushing. A portion of the RO concentrate was aerated and directed to the wetlands while the remainder of the concentrate was discharged to waste. For part of the wetlands operation, sodium bisulfite was added to the concentrate feeding the wetlands to scavenge any remaining residual chlorine. Wetlands effluent was discharged to the wastewater treatment plant waste stream. The pilot testing program was conducted at the OWTP and included the following activities: Equipment installation and functional testing Sampling at influent, effluent, and waste/reject streams of MF, RO and Wetlands Field and laboratory analysis Engineering interpretation of pilot operating data including water quality characterization to observe process efficiencies Microfiltration (MF) System The pilot MF system was provided by Pall Corporation and utilized the membrane modules and the operating configuration as proposed for full-scale. Pilot unit characteristics are shown in Table1. The MF system, when operated at the design criteria shown in Table 1, had a filtrate capacity of approximately 224 L/d (41 gpm). The MF system was utilized to provide particulate free water to the downstream RO process while being an additional pathogen barrier. Figure 2 shows the pilot MF system utilized in this testing.

5 Table 1. Pilot MF Membrane Characteristics, Proposed Design Criteria and Operating Conditions Parameters Units Value Membrane Commercial Designation - Pall Microza UNA-640A Outside Membrane Area ft Module Length inch 84 Module Diameter inch 6.0 Nominal Pore Size µm 0.1 Membrane Material - PVDF Flow Direction - Outside-in Number of Modules - 4 Instantaneous Flux at 20oC gfd 27.2 Terminal TMP Set Point psi 40 Minimum TMP psi 5 MF Membrane Modules MF Feed Tank ` Figure 2. MF System of the Proof Pilot Test Because of the high fouling potential of the OWTP secondary effluent it was important to ensure the cleaning regime for the MF system was adequate to control fouling of the MF membranes.

6 The enhanced flux maintenance (EFM) regime proposed by Pall for the full-scale MF system was to perform a sodium hypochlorite EFM every four days with an average periodicity of no less than three days over a seven day period. Performance criteria from the procurement specifications included the requirement that the MF system operate for at least 30 days between chemical cleanings (Clean-in-places, CIPs) and not exceed a transmembrane pressure (TMP) of 40 psi during the filtration cycle. The specified MF filtrate quality criteria are shown in Table 2. Table 2. MF System Guaranteed Performance Criteria: Filtrate Water Quality Parameter Turbidity 15-minute Silt Density Index (SDI) Free Chlorine Residual Requirement 0.1 NTU, 95 percent of all readings 0.2 NTU, 100 percent of all readings 3, 95 percent of the time 4, 100 percent of the time Non detectable (zero), 100% of the time Reverse Osmosis (RO) System The RO pilot system design mimicked the proposed full-scale RO AWPF design. The RO pilot unit was supplied by the selected RO system supplier, Membrane Systems Inc. (MSI). The pilot unit included membrane elements, pressure vessels, feed, permeate and concentrate headers, system support frames, and numerous valves and instrumentation and is shown in Figure 3. The RO system had three stages where concentrate from the first stage was the feed water to the second stage, and the concentrate from the second stage was the feed water to a third stage. Permeate from the three stages was blended in a final product water. The RO pilot unit was continuously fed with MF filtrate from the break tank using a RO booster pump. Sulfuric acid and antiscalant were added prior to RO booster pumps to minimize precipitation of sparingly soluble salts on the membranes. Sulfuric acid dose was applied to control both calcium carbonate and calcium phosphate precipitation in the RO system. Antiscalant was dosed to control precipitation of a number of sparingly soluble salts, most importantly, silica, calcium carbonate and calcium phosphate.

7 Figure 3. Proof Pilot Plant RO System The RO pilot unit was designed to treat 169 L/d (31 gpm) of filtrate from the MF and operate at 85 percent recovery. The unit was arranged as a three-stage configuration comprising seven 10- cm (4-inch) diameter pressure vessels configured in a 4:2:1 array arrangement using six elements per pressure vessel to simulate the same array arrangement as proposed for the full-scale system. Each vessel was fitted with six 10-cm (4-inch) diameter by 102 cm (40-inch) long spiral-wound membrane elements. Two types of low fouling, low pressure brackish water RO elements were trialed during the pilot. The pilot system had a total permeate output of approximately 142 L/d (26 gpm) based on an average flux of 10.5 gallons per square foot per day (gfd). The first stage permeate piping was fitted with a back-pressure valve and flow meter to allow flux balancing in the first stage. All equipment and components were mounted on a common skid with the exception chemical carboys and housed inside a container. The primary design criteria are shown in Table 4. Table 4. Primary Design Criteria for the RO System Parameter Value Number of Trains 1 Element Size 4-inch diameter x 40-inch long Surface Area of Membrane Element, sq-ft 80 Number of Pressure Vessels 7 Number of Elements per Vessel 6 Array Arrangement 4:2:1

8 Table 4. Primary Design Criteria for the RO System Parameter Value Total Membrane Elements 42 Maximum Overall System Operating Flux, gfd 10.5 Design Recovery, % 85 The quality requirements for the RO permeate are shown in Table 5. These criteria are in accordance Title 22 GWR and the UV system design criteria (UV Transmittance, UVT). Table 5. Required RO Permeate Water Quality Parameter Value TOC (mg/l)* <0.5 TDS (mg/l) <500 Total Nitrogen (mg/l)** <10 Chlorides (mg/l) <250 Min UVT (%) 95 *Assumes a recycled water contribution (RWC) of 100%. **Sum of NH4-N, NO3-N, NO2-N and organic N Constructed Wetlands The use of wetlands to treat and reuse concentrate is receiving increased consideration as a concentrate management technique (WateReuse Foundation, 2006; Bays et al., 2007). Utilizing the natural chemical, biological and physical processes intrinsic to wetlands offers an opportunity to create productive brackish marsh habitat while disposing of concentrate. The City of Oxnard has been investigating the concept of beneficial wetland reuse of concentrate since 2003 (CH2M HILL, 2004b; 2005) has factored in a 0.4 ha (1 ac) wetland demonstration project into the final design of the AWPF. Because earlier pilot studies investigated the wetland treatment and growth response to a lower strength concentrate from reverse osmosis of groundwater, the goal of the constructed wetlands component of the pilot study was to confirm that wetland plants would survive the higher ionic and nutrient strength of the concentrate from AWPF effluent. Assessing the performance of the wetlands in treating the effluent concentrate to meet certain water quality goals was a secondary goal.

9 A portable subsurface flow constructed wetland developed by Mobile Environmental Systems (Irvine, CA) (MES) was utilized for testing. The MES wetland was set up adjacent to the RO pilot plant (Figure 4). During the first month, the wetland plant and bacterial communities were acclimated to first the secondary effluent and to high salinities (5 and 11 g/l). For the first four months of operation, the MES wetland was loaded at the relatively high rate of 317 gallons per day (gpd) (1 L/min), yielding a hydraulic loading rate (HLR) of 5 in/d (12.9 cm/day) and a theoretical residence time of 2.5 days. Thereafter and for a period of 3 months, inflow to the wetlands was halved to 0.5 L/min (190 gpd) and the HLR reduced to 6.45 cm/day with a hydraulic residence time of 5 days. Figure 4. Portable subsurface flow constructed wetland (Mobile Environmental Solutions, MES) From late September 2008 to late January 2009 there were a number of occasions where the MF/RO system shut down. The wetlands were dependent on a continuous flow of water from the MF/RO system and so if it was down for a one to three days, there was no flow to the wetland. During this period, the concentrate was fed directly to the wetland without dechloramination and aeration. RESULTS MF Treatment The MF system membrane cleaning regime included system backwashes with air scouring, EFM cleanings, and CIPs. Air scour and backwashes were conducted every 17 to 20 minutes. The EFM regime was modified throughout the proof pilot study. As presented previously the

10 proposed EFM regime was to perform a sodium hypochlorite EFM every four days with an average periodicity of no less than three days over a seven day period. The proposed CIP periodicity was to be no less than once every 30 days. Figure 5 shows the TMP trend during the optimization run for the MF system. Hypochlorite EFMs were performed at a variable frequency based on the rate of increase in TMP and are shown on Figure 5 by dashed arrows. It is evident from the figure that hypochlorite EFMs performed during the period of July 23rd through August 14th, 2008 were not sufficient to control the rate of fouling to a level that would maintain the TMP below 40 psi over a 30-day period as required by the procurement specification. Consequently, Pall modified the EFM protocol to include a citric acid EFM. Citric acid EFMs were conducted beginning on August 14th, The green arrows on Figure 5 indicate when citric acid EFMs were performed. The citric acid EFM provided a significant reduction in TMP during this period. An evaluation of the fouling rate prior to the citric acid EFMs indicated that the TMP would rapidly increase after about nine days of operation with hypochlorite EFMs only. Therefore, a new EFM protocol was established going forward to perform a hypochlorite EFM every three days and a citric acid EFM every 9 days. Transmembrane Pressure (psi) & Feed Turbidity (ntu) Jul 26-Jul 31-Jul 5-Aug 10-Aug 15-Aug 20-Aug 25-Aug 30-Aug TMP Feed Turb (ntu) Citric Acid EFM Hypo EFM Figure 5. MF System TMP and Feed Turbidity Plot and Cleaning Regime (Initial Optimization Phase) A plot of TMP as a function of run time (expressed as calendar date) is shown in Figure 6 for two of the five test runs conducted. EFM events are shown as solid and dotted lines. Throughout

11 the first run (Run No. 1), the rate of TMP increase between the nine-day period between citric acid EFMs was significant and exceeded the maximum allowable TMP of 40 psi. Various runs were conducted to optimize the cleaning frequency to control the fouling of the MF membranes. Therefore, the cleaning regime was further modified for future runs to perform citric acid EFMs every eight days and maintain the three day cycle for hypochlorite cleanings. Feed turbidity generally ranged from 1 to 10 NTU during these runs but showed excursions to as high as 20 NTU for short periods. Figure 6. MF Plant TMP and Feed Turbidity (test Runs 1 and 2) Although the MF system proof pilot results indicated that the system could meet the procurement specifications, the cleaning regime had to be modified during the testing period. Additional changes to the EFM regime will be further evaluated during the full scale system startup. Pall has indicated that they will perform daily hypochlorite EFMs and the full scale design will include the capability to perform citric acid EFMs. A comparison of the proof pilot performance requirements and the actual results from the testing is shown in Table 6. As indicated all requirements were met except for the 100 th percentile turbidity criterion. Table 6. MF Filtrate Quality Results Versus Performance Criteria

12 Parameter Requirement Actual Met Criterion Turbidity (no. of readings =30,708) 0.1 NTU; 95 percent of all readings 0.2 NTU; 100 percent of all readings 98.7 percent of the readings (10 minutes) were below 0.1 NTU 99.7 percent of the readings were below 0.2 NTU YES NO 15-minute Silt Density Index (SDI) (no. of measurements = 23) 3; 95 percent of the time 4; 100 percent of the time All 15 minute SDI values 3 (average 0.38 with a range of 0.2 to 0.7) YES RO Treatment The required and actual proof pilot performance criteria for the RO unit are shown in Table 7. The RO system optimization resulted in a change to the specified maximum feed pressure used in the full scale design as indicated in Table 7. Table 7. Reverse Osmosis System Performance Criteria Performance Criteria Requirement Actual RO System Recovery (percent) 85 85(average) Maximum Feed water Pressure (psig) < < 225 (revised) Production Capacity-Average Flux (gfd) RO feed water quality data based on samples collected during the pilot plant testing is summarized in Table 8. Table 8. RO Feed Water Quality Summary

13 Parameter, mg/l unless otherwise specified Ave Min Max Number of Samples Standard Deviation Total Dissolved Solids Alkalinity as CaCO Ammonia-N Chloride Fluoride Nitrate-N Nitrite-N <1.1 < Nitrate-N/Nitrite-N n/a Total Nitrogen Reactive SiO Total Kjeldahl Nitrogen Total Organic Carbon Total Suspended Solids ph, units Orthophosphate as P Sulfate Aluminum (μg/l) <40 <40 < Calcium Iron (μg/l) <100 <100 < Manganese (μg/l) Magnesium Silica Selenium (μg/l) Boron (μg/l) Temperature (C) Measured in the RO permeate Permeate from the RO system was required to meet several water quality criteria. During the successful RO system test run, all criteria were met (Table 9). Table 9. RO Permeate Water Quality Parameter Unit Performance Criteria Actual Total Dissolved Solids (TDS) mg/l < 500 Maximum 128 (n=21) Total Organic Carbon (TOC) mg/l < 1 a (<0.5) b <0.5 b (n=16) Total Nitrogen (TN) mg/l < 10 (<5 c ) Maximum 1.35 (n=14) Chloride (Cl) mg/l < 250 Maximum 18.3 (n=16)

14 Table 9. RO Permeate Water Quality Parameter Unit Performance Criteria Actual Ammonia (as N) mg/l None 2.2 UV Transmittance UVT) %T 95 Minimum 97.5 (n=13) a Assuming a recycled water contribution (RWC) of 50%. b Assuming a RWC of 100% c Most recent draft criterion Proof pilot RO feed and permeate TOC results are shown in Figure 7. Note that the red line in the exhibit represents the permeate limit. A single permeate sample had a value that exceeded the TOC limit of 0.5 mg/l (0.61 mg/l); this sample was collected prior to system optimization and the successful RO proof pilot run. All other samples were below the minimum reporting limit (MRL) of 0.5 mg/l. Values shown in Figure 7 that were less than the MRL but greater than the minimum detection limit (MDL) are estimated. On the basis of the estimated values, the average TOC rejection was 97.4 percent with a range of 95.1 to 98.7 percent RO Feed TOC (mg/l) RO Permeate TOC (mg/l) Sep 19-Oct 8-Nov 28-Nov 18-Dec 7-Jan 27-Jan 16-Feb 8-Mar 28-Mar RO Feed RO Permeate Figure 7. Reverse Osmosis Feed and Permeate Water Quality: TOC (the Red Line Represents the Permeate Water Quality Requirement) Total nitrogen results for RO feed and permeate samples are shown in Figure 8. Note that the

15 red line in the exhibit represents the permeate TN limit at 100 percent RWC. All of the permeate samples were below the 5 mg/l limit. Average permeate TN concentration was 0.91 mg/l with a range from 0.31 to 1.35 mg/l. The RO feed TN concentration ranged from 10.2 to 34.5 mg/l with an average of 23.9 mg/l. Average TN rejection by the RO system was 96.1 percent. RO Feed TN (mg/l) RO Permeate TN (mg/l) Sep 19-Oct 8-Nov 28-Nov 18-Dec 7-Jan 27-Jan 16-Feb 8-Mar 28-Mar RO Feed RO Permeate Figure 8. Reverse Osmosis Feed and Permeate Water Quality: TN (the Red Line Represents the Permeate Water Quality Requirement) Results from the sampling for RO feed and permeate TDS are shown in Figure 9. Note that the red line represents the permeate TDS limit. All permeate TDS results were well below the limit and typically less than 50 mg/l or 10 percent of the limit. Average TDS rejection by the RO system was 98.4 percent. Six separate RO feed and permeate samples were collected to assess the removal a number of nitrosamine, including NDMA, s and of 1-4 dioxane by RO treatment of the disinfected and microfiltered effluent. NDMA removal was about 50 percent. In contrast, removal of 1,4- Dioxane averaged about 80 percent. The RO concentrate was characterized as part of the pilot study. This waste stream concentrate had very high concentrations of dissolved solids, alkalinity, hardness and ammonia, and also

16 elevated levels of manganese and TOC. It is also supersaturated with respect to many sparingly soluble salts, including silica, calcium carbonate, calcium phosphate and several sulfate salts. These characteristics make it a challenging source water for remediation by the constructed wetlands. In addition, conveyance of the concentrate to the City s ocean outfall should consider the corrosion and scale characteristics of the concentrate. 4, , RO Feed TDS (mg/l) 3,000 2,500 2,000 1,500 1, RO Permeate TDS (mg/l) Sep 29-Sep 19-Oct 8-Nov 28-Nov 18-Dec 7-Jan 27-Jan 16-Feb 8-Mar 28-Mar RO Feed RO Permeate Figure 9. Reverse Osmosis Feed and Permeate Qualities: TDS (the Red Line Represents the Permeate Quality Requirement) Wetlands Treatment Plant Survival and Growth Plant survival and growth was assessed during an initial study and at the conclusion of the experiment. As a first step toward meeting the study objectives, it was considered prudent to determine if bulrush (Schoenoplectus sp.), a plant typically used in treatment wetlands could tolerate the high levels of TDS and ammonium as in the RO concentrate. At no point in time were indications of stress evident, including tip browning, shoot necrosis, shoot pigment loss, or other indications of plant mortality or injury. The bulrush thrived and grew throughout all conditions. In addition, no foul or unpleasant odors were given off by the water in the aeration tank, indicating no objectionable conditions are likely to occur in the demonstration wetland. Water Quality

17 Data from the period of low HLR and elevated residence time was considered most relevant to the performance of the wetland were collected for the purpose of this report. Table 10 summarizes the analytical results of water quality samples collected during this period. Table 10. Performance of the Pilot Wetland System for Selected Parameters Parameter RO Concentrate; or Wetland Influent (median) Wetland Effluent (median) Percent Change in Median Concentration Nitrate-N (mg/l) % Nitrite-N (mg/l) % TOC (mg/l) % Ammonia (mg/l) % TDS (mg/l) 11,850 14,900-26% A total of six samples collected weekly were analyzed. Median concentrations of TDS increased from 11,850 mg/l to 14,900 mg/l in the passage through the wetland. The 26 percent increase in median TDS is attributable to the concentration of conservative inorganic elements caused by the reduction in water through evaporation and transpiration by the plants, consistent with results observed during the previous study with groundwater-derived concentrate (CH2M HILL, 2004b). The hydraulic residence time for the wetland for the flow rate of 190 gpd was estimated to be on the order of 5 days. In addition, the wetland was completely root-bound and all of the media volume is occupied by the plant roots. This density of plants exerts a significant draw upon the water flow through the system, and the related reduction is considered to be a potential benefit of this method of concentrate management. Median concentrations of TOC decreased 20 percent from 72 mg/l to 58 mg/l in the passage through the wetland. These results suggest an assimilation and transformation of organic carbon in the colored influent to the wetland from a form less bioavailable to a form more typical of a wetland effluent comprised of more biologically available compounds. Mass reductions are estimated to have been 40 percent for TOC. Median concentrations of nitrate decreased 71 percent from 16 mg/l to 5 mg/l. A reduction of this magnitude is only possible in anaerobic or low oxygen conditions and in the presence of liquid-phase organic matter. This reduction is entirely consistent with the expectations, and strongly indicates highly reducing conditions in the wetland substrate (Kadlec and Wallace

18 2008) and has been observed in other wetlands studies (CH2M HILL 2004a), including a study in Orange County, CA where the wetland consistently removed percent of nitrate in runoff water with a nitrate range of mg/l (S. Lyon, pers. comm., 2009) and the addition of an external carbon source. Median concentrations of nitrite decreased 69 percent from 15 mg/l to 5 mg/l. Nitrite concentrations in the wetland effluent ranged from 1.0 to 12.2 mg/l, indicating incomplete denitrification. This may be attributable to insufficient hydraulic residence time, insufficient or unavailable carbon, hydraulic short-circuiting (the lack of oxygenation) in the wetland, or a possible combination of factors. In contrast, median TKN and ammonium concentrations increased in the passage through the wetland. Concentrations of ammonium increased by 29 percent from 98 mg/l to 126 mg/l and TKN increased by 15 percent from 113 mg/l to 130 mg/l. These results are contrary to expectations, given general trends reported in the literature of an average ammonium removal rate of 6.7 lb/ac/d (273 g/m2/yr) reported from 123 horizontal subsurface flow wetlands and a general reduction in removal rate with increased concentration (Kadlec and Wallace 2008). These results should be viewed as preliminary and only generally indicative of the wetland response to this strength of concentrate, given the methodological requirement of diluting inflow and outflow samples for laboratory analysis, and the potential for inflation of values as a function of the dilution factor. However, the increase in organic nitrogen concentration as reflected in TKN values) observed may be attributable to evaporative concentration commensurate with the increase in TDS, where the rate of water consumption by the dense stand of bulrush exceeded the rate of nitrogen assimilation. Mass reductions in nitrate, nitrite, ammonium and TKN are estimated to have been on the order of 78 percent, 74 percent, 4 percent, and 15 percent, respectively, based on this estimate of evaporative loss. Also, the production and export of organic matter by the macrophyte and microbial biomass may account for some export of the organic nitrogen, given the high rate of enrichment and the relatively dense condition of the root mat within the MES wetland. Another possible cause for lack of ammonium removal in this study is the lack of oxygenation in the wetland bed, where the plant root structure may not have fully penetrated the full depth of the wetland. Without continual exposure of the water to the oxygen from the root hairs, a critical step in nitrification could be missed. Although not evaluated during pilot testing, there is possibility that certain compounds might act like inhibitory substrate (i.e., metals, xenobiotics) which can significantly reduce growth rate of inhibitors even at threshold concentrations. It is expected that, this condition will not occur in the AWPF Demonstration Wetlands because there will be mature bulrush with a root structure that penetrates a full three feet into the gravel matrix. In addition, aeration system and wetland hydraulics will be properly designed in fullscale wetland system to ensure adequate oxygen supply to allow the complete nitrification of the RO concentrate.

19 SIGNIFICANCE Proof pilot testing of the MF and RO processes was successfully completed and provided valuable information for that was incorporated into the design of the full-scale AWPF. Additionally, the limited scope wetland test confirmed the viability of the wetlands concept for treatment of a challenging source water (i.e., high salinity, high ammonia) The MF and RO proof pilot tests have shown that the AWPF system design can produce a high quality product water that meets the water quality criteria and operation and guaranteed performance criteria. The proposed MF system fouling management regime, specifically the EFM, did not meet the design criteria and was modified throughout the proof pilot testing. Modifications to the detailed design of the AWPF were made so that EFMs with hypochlorite could be performed daily and with citric acid on an eight-day frequency. A successful 30-day test run was completed using the Hydranautics ESPA2 element. All performance requirements were achieved accept for feed pressure, which was higher than projected. As a consequence, the maximum feed pressure criterion for the full-scale AWPF design was modified. The pilot test yielded information indicating that the wastewater treatment options for indirect potable reuse are technically feasible. The results indicate the following: 1. The MF and RO units met all necessary design and operational criteria after modifications. The Hydranautics ESPA2 membrane elements were selected for full-scale use based on the results of the pilot testing. 2. The RO permeate met the quality criteria, i.e., TOC, TN, TDS and chloride concentrations are less than 0.5 mg/l, 5 mg/l, 500 mg/l and 250 mg/l, respectively. 3. RO treatment produced a permeate having a very high UVT (> 95 percent), which will minimize cost of UV treatment. 4. The NDMA removal is about 51 percent in RO permeate compared to RO feedwater and is about 79 percent for 1,4-dioxane. 5. Bulrush was found to tolerate the high levels of TDS and ammonia in the RO concentrate and demonstrated sustained growth when exposed to this substrate over a period of more than 6 months. 6. Significant reductions in nitrate and nitrite and TOC concentration were measured throughout the wetlands study, and significant reductions in pollutant mass were observed for all parameters through evaporative water loss combined with natural treatment processes.

20 7. The wetland pilot test reinforced the importance of the proposed design to supplement the inflow the demonstration wetland with subsurface aeration to enhance nitrification. ACKNOWLEDGMENTS The authors would like to acknowledge Kathy McKinley of the Applied Science Laboratory of CH2M HILL for coordinating the majority of laboratory tests for this study. Special thanks are due to Dr. Stephen Lyon for his design and implementation of the wetlands pilot study and for the use of the portable wetland system designed by MES. REFERENCES Bays, J., P. Frank, and K. Ortega Oxnard s Membrane Concentrate Pilot Wetlands Project. Session A-5., Proc Water Reuse Symposium, Tampa FL. September CH2M HILL. 2004a. Groundwater Recovery Enhancement and Treatment (GREAT) Program Final Program Environmental Impact Report. SCH # Thousand Oaks CA. Accessed June 7, CH2M HILL. 2004b. Membrane Concentrate Pilot Wetlands Project. Final Report prepared for the City of Oxnard Water Division. Thousand Oaks, CA. CH2M HILL Additional Testing for the Membrane Concentrate Pilot Wetlands Project. Final Report prepared for the City of Oxnard Water Division. Thousand Oaks, CA. Kadlec, R.H. and S. Wallace Treatment Wetlands. CRC Press. Boca Raton, FL pp. WateReuse Foundation. (2006). Beneficial and Non-traditional Use of Membrane Concentrate. Prepared by CH2M HILL and Mickley & Associates, Inc., for the WateReuse Foundation. Project b-01. Alexandria, Virginia.