Scholarly article on topic 'Evaluation of geotextile filtration applying coagulant and flocculant amendments for aquaculture biosolids dewatering and phosphorus removal'

Evaluation of geotextile filtration applying coagulant and flocculant amendments for aquaculture biosolids dewatering and phosphorus removal Academic research paper on "Chemical engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Aquacultural Engineering
OECD Field of science
Keywords
{"Aquaculture effluent treatment" / "Geotextile bag filter" / "Biosolids thickening" / Dewatering / "Nutrient removal" / "Chemical coagulation" / Lime / Alum / "Ferric chloride"}

Abstract of research paper on Chemical engineering, author of scientific article — Mark J. Sharrer, Kata Rishel, Steven Summerfelt

Abstract Wastes contained in the microscreen backwash discharged from intensive recirculating aquaculture systems were removed and dewatered in simple geotextile bag filters. Three chemical coagulation aids (aluminum sulfate (alum), ferric chloride, and calcium hydroxide (hydrated lime)), were tested in combination with a long-chain polymer flocculation aid (HyChem CE 1950 at 25mg/L) to determine the most cost effective and efficient treatment combination. Three different coagulants were tested to determine if coagulant choice impacts nutrient and carbonaceous biochemical oxygen demand (cBOD5) leaching into the filtrate and the final composition of the bag-captured biosolids at the end of each period. If nutrient leaching into the bag filtrate could be minimized through coagulant selection, then geotextile bags could provide a convenient and effective method to dewater waste biosolids and provide them in a form that fish farmers could readily transport, store, or send for disposal. Results from replicate geotextile bag filter tests indicate that when alum, ferric chloride, and hydrated lime (plus a polymer) were amended to a backwash flow, both suspended solids capture and solids thickening were improved; i.e., total suspended solids removal rates of 95.8, 95.1, and 96.0%, respectively, were achieved along with final dewatered filter cake percent solids concentrations of 22.1, 19.3, and 20.9%, respectively. Alum, ferric chloride, and hydrated lime (plus a polymer) amended geotextile bags were not as effective in chemical oxygen demand (COD) and cBOD5 removal, resulting in removal rates of 69.6, 67.2, and 35.3%, respectively, and 56.6, 9.3, and −47.4%, respectively. Further, the use of lime as a coagulant resulted in filtrate COD and cBOD5 concentrations that exceeded inlet concentrations. Total nitrogen removal applying alum, ferric chloride, and lime were also less than effective, resulting in removal rates of 39.1, 46.7, and −8.9%, respectively. Filtrate total nitrogen concentrations were primarily in the inorganic form (total ammonia nitrogen) suggesting mineralization of ammonia as solids were stored within geotextile bags under anaerobic conditions. Alum, ferric chloride, and lime amended bags were moderately efficient at total phosphorus removal, resulting in removal rates of 67.6, 47.0, and 77.3%, respectively. Alum was identified as the most cost effective chemical for coagulation, but hydrated lime was the most effective at dissolved phosphorus precipitation and removal.

Academic research paper on topic "Evaluation of geotextile filtration applying coagulant and flocculant amendments for aquaculture biosolids dewatering and phosphorus removal"

ELSEVIER

Contents lists available at ScienceDirect

Aquacultural Engineering

journal homepage: www.elsevier.com/locate/aqua-online

Evaluation of geotextile filtration applying coagulant and flocculant amendments for aquaculture biosolids dewatering and phosphorus removal

Mark J. Sharrer, Kata Rishel, Steven Summerfelt *

The Conservation Fund Freshwater Institute, 1098 Turner Road, Shepherdstown, WV 25443, United States

ARTICLE INFO

ABSTRACT

Article history: Received 20 May 2008 Accepted 3 October 2008

Keywords:

Aquaculture effluent treatment

Geotextile bag filter

Biosolids thickening

Dewatering

Nutrient removal

Chemical coagulation

Ferric chloride

Wastes contained in the microscreen backwash discharged from intensive recirculating aquaculture systems were removed and dewatered in simple geotextile bag filters. Three chemical coagulation aids (aluminum sulfate (alum), ferric chloride, and calcium hydroxide (hydrated lime)), were tested in combination with a long-chain polymer flocculation aid (HyChem CE 1950 at 25 mg/L) to determine the most cost effective and efficient treatment combination. Three different coagulants were tested to determine if coagulant choice impacts nutrient and carbonaceous biochemical oxygen demand (cBOD5) leaching into the filtrate and the final composition of the bag-captured biosolids at the end of each period. If nutrient leaching into the bag filtrate could be minimized through coagulant selection, then geotextile bags could provide a convenient and effective method to dewater waste biosolids and provide them in a form that fish farmers could readily transport, store, or send for disposal.

Results from replicate geotextile bag filter tests indicate that when alum, ferric chloride, and hydrated lime (plus a polymer) were amended to a backwash flow, both suspended solids capture and solids thickening were improved; i.e., total suspended solids removal rates of95.8,95.1, and 96.0%, respectively, were achieved along with final dewatered filter cake percent solids concentrations of 22.1, 19.3, and 20.9%, respectively. Alum, ferric chloride, and hydrated lime (plus a polymer) amended geotextile bags were not as effective in chemical oxygen demand (COD) and cBOD5 removal, resulting in removal rates of 69.6, 67.2, and 35.3%, respectively, and 56.6, 9.3, and -47.4%, respectively. Further, the use of lime as a coagulant resulted in filtrate COD and cBOD5 concentrations that exceeded inlet concentrations. Total nitrogen removal applying alum, ferric chloride, and lime were also less than effective, resulting in removal rates of 39.1, 46.7, and -8.9%, respectively. Filtrate total nitrogen concentrations were primarily in the inorganic form (total ammonia nitrogen) suggesting mineralization of ammonia as solids were stored within geotextile bags under anaerobic conditions. Alum, ferric chloride, and lime amended bags were moderately efficient at total phosphorus removal, resulting in removal rates of 67.6,47.0, and 77.3%, respectively. Alum was identified as the most cost effective chemical for coagulation, but hydrated lime was the most effective at dissolved phosphorus precipitation and removal.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

1.1. Background

Biosolids generated in fish production systems result from uneaten feed, fish feces, and biological floc sloughed from culture system surfaces and vessels (IDEQ, 1998; Cripps and Bergheim, 2000). Mechanisms such as settling basins, rotating microscreen drum filters, and granular media filters are often used to separate biosolids from process water (Summerfelt, 1999; Bergheim and

* Corresponding author. E-mail address: s.summerfelt@freshwaterinstitute.org (S. Summerfelt).

0144-8609/$ - see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.aquaeng.2008.10.001

Brinker, 2003). Backwash and underflow produced by each of these methods result in a waste stream that is much more concentrated in particulate matter than the fish culture water. However, this backwash is still on the order of 99.5-99.95% water, i.e., 5005000 mg/L total suspended solids. Thus, the biosolids in the backwash flow require nearly a 20-200-fold concentration during an additional dewatering step to reduce hauling costs and allow practical hauling for land application or other off-site disposal option (Ewart et al., 1995; Summerfelt and Vinci, 2008).

Off-line settling basins operated as gravity thickening tanks are the most frequently used technology to dewater and store biosolids found in fish farm backwash flows, probably because of their simplicity (Mudrak, 1981; Westers, 1991; Bergheim et al., 1993,1998; Cripps and Kelly, 1996; Chen et al., 1997, 2002; IDEQ,

1998; Adler and Sikora, 2005; Brazil and Summerfelt, 2006; Summerfelt and Penne, 2007). Other biosolids thickening methods include sand beds (Palacios and Timmons, 2001), created wetland drying beds (Summerfelt et al., 1999), wedgewire sieves, inclined belt filters (Ebeling et al., 2006), bag filters (Ebeling and Rishel, 2006), membrane biological filters (Sharrer et al., 2007), filter presses, centrifuges, and vacuum filters. Each dewatering technology has its own specific advantages and disadvantages. For example, dewatering in gravity thickening tanks is an uncomplicated method to reduce the volume of biosolids before their final disposal. However, storing the captured biosolids in the gravity thickening tank for even a few hours, let alone days or months, will allow leaching of soluble organic matter, nutrients, and fine particulate matter as the biosolids rapidly mineralize. Thus, supernatant exiting gravity thickening tanks will have degraded water quality that will likely require further treatment before discharge (Ebeling et al., 2003; Brazil and Summerfelt, 2006).

Development of practical biosolids thickening, stabilization, and dewatering techniques for fish culture waste is essential to meet stringent effluent guidelines and for proper overall fish culture system management (Lekang et al., 2000; Cripps and Bergheim, 2000; IDEQ, 1998; EPA, 2004). Further, the capacity to reduce sludge quantity can mitigate handling costs associated with storage, transportation, labor, and disposal fees by minimizing volume (Metcalf et al., 1991). Geotextile filter bags are a relatively new technology that can be readily scaled-up and could provide a convenient and effective method to dewater waste biosolids and provide them in a form that fish farmers could readily transport, store, or send for disposal.

1.2. Geotextile material

Geotextile fabric is a woven, porous polyethylene material used for construction site erosion control (Rickson, 2006), improving drainage and enhancing reinforcement of marginally stable slopes (Vishnudas et al., 2006), and as anaerobic lagoon odor control covers (Miner et al., 2003). Double layers of geotextile cloth fabricated into a bag design and filled with various materials have been used to reduce bridge abutment scouring (Korkut et al., 2007) and for beach erosion mitigation (Elko and Mann, 2007; Oh and Shin, 2006; Allan and Komar, 2004). Further, hydraulically loaded geotextile bags are used to dewater dredge slurry (Shin et al., 2002), dairy and swine lagoon waste (Worley et al., 2008; Baker et al., 2002), and sewage sludge in decentralized sites (Wett et al., 2005).

Solids dewatering through hydraulically loaded geotextile bags typically operate by internal bag pressurization with an inlet slurry mixture that is no greater than the depth of slurry within the bag. Filtrate "seeps" through the openings of the geotextile fabric and solids are retained within the bag. Agricultural waste management applications have indicated that geotextile bags loaded with fresh dairy and fresh swine manure without coagulant or polymer amendment resulted in final percent solids concentrations of 18.6 and 3.4%, respectively (Baker et al., 2002). Similarly, Worley et al. (2008) found that dewatered dairy lagoon sludge without chemical amendment and with alum/polymer amendment resulted in final percent solids concentrations of 19 and 16%, respectively. Ebeling and Rishel (2006), assessing fish culture biosolids dewatering capacity during a hanging geotextile bag polymer screening, determined that TSS and reactive phosphorus removal efficiencies were 85.3 and 29.3% without polymer and 99.1 and 61.5% with a high molecular weight polymer. The research indicated low filtrate flux without polymer amendment. Further, this research indicated that pressurized (i.e., pressure exceeded the slurry height within the geotextile bags) pilot-scale bags used to treat fish culture

biosolids over a 1-month period with an alum and polymer amendment resulted in an average TSS removal efficiency of 86% (Ebeling and Rishel, 2006).

1.3. Chemical amendments

Coagulants and flocculants are widely used in the wastewater treatment industry to enhance removal of solids, 5-day carbonaceous biological oxygen demand (cBOD5), and phosphorus (Metcalf et al., 1991). Coagulants act by reducing surface charges on the particulate constituents and result in the formation of complex hydrous oxides (Qasim, 1999). Rapidly mixed coagulants (aluminum sulfate, ferric chloride, ferrous sulfate, ferric sulfate) are then stirred to enhance floc formation and promote subsequent settling (Qasim, 1999). Dissolved phosphorus removal is achieved through formation of particulate phosphorus compounds that can then be settled under those enhanced floc formation conditions. Stoichio-metric relationships describing phosphorus precipitation with aluminum sulfate (alum), ferric chloride, and hydrated lime (calcium hydroxide) are (Metcalf et al., 1991):

Al+3 + H2 PO4 - $ AlPO4 + 2H+ (1)

Fe+3 + H2PO4- $ FePO4 + 2H+ (2)

10Ca+2 + 6PO4-3 + 2OH- $ Ca10(PO4)6(OH)2 (3)

Although effects of alkalinity, pH, and competing reactions are a factor, generally 1 mol alum or ferric chloride will precipitate 1 mol of dissolved phosphorus (Metcalf et al., 1991). Simultaneously, the free acid (H+ in Eqs. (1) and (2)) released will consume approximately 0.45 and 0.55 mg/L of alkalinity (as CaCO3) for every 1 mg/L of alum and ferric chloride, respectively (Ebeling et al., 2003). The precipitation reaction of phosphorus with calcium is more complex. When hydrated lime is added to water, calcium carbonate will precipitate as the lime react with inherent alkalinity (Metcalf et al., 1991). At a pH value >10, hydroxylapatite (Ca10(PO4)6(OH)2) will precipitate as excess calcium ions reacts with phosphate ions (Metcalf et al., 1991).

Subsequent flocculation of coagulated waste can be accomplished through use of long-chained organic molecules (polymers). These high molecular weight polyelectrolytes attach to adsorption sites on the surface of waste particulates and result in bridging and intertwining between particles along the polymer chain (Metcalf et al., 1991). Resulting floc can be more rapidly settled or more readily filtered from wastewater flow.

The effectiveness of commercial coagulation-flocculation polymers and/or alum or ferric chloride for removing both suspended solids and phosphorus from aquaculture wastewater flows has been reported. Replicated jar test studies have determined:

• the optimum alum and ferric chloride dosages (when used separately) and flocculation conditions (e.g., mixing speed and time) required to reduce suspended solids and total phosphorus concentrations in the supernatant overflow from gravity thickening tanks (Ebeling et al., 2003);

• the most advantageous alum or ferric chloride concentrations and conditions for coagulation, flocculation, and settling of suspended solids and phosphorus found in the backwash flow discharged from a microscreen drum filter (Ebeling et al., 2004);

• the most suitable polymer type and the appropriate polymer dose, mixing speed and time, and flocculation conditions required to maximize suspended solids and particulate phosphorus removal from the backwash flow discharged from a microscreen drum filter (Ebeling et al., 2005);

• the optimum combination of alum and polymer concentrations and mixing and flocculation conditions to remove suspended solids and phosphorus from the backwash flow discharged from a microscreen drum filter (Rishel and Ebeling, 2006).

The effectiveness of using a combination of alum and polymer to precipitate dissolved phosphorus and coagulate and flocculate suspended solids was also determined across a commercial inclined belt filter (Ebeling et al., 2006). In summary, these studies determined that application of coagulation-flocculation chemicals can improve the capture of fine solids, phosphorus, and biochemical oxygen demand, which will produce a cleaner and more environmentally sustainable discharge.

Hydrated lime can also be used to condition waste aquaculture biosolids captured at the bottom of a gravity thickening tank (Bergheim et al., 1993,1998). Bergheim et al. (1993,1998) reports that hydrated lime addition to a pH of 12 can be an effective method to kill sludge pathogens, reduce odor problems, increase the removal of phosphorus, and improve solids thickening. The U.S. Environmental Protection Agency's (US EPA) 40 CFR Part 503, Standards for the Use and Disposal of Sewage Sludge, requires processing of residual wastewater biosolids before they can be beneficially used (EPA, 2000,1995). The standard defines two types of biosolids with respect to pathogen reduction: Class A biosolids are those wastewater residuals that have been processed to contain no detectable pathogens; Class B biosolids are those wastewater residuals that have been processed to contain a reduced level of pathogens. Class B biosolids can be produced by lime stabilization of the biosolids slurry when the pH reaches or exceeds 12 after 2 h of contact (EPA, 2000). Class A biosolids can be produced by adding lime to the biosolids slurry until its pH reaches or exceeds 12 for at least 72 h, with a temperature of 52 °C maintained for at least 12 h during this time. Both classes are considered safe, but additional application requirements are necessary with Class B biosolids. If the pH drops below 9.5 during storage, then there is potential for pathogen regrowth (EPA, 2000).

1.4. Objectives

The purpose of this research was to evaluate the capacity of geotextile bags to capture, dewater, and store over an intermediate period (i.e., approximately 3 months) fish culture biosolids found in the combined microscreen drum filter and radial flow settler backwash from intensive fish culture systems. Removal capacity of particulate and dissolved water quality parameters in the geotextile bag filtrate over this intermediate period was assessed. A further objective was to determine how treatment efficiency was influenced by amending the bag inlet flow with a combination of a long-chain polymer flocculant and three different coagulant, i.e., alum, ferric chloride, or hydrated lime. The objective of testing these three coagulants was to determine if coagulant choice would affect nutrient and cBOD5 leaching into the filtrate and the final composition of the bag-captured biosolids at the end of each period. If nutrient leaching into the bag permeate could be minimized through coagulant selection, then geotextile bag could provide a convenient and effective method to dewater waste biosolids and provide them in a form that fish farmers could readily transport, store, or send for disposal.

2. Materials and methods

2.1. Wastewater sources

The geotextile bag experiment was conducted in the greenhouse building at the Conservation Fund Freshwater Institute utilizing fish

waste generated from the facility's commercial-scale fry rearing system, partial reuse system, and a fully recirculating growout system managed for annual production of approximately 35 mton/ yr (80,000 lbs./yr) of rainbow trout (Oncorhynchus mykiss) and a series of six pilot-scale fully recirculating systems also operated for production of rainbow trout. The twelve tank flow-through fry rearing system passed approximately 37.9 lpm(10gpm)per1.37 m (4.5 ft) x 0.61 m (2 ft) circular tank equipped with a bottom-center drain. The partial reuse system recycled 1200-1850 lpm (320490 gpm) through three 3.66 m (12 ft) x 1.1 m (3.5 ft) circular dualdrain tanks operated to mechanically filter 85-90% of the flow through the sidewall drain and into a rotating microscreen drum filter (Model RFM 3236, PRA Manufacturing Ltd., Nanaimo, British Columbia, Canada) equipped with 90 mm filter screens (Summerfelt et al., 2004). Water exiting the fry rearing system tanks and the bottom-center drains of the partial reuse system were mechanically filtered using a rotating microscreen drum filter (Model RFM 4848, PRA Manufacturing Ltd., Nanaimo, British Columbia, Canada) equipped with 60 mm filter screens. Wastewater was also derived from a fully recirculating growout system that recycled 4800 lpm (1250 gpm) through a single 9.1 m (30 ft) x 2.4 m (8ft) circular dual-drain tank that was operated to direct 92-93% of the flow though the sidewall drain and the remaining 7-8% through a bottom center drain and then to a radial-flow settler (Davidson and Summerfelt, 2005). Total system flow was then recombined and the entire recirculating flow filtered though a rotating microscreen drum filter (Model RFM 4848, PRA Manufacturing Ltd., Nanaimo, British Columbia, Canada) equipped with 90 mm filter screens. The radial-flow settler was automatically flushed once every hour using a 7.6 cm (3 in.) pipe diameter pneumatically actuated diaphragm valve (Type 025, Georg Fisher LLC, Tustin, California). The series of six pilot-scale fully recirculating systems each recycled approximately 379 lpm (100 gpm) through individual 2.44 m (8 ft) x 1.22 m (4 ft) circular dual-drain tanks operated to direct 80% of the flow through sidewall drains and the remaining 20% through radial-flow settlers. Each system's total flow was then recombined and the entire recirculating flow filtered through rotating microscreen drum filters (Model HDF501-1P, Hydrotech Vellinge, Sweden) equipped with 60 mm filter screens. All microscreen drum filters backwashed automatically, as required to clear their screens. The microscreen drum filter backwash and radial-flow settler underflow from the fry rearing system, the partial reuse and all fully recirculating fish production systems were collected (as produced) in a common sump immediately prior to chemical amendment and geotextile bag dewatering.

All rainbow trout production systems were maintained under a constant 24-h photoperiod. Timer controlled mechanical feeders on the fry rearing and six pilot-scale recirculating systems fed the fish equivalent portions during 24 daily feeding events, i.e., two feeding events every other hour. Similarly, timer controlled mechanical feeders on the partial reuse and fully recirculating systems fed the fish during eight daily feeding events, i.e., one event approximately every 3 h. The constant lighting and uniform feeding events spaced equally over a given 24-h period produced relatively constant biological respiration and waste production rates. Thus, a relatively consistent backwash was supplied to the geotextile bag filters over a given 24-h period.

2.2. Geotextile filtration and chemical amendment

Accumulated fish waste was pumped to three replicate geotextile bags using three submersible pumps (Model 8-CIM, Little Giant Pump Co., Oklahoma City, OK) programmed to pump every hour for 0.5 min using a Paragon Model EL72 electronic time

controller (Paragon Electrical Products, Downers Groves, IL). Three custom-sized geotextile bags with apparent pore openings of 0.425 mm were constructed using geotextile material (TenCate Geotube, Commerce, GA). Empty bags measured 1.4 m (4.6 ft) x 2.2 m (7.2 ft) resulting in a total surface area of 12.3 m2 (40.3 ft2) per bag. The three geotextile bags were operated at a hydraulic loading rate of 60-70 L/day/m2 geotextile material. Each bag was positioned on a timber-framed gravel surface covered with pond liner material to ensure virtually complete filtrate capture in three approximately 0.74 m3 (195 gal) collection tanks, which facilitated discrete capture of replicate filtrate flows. Geotextile bag replicates were positioned on a 1% grade, atop a PVC-framed plastic screen, and wastewater was administered at a top-center position on the bag in order to maximize filter surface area and facilitate filtrate flow.

Three sets of coagulant/flocculant amended geotextile dewa-tering treatments were conducted independently. The first treatment applied 50 mg/L alum, i.e., A12(sO4)314H2O (594 g/ mol molecular weight; Univar USA Inc., Kirkland, WA), and 25 mg/ L Hychem CE 1950 polymer (Hychem Inc., Tampa, FL) to 37.9 lpm (10 gpm) wastewater flow from 4/19/06 to 7/19/06. Qualitative observations of chemically amended wastewater flow immediately before bag loading indicated that dosing 50 mg/L alum and 25 mg/L polymer generated the desired floc needed to properly settle solids prior to geotextile dewatering; as this dose produced such a consistent floc, it was assumed that this would help maintain permeability through the membrane. The second treatment applied 50 mg/L ferric chloride, i.e., FeCl36H2O (270.3 g/mol molecular weight; Fisher Scientific, Waltham, MA), and 25 mg/L Hychem CE 1950 polymer to 37.9 lpm (10 gpm) wastewater flow from 9/19/06 to 12/15/06. Selection of optimum ferric chloride dose was determined by orthophosphate removal capacity under bench-scale conditions (Ebeling et al., 2004). The third treatment applied 800 mg/L hydrated lime, i.e., Ca(OH)2 (74.1 g/mol molecular weight; Old Castle Stone Products, Easton, PA), and 25 mg/L Hychem CE 1950 polymer to 37.9 lpm (10 gpm) wastewater flow from 2/6/07 to 4/19/07. A side experiment was performed to determine the relationship between concentration of hydrated lime added to the wastewater flow and slurry pH (Fig. 1). This side experiment delineated that 800 mg/L of hydrated lime was required to achieve pH 12 or Class B biosolids status according to EPA (2000) criteria. However, incorporating data from a second trial (also plotted on Fig. 1) determined that between 1100 and 1200 mg/L of hydrated lime was required to achieve a slurry pH of 12, i.e., the 800 mg/L dose did not always produce a slurry pH of 12. Coagulant and flocculant was pumped from individual reservoirs with Masterflex Economy Model digital drive peristaltic pumps (Cole Parmer Instrument Co., Vernon Hills, IL) and added to

7 -I-.-.-.-.-1-1

0 250 500 750 1000 1250 1500 Lime Dose (mg/L)

Fig. 1. Illustrates the hydrated lime dose and corresponding pH of amended rotating microscreen drum filter backwash. (Using regression equation: 1100-1200 mg/L of hydrated lime is required to produce a pH of 12.0).

wastewater flow on the outlet side of the submersible pumps. Peristaltic pump and submersible pump initiation were controlled concurrently with the same electronic time controller described above. Chemical amendment and wastewater mixing was enhanced with static inline mixers and contact time was extended with approximately 30 m (98 ft) of PVC pipe prior to geotextile bag inlet.

2.3. Sampling regimen and water quality parameters analyzed

Two sampling sites were used for each replicate geotextile bag to assess performance. Sampling site one was located immediately after each submersible pump outlet prior to chemical amendment and characterized inlet water quality. Each sampling port was opened during the 08:00 h electronic timer-controlled pumping event and 4 L (1 gal) grab sample of wastewater was diverted into three separate buckets for the 30-s pumping event. After collection, each sample was thoroughly mixed to maintain homogenization and a 500 mL sub-sample was obtained. Sampling site two was collected from each of the filtrate collection tanks, which contained the total filtrate volume from the previous 24 h pumping events. Total collected filtrate was manually homogenized and a 500 mL sub-sample was obtained.

Table 1 summarizes experiment durations and the number of sampling events for each of the three treatments. Sample events occurred 2-4 times per week throughout the duration of each experiment. Collected samples were analyzed for a series of water quality analyses using a Hach DR4000 spectrophotometer (Table 2) to evaluate chemically amended solids dewatering and nutrient retention capacity for each treatment applied. Data were collected and assessed as the mean ± standard error inlet and filtrate. Removal efficiencies of key water quality parameters were calculated (i.e., ((inlet - filtrate)/inlet) x 100). Analysis of variance (ANOVA) and Tukey's HSD (Honestly Significantly Different) tests were performed on relevant data sets to determine significant differences in removal efficiencies between alum, ferric chloride, and lime treatments. An analysis of covariance (ANCOVA) was performed on all relevant data to determine if inlet concentrations influenced filtrate concentrations. Statistical analyses were performed with SYSTAT 11 (2004) statistical software package. At the end of each treatment, geotextile bags were allowed to dewater for 7-10 days inside the greenhouse facility. Homogenized sludge cake samples (1 L) from each dewatered bag were analyzed off-site (Reliance Laboratories, Martinsburg, WV) for percent solids, total nitrogen, total phosphorus, and total potassium.

3. Results and discussion

3.1. Change in dissolved oxygen, alkalinity, pH, and temperature

Mean inlet and filtrate dissolved oxygen concentration, temperature, pH, and alkalinity from the three coagulant/ flocculant amended geotextile bag trials are compiled in Table 3. With respect to oxygen and temperature, passage through the biosolids stored within the geotextile bags warmed (1.03.0 °C) the filtrate and stripped it of dissolved oxygen, i.e., filtrate contained a mean dissolved oxygen concentration of 0.1-0.3 mg/L of dissolved oxygen. Warming, even during winter months, was due to housing the black (sunlight absorbing) geotextile bags in a greenhouse.

With respect to alkalinity and pH, addition of 50 mg/L of alum or ferric chloride is predicted to consume around 25 mg/L of alkalinity. However, the filtrate exiting the alum and ferric chloride amended geotextile bags contained on average 60-95 mg/L more alkalinity than was in the biosolids before amendment addition. A

Duration of each experimental treatment and number of sampling events assessing coagulant/flocculant amended geotextile bags for solids dewatering and phosphorus removal of recirculating aquaculture effluent.

Treatment (+25 mg/L polymer) Coagulant grade (%) Study period (m/d/yr) Experiment Duration (d) Sample events

50 mg/L Alum 100 4/19/06-7/19/06 91 44

50 mg/L Ferric chloride 37-42 9/19/06-12/15/06 87 34

800 mg/L Lime 98 2/6/07-4/19/07 73 29

Table 2

Laboratory methods used to assess inlet and filtrate water quality, solids dewatering, and nutrient retention capacity in coagulant/flocculant amended geotextile bags for

solids dewatering and phosphorus removal of a recirculating aquaculture effluent.

Parameter Method Units

Dissolved oxygen (DO) Hach HQ40d Meter, LDO101-10 Probe mg/L

pH Hach HQ40d Meter, PHC101-10 Probe SU

Temperature (T) Hach HQ40d Meter °C

Alkalinity Standard Methods 2302 mg/L (as CaCO3)

Total suspended solids (TSS) Standard Methods 2560 mg/L

Total volatile solids (TVS) Standard Methods 2560 mg/L

Turbidity Hach Method 8237 NTU

Total phosphorus (TP) Hach Method 8190b mg/L (as P)

Hach Method 10127a c mg/L (as P)

Dissolved reactive phosphorus (DRP) Hach Method 8048a mg/L (as P)

Total nitrogen (TN) Hach Method 10071d mg/L (as N)

Hach Method 10072e mg/L (as N)

Total ammonia nitrogen (TAN) Hach Method 8038a mg/L (as NH3-N)

Nitrite-nitrogen Hach Method 8507 mg/L (as NO2-N)

Nitrate-nitrogen Hach Method 8171 mg/L (as NO3-N)

5-Day carbonaceous biological oxygen demand (cBOD5) Standard Methods 5210 5-day BOD mg/L

Chemical oxygen demand (COD) Hach Method 8000 mg/L

a Adapted from Standard Methods For the Examination of Water and Wastewater. b Low range (0.02-1.10 mg/L-P). c High range (1-100 mg/L-P). d Low range (0.5-25 mg/L-N). e Low range (10-150 mg/L-N).

small portion of the alkalinity increase could be partly due to denitrification in the bags, but the remaining alkalinity gain is unexplained.

Adding hydrated lime to the inlet flow increased pH from 7.58 ± 0.05 to 10.37 ± 0.28 units before the geotextile bags. Note that adding 800 mg/L of hydrated lime did not produce a slurry pH of 12, but of 10.37 on average. After the study was underway, we determined that between 1100 and 1200 mg/L of hydrated lime was required to consistently achieve a slurry pH of 12 (Fig. 1). Thus, it appears that the pH probe used in the 1st lime dosing study had been slightly out of calibration. In comparison, Bergheim et al. (1998) added approximately 12,000 mg/L of hydrated lime to achieve a pH of 12 in biosolids that had already been thickened to 10% dry weight. The biosolids in the Bergheim et al. (1998) study were already dewatered and were 50-60 times more concentrated than the TSS of the slurry

backwash that was treated with hydrated lime in the present study. Bergheim et al. (1998) report that 110 g of hydrated lime must be added for every kilogram of biosolids (dry weight) to lime stabilize the biosolids and produce a sustained pH of 12. However, to maintain the biosolids pH at 12 for extended storage times, Bergheim et al. (1998) reported that more than 110 g of hydrated lime per kilogram of biosolids must be added to achieve an initial biosolids pH of 12.212.3. In hind sight, the dose of hydrated lime used in the present study should have been in excess of 1200 mg/L, which would have been more likely to maintain a pH of 12 for more than 2 h. Maintaining a pH of 12 for more than 2 h is required to meet the Class B biosolid classification for domestic sewage biosolids.

The filtrate exiting the lime amended geotextile bags contained a mean alkalinity of 670 ± 98 mg/L (as CaCO3) and a mean pH of 8.38 ± 0.24. The pH of the lime supplemented filtrate was closer to a

Table 3

Dissolved oxygen, temperature, pH, and alkalinity of geotextile bag inlet and filtrate for each coagulant/flocculant treatment assessing solids dewatering and phosphorus removal capacity of a recirculating aquaculture effluent.

Treatment (+25 mg/L polymer) DO (mg/L) T (°C) pH Alkalinity (mg/L)

50 mg/L Alum

Inlet 7.6 ± 0.3 17.1 ± 0.3 7.55 ± 0.02 303 ± 10

Filtrate 0.1 ± 0.0 20.1 ± 0.4 7.20 ± 0.02 363 ± 16

50 mg/L Ferric chloride

Inlet 6.2 ± 0.2 13.5 ± 0.2 7.06 ± 0.03 287 ± 5

Filtrate 0.1 ± 0.0 15.9 ± 0.6 6.92 ± 0.02 382 ± 2

800 mg/L Lime

Inlet 6.7 ± 0.1 11.7 ± 0.2 7.58 ± 0.05 259 ± 6

Inlet + lime 10.37 ± 0.28

Filtrate 0.3 ± 0.1 12.7 ± 0.6 8.38 ± 0.24 670 ± 98

a ANOVA results indicate a significant difference in total potassium concentration among all treatment groups.

Table 5

Inlet and filtrate concentrations and removal efficiencies for TSS, TVS, turbidity, COD, and cBOD5 for each coagulant/flocculant treatment assessing solids dewatering and phosphorus removal capacity of a recirculating aquaculture effluent.

Percent solids, total nitrogen, total phosphorus, and total potassium concentrations in dewatered geotextile filter cake from a recirculating aquaculture effluent administered a coagulant/flocculant amendment.

Treatment (+ 25 mg/L polymer) Solids concentration (%) Total nitrogen (g/kg) Total phosphorus (g/kg) Total potassium (g/kg)

50 mg/L Alum 22.1 i 1.1 35.6 i 3.2 1.51 i 0.04 0.387 i 0.005a

50 mg/L Ferric chloride 19.3 i 1.0 22.0 i 8.0 1.70 i 0.13 0.646 i 0.041a

800 mg/L Lime 20.9 i 0.5 29.4 i 0.8 1.67 i 0.04 0.908 i 0.038a

Treatment (+25 mg/L polymer) TSS (mg/L) TVS (mg/L) Turbidity (ntu) COD (mg/L) cBOD5 (mg/L)

50 mg/L Alum

Inlet 1874 i 120 1317 i 171 621 i 31 1896 i 125 541 i 58

Filtrate 98 i 4 79 i 2 56 i 3 577 i 20 235 i 25

% Removal 94.8 94.0 95.8 69.6 56.6b

50 mg/L FeCl

Inlet 1889 i 169 1330 i 145 542 i 37 2072 i 180 443 i 59

Filtrate 93 i 5 75 i 4 58 i 2 679 i 29 402 i 27

% Removal 95.1 94.4 91.1 67.2 9.3

800 mg/L Lime

Inlet 1515 i 483 900 i 164 425 i 63 1774 i 224 498 i 67

Filtrate 61 i 5 54 i 11 26 i 3 1147 i 165 734 i 123

% Removal 96.0 94.0 93.9 35.3a -47.4

a Tukey's HSD post hoc analysis indicates significantly lessened COD removal capacity applying lime when compared with alum and ferric chloride treatments. b Tukey's HSD post hoc analysis indicates significantly greater cBOD5 removal capacity applying alum when compared to ferric chloride and lime treatments.

neutral pH (8.38 ± 0.24) than the 10.37 ± 0.28 pH that entered each geotextile bag. Alkalinity of the lime treated filtrate averaged 670 mg/ L, which is lower than the expected 259 mg/L (inlet) plus 1080 mg/L of alkalinity due to addition of 800 mg/L of hydrated lime. Within the bags, alkalinity decreased as hydroxide reacted with phosphate to form hydroxylapatite (Eq. (3)). Alkalinity may have also been destroyed by the release of organic acids as the biosolids mineralized within the geotextile bags.

3.2. Solids thickening

The polymer plus alum, ferric chloride, or hydrated lime amended geotextile bags were operated at hydraulic loading rates of 61.5, 74.2, and 62.8 L/day/m2 geotextile material, respectively. After allowing the geotextile bag filters to dewater for 7-10 days beyond the final loading event, the mean percent solids of sludge cake was highest for alum treated wastewater (22.1%) followed by lime (20.9%) and ferric chloride (19.3%) treatments (Table 4). However, ANOVA results indicated no significance (p = 0.175, a = 0.05) in percent solids concentration. Total nitrogen, phosphorus, and potassium (N:P:K) concentration in the sludge on a dry weight basis for the alum trial were 35.6, 1.5, and 0.4g/kg, respectively, for the ferric chloride trial were 22.0, 1.7, and 0.7 g/ kg, respectively, and for the lime trial were 29.4, 1.7, and 0.9 g/kg, respectively (Table 4). ANOVA results indicated no significant difference in sludge nitrogen (p = 0.475, a = 0.05) or phosphorus (p = 0.267, a = 0.05) concentrations. However, ANOVA results indicated a significant difference in potassium concentrations (p = 0.000, a = 0.05) is evident among all treatments.

A similar solids dewatering option comparable to geotextile dewatering might involve off-line settling basins. Bergheim et al. (1993) indicated that captured fish culture solids with <2% total solids content can be thickened (without coagulant and flocculant amendment) to 5-10% total solids. And, although requiring greater capital and operating costs, belt filter technology utilizing coagulant and flocculant amendment can achieve final percent

solids concentrations of 12.6% (Ebeling et al., 2005). However, a belt filter rapidly separates and removes biosolids from the discharge, which minimizes the leaching of nutrients (nitrogen and phosphorus) and cBOD5 into the filtrate.

Sludge cake produced by geotextile bag filtration with the aforementioned percent solids and N:P:K characteristics are amenable for either land application or composting. Typical fish biosolids contain more nitrogen and phosphorus than manure removed from cattle, pig, and sheep farms, but not as much nitrogen and phosphorus concentration as found in poultry litter (Olson, 1991). In addition, typical solids concentration of dewatered sludge for land application is 15-30% (Metcalf et al., 1991). And, an important advantage of dewatered sludge is the capacity for farmers to use their own land application equipment, further reducing hauling, storage, and spreading costs (Metcalf et al., 1991). Composting process options include aerated static pile, windrow, and in-vessel systems (Metcalf et al., 1991; Outwater, 1994; Adler and Sikora, 2005).

3.3. TSS and turbidity removal

Results indicated that the geotextile bags under all treatment conditions efficiently removed TSS from the drum filter backwash. Removal rates for alum, ferric chloride, and lime were 94.8, 95.1, and 96.0%, respectively, indicating comparable dewatering capacity under all conditions tested (Table 5). ANOVA results indicated no significant difference (p = 0.113, a = 0.05) in TSS removal efficiencies for the treatments. In addition, ANOVA results indicated no significant difference (p = 0.188, a = 0.05) in turbidity removal efficiencies. Covariate analyses indicated that inlet concentration did not affect filtrate concentration. Losordo et al. (2006) reported a TSS removal efficiency of 96% when dewatering polymer-amended fish culture biosolids through a geotextile bag operated for 230 days.

Examination of filtrate concentrations over the course of each treatment indicated that TSS release increases over the first 20-30

i> U°

,§, 120

« 100

K 80 ai

0 10 20 30 40 50 60 70 80 90 100 Day

Fig. 2. Recirculating aquaculture drum filter backwash dewatered through coagulant/flocculant amended geotextile bags. Filtrate TSS concentrations for each treatment (alum + polymer, FeCl + polymer, and lime + polymer) as a function of time.

days, but then TSS capture showed no trend over the remaining 90 days (Fig. 2). This may reflect the mineralization and eventual release of some of the biosolids that collect in the geotextile bag filters.

3.4. Chemical and biological oxygen demand removal

The geotextile bag filters did not remove COD or cBOD5 as effectively as they removed TSS. Mean filtrate concentrations of COD and cBOD5 were high under all treatments, ranging from 577 to 1147 mg/L and from 235 to 734mg/L, respectively (Table 5). Lowest mean COD removal efficiency was for lime amended wastewater (35.3%) followed by ferric chloride (67.2%) and alum (69.6%). ANOVA results indicated a significant difference in removal efficiencies exists (p = 0.000, a = 0.05). Closer examination applying Tukey's post hoc analysis indicated lime treatment results in a statistically significant lower COD removal rate compared to alum (p = 0.000, a = 0.05) and ferric chloride (p = 0.006, a = 0.05), but no significant difference in COD removal efficiency is evident when comparing alum with ferric chloride treatments (p = 0.160, a = 0.05). Least effective cBOD5 removal efficiencies were for lime (-47.4%) and ferric chloride (9.3%) treated wastewater. The alum amended dewatering process performed moderately better (56.6.%) ANOVA results indicated a significant difference in cBOD5 removal efficiency among treatments (p = 0.001, a = 0.05). Tukey's post hoc analysis indicated alum application results in a significantly higher cBOD5 removal

0 10 20 30 40 50 60 70 80 90 100 Day

Fig. 3. Recirculating aquaculture drum filter backwash dewatered through coagulant/flocculant amended geotextile bags. Filtrate cBOD5 concentrations for each treatment (alum + polymer, FeCl + polymer, and lime + polymer) as a function of time.

rate when compared to ferric chloride (p = 0.011, a = 0.05) and lime (p = 0.002, a = 0.05). No statistical difference is evident in cBOD5 removal efficiencies when comparing ferric chloride and lime treatments (p = 0.454, a = 0.05). Inlet COD and cBOD5 (Fig. 3) concentrations did not act as a covariate on filtrate concentrations. Concentrations of COD and cBOD5 in filtrate remained relatively constant over time during alum and ferric chloride treatments. When lime was used, however, filtrate concentrations of COD and cBOD5 (Fig. 3) increased over time and the mean concentration of cBOD5 in the filtrate was higher than in the inflow to the bag. Retaining biosolids under anaerobic conditions within the geotextile bag filters most likely mineralized the organic matter and produced a more readily biodegradable organic carbon as measured by the cBOD5 test, which would explain the increase in cBOD5 concentration across the bag filters.

3.5. Nitrogen removal

The capacity of geotextile bags to remove nitrogen from drum filter backwash over an intermediate or extended period of time was limited. Poor TN removal efficiency was evident with the lime treatment (-8.9%) followed by moderate removal with alum (39.1%) and ferric chloride (46.7%) (Table 6). ANOVA results indicated a significant difference in TN removal efficiencies exists among treatments (p = 0.000, a = 0.05). Tukey's post hoc analysis indicated no significant difference in TN removal efficiency when comparing alum and ferric chloride treatments (p = 0.810,

Table 6

Inlet and filtrate concentrations and removal efficiencies for total nitrogen, total ammonia nitrogen, nitrite-nitrogen, nitrate-nitrogen, total phosphorus, and dissolved reactive phosphorus concentrations for each coagulant/flocculant treatment assessing solids dewatering and phosphorus removal capacity of a recirculating aquaculture effluent.

Treatment (+25 mg/L polymer) TN (mg/L) TAN (mg/L) NO2-N (mg/L) NO3-N (mg/L) TP (mg/L) DRP (mg/L)

50 mg/L Alum

Inlet 61.9 ± 3.7 1.8 ± 0.1 0.26 ± 0.03 2.2 ± 0.1 40.1 ± 2.4 1.0 ± 0.1

Filtrate 37.7 ± 1.8 28.1 ± 1.4 0.01 ± 0.00 1.4 ± 0.1 13 ± 0.6 11.1 ± 0.5

% Removal 39.1 — 1461 96.2 36.4 67.6 —1010

50 mg/L Ferric chloride

Inlet 82.6 ± 6.8 1.4 ± 0.1 0.306 ± 0.030 3.3 ± 0.2 42.1 ± 3.7 1.0 ± 0.1

Filtrate 44.0 ± 2.8 28.8 ± 2.1 0.002 ± 0.001 1.8 ± 0.2 22.3 ± 0.9 20.0 ± 1.1

% Removal 46.7 — 1957 99.3 45.5 47.0b — 1900

50 mg/L Lime

Inlet 79.7 ± 4.7 1.4 ± 0.2 0.799 ± 0.164 15.1 ± 3.7 33.9 ± 3.2 1.6 ± 0.2

Filtrate 86.8 ± 7.7 59.4 ± 6.5 1.139 ± 0.756 2.7 ± 0.6 7.7 ± 1.6 4.7 ± 1.0

% Removal —8.9a —4142 —42 82.1 77.3 — 194

a Tukey's HSD post hoc analysis indicates significantly lessened TN removal capacity applying lime when compared with alum and ferric chloride treatments. b Tukey's HSD post hoc analysis indicates significantly lessened TP removal capacity applying ferric chloride when compared with alum and lime treatments.

Fig. 4. Recirculating aquaculture drum filter backwash dewatered through coagulant/flocculant amended geotextile bags. Filtrate TN concentrations for each treatment (alum + polymer, FeCl + polymer, and lime + polymer) as a function of time.

Fig. 5. Recirculating aquaculture drum filter backwash dewatered through coagulant/flocculant amended geotextile bags. Filtrate DRP concentrations for each treatment (alum + polymer, FeCl + polymer, and lime + polymer) as a function of time.

a = 0.05), but that TN removal efficiency was significantly less with lime application when compared to either alum (p = 0.000, a = 0.05) or ferric chloride (p = 0.000, a = 0.05). Inlet TN concentration did not act as a covariate on filtrate concentration. Mean bag inlet TN and TAN concentrations ranged between 61.982.6 mg/L and 1.4-1.8 mg/L, respectively (Table 6). Mean filtrate TN and TAN concentrations ranged between 37.7-86.8 mg/L and 28.1-59.4 mg/L, respectively. Thus, inlet TN concentrations were comprised primarily of organically bound nitrogen, which was captured in each geotextile bag filter with the TSS captured. Filtrate TN concentrations were principally in the inorganic fraction (TAN), which suggests that the TN in captured biosolids was mineralized to ammonia during storage under anaerobic conditions in the geotextile bag. In addition, the TN concentration in the filtrate tended to increase over time for all treatments (Fig. 4). In comparison, Stewart et al. (2006) describe rapid leaching of nitrogen into overlying water from solids settled in raceway quiescent zones within the first 24 h and continual leaching after 7 days. The trial in which lime was used as a coagulant showed particularly poor TN capture (Fig. 4). The data suggest that the pH increase seen in lime application correlated to an increased release of TN (primarily as ammonia) from the accumulated biosolids.

3.6. Phosphorus removal

Coagulant and polymer amended geotextile bag filters removed 67.6, 47.0, and 77.3% of TP, on average, when treated with alum, ferric chloride, and lime, respectively (Table 6). ANOVA results indicated a significant difference in TP removal efficiencies among the three treatments (p = 0.000, a = 0.05). Tukey's post hoc analysis shows no significant difference in TP removal efficiency between the alum and lime treatments (p = 0.980, a = 0.05), but indicated a significantly lessened removal efficiency when applying ferric chloride as compared to alum (p = 0.000, a = 0.05) and lime (p = 0.000, a = 0.05). No inlet covariate influence was observed. Inlet TP and dissolved reactive phosphorus (DRP) concentrations ranged between 33.9-42.1 mg/L and 1.0-1.6 mg/L, respectively (Table 6), indicating the bulk of the TP was bound in the biosolids entering the geotextile bags. However, filtrate TP and DRP concentrations ranged from 7.7-22.3 mg/L to 4.7-20.0 mg/L, indicating that mineralization within the geotextile bag filters released a large portion of the TP as DRP.

Dissolved phosphorus is precipitated with alum, ferric chloride, and hydrated lime (Eqs. (1)-(3)). However, as biosolids within the geotextile bags began to mineralize under anaerobic conditions, TP in the biosolids began leaching into the filtrate as DRP, especially in the ferric chloride and alum treatments. These results indicated that the alum and ferric chloride dosages used were only sufficient to

bind the DRP that entered the geotextile bag filters. Alum and ferric chloride dosages were not sufficient to bind the large amount of DRP that was released after the biosolids mineralized under anaerobic conditions. Alternatively, alum and ferric chloride may lose their capacity to bind DRP under anaerobic conditions. In contrast, DRP leaching into the filtrate was much lower in the lime treatment than in the alum and ferric chloride treatments; DRP concentrations in the filtrate were consistently <3.0 mg/L throughout the first 40 days of lime application (Fig. 5). Over time, as biosolids captured within the bag began to mineralize, the release of organic acids resulted in a gradual decrease in filtrate pH during the lime trial (Fig. 6). By day 40 of the trial, the pH of the lime amended filtrate had dropped below 8 and phosphorus retention was reduced. Sedlak (1991) describe two types of lime based phosphorus-removal system: a single stage, low lime system (pH <9.5) that can achieve filtrate TP concentrations of 1-2 mg/L and a two-stage, high lime system (pH >11.5) that can achieve TP concentrations of <0.1 mg/L. It is likely that lime dosing conditions prior to geotextile filtration were sufficient for phosphorus precipitation and subsequent capture in the geotextile bag under the single stage, low lime model. If higher lime dosages had been applied to achieve an inlet pH of 12, it is more likely that DRP concentrations in the filtrate would have remained low throughout the trial.

3.7. Cost analysis

The unit costs of polymer, alum, ferric chloride, and hydrated lime were obtained from bulk chemical supplier at US $2.88/kg, $0.55/kg, $2.38/kg, and $0.42/kg (US $), respectively (Table 7). Combining these chemical costs with the volume of daily backwash treated across each pilot-scale geotextile bag filter,

Fig. 6. Recirculating aquaculture drum filter backwash dewatered through coagulant/flocculant amended geotextile bags. Filtrate pH for each treatment (alum + polymer, FeCl + polymer, and lime + polymer) as a function of time.

The calculated costs of amending biosolids entering the pilot-scale geotextile bag filters with polymer, alum, ferric chloride, and hydrated lime. In addition, an example is provided on the requirements for a geotextile bag filter sized to capture and dewater backwash from a hypothetical 454 ton/yr fish production facility, which includes estimates of the biosolids bound in the backwash flow, polymer and alum requirements and annual costs.

Alum Ferric chloride Hydrated lime Commercial example (454 mton/yr)

Filter inlet TSS, mg/L 621 542 600 697a,c

Inlet flow, L/day 379 455 450 500,000ac

Biosolids loading, kg/yr 85.91 90.01 98.55 127,273a

Polymer dose, mg/L 25 25 25 25

Polymer annual use, kg/yr 3.458 4.152 4.106 4,563

Polymer unit cost, $/kg 2.88 2.88 2.88 2.88

Polymer annual cost, $/yr 10 12 12 13,140

Coagulant dose, mg/L 50 50 800 50b

Coagulant annual use kg/yr 6.917 8.304 131.400 9,125c

Coagulant unit cost, $/kg 0.55 2.38 0.42 0.55b

Coagulant annual cost, $/yr 4 20 55 5,019b

Annual Coagulant + Polymer cost, $/yr 14 32 67 18,159

Coagulant + Polymer cost per dry weight biosolids, $/kg 0.160 0.352 0.680 0.143

Coagulant + Polymer cost per unit feed feda, $/mton feed 32.05 70.39 136.00 28.54

a Assuming that 0.20 kg TSS captured per 1.0 kg feed fed. b Assuming alum is used.

c Assuming a 0.5% backwash flow (worst case scenario), feed conversion rate of 1.4, and 0.2 kg TSS captured per kg feed fed.

we estimated the annual cost of polymer plus alum, ferric chloride, or hydrated lime amendment (Table 7), i.e., US $14, $32, or $67, respectively. This produces a cost for the polymer plus alum, ferric chloride, or lime amendments of US $0.16, $0.35, or $0.68 per kilogram (dry weight) of TSS treated, respectively. Assuming that 0.20 kg TSS is captured for every 1.0 kg feed fed (Davidson and Summerfelt, 2005), then the cost for the polymer plus alum, ferric chloride, or lime amendments would be US $0.032, $0.070, or $0.136 per kilogram of feed fed, respectively (Table 6). Alum is clearly the most cost effective amendment, while lime is the most expensive. However, the cost of polymer, which is required to maintain hydraulic flux through geotextile bag filter, is more than twice the cost of the required alum. Thus, even if alum were not used, the cost of adding polymer alone would only reduce treatment chemical costs by about 30%.

An estimate of the costs required to apply a large geotextile bag filter at a hypothetical commercial scale aquaculture facility producing 454 mt of fish annually is summarized in Table 7. This bag filter would be loaded with approximately 127 mt of TSS annually, assuming that 0.2 kg of TSS are captured as part of drum filter backwash per kg feed fed, and a mean facility feed conversion rate of 1.4. The total backwash flow to be treated would conservatively be <500 m3/day, assuming a total recirculating water flow of100,000 m3/day and a backwash that is 0.5% of the total recycle flow (i.e., 100,000 x 0.005 = 500 m3/day). Estimates of annual alum and polymer cost would be US $18,159 (Table 7). Assuming the 65 L/day/m2 of bag area hydraulic loading rate applied as in the present study, four geotextile bags sized to provide a surface area of approximately 1925 m2 each is required for bag removal and replacement events to occur on a quarterly basis. This size requirement could be met with geotextile bags measuring 27.4 m (90 ft) in circumference and approximately 69.8 m (229 ft) long, which would cost approximately US $20,000 per bag. Additionally, the geotextile bag filter needs to rest on an impermeable, drained surface to capture the filtrate. The filtrate would also require further treatment, i.e., additional capital equipment and operating cost, if it were to be discharged or reclaimed and reused.

4. Conclusions

Geotextile bag filters can consistently remove approximately 95% of the TSS contained in aquaculture backwash flows when

loaded at approximately 60-70 L/day/m2 bag surface area and amended with polymer plus alum, ferric chloride, or lime. Alum was found to be the most cost effective coagulant amendment ($0.16/kg dry weight biosolids) and hydrated lime the most expensive ($0.68/kg dry weight biosolids), with cost for ferric chloride mid-way between alum and lime. Geotextile bag filters also provide good solids dewatering, producing 19-22% biosolids concentrations 7-10 days after wastewater treatment additions were discontinued. As a result, cost associated with handling and disposal of fish culture wastes can be mitigated, and sludge produced is suitable for land application or composting. Conversely, mineralization and leaching releases large concentrations of dissolved wastes as the captured biosolids age within the geotextile bag filters. Of note, COD and cBOD5 removal under all treatments were negligible, and lime application appeared to promote release of organic acids, which produced no net removal of cBOD5 across the bag filters. Further, it was evident that mineralization of organic nitrogen to TAN occurred over time, producing mean TAN concentrations in the filtrate of 28-60 mg/L. Also, TP precipitation and capture was limited with the alum and ferric chloride treatments. However, TP precipitation and capture with lime amendment was evident throughout the first 40 days that the bags were hydraulically loaded but after that declined. Addition of sufficient hydrated lime to create an inlet pH of at least 12 would likely have further improved TP capture and storage within the geotextile bag filters.

The significant amount of COD, cBOD5, DRP, and inorganic nitrogen that leached into the filtrate is a grave disadvantage that must be considered before applying geotextile bag filters in typical dewatering and effluent treatment application. However, geotex-tile bag filters could provide an excellent pretreatment in an application where the TSS must be dewatered for disposal, but leaching of dissolved organic carbon and cBOD5 from these biosolids is desired to drive a denitrification process removing nitrate from another effluent. Geotextile bag filters could also provide an excellent pretreatment in an application where the TSS must be dewatered for disposal, but leaching of inorganic nitrogen and DRP from these biosolids is desired to feed nutrients to a downstream hydroponic/aquaponic operation or used as irrigation for field crops.

The captured biosolids that have been removed from the geotextile bag filters can serve as a soil amendment due to their

2.2-3.6% nitrogen, 0.15-0.17% phosphorus, and 0.4-0.9% potassium, when field applied at agronomic rates. In addition, field application of lime amended biosolids can increase soil fertility, increase soil pH, and make metals more insoluble, which can minimize plant uptake and movement of metals to groundwater (EPA, 2000).

Future work with geotextile bags will explore comparisons with other settling (i.e., gravity thickening tank) and filtration techniques (i.e., inclined belt filter) in terms of performance and cost. Additional research will also examine the capacity to incorporate a modified activated sludge process to promote nitrification, denitrification, and biological phosphorus removal of geotextile bag filtrate in a sequencing batch reactor system.

Acknowledgments

This work was supported by the United States Department of Agriculture, Agricultural Research Service under grant Agreement No. 59-1930-1-130. We would like to thank Brain Mason, Daniel Coffinberger, and Fred Ford for their assistance constructing the geotextile bag filter system.

References

Adler, P.R., Sikora, L.J., 2005. Mesophilic composting of Arctic char manure. Compost. Sci. Util. 13, 34-42. Allan, J.C., Komar, P.D., 2004. Environmentally compatible cobble berm and artificial

dune for shore protection. Shore Beach 72 (1), 9-18. Baker, K.B., Chastian, J.P., Dodd, R.B., 2002. Treatment of lagoon sludge and liquid animal manure utilizing geotextile filtration. In: ASAE Annual International Meeting, Chicago, IL. Bergheim, A., Kristiansen, R., Kelly, L., 1993. Treatment and utilization of sludge from landbased farms for salmon. In: Wang, J.K., Joseph, St. (Eds.), Techniques for Modern Aquaculture. American Society of Agricultural Engineers, Michigan, pp. 486-495.

Bergheim, A., Cripps, S.J., Liltved, H., 1998. A system for the treatment of sludge from

land-based fish-farms. Aquat. Living Resour. 11, 279-287. Bergheim, A., Brinker, A., 2003. Effluent treatment for flow through system and

European environmental regulations. Aquacult. Eng. 27, 61-77. Brazil, B.L., Summerfelt, S.T., 2006. Aerobic treatment of gravity thickening tank

supernatant. Aquacult. Eng. 34, 92-102. Chen, S., Coffin, D.E., Malone, R.F., 1997. Sludge production and management for

recirculating aquaculture systems. J. World Aquacult. Soc. 28, 303-315. Chen, S., Summerfelt, S.T., Losordo, T., Malone, R., 2002. Recirculating systems, effluents, and treatments. In: Tomasso, J.R. (Ed.), Aquaculture and the Environment in the United States. U.S. Aquaculture Society, a Chapter of the World Aquaculture Society, Baton Rouge, LA, pp. 119-140. Cripps, S.J., Bergheim, A., 2000. Solids management and removal for intensive land-

based aquaculture production systems. Aquacult. Eng. 22, 33-56. Cripps, S.J., Kelly, L.A., 1996. Reductions in wastes from aquaculture. In: Baird, D.J., Beveridge, M.C.M., Kelly, L.A., Muir, J.F. (Eds.), Aquaculture and Water Resource Management. Blackwell Science, Inc., Cambridge, MA, pp. 166-201. Davidson, J., Summerfelt, S.T., 2005. Solids removal from a coldwater recirculating system—comparison of a swirl separator and radial-flow settler. Aquacult. Eng. 33, 47-61.

Ebeling, J.M., Rishel, K.L., Sibrell, P.L., 2005. Screening and evaluation of polymers as flocculation aids for the treatment of aquacultural effluents. Aquacult. Eng. 33 (4), 235-249.

Ebeling, J.M., Sibrell, P.L., Ogden, S., Summerfelt, S.T., 2003. Evaluation of chemical coagulation-flocculation aids for the removal of phosphorus from recirculating aquaculture effluent. Aquacult. Eng. 29, 23-42. Ebeling, J.M., Ogden, S., Sibrell, P.L., Rishel, K.L., 2004. Application of chemical coagulation aids for the removal of suspended solids (TSS) and phosphorus from the microscreen effluent discharge of an intensive recirculating aquaculture system. N. Am. J. Aquac. 66, 198-207. Ebeling, J.M., Welsh, C.F., Rishel, K.L., 2006. Performance evaluation of the Hydro-tech belt filter using coagulation/flocculation aids (alum/polymers) for the removal of suspended solids and phosphorus from intensive recirculating aquaculture microscreen backwash effluent. Aquacult. Eng. 35, 61-77. Ebeling, J.M., Rishel, K.L., 2006. Performance evaluationofgeotextile tubes. In: Paper Presented at the 6th International Recirculating Aquaculture Conference, Roanoke, VA.

EPA (United States Environmental Protection Agency, 2004. Effluent Limitations Guidelines and New Source Performance Standards for the Concentrated Aquatic Animal Production Point Source Category, Final Rule. 40 CFR Part 451. EpA, Washington, DC.

EPA (United States Environmental Protection Agency, 2000. Biosolids Technology Fact Sheet: Alkaline Stabilization of Biosolids. EPA 832-F-00-052. EPA Office of Water, Washington, DC.

EPA (United States Environmental Protection Agency, 1995. Standards for the Use or Disposal of Sewage Sludge. 40 CFR Parts 403 and 503. EPA, Washington, DC.

Elko, N.A., Mann, D.W., 2007. Implementation of geotextile T-groins in Pinellas County, Florida. Shore Beach 75 (2), 2-10.

Ewart, J.W., Hankins, J.A., Bullock, D., 1995. State Policies for Aquaculture Effluents and Solid Wastes in the Northeast Region Bulletin No. 300. University of Massachusetts: Northeastern Regional Aquaculture Center, North Dartmouth.

Idaho Division of Environmental Quality, 1998. Idaho Waste Management Guidelines for Aquaculture Operations. Idaho Department of Health and Welfare, Division of Environmental Quality, Twin Falls, Idaho.

Korkut, R., Martinez, E.J., Morales, R., Ettema, R., Barkdoll, B., 2007. Geobag performance as scour countermeasure for bridge abutments. J. Hydraul. Eng. 133 (4), 431-439.

Lekang, O.I., Bergheim, A., Dalen, H., 2000. An integrated wastewater treatment system for land-based fish-farming. Aquacult. Eng. 22 (3), 199-211.

Losordo, T.M., Delong, D.P., Guerdat, T.C., 2006. Advancements in solids waste capture and wastewater treatment in intensive tank based production systems. In: Paper Presented at the Annual Meeting of the World Aquaculture Society, Florence, Italy.

Metcalf and Eddy Inc., 1991. Wastewater Engineering, Treatment/Disposal/Reuse, third edition. McGraw Hill, New York, NY.

Miner, J.R., Humenik, F.J., Rice, J.M., Rashash, D.M.C., Williams, C.M., Robarge, W., Harris, D.B., Sheffield, R., 2003. Evaluation of a permeable, 5 cm thick polyethylene foam lagoon cover. Trans. ASAE 46 (5), 1421-1426.

Mudrak, V.A., 1981. Guidelines for economical commercial fish hatchery waste-water treatment systems. In: Allen, L.J., Kinney, E.C. (Eds.), Proceedings of the Bio-engineering Symposium for Fish Culture. American Fisheries Society, Fish Culture Section, Bethesda, Maryland, pp. 174-182.

Oh, Y.I., Shin, E.C., 2006. Using submerged geotextile tubes in the protection ofthe E. Korean shore. Coast Eng. 53 (11), 879-895.

Olson, G.L., 1991. The use of trout manure as a fertilizer for Idaho crops. In: Blake, J., Donald, J., Magette, W. (Eds.), National Livestock, Poultry and Aquaculture Waste Management. American Society of Agricultural Engineers, St. Joseph, Michigan, USA, pp. 198-205.

Outwater, A.B., 1994. Reuse of Sludge and Minor Wastewater Residuals. CRC Press, Boca Raton, FL.

Palacios, G.L., Timmons, M.B., 2001. Determining design parameters for recovery of aquaculture wastewater using sand beds. Aquacult. Eng. 24, 289-299.

Qasim, S.R., 1999. Wastewater Treatment Plants: Planning, Design and Operation. Technomic Publishing Company, Inc., Lancaster, PA.

Rickson, R.J., 2006. Controlling sediment at source: an evaluation of erosion control geotextiles. Earth Surf. Proc. Land. 31 (5), 550-560.

Rishel, K.L., Ebeling, J.M., 2006. Evaluation of coagulation-flocculation aids for the removal of suspended solids from intensive recirculating aquaculture effluent discharge. J. World Aquacult. Soc. 37 (2), 191-199.

Sharrer, M.J., Tal, Y., Ferrier, D., Hankins, J.A., Summerfelt, S.T., 2007. Membrane biological reactor treatment of a saline backwash flow from a recirculating aquaculture system. Aquacult. Eng. 35, 159-176.

Sedlak, R.I., 1991. Phosphorus and Nitrogen Removal from Municipal Wastewater— Principles and Practice, second edition. CRC Press, Boca Raton, FL.

Shin, E.C., Ahn, K.S., Oh, Y.I., Das, B.M., 2002. Construction and monitoring of geotubes. In: Proceedings of the International Offshore and Polar Engineering Conference, vol. 12. pp. 469-473.

Stewart, N.T., Boardman, G.D., Helfrich, L.A., 2006. Characterization of nutrient leaching rates from settled rainbow trout (Oncorhynchus mykiss) sludge. Aquacult. Eng. 35 (2), 191-198.

Summerfelt, S.T., Davidson, J.T., Waldrop, T., Tsukuda, S., Bebak-Williams, J., 2004. A partial reuse system for coldwater aquaculture. Aquacult. Eng. 31, 157-181.

Summerfelt, S.T., 1999. Waste handling systems. In: Bartali, E.H., Wheaton, F.W. (Eds.), CIGR Handbook of Agricultural Engineering Volume II: Animal Production and Aquacultural Engineering. American Society of Agricultural Engineers, St. Joseph, MI, pp. 309-350.

Summerfelt, S.T., Adler, P.R., Glenn, D.M., Kretschmann, R., 1999. Aquaculture sludge removal and stabilization within created wetlands. Aquacult. Eng. 19 (2), 81-92.

Summerfelt, S.T., Vinci, B.J., 2008. Better management practices for recirculating systems. In: Tucker, C.S., Hargreaves, J.A. (Eds.), Environmental Best Management Practices for United States Aquaculture. Blackwell Publishing, Ames, Iowa.

Summerfelt, R.C., Penne, C.R., 2007. Septic tank treatment of the effluent from a small-scale commercial recycle aquaculture system. N. Am. J. Aquacult. 69 (1), 59-68.

Vishnudas, S., Savenije, H.H.G., Van der Zaag, P., Anil, K.R., Balan, K., 2006. The protective and attractive covering of a vegetated embankment using coir geotextiles. Hydrol. Earth Syst. Sci. 10 (4), 565-574.

Westers, H., 1991. Operational waste management in aquaculture effluents. In: Cowey, C.B., Cho, C.Y. (Eds.), Nutritional Strategies and Aquaculture Waste. Fish Nutrition Research Laboratory, Department of Nutritional Sciences, University of Guelph, Guelph, Ontario, pp. 231-238.

Wett, B., Demattio, M., Becker, W., 2005. Parameter investigation for decentralised dewatering and solar thermic drying of sludge. Water Sci. Technol. 41 (10), 65-73.

Worley, J.W., Bass, T.M., Vendrell, P.F., 2008. Use of geotextile tubes with chemical amendments to dewater dairy lagoon solids. Bioresour. Technol. 91,4451-4459.