Bioresource Technology 102 (2011) 641–646

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Pathogen removal in farm-scale psychrophilic anaerobic digesters processingswine manure

Daniel Massé a,*, Yan Gilbert a, Edward Topp b

a Dairy and Swine Research and Development Centre, Agriculture and Agri-Food Canada, 2000, College Street, Sherbrooke, QC, Canada J1M 0C8b Southern Crop Protection and Food Research Centre, Agriculture and Agri-Food Canada, 1391 Sandford Street, London, ON, Canada N5V 4T3

a r t i c l e i n f o

Article history:Received 21 April 2010Received in revised form 29 July 2010Accepted 5 August 2010Available online 10 August 2010

Keywords:Anaerobic digestionPathogensSwine manurePsychrophilic

0960-8524/$ - see front matter Crown Copyright � 2doi:10.1016/j.biortech.2010.08.020

* Corresponding author. Tel.: +1 418 565 9174; faxE-mail address: daniel.masse@agr.gc.ca (D. Massé)

a b s t r a c t

This study assessed the efficiency of commercial-scale psychrophilic anaerobic digestion in sequencingbatch reactors (PADSBRs) for pathogen removal from pig manure. The impact of treatment cycle lengthand of hydraulic flow regimes on pathogen removal efficiency was investigated. Two conventionallyoperated SBRs (BR1 and BR2) and two SBRs simultaneously fed during the draw step (BR3 and BR4) weremonitored over a two-year period. PADSBRs significantly decreased the concentration of coliforms, Sal-monella, Campylobacter spp., and Y. enterocolitica, respectively from about 106, 103 CFU g�1, 103, and104 CFU g�1 to undetectable levels in most samples. Densities of the gram-positive Clostridium perfringensand Enterococcus spp. remained high (105 CFU g�1) in the digesters throughout treatment. The PADSBRsmaintained the same level of pathogen removal when the treatment cycle length was reduced from 2 to 1week. Mass balances on volatile fatty acids (VFAs) revealed short-circuits of inlet flow respectively aver-aging 6.3% and 6.4% for BR3 and BR4, significantly reducing the overall performance of these reactorsregarding pathogens removal. The results obtained in this study show the ability of low temperatureanaerobic digestion to remove or significantly reduce indicator and pathogen concentration from rawpig manure.

Crown Copyright � 2010 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Canada is among the ten most important hog producers world-wide, based on mass of pig meat produced (FAO, 2009). Moreover,during the past 20 years, the Canadian pig industry has changed tomore specialized facilities and intensified production systems. In2006, there was three times fewer farms than in 1986, while theaverage number of pigs increased by more than 400% in swinefacilities (CCP, 2008). The quantity of pigs produced in the countryduring this period has more than doubled, raising concerns aboutthe environmental impacts of the industry, particularly with re-spect to swine manure management (CCP, 2008).

Manure is an excellent source of crop nutrients and will im-prove soil structure through provision of organic matter. However,over-application will increase the risk of movement of nitrogenand phosphorus to groundwater and adjacent surface water(Sharpley et al., 2002). Most environmental studies concerningswine manure management either have focused on the effects ofnutrients on water quality, or odour problems and air quality.Nonetheless, the microbial quality of manure should not be ne-glected since many outbreaks of gastroenteritis related to livestock

010 Published by Elsevier Ltd. All r

: +1 418 564 5507..

operations have been reported (Guan and Holley, 2003; Pell, 1997;Spencer and Guan, 2004).

The most prevalent pathogenic microorganisms from manureare the bacteria Salmonella, Escherichia coli, Yersinia, Campylobacter,and the protozoa Giardia and Cryptosporidium (Bicudo and Goyal,2003; Hutchison et al., 2005), but Clostridium perfringens, Listeriamonocytogenes, and Treponema hydrosenteriae have also been re-ported as causative agents of human infections related to livestock(Colleran, 2000). The persistence of enteric pathogens in manurewill vary according to manure handling practices, storage manage-ment, ambient conditions and duration, type of slurry, and patho-gen type (Bicudo and Goyal, 2003). There is an urgent need forcost-effective methods to address environmental issues of pigmanure by reducing microbial and chemical contaminants.

Côté et al. (2006) studied the efficiency of a low temperatureanaerobic laboratory-scale digester to reduce viable populationsof indicator microorganisms (total coliforms, E. coli) and selectedpathogens (Salmonella, Yersinia enterocolitica, Cryptosporidium andGiardia) in swine slurries. Forty liter anaerobic intermittently-fedsequencing batch reactors (SBR) operated at 20 �C during 20 daycycles reduced total coliforms by 1.6 to 4.2 log10 units (CFU g�1)(97.9–100%), and E. coli by 2.5–4.2 log10 units (99.7–100%). Salmo-nella, Cryptosporidium, and Giardia concentrations decreased toundetectable levels (<100 CFU g�1). Even though the latter study

ights reserved.

642 D. Massé et al. / Bioresource Technology 102 (2011) 641–646

showed the efficiency of lab-scale psychrophilic anaerobic diges-tion regarding pathogen and indicator removal, no study reportsthe performance of commercial-scale digesters for the removal ofthese microorganisms. In the study reported here, the long termperformance and reliability of farm-scale (165 and 450 m3) lowtemperature digesters for the reduction of pathogens concentra-tion in raw swine manure was investigated.

2. Methods

2.1. Reactor operation

Four farm-scale cylindrical digesters (BR1–BR4) were used forpsychrophilic anaerobic digestion in sequencing batch reactors(PADSBRs) of pig manure, and monitored over a two years period(June 2007–2009). The effective volume was 165 m3 (OutsideDiameter (OD) � 7 m) for BR1 and BR2, and 450 m3 (OD � 11 m)for BR3 and BR4. The outlet consisted of a flexible pipe (InsideDiameter (ID) = 0.101 m) attached to a steel wire, allowing sludgeheight control during the draw step. Feeding was performedthrough a polyvinyl chloride (PVC) pipe (ID = 0.101 m) at the bot-tom of the reactor. An inflatable device was located at the top ofthe reactor to allow for gas and liquid displacement without airentering the system.

Three reactors (BR1, BR2, and BR3) in operation since 2001were located on the same farm and operated under a 7 day cy-cle, while the fourth (BR4) in operation since 2004 was operatedunder a 14 day cycle on another farm. BR1 and BR2 were oper-ated as conventional SBRs, i.e. after sludge settling, supernatantwas removed before feeding with fresh manure. Rather thanbeing removed before feeding, BR3 and BR4’s supernatant wasforced outside the bioreactors by simultaneously feeding withfresh manure at the bottom of the digesters. Approximately fromthe 500th day to the end of the monitoring period, BR3 wasoperated as a conventional SBR due to operational constraints.Due to solids accumulation inside BR3, 46 m3 of sludge were re-moved at the 665th day of monitoring, before the feed step. Thetemperature of all reactors was maintained throughout the yearat an average of 24 �C, with fluctuations ranging from 20 to26.5 �C. Composite samples were collected from the effluent foranalysis during the drawing step. Raw manure was sampled dur-ing the filling period. All samples were kept at 4 �C until theanalysis, which was done less than 24 h after sampling. Table1 summarizes the operating conditions during the study andmanure characteristics.

Table 1Operational conditions of full-scale PADSBRs and inlet characteristics of manure fed to bio

Parameters BR1 B

Operational conditionsLoading rate (kg TCOD m�3 d�1) 1.47 (0.57)a 1Total cycle length (day) 7 7Reactor effective volume (m3) 120 1

Manure characteristicsTCOD (g O2 L�1) 115 (44) 1SCOD (g O2 L�1) 41 (12) 3pH 6.88 (0.14) 6Alkalinity (mg CaCO3 L�1) 17,300 (4100) 1Acetic acid (mg L�1) 9300 (3100) 8Propionic acid (mg L�1) 3600 (1300) 3Butyric acid (mg L�1) 5400 (2300) 5TS (%) 7.20 (2.02) 6VS (%) 5.63 (1.65) 5TKN (mg N L�1) 6600 (1600) 6NH4–N (mg N L�1) 4600 (1200) 4

a Values in parenthesis represent the standard deviation of observations (Number of

The average organic loading rate (OLR) was based on theamount of total chemical oxygen demand (TCOD) fed per volumeof sludge present at the start of a cycle and the cycle length. Itwas calculated as follows:

Lf ¼Vf Cf

Vitcð1Þ

where Lf is the loading rate based on total cycle length (kgT-COD m3 day�1), Vf is the volume of feed (m3), Cf is TCOD concentra-tion in the feed (kg m�3), Vi is the volume of sludge in the reactor atthe beginning of the cycle (m3), and tc is the total duration of the cy-cle (day).

2.2. Physico-chemical analysis

Soluble chemical oxygen demand (SCOD) of fresh manure andeffluent samples was determined by analyzing the supernatant ofcentrifuged slurry. Total and soluble COD were determined accord-ing to the method developed by Knechtel (1978). Alkalinity, pH, to-tal solids (TS), volatile solids (VS), total Kjeldahl nitrogen (TKN) andammonia nitrogen (NH4–N) were determined using standardmethods (APHA, 1992). TKN and NH4–N were analyzed using aKjeltec auto-analyzer model TECATOR 1030 (Tecator AB, Hoganas,Sweden).

Before VFAs quantification, samples were conditioned. Tengrams of sample was centrifuged at 41,700g, 30 min at 15 �C.One microliter of 0.5 M H2SO4 was added to 5 mL of the superna-tant and centrifuged at 21,800g, 15 min at 15 �C. An internal stan-dard (2-ethylbutyrate at 2 g L�1) was added (0.5 mL) to a tube with0.5 mL of the acidified–centrifuged supernatant and 0.1 g of DOW-EX 50WX8 resin (The Dow Chemical Company, Midland, MI) andvortexed. VFAs were measured with a Perkin Elmer gas chromato-graph model 8310 (Perkin Elmer, Waltham, MA), mounted with aDB-FFAP high resolution column. Results were analyzed using Tur-boChrom version 6.2.1 software (Perkin Elmer).

2.3. Microbiological analysis

For pathogen quantification, raw manure and anaerobic diges-ter effluent samples were prepared by diluting 1:10 in sterile so-dium metaphosphate buffer (2 g L�1; pH 6.8). Total coliformswere enumerated by direct plating on mEndo–LES agar (Difco,Mississauga, ON, Canada) and incubated at 37 �C for 18–20 h. Col-onies which produced a distinctive metallic green sheen werecounted as total coliforms. Fecal coliforms were quantified by

reactors.

R2 BR3 BR4

.48 (0.59) 1.31 (0.45) 1.65 (0.64)7 14

20 425 425

06 (41) 97 (34) 98 (39)8 (12) 35 (10) 33 (15).90 (0.16) 6.95 (0.17) 7.33 (0.21)6,300 (4100) 15,600 (3400) 20,300 (5000)500 (2900) 7900 (2700) 8600 (3700)300 (1100) 3100 (1100) 2900 (1000)400 (2200) 4800 (1800) 2600 (1700).72 (2.01) 6.20 (1.69) 6.99 (2.23).21 (1.64) 4.77 (1.40) 5.24 (1.86)300 (1700) 5900 (1300) 6600 (1900)400 (1300) 4100 (1000) 5100 (1500)

observations for BR1 = 81, BR2 = 89, BR3 = 99, BR4 = 30).

D. Massé et al. / Bioresource Technology 102 (2011) 641–646 643

direct plating on mFC agar (Difco) and incubated for 18–20 h at44.5 �C. Colonies producing a distinctive indigo blue color wereenumerated as fecal coliforms. Enterococcus spp. were counted bydirect plating onto mEnterococcus agar (Difco) and incubated at37 �C for 48 h. Colonies producing a burgundy color were countedas Enterococcus sp. E. coli were enumerated by direct plating onmFC basal medium (Difco) supplemented with 3-bromo-4-chloro-5-indolyl-b-D-glucuronide (100 mg L�1, Med-Ox Diagnos-tics, Ottawa, ON) and incubated for 18–24 h at 44.5 �C. Coloniesproducing the characteristic blue color indicative of b-glucuroni-dase activity were enumerated as E. coli. C. perfringens was quanti-fied on mCP agar (Med-Ox Diagnostics) incubated at 44.5 �C for24 h under anaerobic conditions. Yellow colonies with a yellowhalo and which turned magenta after exposure to ammoniumhydroxide fumes were expected to be C. perfringens. Presumptivecolonies were confirmed by inoculation into skim milk broth. Stor-my fermentation of the broth after 24 h was considered a positiveconfirmation for C. perfringens. This protocol allowed the quantifi-cation of vegetative cells and spores together.

Y. enterocolitica were enumerated by direct plating onto Cefsu-lodin–Irgasan–Novobiocin agar (CIN, Difco) and incubated for 18 hat 30 �C. Colonies that were less than 2 mm in diameter andformed a red bulls-eye were considered presumptive colonies forY. enterocolitica. Five presumptive colonies were picked and inocu-lated into lysine arginine iron agar slants (LAIA) and considered po-sitive if LAIA results were alkaline slant, acid butt, H2S negative andgas negative (Weagant, 1983).

Campylobacter spp. were counted by direct plating onto Char-coal Cefoperazone Desoxycholate Agar (CCDA, Oxoid, Nepean,ON) supplemented with 0.032 mg mL�1 of cefoperazone and0.01 mg mL�1 of amphotericin B (Oxoid) and Campy-Line Agar(CLA, (Line, 2001)). Plates were incubated at 42 �C in microaero-philic conditions for 48 h in an anaerobic jar with GasPakTM EZCampy (6–16% O2, 2–10% CO2, BD, Oakville, ON). Campylobacter-like colonies were quantified on both media. On CLA agar, pre-sumptive colonies of Campylobacter are smooth, convex, burgundycolored and are approximately 2–4 mm in diameter. On CCD agar,presumptive colonies of Campylobacter are smooth, convex, andtranslucent colorless to cream colored and are approximately 2–4 mm in diameter. Suspected colonies were also observed undermicroscope and tested for Gram coloration, oxidase and catalase,(Sigma–Aldrich, Oakville, ON) after inoculation onto Columbiablood agar and incubate 24 h at 35 �C under microaerophilic condi-tions. Results were confirmed using Campylobacter Latex kit as de-scribed by the manufacturer (Oxoid).

Salmonella spp. were quantified by direct plating onto Salmo-nella chromogenic agar (Oxoid) and XLD plates (Difco). Plates wereincubated at 42 �C for 24 h before being enumerated. Salmonella-like colonies were quantified on both media. On XLD agar, pre-sumptive colonies of Salmonella are smooth, convex, and red-col-

Table 2Performance of PADSBRs regarding pathogen and indicator removal.

Parameters Concentration range (CFU g�1) Log10 Rem

BR1

SBR

Total coliforms 6.0 � 102–1.5 � 106 2.9Fecal coliforms 5.0 � 102–1.6 � 106 2.8Escherichia coli 4.0 � 102–5.8 � 105 2.9Enterococcus spp. 1.8 � 104–1.7 � 106 0.8Clostridium perfringens <1.0 � 103–3.7 � 106 0.2Yersinia enterocolitica <1.0 � 102–3.3 � 105 2.1Salmonella spp. <1.0 � 102–5.0 � 104 1.2Campylobacter spp. <1.0 � 102–5.2 � 104 1.3

a Values were calculated using mean inlet and outlet concentrations (Number of obse

ored with a black bulls-eye and are approximately 2–4 mm indiameter. Colonies may also produce a red halo in the media. OnSalmonella chromogenic agar, presumptive colonies of Salmonellaare mauve to magenta in color and are approximately 1.5–3 mmin diameter. Suspected colonies were picked and streaked on BHIagar and plates were incubated at 35 �C for 24 h. A well isolatedcolony was inoculated into the Rapid Salmonella Latex test as de-scribed by manufacturer (Oxoid). When further identificationwas necessary, a suspect colony was inoculated onto a fresh nutri-ent agar and incubated 24 h at 35 �C. Gram stain, oxidase and cat-alase tests were performed and an API 20E strip was inoculated forultimate identification.

2.4. Statistical analysis

Statistical data analysis was carried out using the software JMP,version 7.0.1 (SAS Institute Inc., Cary, NC). Homogeneity of vari-ances was analysed with the Levene test statistic. Means, standarddeviations and frequency distributions of the data were deter-mined. Differences between means were tested with comparativestatistics, using one-way ANOVA. After obtaining F-ratios, Tukey’spost hoc tests and Student t-test were performed to determine sig-nificantly different pairs of data. The level of significance was set atP < 0.05. The value ‘‘99” was used for statistical analysis whenpathogen and indicator concentrations were below the methoddetection limit (<100 CFU g�1).

3. Results and discussion

3.1. Manure characteristics

Four farm-scale psychrophilic anaerobic digesters were moni-tored over a two-year period to assess their reliability and stabilitywith respect to removal of enteric pathogens. No significant differ-ence (P > 0.05) was found between raw manure samples collectedat bioreactor inlet for all tested parameters (Table 1). The averageOLR applied to the bioreactors ranged from 1.31 to 1.65 kg TCODm�3 d�1, with a TCOD concentration around 100,000 mg O2 L�1.The pH of raw manure was near neutrality, even though high VFAsconcentrations were observed, mainly because of the high amountof alkalinity in the manure.

Pathogen and indicator concentrations in raw pig manure werehighly variable (Table 2), but within the range of reported values(Hill and Sobsey, 1998; Hutchison et al., 2005; Kearney et al.,1993a). The mean concentration of total and fecal coliforms,Enterococcus spp. and E. coli was 2.7 � 105, 1.9 � 105, 3.8 � 105,and 8.8 � 104 CFU g�1, respectively. Lower mean concentrationswere observed for Y. enterocolitica, Salmonella and Campylobacterspp., with 2.0 � 104, 2.2 � 103, and 3.0 � 103 CFU g�1, respectively.Y. enterocolitica, Salmonella and Campylobacter concentrations

ovala

BR2 BR3 BR4

SBR Non-SBR SBR Non-SBR SBR

2.5 1.0 2.4 1.2 2.72.6 0.9 3.0 1.2 2.82.8 0.9 2.8 0.9 2.81.0 0.7 0.6 0.8 1.00.0 0.2 �0.1 0.2 �0.12.1 1.1 1.9 1.3 2.41.4 0.9 1.3 1.1 1.31.5 1.1 1.4 1.1 1.3

rvations for BR1 = 48, BR2 = 48, BR3 = 45, BR4 = 45).

644 D. Massé et al. / Bioresource Technology 102 (2011) 641–646

equal of below the method detection limit (100 CFU g�1) were de-tected in 10%, 43% and 24%, respectively of the samples from theinlets of BR1–BR3 inlet. For BR4 inlet, 10%, 45% and 53% of samplesshowed concentrations equal or below the detection limit, forthese microorganisms, respectively.

3.2. Bioreactor performance

Since biogas production was not monitored on a regular basison these full-scale reactors during digestion, the overall bioreactorperformance was verified by monitoring TCOD removal. Conven-tionally operated SBRs (BR1 and BR2) showed TCOD removal above70%, while those being simultaneously fed with fresh manure dur-ing the drawing step (BR3 and BR4) only reached 61.6% and 56.3%(Fig. 1C and D), respectively. Near the end of the two-year monitor-ing period (500th day), BR3 performance decreased significantly asmeasured by TCOD removal. This was principally due to typical so-lid accumulations inside the reactor which increased the solidsconcentration at the outlet. This was indicated by an increase ofthe outlet VS/TS ratio (Fig. 1C). Similar results were obtained forBR4 at the end of the monitoring period (Fig. 1D). For BR3, satisfac-tory performance was regained after removing solids from thereactor.

The low temperature AD treatment resulted in significantlylower concentrations of pathogens and indicators in BR1 and BR2effluents. An average reduction of 2.6–2.9 log10 units and 1.2–1.4 log10 units was observed for fecal coliforms and Salmonellaspp. for an HRT of 7 days (Table 2), leading to undetectable levelsin more than 81% and 96% of outlet samples, respectively. Otherstudies obtained longer pathogen and indicator removal times forsimilar or higher temperatures during digestion. Kumar et al.(1999) used an ampicillin-resistant E. coli strain to study thepersistence of this organism in cattle dung slurry during anaerobicdigestion. The survival was 25 days at room temperature(18–25 �C) and 15 days at 35 �C. In the same study, the completeelimination of Salmonella at 35 �C occurred on the 15th day

0

5

10

15

20

25

30

35

40

0 200 400 600 800Time (d)

Shor

t circ

uit (

%)

Conventional SBR

0

50000

100000

150000

200000

250000

0 200 400 600 800

Time (d)

TCO

D (m

g O

2 L-1

)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

VS/T

S

-1

A B

C D

Fig. 1. Percentage of VFAs going through BR3 (A) and BR4 (B) by short circuiting of hyperiods when bioreactors were operated as conventional SBRs. Dashed line representsEffluent ( ); VS/TS (N).

whereas 25 days were necessary at room temperature. Berg andBerman (1980) reported a reduction of 2.0 and 1.8 log10 units ofindigenous total and fecal coliforms, respectively, during the mes-ophilic digestion (35 �C; SRT 20 days) of municipal sludge. In an-other study, Salmonella and E. coli, which were added to liquidpig manure, and coliforms of the indigenous flora, were reducedby 1.0–2.1 log units when subjected to anaerobic filter treatmentat 35 �C in two pilot-plant reactors operated at 0.8–4.2 days HRT(Olsen, 1988).

In an investigation studying the survival of indigenous patho-gens in a full-scale anaerobic digester which was daily fed withcattle farm wastes and operated at 28 �C, Kearney et al. (1993a)found that Campylobacter jejuni was the most resistant bacteriumtested among E. coli, Salmonella enterica Typhimurium, Y. enterocol-itica, and L. monocytogenes. Its mean T90 (time to reduce itsconcentration of 90%) value was 438.6 days, compared toY. enterocolitica that showed a T90 of 18.2 days. Mean T90 observedunder the operational conditions tested in their study suggestthat the process did not allow the complete removal of testedpathogens. In comparison, undetectable levels of Y. enterocoliticaand Campylobacter were obtained in our study in more than87% of BR1 and BR2 outlet samples for only 7 days treatment.The other 13% showed an average concentration of7.5 � 102 CFU g�1 for these microorganisms. It seems that forthe same species or genera, various removal efficiencies are ob-served, depending on the substrate being digested. For example,90% of Salmonella spp. was removed in 34.5 days when digestingcattle slurry mixed with hen, pig and potato wastes at 28 �C,whereas it was completely removed in 15 days when digestingcattle dung slurry at 35 �C (Kumar et al., 1999), in 21 days whendigesting vegetable, fruit, and garden waste in mesophilic condi-tions (Termorshuizen et al., 2003), and in only 7 days during thepsychrophilic digestion (24 �C) of pig manure in this study.According to Smith et al. (2005), efficient mixing and organicmatter stabilisation are the main factors controlling the rate ofinactivation under mesophilic conditions and not a direct effect

0

5

10

15

20

25

30

35

40

0 200 400 600 800

Time (d)

Shor

t circ

uit (

%)

0

50000

100000

150000

200000

250000

0 200 400 600 800Time (d)

TCO

D (m

g O

2 L)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

VS/T

S

draulic flow, TCOD removal and VS–TS ratio in BR3 (C) and BR(4). Arrows indicatethe cycles when BR3 operation was switched toward conventional SBR. Feed (�);

D. Massé et al. / Bioresource Technology 102 (2011) 641–646 645

of temperature. Thus, the observed differences regarding patho-gen removal from various matrices could be due to the level ofstabilization reached for each of the digested substrates.

C. perfringens and Enterococcus spp. were not significantly re-moved by psychrophilic anaerobic digestion (Table 2). Negligibleremoval of 0.2 log10 unit was reported for C. perfringens spores dur-ing the treatment of swine manure in an anaerobic lagoon in NorthCarolina (Hill and Sobsey, 1998). Similarly, no significant C. perfrin-gens removal was observed during the mesophilic digestion(HRT = 20 days) of sewage sludge in a full-scale anaerobic reactor,measured both as spores and total counts (Chauret et al., 1999).The same study also showed that Enterococcus spp. was not signif-icantly removed from the sludge during digestion. Effenbergeret al. (2006) also reported low removal efficiencies for fecal entero-cocci (about 1.0 log10 unit) after the first stage of a mesophilic–thermophilic–mesophilic anaerobic digestion process treatingliquid dairy cattle manure. These data are in agreement with ourresults since Enterococcus spp. removal was at most 1.0 log10 unitwith an average concentration in bioreactors’ outlet around105 CFU g�1. Enterococcus is a facultative anaerobic microorganismthat is present in the intestinal tract of various animals, Enterococ-cus faecium and Enterococcus faecalis being the most frequentlyencountered species in pig’s (Klein, 2003). This genus is recognizedas being able to survive a range of stresses and hostile environ-ments, including those of extreme temperature (5–65 �C), pH(4.5–10.0) and high salt concentrations, enabling them to colonizea wide range of niches (Fisher and Phillips, 2009). C. perfringens isan anaerobic spore-forming bacterium that can withstand stressessuch as low oxidation–reduction potential (ORP). Therefore, it islikely to survive anaerobic manure digestion. Many Clostridiumspecies are commonly observed in anaerobic digesters, beingresponsible with other microorganisms for acetate production(Diekert and Wohlfarth, 1994). They can also produce lower fattyacids from acetate or ethanol when the concentration of hydrogenis high, thereby reversing the reactions of the syntropic bacteria.This may be an indicator of reactor instability (O’Flaherty et al.,2006). It is thus not surprising that high concentrations of Entero-coccus and Clostridium are observed in outlets of anaerobicdigesters.

Temperature and retention time are critical parameters for indi-cator and pathogen survival during anaerobic digestion of wastes(Arthurson, 2008; Sahlström, 2003). The elimination of E. coli andSalmonella was faster at 35 �C than at room temperature duringanaerobic digestion of cattle slurry (Kumar et al., 1999). S. typhimu-rium, Y. enterocolitica and L. monocytogenes also declined more rap-idly at 17 �C than at 4 �C during anaerobic digestion of thissubstrate (Kearney et al., 1993b). According to Olsen and Larsen(1987), increasing retention time in mesophilic digesters enhancedpathogen removal efficiencies toward those encountered for anaer-obic thermophilic treatment. However, Kearney (1991) reportedthat S. typhimurium had a T90 value of 2.1 days at a 25 day HRT,but after the HRT was reduced to 5 days the T90 value decreasedto 1.2 days. It was hypothesized that higher concentrations of VFAsat shorter HRTs would be correlated with a decline in viable patho-gens number. Since their results did not show significant correla-tion between VFAs and pathogen concentrations, they speculatedthat other factors such as the source and type of slurry, pH andtemperature of the anaerobic digestion process could also have asignificant effect on the pathogen populations. Regarding S.typhimurium removal, Gadre et al. (1986) reported a T90 of approx-imately 7.7 days during mesophilic (37 �C) anaerobic digestion ofcattle dung, while Olsen and Larsen (1987) found a T90 of 2.0 daysduring the digestion of pig slurry in similar conditions. Eventhough T90 were not specifically calculated during our study, re-sults suggest that T90 values of analyzed pathogens and indicatormicroorganisms were lower than 7 days under the conditions

encountered in the psychrophilic anaerobic digesters. Since tem-perature was lower in these reactors than in mesophilic ones,parameters other than temperature are probably responsible forthe elimination of pathogens.

Côté et al. (2006) studied the removal of pathogens and indica-tor microorganisms during PADSBRs of swine manure in 28 dayHRT laboratory-scale digesters (42 L) operated at 20 �C. At theend of the treatment, 47%, 75% and 100% of the collected sampleswere free of total coliforms, E. coli and Salmonella, respectively.They observed a diminution of 1.6–4.2 log10 units and of 2.5–4.2log10 units in the residual samples still having detectable concen-trations of total coliforms and E. coli, respectively. Farm-scale reac-tors operated in our study showed comparable performances withlab-scale regarding pathogen removal for a similar temperature, ata much lower HRT (7 days). This process has already proven suc-cessful for pig manure treatment in full-scale reactors, regardingbiogas production and COD removal (Masse et al., 2009). Resultsobtained in this study show that scaled-up PADSBR was not onlysuccessful in stabilizing the organic matter, but was also able to de-crease pathogen and indicator concentrations to undetectablelevels.

Lower performances regarding pathogen and indicator removalwere also observed for BR3 and BR4 when operated as non-con-ventional SBRs (Table 2). BR1 and BR2 removed more than2.5 log10 units of total and fecal coliforms, and E. coli, correspond-ing to more than 99.7% removal. Only 0.9–1.2 log10 unit of thesemicroorganisms was removed by BR3 and BR4, corresponding to86.1–94.0% removal. Y. enterocolitica removal averaged 2.1 log10

units for BR1 and BR2, and ranged from 1.1 to 1.3 log10 unit forthe other reactors. Salmonella and Campylobacter spp. removal byBR1 and BR2 ranged from 1.2 to 1.5 log10 unit, corresponding to94.1–96.6%, while their concentration decreased by 0.9–1.1 log10

unit in BR3 and BR4. Enterococcus spp. and C. perfringens concentra-tion remained high in all effluents, with average concentrations of5.6 � 104 and 2.7 � 105 CFU g�1, respectively. Since filling was per-formed simultaneously with drawing in BR3 and BR4, it washypothesized that there could be a short circuit in the hydraulicflow. Mass balances were performed to estimate which proportionof influent entering the reactor during filling was going out withthe supernatant overflow. VFAs were chosen as short circuitingindicators, because of their low concentration after digestion andof a more precise and reliable quantification method comparedto TCOD and bacteria. It appeared that an average of 6.42% (stan-dard deviation (SD) = 5.50) and 6.33% (SD = 6.29) of VFAs enteringthe system were directly going to the bioreactor’s outlet while fill-ing BR3 and BR4, respectively. Sporadically, for diverse operationalconstraints, BR3 and BR4 were fed after the drawing step, as forconventional SBRs, indicated by arrows in Fig. 1A and B. Moreover,BR3 operational conditions were switched toward those of a con-ventional SBR near the 500th day of this study (Fig. 1). Results ob-tained clearly show that the presence of VFAs in the bioreactoroutlet is due to short circuiting, since negligible VFA concentra-tions were observed in the outlet when filling and drawing wereperformed separately. The increase of TCOD in BR3 and BR4 efflu-ent (Fig. 1C and D) due to short circuiting was significant comparedto conventionally operated SBRs (BR1 and BR2; ca. 25,000 mgT-COD L�1). The presence of short circuits could lead to an incom-pletely stabilized digestate, causing odor problems andgreenhouse gas production during storage of the effluents fromthe bioreactors. The difference was also significant for pathogensand indicators.

Two hydraulic retention times (7 and 14 days) were tested inthis study, in BR3 and BR4, respectively. Even though these reac-tors showed higher TCOD and pathogen and indicator concentra-tions in effluents compared to BR1 and BR2, mainly because ofshort-circuiting of hydraulic flow, no significant impact was ob-

646 D. Massé et al. / Bioresource Technology 102 (2011) 641–646

served on pathogen and carbon removal when the treatment cyclelength was reduced from 2 to 1 week (P > 0.05). Moreover, as rep-resented by arrows in Fig. 1, a return to conventional SBR opera-tions for BR3 and BR4 allowed the elimination of hydraulic shortcircuiting and permitted the comparison of conventionally oper-ated SBRs under two HRTs for some cycles. Under these operationalconditions, similar pathogen and indicator removal efficiencies asfor BR1 and BR2 were reached for all tested organisms (Table 2),again showing that these full-scale reactors performed similarlywhen the treatment cycle length was reduced from 2 to 1 week.

4. Conclusion

Farm-scale psychrophilic anaerobic digestion in sequentialbatch reactors (PADSBRs), operated at 7 or 14 days HRT and24 �C, significantly decreased the concentration of total and fecalcoliforms, E. coli, Salmonella, Campylobacter spp., and Y. enterocoli-tica to undetectable levels in most samples. The concentration ofC. perfringens did not decrease significantly during treatment whileEnterococcus spp. concentration diminished by less than 1.0 log10

unit, still remaining above 105 CFU g�1. All these results were sim-ilar to those obtained from laboratory-scale PADSBRs, which were3000 to 10,000 times smaller in volume than the full-scale digest-ers. No significant impact was observed on pathogen and carbonremoval when the treatment cycle length was reduced from 2 to1 week. This allows the use of smaller reactors, significantly reduc-ing installation and operational costs of this particular process.

Acknowledgements

We thank Bio-Terre for providing access to on-farm full-scalebioreactors and for samples collection. We are also grateful to De-nis Deslauriers and Andrew Scott for their dedication to this pro-ject and for providing their technical expertise.

References

APHA, 1992. Standard methods for the examination of water and wastewater, 18thed. American Public Health Association.

Arthurson, V., 2008. Proper sanitization of sewage sludge: a critical issue for asustainable society. Applied and Environmental Microbiology 74 (17), 5267–5275.

Berg, G., Berman, D., 1980. Destruction by anaerobic mesophilic and thermophilicdigestion of viruses and indicator bacteria indigenous to domestic sludges.Applied and Environmental Microbiology 39 (2), 361–368.

Bicudo, J.R., Goyal, S.M., 2003. Pathogens and manure management systems: areview. Environmental Technology 24 (1), 115–130.

CCP, 2008. Statistics & Market Reports, Canadien Pork Concil. Ottawa, ON.Chauret, C., Springthorpe, S., Sattar, S., 1999. Fate of Cryptosporidium oocysts, Giardia

cysts, and microbial indicators during wastewater treatment and anaerobicsludge digestion. Canadian Journal of Microbiology 45 (3), 257–262.

Colleran, E., 2000. Hygienic and sanitation requirements in biogas plants treatinganimal manures or mixtures of manures and other organic wastes. In:Ørtenblad, H. (Ed.), Anaerobic Digestion: Making energy and solving modernwaste problems. Herning Municipal Authorities, Denmark, pp. 77–86. AD-NETT.

Côté, C., Massé, D.I., Quessy, S., 2006. Reduction of indicator and pathogenicmicroorganisms by psychrophilic anaerobic digestion in swine slurries.Bioresource Technology 97 (4), 686–691.

Diekert, G., Wohlfarth, G., 1994. Metabolism of homoacetogens. Antonie vanLeeuwenhoek 66 (1), 209–221.

Effenberger, M., Bachmaier, J., Garcs, G., Gronauer, A., Wilderer, P., Lebuhn, M., 2006.Mesophilic–thermophilic–mesophilic anaerobic digestion of liquid dairy cattlemanure. Water Science and Technology 53 (8), 253–261.

FAO, 2009. FAO Statistical Yearbook 2007–2008, Food and Agriculture Organizationof the United Nations.

Fisher, K., Phillips, C., 2009. The ecology, epidemiology and virulence ofEnterococcus. Microbiology 155 (6), 1749–1757.

Gadre, R.V., Ranade, D.R., Godbole, S.H., 1986. A note on survival of salmonellasduring anaerobic digestion of cattle dung. Journal of Applied Microbiology 60(2), 93–96.

Guan, T., Holley, R.A., 2003. Pathogen survival in swine manure environments andtransmission of human enteric illness – a review. Journal of EnvironmentalQuality 32 (2), 383–392.

Hill, V., Sobsey, M., 1998. Microbial indicator reductions in alternative treatmentsystems for swine wastewater. Water Science and Technology 38 (12), 119–122.

Hutchison, M.L., Walters, L.D., Avery, S.M., Munro, F., Moore, A., 2005. Analyses ofLivestock Production, Waste Storage, and Pathogen Levels and Prevalences inFarm Manures. Applied and Environmental Microbiology 71 (3), 1231–1236.

Kearney, T., 1991. Survival of pathogenic bacteria in anaerobic digesters. TheQueen’s University of Belfast, Belfast, UK.

Kearney, T.E., Larkin, M.J., Frost, J.P., Levett, P.N., 1993a. Survival of pathogenicbacteria during mesophilic anaerobic digestion of animal waste. Journal ofApplied Microbiology 75 (3), 215–219.

Kearney, T.E., Larkin, M.J., Levett, P.N., 1993b. The effect of slurry storage andanaerobic digestion on survival of pathogenic bacteria. Journal of AppliedMicrobiology 74 (1), 86–93.

Klein, G., 2003. Taxonomy, ecology and antibiotic resistance of enterococci fromfood and the gastro-intestinal tract. International Journal of Food Microbiology88 (2–3), 123–131.

Knechtel, J.R., 1978. A more economical method for the determination of chemicaloxygen demand. Water and Wastes Engineering 14 (4), 25–28.

Kumar, R., Gupta, M.K., Kanwar, S.S., 1999. Fate of bacterial pathogens in cattle dungslurry subjected to anaerobic digestion. World Journal of Microbiology andBiotechnology 15 (3), 335–338.

Line, J.E., 2001. Development of a selective differential agar for isolation andenumeration of Campylobacter spp.. Journal of Food Protection 64, 1711–1715.

Masse, D.I., Masse, L., Xia, Y., Gilbert, Y., 2009. Potential of low temperatureanaerobic digestion to address current environmental concerns on swineproduction. Journal of Animal Science 88 (13-electronic-suppl), E112–E120.

O’Flaherty, V., Collins, G., Mahony, T., 2006. The microbiology and biochemistry ofanaerobic bioreactors with relevance to domestic sewage treatment. Reviews inEnvironmental Science and Biotechnology 5 (1), 39–55.

Olsen, J.E., 1988. Studies on the reduction of pathogenic and indicator bacteria inliquid pig manure treated by sedimentation and anaerobic filter digestion formethane generation. Biological Wastes 24 (1), 17–26.

Olsen, J.E., Larsen, H.E., 1987. Bacterial decimation times in anaerobic digestions ofanimal slurries. Biological Wastes 21 (3), 153–168.

Pell, A.N., 1997. Manure and microbes: public and animal health problem? Journalof Dairy Science 80 (10), 2673–2681.

Sahlström, L., 2003. A review of survival of pathogenic bacteria in organic wasteused in biogas plants. Bioresource Technology 87 (2), 161–166.

Sharpley, A., Meisinger, J.J., Breeuwsma, A., Sims, J.T., Daniel, T.C., Schepers, J.S.,2002. Impacts of animal manure management on ground and surface waterquality. In: Hatfiels, J.L., Stewart, B.A. (Eds.), Animal Waste Utilization: EffectiveUse of Manure as A Soil Resource. CRC Press.

Smith, S.R., Lang, N.L., Cheung, K.H.M., Spanoudaki, K., 2005. Factors controllingpathogen destruction during anaerobic digestion of biowastes. WasteManagement 25 (4), 417–425.

Spencer, J.L., Guan, J., 2004. Public Health Implications Related to Spread ofPathogens in Manure From Livestock and Poultry Operations. In: Spencer, J.F.T.,Ragout de Spencer, A.L. (Eds.), Public Health Microbiology: Methods andProtocols. Humana Press, Totowa, NJ, pp. 503–515.

Termorshuizen, A.J., Volker, D., Blok, W.J., Brummeler, E.t., Hartog, B.J., Janse, J.D.,Knol, W., Wenneker, M., 2003. Survival of human and plant pathogens duringanaerobic mesophilic digestion of vegetable, fruit, and garden waste. EuropeanJournal of Soil Biology 39 (3), 165–171.

Weagant, S.D., 1983. Medium for presumptive identification of Yersiniaenterocolitica. Applied and Environmental Microbiology 45 (2), 472–473.