In-storage psychrophilic anaerobic digestion: acclimated microbial kinetics

Download In-storage psychrophilic anaerobic digestion: acclimated microbial kinetics

Post on 19-Mar-2017

212 views

Category:

Documents

0 download

TRANSCRIPT

  • This article was downloaded by: [University of Guelph]On: 14 November 2014, At: 02:24Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

    Environmental TechnologyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/tent20

    In-storage psychrophilic anaerobic digestion:acclimated microbial kineticsSusan King a , Pierre Courvoisier a , Serge Guiot b & Suzelle Barrington aa Department of Bioresource Engineering , Macdonald Campus of McGill University , 21 111Lakeshore, Ste Anne de Bellevue , QC , H9X 3V9 , Canadab Department of Environmental Bioengineering , Biotechnology Research Institute, NationalResearch Council of Canada , 6100 Royalmount, Montreal , QC , H4P 2R2 , CanadaAccepted author version posted online: 09 Dec 2011.Published online: 24 Jan 2012.

    To cite this article: Susan King , Pierre Courvoisier , Serge Guiot & Suzelle Barrington (2012) In-storagepsychrophilic anaerobic digestion: acclimated microbial kinetics, Environmental Technology, 33:15, 1763-1772, DOI:10.1080/09593330.2011.644867

    To link to this article: http://dx.doi.org/10.1080/09593330.2011.644867

    PLEASE SCROLL DOWN FOR ARTICLE

    Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions and views of the authors, andare not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon andshould be independently verified with primary sources of information. Taylor and Francis shall not be liable forany losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use ofthe Content.

    This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

    http://www.tandfonline.com/loi/tent20http://www.tandfonline.com/action/showCitFormats?doi=10.1080/09593330.2011.644867http://dx.doi.org/10.1080/09593330.2011.644867http://www.tandfonline.com/page/terms-and-conditionshttp://www.tandfonline.com/page/terms-and-conditions

  • Environmental TechnologyVol. 33, No. 15, August 2012, 17631772

    In-storage psychrophilic anaerobic digestion: acclimated microbial kinetics

    Susan Kinga, Pierre Courvoisiera, Serge Guiotb and Suzelle Barringtona

    aDepartment of Bioresource Engineering, Macdonald Campus of McGill University, 21 111 Lakeshore, Ste Anne de Bellevue QC H9X3V9 Canada; bDepartment of Environmental Bioengineering, Biotechnology Research Institute, National Research Council of Canada,

    6100 Royalmount, Montreal QC H4P 2R2 Canada

    (Received 20 June 2011; final version received 23 November 2011 )

    In-storage psychrophilic anaerobic digestion develops by microbial acclimation in covered swine-manure storage tanks,producingCH4 and stabilizing organicmatter. To optimize the systems performance, the process kineticsmust be understood.The objective of this study was to evaluate kinetic parameters describing the major stages in the digestion process, and toinvestigate the effect of temperature acclimation on these parameters. Specific activity tests were performed using manureinocula and five substrates at three incubation temperatures. Extant substrate activities were determined analytically foreach case, and intrinsic kinetic parameters for glucose uptake were estimated by grid search fitting to the Monod model.The results demonstrate that this acclimated microbial community exhibits different kinetic parameters to those of themesophilic communities currently modelled in the literature, with increased activity at low temperatures, varying withsubstrate and temperature. For glucose, the higher uptake is accompanied by lower microbial yield and half-saturationconstant. Decomposing these values suggests that active psychrophilic and mesophilic microbial populations co-exist withinthe community. This work also confirms that a new method of assessing microbial substrate kinetics must be developed formanure microbial communities, separating microbial mass from other suspended organics.

    Keywords: psychrophilic anaerobic digestion; specific substrate uptake;manure treatment;microbial kinetics; swinemanure

    IntroductionIn-storage psychrophilic anaerobic digestion (ISPAD) isa combined manure treatment and storage system devel-oped for Canadian pork producers [1]. The process occursin manure storage tanks with air-tight covers when theanaerobic microbial community in the manure has accli-mated to the ambient operating conditions; manure solidsare reduced by 24%, and 63% of the potential methane isreleased [2]. This performance could be optimized usingkinetic modelling, if the appropriate parameters are eval-uated. Current modelling of manure decomposition, usingMonod kinetics and Arrhenius temperature correction fac-tors, requires substrate uptake parameters for eachmicrobialpopulation [3,4].

    The substrate uptake kinetics of anaerobic microbialcommunities may be assessed using specific activity assays[5]. Substrates represent the major stages of anaerobicdigestion: glucose is the model substrate for acidogen-esis; propionate and butyrate are used for acetogenesis;acetate for aceticlastic methanogenesis; and H2/CO2 forhydrogenotrophic methanogenesis and homoacetogenesis[6]. Microbial communities acclimated to psychrophilicconditions exhibit increased substrate uptake at lower tem-peratures, compared with non-acclimated communities,but the increases are not uniform across substrates and

    Corresponding author. Email: Susan.King@mail.mcgill.ca

    temperatures [7]. For example, after cultivation of granularsludge for 306 days at 10 C, activities on acetate,propionate and butyrate increased by factors of 3.65, 1.45and 4.1, respectively at 10 C, and 2.44, 1.20 and 2.61respectively at 30 C [8]. Similarly, after operating at 15 Cfor 625 days, activities on acetate and butyrate increasedat both 15 and 37 C, whereas propionate activity remainedlow, and H2/CO2 activity increased continually throughoutthe experimental period [9]. Propionate activity appears tobe the most sensitive to temperature change and the slow-est to acclimate [10,11], while the highest acclimated VFAactivity is reported for butyrate [12].

    When a mesophilic anaerobic microbial communityacclimates to psychrophilic conditions, such as those occur-ring in ISPAD, for at least a year, the component popula-tions generally exhibit maximum substrate uptake at 35 C,which is taken to mean that they are still mesophilic [3,9].At the same time, the greater increases in activity at lowtemperatures indicate that these communities are psychro-active [13]. Occasionally a fully psychrophilic microbialpopulation with a temperature optimum near 15 C is foundwithin the community [14,15]. However, psychrophilic andmesophilic populations consuming a single substrate mayalso coexist in a single community [16,17], in which casethe substrate uptake data may exhibit a bi-modal form with

    ISSN 0959-3330 print/ISSN 1479-487X online 2012 Taylor & Francishttp://dx.doi.org/10.1080/09593330.2011.644867http://www.tandfonline.com

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • 1764 S. King et al.

    mesophilic optimum, composed of two superposed uptakecurves [18].

    Experimental substrate uptake datamay be fittedmathe-matically to amodel equation to estimate kinetic parameterssuch as the Monod half-saturation constant, Ks [19,20].The goodness of fit is used to compare the suitability ofdifferent relationship forms to the process represented bythe data [21]. Using this approach, a study of aceticlas-tic methanogenesis data found that both the Monod andHaldane models were accurate at 30 C, but the Haldanemodel produced a better fit between 6 and 22 C [22]. Asimilar study concluded that the Haldane model was pre-ferred at 22 C, while at 11 C both the Haldane and anon-competitive model fit the data equally well, proposingthat differences could be attributed to the representation ofinhibition in each model [23]. However, a key difficulty inassessing and comparing kinetic parameters from substrateuptake data is the evaluation of the size of the active micro-bial population involved [8]. In addition, the endogenousproduction of substrate must be accounted for when eval-uating the substrate uptake behaviour of manure microbialcommunities [24].

    The relationship between the maximum substrateuptake kinetic parameter, qmax, and temperature is usuallydescribed using the exponential Arrhenius equation [25].Accordingly, in a study using granular sludge adapted to10 C for 235 days, the temperature dependence of acetateconversion was well described by an Arrhenius model;however, propionate, butyrate and mixed volatile fattyacid (VFA) activities were better described by a square-root formulation [26]. Conversely, when digested sewagesludge was adapted to 20 C for four months, the tempera-ture dependence of the resulting methane production fromacetatewas poorly described by both theArrhenius andHal-dane equation forms; no better-fitting model was proposedin this study [23].

    These results illustrate that acclimated anaerobic com-munities may not always be treated as mesophilic, and thatparameters must be evaluated for each intermediate sub-strate. In addition, the assumptions of Monod kinetics andArrhenius temperature dependence may not apply. As such,the substrate uptake kinetics of an ISPAD community can-not be estimated from known data, and must be defined inthe laboratory. Therefore the objective of this study wasto define the kinetic parameters and temperature depen-dence relationships of the main microbial populations inan ISPADmicrobial community. To accomplish this, swinemanure inocula from a three-year-old ISPAD installationwere evaluatedusing specific substrate activity tests.Assayswere performed at three temperatures, 8, 18 and 35 C,using glucose, acetate, propionate, butyrate and H2/CO2 assubstrates. Inocula from a similar uncovered storage tank,and freshly produced manure were evaluated as controls.Extant substrate uptake activity and optimal temperaturewere defined for each population. Intrinsic activity wasestimated by curve fitting to the Monod model.

    Materials and methodsManure inoculaIn 2004, a full-scale ISPAD system was established in St.Francois Xavier Quebec, Canada. The facility used a con-crete tank, 30m in diameter by 3.66mdeep, covered withan air-tightmembrane (GTI, Fredericton, NB,Canada). Thetank received swinemanure on a regular basis. The contentswere removed for land-spreading twice yearly, except for adepth of 0.30.6m.Manure from this facility (Xf ) was usedto represent ISPAD in this study. The two controls, freshlyproduced manure (Xm) and one-year-old manure containedin an uncovered storage tank (Xo), were obtained from theswine research facility of theMcGill UniversityMacdonaldCampusExperimental Centre, located inMontreal, Quebec,Canada. Manures from the two operations were consideredcomparable in terms of solids and nutrients, as they are pro-duced by hogs fed a standard corn and soybean-based ration[27]. Samples of each manure were collected in June 2007as described previously [28].

    Manure characterizationSub-samples of all three manures (Xf , Xo and Xm) wereanalysed according to standard methods [29] to estab-lish: solids, chemical oxygen demand (total and soluble)and pH. The quantity of active microbial biomass ineach manure sample was estimated using the Luminultrawastewater ATP kit (Luminultra, NB, Canada) and a lumi-nometer (Sirius, model V3.2, Bethold Detection Systems,TN, USA).

    Specific substrate activity testsThree sets of substrate activity tests (SAT) were performed[30]: one using the microbial community contained in theISPAD manure as active biomass (Xf ), one using the com-munity in the uncovered tank manure (Xo) and the thirdusing the community in fresh manure (Xm). Each set com-prised three batches, one each at 8, 18 and 35 C. Each batchincluded five individual substrate assays: glucose, acetate,propionate and butyrate were the liquid substrates, andH2/CO2 was the gaseous substrate used. All combinationswere run in triplicate.

    For each batch, twelve 120mL bottles (for liquid sub-strates) and three 60mL bottles (for gaseous substrate) wereprepared. Manure inoculum was added to each bottle, toprovide 5 gVSS L1 for liquids and 2 gVSS L1 for thegaseous substrate. Phosphate buffer was added, bringingthe volume of liquid in each bottle to 20mL. The bot-tles were sealed, flushed with N2/CO2 gas (80%/20%)and placed in a shaker (New Brunswick Scientific, Edi-son, NJ, USA) operating at 100 rpm (400 rpm for gaseoussubstrates), in a thermostatically controlled environment, inthe dark. The bottles were allowed three or four days underthese conditions for acclimation to the assay conditions.

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • Environmental Technology 1765

    Following the acclimation period, liquid substrate wasinjected through the cap of each 120mL bottle, and the60mL bottles were flushed and pressurized to 140 kPa withH2/CO2. The first sample was immediately taken: 0.5mLof liquid (for liquid substrate bottles) or 100l of headspacegas (for H2/CO2 bottles). Liquid samples were centrifugedto remove solids. Sub-samples of supernatant were anal-ysed for glucose in the bottles with this substrate. For thebottles fed acetate, butyrate or propionate, sub-samples ofsupernatant were diluted fivefold for VFA analysis. Forthe H2/CO2 bottles, the headspace gas samples were anal-ysed immediately by gas chromatography. Sampling wasrepeated at regular intervals during each assay period. Atthe end of each assay, the bottle contents were analysed todetermine solids and pH.

    Analytical methodsGlucose was measured using high-pressure liquid chro-matography (Waters Chromatography Division, Milford,MA, USA) equipped with an injector (model 717+), photo-diode array detector (model 2996), pump (model 600) andrefractive indexdetector (model 2414). The columnused forthe separation was an ICSep IC ION-300 column (Transge-nomics, San Jose, CA, USA) of 300mm 7.8mm i.d. andan ion guard GC-801 column (Transgenomics). The mobilephase consisted of 0.035N H2SO4 at a pH of 4, flowing ata rate of 0.4mLmin1. The measurements were conductedusing a wavelength of 210 nm. Analysis was carried out at35 C.

    Acetic, propionic and butyric acid concentrations weremeasured using a gas chromatograph (Agilent, model 6890,Wilmington, DE, USA) equipped with a flame ionizationdetector of 0.2L. The samples were fortified at a ratio of1:1 (by volume) using an internal standard of iso-butyricacid dissolved in 6% formic acid. These were directlyinjected into a glass column of 1mm 2mm Carbopack C(60 to 80 mesh) coated with 0.3% Carbowax 20M and0.1% H3PO4. The column was held at 130 C for fourminutes, and helium as carrier gas was injected at a rateof 20mLmin1. The injector and the detector were bothmaintained at 200 C.

    To quantify hydrogen consumption, biogas composition(H2, N2 + O2, CH4, CO2) was measured using a gas chro-matograph (Hewlett Packard, 6890 Series, Wilmington,DE, USA) equipped with a thermal conductivity detec-tor and a 900mm 3mm 60/80 mesh Chromosorb 102column (Supelco, Bellefonte, PA, USA).

    ComputationsAll gas measurements were corrected to standard temper-ature and pressure of 0 C and 101.3 kPa, according to theideal gas law. For each bottle in each assay, substrate con-centration was plotted versus time. The maximum slopeof this curve within the initial period was defined as the

    Table 1. Initial periods for determination of extant substrateactivity from SAT assays.

    Initial period (hours)

    qa,bmax Y a,c 35 C 18 C 8 C

    Glucose 30 0.10 6 18 36Butyrate 20 0.06 14 45 91Propionate 13 0.04 32 104 209Acetate 8 0.05 42 136 272H2 35 0.06 8 26 52

    aBatstone et al. [3].bMaximum substrate uptake rate; COD COD1 d1.cMicrobial yield; COD COD1.

    extant substrate uptake rate. This rate, divided by the VSSconcentration in the bottle, gave the extant activity, qext. Incases where the curve had a suitable form, it was fitted toMonod equations, extracting intrinsic kinetic parameters,Ks, Y and qmax.

    Initial period definitionTo avoid microbial growth during a specific activity assay,samplingmust be restricted to thefirst fewhours of the assay[5,6,31]. Activity measured this way may be called extant,while activity during the growth phase is termed intrinsic[32,33]. To assess extant activities in this study, the initialperiod was defined as equal to one doubling time for eachpopulation, at the temperature of the assay.

    The doubling time values for 35 C were determinedbased on ADM1 model parameters [3]. The appropriatevalues for 18 and 8 C were extrapolated by applying theQ10 assumption to the maximum substrate uptake rateqmax(T) = qmax(35 C) 1.072(T35) [34]. The doubling timefor each temperature (T) is then d(T) = ln(2)/(qmax(T) Y ).The values of qmax(35 C) and Y and the calculated uptakerates and doubling times for each population are presentedin Table 1.

    Estimation of kinetic parametersKinetic parameters were estimated by fitting the substrateuptake data to the Monod-type equations for populationgrowth and substrate uptake [25]:

    x = xi + Y (Si S)dS/dt = (qmax S x)/(Ks + S)

    where x is the concentration of active microbial biomass ing L1, Y is the yield ofmicrobial biomass from the substratein g biomass (g substrate)1, S is the concentration of sub-strate in g L1, qmax is themaximum substrate uptake rate ing substrate (g biomass)1 d1, and Ks is the half-saturationconstant in g substrate L1. The subscript i indicates theinitial value of a parameter. Themethod of fittingwas devel-oped using the means of the triplicate glucose consumption

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • 1766 S. King et al.

    Table 2. Initial parameter estimates and boundary values forglucose consumption at 35 C.

    Minimum Starting value Maximum

    qa,bmax 38.5 39.8 165.9Y a,c 0.008 0.075 0.128Ka,ds 0.022 0.469 0.591xei 0.1 1.0 10.0

    aBatstone et al. [3].bMaximum substrate uptake rate; g glucose (g biomass)1 d1.cMicrobial yield; g biomass (g glucose)1.dHalf-saturation constant; g glucose L1.eBiomass concentration; % VSS.

    data for the uncovered tank biomass,Xo, which had a clearlydefined Monod-type form at each assay temperature. Oncethe method was perfected, it was applied with the samealgorithms, starting points, step sizes and assumptions tothe glucose consumption data sets for the fresh manure andthe ISPAD tank communities. For each optimization, thestarting and boundary values for the optimized parametersare presented in Table 2.

    The ADM1 literature values of Ks, Y and qmax for glu-cose consumers were used with the 35 C data to find astarting value for xi. Values for all four variables were thenoptimized using the grid search method, minimizing thesum of squares error (SSE) between the calculated valueand the experimental data. This method finds local minimabut can be trapped in a local minimum and miss the globalminimum; however, because the fitted valueswere expectedto be of the same order of magnitude as the ADM1 values,this was considered acceptable. The first iteration step sizewas 15%of the initial parameter value. For each subsequentiteration, the step size was halved. Because the data con-tained a small number of points, the algorithm tended tocome up with extreme values that fitted the data well, butonly reduced the SSE by negligible amounts. These valueswere also unrealistic in terms of the shape of the curve andthe difference from the starting values in the literature. Tocontrol this behaviour, boundary values were added to thealgorithm.Boundary values forqmax,Y andKs were selectedfrom those used by the ADM1 model [3], while xi was keptwithin one order of magnitude from the starting value. Theiterative process was then stopped when the change in SSEfrom the previous iteration was less than 5% of the previousSSE value.

    The optimized parameter values for the 35 C data werethen used as starting values for the 18 and 8 C data sets tooptimize Ks and qmax for each temperature, with the valueof Y and the ratio of xi to VSS kept constant throughout.

    Statistical analysisAll statistical analyses were performed using SAS9.0 (2009, SAS Institute inc., Cary NC, USA). The

    characterization data were collected in triplicate, onsub-samples from the composite sample taken from eachtank. Significance of differences between these measuredvalues was evaluated by the Student NewmanKeulsmethod in a simple analysis of variance based on a com-pletely randomized design. The specific substrate activityassays used a randomized complete block design, consid-ering microbial community type (Xf , Xo and Xm) as thetreatment factor, temperature (35, 18 and 8 C) as the blockfactor, with substrate uptake rate (qext) as the dependantvariable. Treatmentswere assigned randomly to experimen-tal units (bottles), and all treatmentblock combinationswere completed in triplicate.

    Results and discussionCharacteristics of manure inoculaTable 3 lists the characteristics of the three manures usedas sources of active biomass in the specific activity assays.The total solids concentrations illustrate differences inwatercontent: the ISPAD manure, Xf , and the uncovered tankmanure, Xo, are respectively 19% and 52% less concen-trated than the fresh manure, Xm. This is due to wash waterentering both storage tanks, as well as rain in the uncov-ered installation. Evidence of anaerobic digestion in theISPAD tank is seen in the volatile solids (VS), volatile dis-solved solids (VDS) and soluble chemical oxygen demand(sCOD) contents of Xf . These are, respectively, 66%, 5%and 4% of total solids (TS), reduced from 72%, 20% and32% in fresh manure. Hydrolysis of VSS to VDS withoutmethanogenesis is evident in the uncovered tank, where VSand sCOD are equal to those of fresh manure in terms ofpercentage of TS, while VDS has increased to 32%. Forall three manures, pH is within the normal range [35], andthe active microbial communities account for less than 2%

    Table 3. Characteristics of experimental manures.

    Fresh Uncovered ISPADmanure, Xm tank, Xo tank, Xf

    TS (g L1) 48.0 (0.3) 22.7 (0.2) 38.7 (0.4)VS (g L1) 34.3 (0.2) 16.5 (0.1) 25.4 (0.2)FS (g L1) 13.7 (0.1) 6.3 (0.1) 13.3 (0.2)VSS (g L1) 27.4 (1.2) 11.2 (0.6) 24.0 (0.2)VDS (g L1) 7.0 (0.7) 5.3 (0.4) 1.4 (0.2)pH 6.9 (0.1) 7.3 (0.1) 7.5 (0.1)Total COD

    (g gVS1)2.4 (n.d.) 2.1 (0.1) 2.0 (0.1)

    Soluble COD(g gVS1)

    0.9 (n.d.) 0.7 (0.01) 0.1 (0.02)

    ATP(g gVSS1)

    12.0 10.7 16.7

    Active biomass(% VSS)

    0.41.2 0.31.0 0.51.6

    Note: Values represent the average of three replicates; standarddeviations in brackets.FS = fixed solids, n.d. = not determined.

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • Environmental Technology 1767

    Table 4. Form of substrate uptake response curve for each SATassay.

    Fresh manure, Uncovered tank, ISPAD tank,Xm Xo Xf

    Glucose35 C E E E18 C E E E8 C L/E E EButyrate35 C A/E L L18 C A L L8 C A L LPropionate35 C A/E L L/E18 C A A E8 C A A LAcetate35 C A/E L/E L18 C A A L8 C A A LH2/CO235 C E E L18 C E E L/E8 C E L E

    E = exponential, L = linear, A = accumulating.

    of the VSS. The similar community size allows activitiesfor the three inocula, presented in terms of total VSS, to becompared within the study.

    Substrate activity test (SAT) data setsPlots of the 45 SAT data sets displayed three distinct forms:exponential, linear and accumulation. This categorization,summarized in Table 4, dictated the kinetic parametersthat could be extracted from each set. Exponential curves,such as those illustrated in Figure 1, allow the most com-plete analysis. The initial slope represents qext; in addition,these curves may be accurately fitted to the Monod equa-tions, which estimate qmax and Ks. Linear data, illustratedin Figure 2, defines qext, though qmax and Ks cannot beevaluated. A combined linear-exponential response, suchas that seen in Figure 3, allows the same determinations asthe linear type, because the mild exponential portion doesnot permit an acceptable fit to the Monod equations. Accu-mulation of a substrate is illustrated in Figure 4. Wherethis is the complete response, no activity can be calcu-lated, although some may be occurring concurrently withthe production of substrate from manure organic matter.However, in cases where a linear or exponential patternfollowed initial accumulation, qext could be determined.

    Extant substrate uptake activityThe extant substrate uptake activities, qext, for all substratebiomass combinations are presented in Table 5. The ISPAD

    Glu

    cose

    (m

    g L

    1)

    Figure 1. Exponential glucose uptake by uncovered tankbiomass, Xo. Data points represent the average of three replicates.Error bars represent 1 standard deviation.

    But

    yrat

    e (m

    g L

    1)

    Figure 2. Linear butyrate uptake by ISPAD tank biomass, Xf .Data points represent the average of three replicates. Error barsrepresent 1 standard deviation.

    Prop

    iona

    te (

    mg

    L1

    )

    Figure 3. Linearexponential propionate uptake by ISPAD tankbiomass, Xf . Data points represent the average of three replicates.Error bars represent 1 standard deviation.

    microbial community, Xf , is highly active on all substratestested and at all assay temperatures. This confirms its classi-fication as a fully acclimated balanced anaerobic microbialcommunity [2,7]. The fresh manure community, Xm, whichshowsmeasureable activity on glucose andH2/CO2 only, isnot acclimated to anaerobic conditions. The community inXo is more active than this, consuming butyrate at all assaytemperatures, as well as acetate and propionate at 35 C.Because butyrate activity is developed preferentially, fol-lowed by acetate and then propionate activity later in theacclimation process, the uncovered tank communitymay beclassified as semi-acclimated to anaerobic conditions [12].

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • 1768 S. King et al.

    Ace

    tate

    (m

    g L

    1)

    Figure 4. Accumulating and accumulating-linear acetate uptakeby fresh manure biomass, Xm. Data points represent the averageof three replicates. Error bars represent 1 standard deviation.

    Table 5. Extant substrate activities, qext, from SAT assays.

    Fresh manure, Uncovered tank, ISPAD tank,Xm Xo Xf

    Glucose35 C 55 (68) 307 (243) 128 (100)18 C 36 (23) 87 (12)8 C 12 (1) 10 (3) 7 (7)Butyrate35 C 18 (29) 204 (37)18 C 81 (1) 105 (23)8 C 12 (8) 10 (2)Propionate35 C 24 (35) 32 (10)18 C 7 (9)8 C 12 (5)Acetate35 C 83 (254) 316 (135)18 C 28 (9)8 C 47 (10)H2/CO235 C 14 (7) 131 (225)18 C 8 (1) 59 (24)8 C 52 (5) 1 (0.3)

    Note: Values represent the average of three replicates; standarddeviations in bracketsqext; mg substrate (gVSS)1 d1, = no measurable activity.

    As reported by other researchers, the variations betweencommunities, substrates and temperatures are not uni-form, and the values tend to support the mesophilic opti-mum assumption, with maximum activities at 35 C. TheH2/CO2 activity of Xo, which increases with decreasingtemperature, is the exception. In this case, the test pro-cedure may have been a limiting factor, or it may be afurther indication of acclimation occurring in the uncov-ered tank. Whereas homoacetogens were reported to havea larger role in low-temperature digestion than in themesophilic process [36], it has been proposedmore recentlythat homoacetogen activity may be highest during the

    Table 6. Fitted kinetic parameters, biomass concentration anderror for glucose consumption.

    Fresh manure, Uncovered tank, ISPAD tank,Xm Xo Xf

    qamax35 C 39.12 39.12 51.8418 C 10.01 12.02 31.928 C 3.10 4.47 5.90Y b35 C 0.128 0.128 0.08218 C 0.128 0.128 0.0828 C 0.128 0.128 0.082Kcs35 C 0.238 0.022 0.02218 C 0.776 0.574 0.0298 C 0.679 0.469 0.029xdi35 C 7.16 5.26 7.4118 C 7.16 5.26 7.418 C 7.16 5.26 7.41e2e35 C 0.021 0.433 0.07618 C 0.036 0.192 0.0038 C 0.013 0.049 0.003

    Fitting done using the average of three replicates.aMaximum substrate uptake rate; g glucose (g biomass)1 d1.bMicrobial yield; g biomass (g glucose)1.cHalf-saturation constant; g glucose L1.dBiomass concentration; mg biomass L1.eSum of squares error (goodness of fit); (g glucose L1)2

    acclimation process and be reduced in a fully acclimatedpsychro-active community [7].

    Monod kinetic parametersSince the Monod equations may only be fitted to data setswith an exponential form, the glucose SAT data sets wereanalyzed in this way. Table 6 presents the resulting fittedkinetic parameters qmax, Y and Ks, as well as the populationsize estimate, xi, for glucose consumption by each of thethree studied microbial communities. The population sizes,representing 0.00.05% VSS, are much smaller than thosefrom the ATP analysis (Table 3), which is to be expectedsince only a small number of the microbial species presentare expected to be glucose consumers. However, becausethe percentage VSS values obtained are similar for the threecommunities, the resulting activity valuesmay be comparedon this basis.

    For Xo and Xm, the maximum substrate uptake rate,qmax, at 35 C remains equal to the literature value pre-sented in Table 2, at 39.12 g substrate (g biomass)1 d1,while Xf exhibits a 33% increase to 51.84 g substrate(g biomass)1 d1. This increase in the processing effi-ciency of the glucose-consuming population within Xfillustrates acclimation to the substrate. This is furtherdemonstrated at 18 and 8 C, where Xf shows increases

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • Environmental Technology 1769

    of 220% and 90%, respectively, in comparison with thevalues calculated for Xm at these temperatures. Thesegreater increases at lower temperatures indicate acclima-tion to the psychrophilic operating temperature of ISPAD,as reported by other researchers [8,9]. In contrast, qmax forthe semi-acclimatedXo remains equal to that ofXm at 35 C,showing modest increases of 20% and 44% at 18 and 8 C,respectively.

    The microbial yield, Y , exhibits a similar pattern,with the Xf value 35% lower than that of Xo and Xm,which remain identical. This indicates a change in cellularmetabolic processing, as the glucose-consuming populationin Xf produces less microbial biomass from each gram ofsubstrate consumed, while concurrently consuming moresubstrate. The excess consumed substrate, then, is convertedto intermediate products in the anaerobic digestion process.This contributes eventually to the increase in methane pro-duction previously reported for the ISPAD biomass [2].Because Y is expected to resist change for any given species[37], this effect may represent a change in the predominanceof species making up the glucose-consuming population asit acclimates to the ISPAD operating conditions. A follow-up molecular biology investigation of the ISPADmicrobialcommunity is planned, to explore this possibility.

    The apparent half-saturation constant, Ks, indicates theconcentration at which the population is able to process thesubstrate at half its maximum rate, which represents theaffinity of the microbial population for the substrate. Theimplication is that, at concentrations lower thanKs, process-ing is slow and inefficient and may result in accumulation.Continuing the pattern seen with qmax and Y , the Ks val-ues obtained for Xm remain similar to the original literaturevalues representing a standard mesophilic population. Thevalues of Ks representing Xf are 9096% lower, showingthat the glucose-consuming population in Xf is able to effi-ciently utilize substrate even at concentrations a full orderof magnitude lower than the standard mesophilic popula-tion. This indicates that the ISPAD microbial communityin Xf is acclimated not only to temperature and substratebut also to the low substrate concentrations which occur inthe lightly fed ISPAD tank. The uncovered tank commu-nity in Xo, with a low Ks comparable to that of Xf at 35 C,illustrates the beginning of the acclimation process, whilethe higher values similar to Xm at 18 and 8 C indicate thatthe community in Xo remains only semi-acclimated andprimarily mesophilic.

    Temperature dependenceFigure 5 illustrates the relationship between the fitted valuesof qmax and temperature. These points describe an expo-nential Arrhenius-type curve for Xo and Xm, as expectedfor mesophilic populations. In contrast, for Xf the curveis reversed. This suggests the coexistence of psychrophilicand mesophilic populations [18]. Assuming that the qmaxvalues for Xm represent an average mesophilic population,

    Figure 5. Maximumglucose uptake rates, qmax, for freshmanurebiomass, Xm, uncovered tank biomass, Xo, and ISPAD tankbiomass, Xf . Data points represent the average of three replicates.Data are derived from curve fitting. See Table 6 for error data forthe fitting process.

    these values were subtracted from the qmax values of Xfto produce psychrophilic values. These are illustrated inFigure 6, and represent the effect of the temperature accli-mation process on the ISPAD microbial community. Withthe clear peak near 18 C, these psychrophilic qmax valuesfor Xf illustrate the development of a true psychrophilicpopulation coexisting with the original mesophilic popu-lation. This concept is corroborated by recent reports ofsimilar results [16,17]. Assessing qmax at other intermediatetemperatures could further define the relationship.

    Extant activity interpretationModest activity in the initial period of an SAT assay issometimes dismissed as a lag period, caused by microbialacclimatization to the test conditions; however, these SATbottles were allowed three to four days for acclimatizationprior to the injection of specific substrates. This period didnot constitute starvation conditions, because of the pres-ence of manure organic matter. Strong activity immediatelyfollowing substrate injection was evident in many bottles,confirming that acclimatization had occurred. Therefore,any lag due to acclimatization had occurred prior to sub-strate injection, and activity during the initial period afterinjection was considered to represent the original biomass.Increasing activity following the initial period was assumedto represent microbial growth due to substrate excess in thebottle.

    The extant activities reported in Table 5 exhibit a widevariety of standard deviations. This was understood asa compound error, because substrate concentration, timeelapsed and VSS concentration are all measured factors

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • 1770 S. King et al.

    Figure 6. Maximum glucose uptake rates, qmax, for possiblecoexisting mesophilic and psychrophilic populations within theISPAD tank biomass, Xf . Data points represent the average ofthree replicates. Data are derived from curve fitting. See Table 6for error data for the fitting process.

    exposed to experimental error, which combine to deter-mine each specific substrate activity. Thus the errors, insome cases, may be additive and, in others, may be subtrac-tive, resulting in the wide variety of deviations reported.Despite this variation, the data illustrate trends that havebeen observed by other researchers, which strengthen theirvalidity. In fact, substrate uptake activities are known to behighly variable, as many related parameters used by ADM1are reported to have up to 300% variability [3].

    Monod kinetics interpretationThe extant substrate uptake rate may, in cases where theSAT response form was linear, be equivalent to the Monodmaximum uptake rate. However, further assays using dif-ferent initial substrate concentrations would be required toconfirm this. Therefore, the Monod parameter was onlyassessed for those assays demonstrating an exponentialuptake response. Likewise, assessment of the Monod half-saturation constant,Ks, from SAT data requires that the rateof substrate uptake varies during the assay. As a result, thisparameter could only be estimated for the data sets withan exponential response pattern, as part of the curve-fittingprocedure.

    The SAT assay protocol is designed for specificbiomass-substrate ratios. The non-exponential responsesseen in this study may have been caused by variations inthis ratio, caused by both the presence of extremely smallanaerobic microbial populations in this biomass and thepresence of manure organic matter in the bottles, the degra-dation of which may produce VFAs and mask concurrentconsumption by the appropriate population. Itmust be noted

    that the assay results for the ISPAD biomass were moreaccurate in many cases than those from the two controlsamples, owing to the minimal concentration of dissolvedsolids in the ISPAD-treated manure.

    Both accumulation andminimal activity responses wereseen primarily in the data from assays using fresh manure.Unlike the dairy manure used in other anaerobic digestionstudies, this seed community is not expected to harboura fully developed anaerobic population; pigs are non-ruminants and have a dietary transit time of approximatelyone day, which is less than the doubling time of some keyanaerobic populations [38]. Therefore these response typesare understood to be due to the exceptionally small popula-tion, as well as the presence of other organic substrates inthe manure.

    ISPAD kineticsThe kinetic impact of acclimation may be inferred by com-paring the ISPAD qext values in Table 5 to the mesophilicqmax values reported in Table 1. This can only be doneby ranking the activities in each case, owing to both thepopulation size and experimental uncertainties. Mesophiliccommunities are most active on hydrogen, followed by glu-cose, butyrate propionate and acetate in decreasing orderof magnitude. The ISPAD community, however, is mostactive on acetate, followed by butyrate, hydrogen, glucoseand propionate, at 35 C; the rankings are led by butyrateat 18 C, and glucose at 8 C, with variations in the fol-lowing rankings as well. This comparison confirms that theimpact of acclimation on the kinetic parameters varies withsubstrate and temperature.

    TheMonod kinetic parameters estimated for the ISPADglucose-consuming population reveal a higher qmax, lowerY and lower Ks than a corresponding mesophilic popu-lation, which illustrates acclimation to the psychrophilicoperating temperature as well as the low substrate concen-trations found in the ISPAD tank. The fitting process usedin the study is considered to be accurate because, for the twocontrol inocula, the algorithm selected parameters compa-rable to the mesophilic literature data recommended by theADM1.

    Decomposing the Monod uptake rates shows develop-ment of coexisting mesophilic and psychrophilic glucoseconsumers,which is an exciting development and confirma-tion of the latest understanding of temperature acclimation.Further study will be undertaken to see if this effect appliesto other substrates, and to refine estimates of the associatedparameters. In addition, this poses new questions on thefactors controlling the interaction of two such populations,and the requirements for modelling such processes.

    Follow-upTo evaluate the remaining ISPAD kinetic parameters, anewmanure-SATprotocolmust be developed. The standard

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • Environmental Technology 1771

    SAT assay requires an optimal ratio of substrate to biomass,which cannot be obtained owing to the small proportionof biomass in the manure and the confounding substrateconcentrations from the manure organic matter. This mustbe overcome in developing the manure-SAT, which willinclude measurement of the biomass concentration usingmolecular biology methods and separation of the biomassfrom all or part of the manure organic matter. A more elab-orate series of control assays could be used to separateproduction and consumption, or tracer studies could be used[24].

    The accurate kinetic parameters resulting from use ofthe m-SAT could then be used in modelling ISPAD. How-ever, because psychrophilic and mesophilic populations,consuming a single substrate, coexist in ISPAD, currentmodels would have to bemodified to include this condition.One approach tomodellingmicrobial diversity that could beappropriate, using multiple microbial populations for eachsubstrate, has recently been applied to the ADM1 [39].

    ConclusionsThe objective of this study was to measure the substrateuptake activities of the main microbial populations in anISPAD microbial community and to define from thesethe kinetic parameters required for modelling the pro-cess. Extant substrate uptake rates were determined forthe ISPAD biomass on five substrates at three tempera-tures. Half-saturation constants, maximum substrate uptakerate and microbial yield were estimated by mathematicalcurve-fitting for glucose at all three temperatures. The studyconcluded that:

    1. The ISPAD biomass is described by kinetic param-eters that are significantly different from those ofan equivalentmesophilic community. These changesvary with substrate and temperature.

    2. For glucose consumption, the fitted higher maxi-mum uptake rate, lower microbial yield and lowerhalf-saturation constant illustrate acclimation to boththe psychrophilic operating temperature and lowsubstrate concentrations found in the ISPAD tank.

    3. The glucose uptake rates can be decomposed toshow the development of coexisting psychrophilicand mesophilic populations consuming glucose.

    4. To accurately evaluate kinetic parameters of manuremicrobial communities for mathematical modellingrequires development of a new substrate uptakeassay protocol, which evaluates the size of themicro-bial community and removes the remaining manureorganic matter.

    AcknowledgementsThe authors wish to acknowledge the financial contributions of theNatural Science and Engineering Research Council of Canada,

    GTI of Fredericton, New Brunswick, Canada, and the Biotech-nology Research Institute of the National Research Council ofCanada, as well as the personal contributions of Luis Ortega,Helene Leblanc, Grant Clark and Dominic Frigon.

    References[1] J. Abou Nohra, S. Barrington, J.C. Frigon, and S.R. Guiot,

    In storage psychrophilic anaerobic digestion of swine slurry,Resour. Conserv. Recycl. 38 (2003), pp. 2337.

    [2] S.M. King, S. Barrington, and S.R. Guiot, In-storage psy-chrophilic anaerobic digestion: Acclimation of themicrobialcommunity, Biomass Bioenergy 35 (2011), pp. 37193726.

    [3] D.J. Batstone, J. Keller, I. Angelidaki, S.V. Kalyuzhnyi, S.G.Pavlostathis,A.Rozzi,W.T.M. Sanders,H. Siegrist, andV.A.Vavilin, Anaerobic Digestion Model No. 1 (ADM1), IWAScientific and Technical Report No. 13, IWA, London, 2002.

    [4] A. Gal, T. Benabdallah, S. Astals, and J. Mata-Alvarez,Modified versionofADM1model for agro-waste application,Bioresour. Technol. 100 (2009), pp. 27832790.

    [5] E. Colleran, F. Concannon, T. Golden, F. Geoghegan, B.Crumlish, E. Killilea, M. Henry, and J.D. Coates, Useof methanogenic activity tests to characterize anaerobicsludges, screen for anaerobic biodegradability and deter-mine toxicity thresholds against individual anaerobic trophicgroups and species, Water Sci. Technol. 25 (7) (1992),pp. 3140.

    [6] M. Soto, R.Mendez, and J.M. Lema,Methanogenic and non-methanogenic activity tests. Theoretical basis and experi-mental set up, Water Res. 27 (1993), pp. 13611376.

    [7] O.R. Kotsyurbenko, Trophic interactions in themethanogenic microbial community of low-temperature ter-restrial ecosystems, FEMS Microbiol. Ecol. 53 (2005),pp. 313.

    [8] S. Rebac, S. Gerbens, P. Lens, J.B. van Lier, A.J.M. Stams,K.J. Keesman, and G. Lettinga, Kinetics of fatty aciddegradation by psychrophilically grown anaerobic granularsludge, Bioresour. Technol. 69 (1999), pp. 241248.

    [9] S. Connaughton,G.Collins, andV.OFlaherty,Developmentof microbial community structure and actvity in a high-rate anaerobic bioreactor at 18 C, Water Res. 40 (2006),pp. 10091017.

    [10] S. McHugh, M. Carton, G. Collins, and V. OFlaherty, Reac-tor performance and microbial community dynamics duringanaerobic biological treatment of wastewaters at 1637 C,FEMS Microbiol. Ecol. 48 (2004), pp. 369378.

    [11] Z. Arbeli, A. Brenner, and A. Abeliovich, Treatment of high-strength dairy wastewater in an anaerobic deep reservoir:Analysis of the methanogenic fermentation pathway and therate-limiting step, Water Res. 40 (2006), pp. 36533659.

    [12] A.N. Nozhevnikova, O.R. Kotsyurbenko, S. Rebak, S.N.Parshina, C.Holliger, andG. Lettinga,Anaerobic productionand degradation of volatile fatty acids in low temperatureenvironments,Water Sci. Technol. 41 (12) (2000), pp. 3946.

    [13] S. McHugh, G. Collins, and V. OFlaherty, Long-term, high-rate anaerobic biological treatment of whey wastewaters atpsychrophilic temperatures, Bioresour. Technol. 97 (2006),pp. 16691678.

    [14] G. Akila, and T.S. Chandra, Performance of an UASBreactor treating synthetic wastewater at low-temperatureusing cold-adapted seed slurry, Process Biochem. 42 (2007),pp. 466471.

    [15] R.M. McKeown, C. Scully, T. Mahony, G. Collins, andV. OFlaherty, Long-term (1243 days), low-temperature(415 C), anaerobic biotreatment of acidified wastewaters:

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

  • 1772 S. King et al.

    Bioprocess performance and physiological characteristics,Water Res. 43 (2009), pp. 16111620.

    [16] R.M. McKeown, C. Scully, A.M. Enright, F.A. Chinalia,C. Lee, T. Mahony, G. Collins, and V. OFlaherty, Psy-chrophilic methanogenic community development duringlong-term cultivation of anaerobic granular biofilms, ISMEJ. 3 (2009), pp. 12311242.

    [17] A.N. Nozhevnikova, K. Zepp, F. Vazquez, A.J.B. Zehnder,and C. Holliger, Evidence for the existence of psychrophilicmethanogenic communities in anoxic sediments of deeplakes, Appl. Environ. Microbiol. 69 (2003), pp. 18321835.

    [18] N.S. Panikov, Population dynamics in natural habitats, inMicrobialGrowthKinetics, Chapman&Hall, London, 1995,pp. 275288.

    [19] C.T. Goudar, J.R. Sonnad, and R.G. Duggleby, Param-eter estimation using a direct solution of the integratedMichaelis-Menten equation, Biochim.Biophys.ActaProteinStruct. Mol. Enzymol. 1429 (1999), pp. 377383.

    [20] J.A. Robinson and J.M. Tiedje, Nonlinear estimation ofMonod growth kinetic parameters from a single substratedepletion curve, Appl. Environ. Microbiol. 45 (1983),pp. 14531458.

    [21] P. Bhunia and M.M. Ghangrekar, Analysis, evaluation,and optimization of kinetic parameters for performanceappraisal and design of UASB reactors, Bioresour. Technol.99 (2008), pp. 21322140.

    [22] L.Y. Lokshina, V.A. Vavilin, R.H. Kettunen, J.A. Rintala,C. Holliger, and A.N. Nozhevnikova, Evaluation of kineticcoefficients using integrated monod and haldane models forlow-temperature acetoclastic methanogenesis, Water Res.35 (2001), pp. 29132922.

    [23] R.H. Kettunen and J.A. Rintala, The effect of low tem-perature (529 C) and adaptation on the methanogenicactivity of biomass, Appl. Microbiol. Biotechnol. 48 (1997),pp. 570576.

    [24] H.B. Nielsen, Z. Mladenovska, and B.K. Ahring, Kinet-ics of propionate conversion in anaerobic continuouslystirred tank reactors, Environ. Technol. 29 (2008),pp. 151160.

    [25] B.E. Rittmann and P.L. McCarty, Environmental Biotech-nology: Principles and Applications, McGraw-Hill, Boston,MA, USA, 2000.

    [26] S. Rebac, J. Ruskova, S. Gerbens, J.B. van Lier, A.J.M.Stams, and G. Lettinga, High-rate anaerobic treatmentof wastewater under psychrophilic conditions, J. Ferment.Bioeng. 80 (1995), pp. 499506.

    [27] K.L. Conn, E. Topp, and G. Lazarovits, Factors influencingthe concentration of volatile fatty acids, ammonia, and othernutrients in stored liquid pig manure, J. Environ. Qual. 36(2007), pp. 440447.

    [28] S.M. King, S. Barrington, and S.R. Guiot, In-storage psy-chrophilic anaerobic digestion: Acclimation of themicrobialcommunity, Biomass Bioenergy 35(8) (2011), pp. 37193726.

    [29] APHA, AWWA, andWEF, Standard Methods for the Exam-ination of Water and Wastewater, revised ed., AmericanPublic Health Association, Washington, DC, 2005.

    [30] S.R. Guiot, B. Safi, J.C. Frigon, P. Mercier, C. Mulligan, R.Tremblay, andR. Samson,Performances of a full-scale novelmultiplate anaerobic reactor treating cheese whey effluent,Biotechnol. Bioeng. 45 (1995), pp. 398405.

    [31] J.D. Coates, M.F. Coughlan, and E. Colleran, Simple methodfor the measurement of the hydrogenotrophic methanogenicactivity of anaerobic sludges, J. Microbiol. Methods 26(1996), pp. 237246.

    [32] C.P.L. Grady, B.F. Smets, and D.S. Barbeau, Variability inkinetic parameter estimates: A review of possible causes anda proposed terminology, Water Res. 30 (1996), pp. 742748.

    [33] K. Kovarova-Kovar and T. Egli, Growth kinetics of sus-pended microbial cells: From single-substrate-controlledgrowth to mixed-substrate kinetics, Microbiol. Mol. Biol.Rev. 62 (1998), pp. 646666.

    [34] R.C. Loehr, Pollution Control for Agriculture, AcademicPress, Orlando, FL, USA, 1984.

    [35] ASAE, Manure production and characteristics, ASAED384.2 MAR2005, American Society of Agricultural Engi-neers, St Joseph, MI, USA, 2005.

    [36] V.A.Vavilin, L.Y.Lokshina, S.V.Rytov,O.R.Kotsyurbenko,and A.N. Nozhevnikova, Description of two-step kinet-ics in methane formation during psychrophilic H2/CO2and mesophilic glucose conversions, Bioresour. Technol. 71(2000), pp. 195209.

    [37] D.B. Nedwell, Effect of low temperature on microbialgrowth: Lowered affinity for substrates limits growth at lowtemperatures, FEMS Microbiol. Ecol. (1999), pp. 101111.

    [38] B.E. Rittmann, and P.L. McCarty, Environmental Technol-ogy: Principles and Applications, McGraw-Hill, Boston,MA, USA, 2001.

    [39] I. Ramirez, E.I.P. Volcke, R. Rajinikanth, and J.P. Steyer,Modelingmicrobial diversity in anaerobic digestion throughan extended ADM1 model, Water Res. 43 (2009), pp. 27872800.

    Dow

    nloa

    ded

    by [

    Uni

    vers

    ity o

    f G

    uelp

    h] a

    t 02:

    24 1

    4 N

    ovem

    ber

    2014

Recommended

View more >