rAQUACULTURE- A BIOTECHNOlOGYIN PROGRESS.1989N.De Pauw, E. Jaspers, H. Ackefors, N. Wilkins (Eds)European Aquaculture Society, Bredene, Belgium

Practical basis of nitrification in aquaculturewaste-water

P. Poquillon and J. Petit

Laboratoire de Physiologie des Poissons, INRA

Campus de Beaulieu, F-35042 Rennes Cédex, France

Abstract

Nitrification is an importanl process in the chain of waler Ireatmenl procedures required 10operale a recyclingunit for fish culIure. Experimenls were performed with reconstituled fish breeding effluent on an experimental

submerged fixed-bed reactor under ~ressure. The filter medium consisted of a calcinid clay (Biogrog, 1.2-1.5mm)with a specific surface area (106m .m-1. The treatment efficiency reached 9.6kgN-NH4.m'3.ól al 20'C and4.9kgN-NH4.m-3.ól al 13'C forhydraulic load of 644m3.m-2.ól. Reduction of the ammonia amounl ceased when!he dissolved oxygen concentration was equa! or inferior 103mg02J 1.The nitrification's natural slarting up phaselasted 42 days. This process could be reduced 1024 days by injection of commerciallyophilized bacteria into thefiller.

KEYWORDS: Nitrification, Wastewater, Water quality, Purification.

Introductionii Fish culture in recycled water can considerably

I

reduce water requirements and possiblycalories. It is especially suited for accelerating

! the hatching of eggs and the growth of fry,I breeding for warm-water fish (Anguilla, tilapia,I etc.), and for stocking or breeding fish in towns,

to replace natural spawning.

Nitrification is an important component of thewater treatment process of a recycling unit forfish breeding. Through nitrification, toxicammonia is oxidized to nitrate which is less toxicfor fish. Bacterial nitrification is a biologicalphenomenonof which the kineticsdepend onmany factors and particularly on availablesubstrates. A number of authors (Roques et al.,

1976; Grasmick et al., 1979; Roques et al., 1982;Gonenc and Harremoes, 1985) have proposedmodels of its kinetics. We find that Monod model(Sharma and Ahlert, 1977) is a special case ofsuch modeis. (Fig. 1). On the basis of thisformula, Faup et al. (1982) found that theammonia concentration of the effluent had amuch altered Ks value (1.5mgN-NH4.r\ Thiswas true with industrial waste water. But inaquaculture waste waterthe amount of ammoniawas not very much altered. Many authors(Csavas and Varadi, 1981; Kujal, 1981; 'Scottand Allard, 1984; Wickins, 1985) studiedbacterial nitrification in aquaculture situations butsome did not make allowance for all theparameters which modify nitrification kinetics, in

dSn Sn= Yn X ~ax X Xn X

dt Ks+Sn

Fig. 1. Monoe! equation. (S) Substrate concenlration, massIvolume; (I) time; (Y) yield coefficienl, mass oforganism grownlmass of subslrale ulilized; (J1max)maximum specific growth rate, time'\ (X)concentration of microorganisms, masslvolume; (Ks) Monoe! half velocity constant, massIvolume; (n)a specific type of microorganism. (Sharma and Ahlert, 1977).

j

1116

particu/ar the amount of dissolved oxygen(Sharma and Ahlert, 1977; Stenstrom andPoduska, 1980), pH (Martin, 1979; Quinlan,1984), temperature, BOD (paolini and Variali,1982) and phosphate levels.

Our approach consisted in studying biologicaldepuration in conditions as near as possible toaquaculture. A purification technique, then afilter medium were chosen, and experimentswere carried out to determine the treatmentefficiency of the filter medium, the oxygenrequirement and the means of commencingnitrification in an environment comparable to theexploitation environment.

Materlals and methods

Cholce of a purification technique

Among differenttechniques available, and whichare utilized in fish-farm waste-water treatment,the pressurized and submerged fixed bedreactor is, for a given nitrification efficiency, morecompact than trickling filters or a rotating discsystem. lts disadvantage is that the oxygennecessary for nitrification is brought on by waterand not by air as in the other process. We chosethe technique of pressurized and submergedfixed bed reactor in our studies.

Cholce of the filter medium

'n the literature, the best media (measured bykgN-NH4+NH3 epurated.m-3.d-1) for nitrificationon submerged filters were biolite (Faup et al.,1982), then pouzzolane (Richard et al., 1979),marble (Kowalski and Lewandowski, 1983),

Photo 1. Biogrog grain, 1mm;(x32).

~

P. Poquillon et al.

gravel (Kowalski and Lewandowski, 1983;Wickins, 1985), foam polystyrene (Kujal, 1981)and plastic (Boller and Gujer, 1986; Hall, 1986):

We chose Biogrog, manufactured by the SociétéArgiles et Minéraux AGS (Clerac, 17270Montguyon, France). Biogrog is a calcined clay.It is made by cooking and milling kaolinite. It isvery resistant to abrasion and does notdisaggregate after filter washing. It does notneed to be replaced or regenerated.

lts physical characteristics are:

granulometry: 1.2-1.5mmapparent density:1.5t.m-3porosity:0.3-0.4m3.m -3specific area: 106m2.m-3

Biogrog grains (Photo 1) are variabie. Their areais uneven and there are many holes on thesurface; th is explains the importance of thespecific area. Nitrification bacteria (Photo 2)settled in the holes and fracture area of theBiogrog grains.

Experi ments

Pilot unit

Experiments were carried out on a pilot unit (Fig.2) consisting of a biological filter (height 2m,diameter 160mm) with seven sampling points,an oxygenator, a thermoregulation unit and a250 I tank. The height of the Biogrog filterbedwas 1.5m. The flow in the biofilter can beascending or descending.

Photo 2. Nitrificationbacteria, 1J1in;(x10 000).

.-i

,...-

Nitrification in aquaculture waste-water

rCD

~

1117

*-J0 0

Fig. 2. Diagram of pilot unit. (1) Biological filter (2m height, 100mm diameter); (2) three-ways valve; (3) washingwater and air inlet; (4) oxygen inlet; (5) oxygenator; (6) pump and bypass; (7) 250 I tank; (8)thermoregulation unit.

The temperature was controlled, and 02, pH,NH4, ND2 were measured.

Experimental conditions were as follows:

- temperature 18:t1'C

- pH 7-7.2 UpH

- dissolved oxygen >5mg02.r1

Reconstituted fish breeding effluent

In orderto limitthe parameters which might affectthe experiments, we worked with a reconstitutedeffluent having a constant composition. deKinkelin et al. (1986) reported that troutnitrogenous excreta consist of 80% ammoniaand 20% faeces and urea. Therefore, thereconstituted effluent was composed of potabiewater and nitrogen (80% in ammonia form and20% in artificial faeces form). Ammonia wasinjected as (NH4)2S04. Artificial faeces were

specially made by the experimental food factoryof INRA according to Luquet (pers. comm.).

Start-up of nitrification

Natural start-up and start-up by injection ofcommercial Iyophilized bacteria (Bioariummanufactured by Lobial SA, 53000 Laval.France) were studied.

The pilot unit was washed with a solution ofhydrochloric acid. then the filter was filled withBiogrog up to 1.5m, and the whole unit wasrinsed with potable water for 7 days. Thepollutant [O.8g (NH4)2S04 and 8.4greconstituted faeces] and, if required, Bioariumwere injected, then ammonia and nitriteconcentrations in the tank were recorded eachday. When these parameters were stabie thefilter was seeded. Samples (5g) of Biogrog werecollected at the top, the middle, and the bottomof the filter on days 1, 28, and 40 of naturalnitrificationstartingup.

f".,.

j!

'11

11:i'

:1'i.."

jj.!:i

!i

1118

Sampling of the filter medium were kept at -20'Cin a mixture of 50% water and 50% glycerol.Observation of bacteria on the medium wasmade, after fixation, with a scanning electronmicroscope. Samples were fixed according tolaf rance et al. (1983). Then they were dried in acritical point desiccator Balzer, oovered with goldand three grains of each sample were observedwith a scanning electron microscope JeolJSM35 of the Faculté des Sciences de Rennes.The magnification was up to x 20 000.Observations were qualitative only.

Treatment efficiency

Biogrog efficiency was determined whennitrification was started. The flow rate in a filterdescended but oxygen and pH were not limiting.Retention time of the treated medium was from200s to 250s. Pollutant was injected once a dayand efficiency was calculated from the amountof ammonia oxidized between the entry and theexit of the filter. Efficien~ was expressed inkg N-NH4 oxidized per m3 Biogrog per day.

Influence of dissolved oxygen

In the experiments we wished to determine thedissolved oxygen available for nitrification underaquacultural conditions. Oxygen injection wasstopped after injection of the pollutant and theevolution of dissolved oxygen, ammonia andnitrite in the tank of the pilot unit was recorded.

Results and discussion

Start-up of nltriflcaUon

Natural nitrification start-up was slow (Koillerand Avtailion, 1985). It lasted 42 days (Fig. 3).Injection of Bioarium reduced it to 24 days (Fig.4). The effect of Bioarium was fast. Twenty-fourh after injection we observed the appearance ofnitrites and a decrease of ammonia. But 3 daysaft er injection ammonia concentrationincreased. This nitrogen production could haveresulted from bacterial lysis. Bernard andChaoot (1986) have shown, for examp/e, thatIyophilized bacteria cannot attach themselves tothe filter and die. However wh en injection ofBioarium was stopped, ammonia decreased 5days after the last injection. Previous studies(Usel and leffemberg, 1977; Bower andTurner, 1984; Bernard and Chabot, 1986) did notreport acceleration of the process with injection

P. Poquillon et al.

of Iyophilized bacteria into the filter. Sutterlin elal. (1984) and Soott and Allard (1984) usedIyophilized bacteria to reduce nitrificalionstarting up.

The observations with the scanning electronmicroscope showed the evolution of the organicmateriaion the Biogrog surface. Early innitrification start up (day 1), organisms werescarce and located in sheltered areas. Weobserved bacteria and fungal spores (Phoio 3).Later (day 28) the number of bacteria seemed 10increase, amoebae (Photo 4) and eucaryolicfungi appeared. When the filter was seeded (day40) ;there were mainly only two species ofbacteria present (Photo 2 and 5). But we couldnot identify them because their morphologydepends on their environment (Johnson andSieburth, 1976).

In fish breeding, filters must be disinfected allereach breeding cycle and the nitrification restartis expensive in terms of time and energy. Theuse of Bioarium reduces the duration and theoost of this operation.

Two other techniques to acee/erate nitrificationstart-up will be tested in our next experiment.These are doping with metaJ supply (iron andoopper) and preseeding the Biogrog.

Treatment efficiency

Biogrog efficiency was more than 9kgN-NH4.m,3.d-1 (Tabie I). We obtained 4.9kgN-NH4.m,3.ó1 when these results wereextrapolated to 13'C (according to the formulareported by Martin, 1979). These results were

better than those observed previous~ ~ Faupet al. (1982) on biolite (1.1kgN-NH4m' .d- ). This~ifference could be explained by the g,reateslspecific area of Biogrog and by the experimentalconditions. These results were obtained withsequential injection of pollutant. Biogrogefficiency wiJl be tested with a continuousinjection of pollutant and with a weaker hydraulicburden. Indeed the hydraulic burden utilized inour experimentations oould be employed in fishbreeding only if there was a very efficientdecantor.

Influence of dlssolved oxygen

Fig. 5 shows that if the injectionof oxygen isstopped, the reduction of total ammonia

~

---

I

j

"Nitrification in aquaculture waste-water 1119

.gJ</LTotol anmoniaNi tri tes

4 A

n'

.:-'.'-,,'

10 15 20 25 30 35 40' 45 days

Fig. 3. Natural start-up of nitrification. (*) Pollutant injection; (A ) Biogrog sampling; (-) total ammonia intank; ( ) nitrites in tank.

mgN/L

t

: \"., ': "

.. "

,,.,..,.."

, ..0 5 10 15 20 25 30 35 days

Fig. 4. Start-upwilh injection of Bioarium. (*) Pollutant injection; (t ) Bioarium injection; (-) tata! ammonia intank; ( . . . ) nitrites in tank.

J

rI'iII:

k

1120

~i

Photo3. Mushroom spore, 1J1.ll1;{x10000).

Photo 4. Amoeba, 1J1.ll1;{x6000).

concentration decreases when the amount ofdissolved oxygen is less than 4mg02.r1, andstops when it is less than 3mg02.r1. As soon asoxygen was injected the proc,ess of nitrification

Table I. Efficiency of Biogrog

Iq

P. Poquillon et al.

Photo 5. Nitrification bacteria, 1J1.ll1;(x6 009). .--

restarted and recovered its maximal efficiencyafter 48h. Oxidizing of 19 ammonia required4.5goxygen. So, nitrification depended on availableoxygen. According to Sharma and Ahlert (1977)there was a threshold value of dissolved oxygenbelow which there was no bacterial growth.However according to Quinlan (1984) andStenstrom and Poduska (1980) the influenceofdissolved oxygen on nitrification folio wed aMichaelis law.

The most probable explanation of our resultswas that when dissolved oxygen decreasedbelow 4mg02.r1, nitrification was very reducedbut not stopped. On the other hand,ammonification appeared and resulted in astagnation of ammonia and dissolved oxygen inthe unit.

So the oxygen quantity available for bacterialepuration of aquaculture waste water is the

~o: amount of total ammonia in filter entry.Experimental filter: height 2m, diameter 160mm.

Biogrog: height 1.5m.

-.

Na Temp. Hydraulic lood Retention time Efficiency0

(mgJ') (oC) (m3.m'2.d'1) (kgN-NH4.m'3.d.1)

1.5-2.5 20:Î:1 515 4min 10s 9.18:t2.44

0.3-0.5 2oI1 644 3min 20s 9.60:t2.14

,- ~

Nitrification in aquaculture waste-water 1121

mgl'j/L

fOxygen injection

m".02/L

4

000 10 days

Fig. 5. Influenoe of dissolved oxygen on nibifJCation. (0) Dissolved oxygen in tank; (0) total ammonia in tank;(0) nitrites in tank.

f1'1

!:

I,"

1122

dissolved oxygen quantity int he water less4mgCh.r1.

Conclusion

Biogrog appeared to be an excellent medium fornitrification in a submerged filter. lts high specificarea increased its efficiency therefore reducingfilter litter. These experiments were performedwith a single dose pollutant loading. In our nextexperiments we wil! test biogrog with continuousinjection of pollutant (ammonia andreconstituted faeces) and with variations ofloading.

In the present experiments, the oxygen quantityavailable for nitrification was found to be thedissolved oxygen quantity in the water, less4mgCh.r1. this means that, if the amount ofdissolved oxygen is 15mgO2.rl, the oxygenquantity available for nitrification is 11mgO2.r1.

The start-up of nitrification could be shortened to24 days by the injection of Bioarium. Thisreduces filter starting oosts and could reduce theeHeets of an accidental pollution of the biologicalfilter.

ReferencesBemard S.C. and M. Chabot 1986. Les activateurs

biologiques et leur utilisations dans les eauxrésiduaires. L'eau, I'industrie, les nuisances101:47-48.

Boller M. and W. Gujer. 1986. Nitrification in tertiarytrickling fihers followed by deep-bed filters. WaterRes.2O(11):1363-1373.

Bower C.E. and D.T. Turner. 1984. Evaluation of twocommercial nitrification accelerators in dosed

seawater culture systems. Aquaculture41:155-159.

Csavas I. and L Varadi. 1981. Design and operationof a large scale 9xperimental recycling systemheated wilh geolhermal energy at !he fish cultureresearch institute, Szarvas, Hungary.FAO-EIFAC Symp. Stavanger 2:445-454.

de Kinkefin P., C. Michel, and P. Guittino, 1986. Précisde pathologie des poissons. INRA-OIE, Paris.348p.

Faup G.M., A. Leprince, and M. Pan nier. 1982.Biologica! nitrification in a up flow fixed bedreactor (UFBR) Water Sci. Tech. 14:795-810.

Gonenc E. and P. Harremoes. 1985. Nitrification in

rotating cis systems. I. Criteria for transition tromoxygen to ammonia rate fimitation. Water Res.19(9)1119-1127.

--...........

P. Poquillon et al.

Grasmick A., S. Elmaleh, and R Ben Aim. 1979

Théorie de I'épuration par filtration biologiqu~immergée. Water Res. 13:1134-1137.

Hall D.J. 1986. Nitrification options evaluated usingpilot filters. Water Pollut. Contr. 431-439.

Johnson P.W. and J.McSieburth. 1976. In situmorphology on nitrifying-like bacteria inaquaculture systems. Appl. Environ. Microbiol.31(3):423-432.

Koiller M. and RR Avtalion. 1985. A laboratory scalerecycling water unit for tilapia breeding. AquacultEng. 4(4):235-246.

Kowalski E. and Z. Lewandowski. 1983. Nitrification

process in a packed bed reactor wilh chemicallyactive bed. Water Res. 17:157-160.

Kujal B. 1981. Pre-treatrnent of water for intensiveaquaculture by filters with invented water flow.FAO-EIFAC Symp. Stavanger 1:383-394.

Lafrance P., M. Mazet and D. Villessot. 1983.Croissance bactérienne sur charbon acti!

granulé, investigation au microscope à balayage.Water Res. 17(10):1467-1470.

Lésel Rand F. Leffemberg. 1977. Etude de lanitrification dans les circuits d'eau recydée. ,.Oxydation des nitrites. Bull. Cent. Etud. Rech.Sci. Biarritz 11(4):441-459.

Martin G. 1979. Le problème de I'azote dans I'eau.Technique et Documentation Lavoisier, Paris.279 p.

Paolini A.E. and G. Variali. 1982. Kinetic consideration

on Ihe performance of activated sludge reactorsand rotating biological contactors. Water Res.16(2):155-160.

Quinlan A.V. 1984. Prediction of Ihe optimum pH lorammonia N oxidation by Nitrosomonas europaeain weil aerated natural and domestic waste water.

Water Res. 18(5):561-566.RichardY., L Brener, and G. Martin. 1979. Nitrification

des eaux souterraines et des eaux clarfiées. TSM

L'eau 74(3):157-163.Roques H., L. Masbermat, and J. Tsacoyannis. 1976.

Essais de mise au point d'un moÇèle général decroissance des microorganismes englOOant lesphénomènes de nitrification et la métabolisationdes matières hydrocarbonées. Water Res.10(3):265-270.

Roques H.. S. Yue, S. Saipanich, and B. Capdeville.1982. Faut-il abandonner le formalisme de

MONOD pour la modélisation des processur parvoie biologique. Water Res. 16:839-847.

Scolt K.R and L. Allard. 1984. A four-tank water

recirculation system wilh a hydrocyclone prefillerand a single water reconditioning unit. Prog. Fish-Cult. 64(4):254-261.

Sharma B. and RC. Ahlert. 1977. Nitrificalion and

nitrogen removal. Water Res. 11:897-925.Stenstrom M.K. and RA Poduska. 1980. The effect

J

Nitrification in aquaculture waste-water

of dissolved oxygen concentration on nitrification.Water Res. 14:643-649.

StUtteriin A.M., C.V. Couturier, and T. Devereaux.1984. Arecirculation system using ozone for theculture of Atlantic salmon. Prog. Fish - Cult.46(4):239-244.

-

..1123

Wickins J.F. 1985. Ammonia production and oxidationduring the culture of marine prawns and lobstersin laboratory recirculation systems. Aquacult.Eng. 4(3):155-174.

._-

j