Fisheries Science 60(2), 133-136 (1994)

Ammonia Oxidation in Marine Biological Filters with Plastic Filter Media

Kotaro Kikuchi, Haruo Honda, and Michiyasu Kiyono

Abiko Research Laboratory, Central Research Institute of Electric Power Industry, Abiko, Chiba 270-11, Japan

(Received September 10, 1993)

Ammonia oxidation activities were compared across 5 kinds of plastic filter media and one type of earthenware ball filter medium in submerged biological filters.

With the progress of ammonia oxidation, the pH and alkalinity of the recirculating seawater decreased. Ammonia oxidation was inhibited when the pH and alkalinity reached 6.0 and 0.5 meq./l, respectively. A significant linear relationship was observed between the amount of biologically oxidized ammonia and the decrease in alkalinity. Among well-conditioned biological filters with 6 types of filter media, the filters with a net or fibriform filter media showed higher ammonia oxidation. There was no significant linear relationship between the ammonia oxidation rate in the filter and the surface area of the filter medium. Daily loading of an organic substance for 92 days reduced the ammonia oxidation rates in the filters by 28 to 52%. The maximum ammonia oxidation rate (0.55 g-N/m2/day) after loading the organic substance was obtained in the filter with a net filter medium.

Key words: closed recirculating fish culture, marine biological filter, ammonia oxidation rate, alkalinity, plastic filter media

The removal of ammonia is one of the most important factors in maintaining conditions in recirculating fish cultures. This is because ammonia is the major excretory product harmful to fish. Ammonia is generally reduced by bioligical treatment, which utilizes bacterial metabolism to oxidize ammonia to nitrate. Sands and gravels are widely used as filter media, however, they are heavy and require frequent backwashing to prevent blockage. Recently, a variety of plastic filter media have been introduced to water treatment processes. Plastic media have several advantages such as high specific surface area for bacterial adhesion, great porosity, and low specific gravity.

To design effective closed recirculating culture systems, it is important to examine the nitrifying capacity of biological filters. Although several studies1-4) have been conducted on nitrification in filter beds, there is little information available on the nitrification capacity. In freshwater, Saekisl reported that the ammonia oxidation rate in 10 g of sand was 0.14 mg-ammonia/day. Rijn and Rivera6) showed that the maximum removal rate of ammonia in a trickling filter with PVC material as a filter medium was 0.43 g-N/m2 surface area of the filter medium/day. In seawater, Nijhof and Bovendeur7) reported that the maximum nitrification capacity in a filter with a plastic filter medium was 0.28 g-N/m2/day. The ammonia oxidation rate in the filter is considered to be affected by the type of filter media as well as the specific surface area.

In this study, ammonia oxidation activities in marine biological filters were compared across 5 kinds of plastic filter media with different surface areas and/or forms, and one type of earthenware ball filter medium.

Materials and Methods

The experimental seawater recirculating system is shown in Fig. 1. The total water volume was 101. Flow to the filters were supplied by a magnet pump (MD 6, Iwaki Co., Ltd.), and the flow rate was maintained in the

range 0.3 to 0.5l/min by a flow meter (Nihon Tokushu Seiki Co., Ltd.).

The water circulation and air supply in the reservoir were provided by an

air lift to keep the oxygen concentration at more than 90% of the saturation

level. All experiments were conducted in a dark room at 20•Ž. Duplicate

experimental recirculating systems were prepared for 6 kinds of filter media

(twelve in total). The filter media used in the experiments are shown in Table I and Fig. 2. The volume of each filter medium was 11. The

earthenware ball filter medium was almost spherical and non porous. The

Fig. 1. Diagram of the experimental seawater recirculating system.The total water volume was 10 1. The air was supplied produced

by an air lift in the reservoir.

Table 1. Filter media used in the experiment

* The filter consists of filter bed walls , polyvinyl chloride pipes and I l volume of filter media. Effective surface area for bacterial adhesion of the filter without

filter media is 0.13 m2.

134 Kikuchi et al.

Fig. 2. Filter media used in the experiment.

(a) Earthen ware balls; (b) Honeycomb tube (cell size: 8 mm); (c) Net filter medium; (d) Fibriform filter medium. Scale bars represent 1 cm.

honeycomb tube (Shin-Nihon Core Co., Ltd.) was made of thin polyvinyl chloride sheets in a honeycomb shape. The net filter medium (Eishin Kogyo Inc.) was a set of six 5cm square polyethylene nets piled on each other, and 30 pieces were used in each filter. The fibriform filter medium (Nihon Sangyo Kikai Co., Ltd.) was made of fine vinylidene chloride fibers, and 52 elements each 10 cm long were used in each filter.

The study was divided into three serial experiments (Experiments 1, 2, and 3). Before the start of each experiment, all filters with new filter media were conditioned by adding a 50 ml nitrifying bacteria solution (Fritz-zyme No. 9, Fritz Chemical Company), as well as by the daily addition of 10 mg ammonia-N for 60 days. At the end of each experiment, the reservoir was washed with tap water to prevent the growth of nitrifying bacteria, and the seawater in the system was completely replaced. The seawater used in the experiment was collected at Onjuku in Chiba Prefecture (salinity: 32-34). In all experiments, ammonium chloride was used as the ammonia source.

Changes in Water Quality by Ammonia Oxidation (Experiment 1)Ten mg ammonia-N was added daily to each recirculating system for

30 days to examine the change in water quality due to ammonia oxidation. Ammonia, nitrite, and nitrate concentrations were analyzed every 5 days, and the pH and alkalinity were measured every 2 or 5 days. To prevent a decrease in pH and alkalinity, two pieces of oyster shell about 10 cm in length were put into the reservoir on the 25th day.

Ammonia Oxidation Rate in the Filters after Loading Ammonia-N for 90 Days (Experiment 2)

After Experiment 1 was completed, 20 mg ammonia-N was added daily to each system for 90 days to make well-conditioned biological filters. To determine the ammonia oxidation rate in each filter, 1000 mg ammonia-N was added to each recirculating system at the start of Experiment 2, and ammonia, nitrite, and nitrate concentrations were measured every 12h thereafter. Sodium bicarbonate was added, when necessary, to maintain the pH abouve 7.0. The ammonia oxidation rate in each filter was calculated by the least squares method based on the data until the 96th hour of the experiment. After transferring each filter medium from the well-conditioned biological filter to a new one (thus removing background nitrification), the ammonia oxidation rate in each filter medium was estimated in the same manner. The ammonia oxidation rates in each filter and filter media alone were measured repeatedly (4 data for each rate determination).

Ammonia Oxidation Rate in the Filters after Loading Ammonia-N and Organic Substance for 92 days (Experiment 3)

After conducting Experiment 2, 50 mg ammonia-N and 500 mg Ehrlich meat extract (Kyokuto Pharmaceutical Industries) were added daily to each system for 92 days. The ammonia oxidation rate in each filter was measured by the method described before to examine the effect of the organic load on the ammonia oxidation rate. In Experiment 3, duplicate experimental systems were used for filter Nos. 1, 3, 5 and 6 (eight in total). One gram of Ehrlich meat extract contains 96 mg of organic nitrogen.

AnalysisAmmonia and nitrite were analyzed by the indophenol method8) and

the N-(1-naphthyl)-ethylenediamine method9) respectively. Nitrate was measured using an ion chromatographic analyzer (IC-500, Yokogawa

Fig. 3. Changes in ammonia, nitrite, and nitrate concentrations, pH

and alkalinity in a closed seawater system (Experimental system

No. 4).

Ten mg ammonia-N was added daily until the 30th day. Two pieces

of oyster shell (shell length; ca. 10 cm) were placed in the reservoir

on the 25th day to maintain pH and alkalinity The experiment was

conducted in a dark room at 20•Ž.

Electric Corp.) and the alkalinity was determined by the method of Almgren

et al.10)

Results and Discussion

Experiment 1The results of experimental filter No. 4 are shown in Fig.

3. Although nitrite accumulated immediately after starting the experiment, this had ceased by the 20th day. Ammonia began to accumulate on the 20th day when the pH and alkalinity decreased to 6.0 and 0.5 meq./l, respectively. However, the addition of oyster shells on the 25th day decreased the ammonia to a level which could not be detected on the 35th day. Similar trends in the water quality change were observed among all the systems .

There was a significantly close relationship between the amount of oxidized ammonia and the decrease in alkalinity (Fig. 4), as described by the following linear equation;

Ammonia Oxidation in Marine Biological Filters 135

Fig. 4. Relationship between the amount of biologically oxidized ammonia and alkalinity decrease in recirculating seawater in a closed system.

where Y is the decrease in alkalinity (meq./1) and X is biologically oxidized ammonia (mg-N/1).

This result shows that the alkalinity of recirculating seawater was mostly depleted when 20 mg-N/l of ammonia was biologically oxidized, since the alkalinity of natural seawater is usually in the range of 2.2 to 2.5 meq./l.11) Calcareous filter media such as gravel, crushed oyster shell and limestone supply alkalinity to recirculating water. However, Siddall12) found that an acid-insoluble, non-organic substance growing on the surface of filter gravels might reduce the buffering activity. Bower et al. 13) reported that in the closed culture of Fundulus spp., the calcareous filter media prevented the rapid acidification of the culture water. However, buffering agents such as sodium bicarbonate or sodium carbonate must be regularly added to maintain the pH above 8.0. The decrease in alkalinity according to the nitrification obtained in this study (5.51 mg CaCO3/mg ammonia-N) is lower than the theoretical value of 7.13 mg CaCO3.

The effect of a low pH on the ammonia oxidation rate shown in this study was previously reported by Hirayama, 1) and his results showed that the nitrifying activity of a filter is restricted to some extent without ceasing completely even if the pH and alkalinity are less than 6.0 and 0.05 meq./l, respectively. Forster2) also reported that nitrification in seawater is affected by low pH and ceases completely at pH 5.5.

Experiment 2The ammonia oxidation rates in well-conditioned bio

logical filters and filter media alone are shown in Table 2. The ammonia oxidation rates in filters were higher in filters with a net or fibriform filter medium (Nos. 5 or 6), and the rate in No. 6 was significantly higher than the rate in the others (Mann-Whitney test, p <0.05). Similar trends were observed in the rates in filter media alone. When comparing Nos. 2 to 4 (honeycomb tube), in spite of the difference in surface areas of the filter media, there was no significant difference in the rates both in the filters and in the filter media alone. The average rate of ammonia oxidation per square meter of surface area in the filter media alone was 17.8 (No. 1), 33.1 (No. 2), 12.6 (No. 3),

Table 2. The ammonia oxidation rates in the filters and filter

media in well-conditioned biological filters

Figure in the same column having the same superscript are not significantly different

(p> 0.05). The determination of each rate were conducted with duplicate experimental

systems and duplicate measurements spiked 1000 mg of ammonia-N (4 data for each

rate determination). The experiment was carried out in a dark room at 20•Ž. Twenty

mg ammonia-N was added daily for 90 days before the measurement of oxidation rate.

6.6 (No. 4), 26.6 (No. 5), and 7.3 mg-N/m2/h (No . 6), and

the rate seemed to decrease as the surface area of the filter

media increased.

The ammonia oxidation rates in the filters without filter

media (calculated by subtracting the rate in the filter media

from that of the entire filter) ranged from 5.9 •} 1.6 to

7.0•} 1.6 mg-N/0.13 m2/h, and were not significantly differ

ent to each other. These results show that the filter bed

walls and polyvinyl chloride pipes in the filter provide an

effective area for bacterial adhesion in spite of their small

surface area, and are considered to play an important role

in ammonia oxidation. The ratios of ammonia oxidation

rates in filters without filter media to those with entire

filters were high (48 to 63%) in Nos. 1 to 4, and low (38

and 39%) in Nos. 5 and 6.

Experiment 3

Table 3 shows the ammonia oxidation rates in the filters

after daily loading of an organic substance for 92 days. In

all filters tested, the ammonia oxidation rates were lower

than those of Experiment 2 by 28 to 52%. The ammonia

oxidation rates in the filters were significantly higher in Nos.

5 and 6 (a net and fibriform filter media) than in the others

136 Kikuchi et al.

Table 3. The Ammonia oxidation rates in the filters after loading

an organic substance for 92 days

Experimental conditions are shown in the footnote of Table 2. Figures in the same column having the same superscript are not significantly different (p>0.05).

Fifty mg ammonia-N and 500mg Ehrlich meat extract were added daily for 92 days before the measurement of oxidation rate.

(p <0.05).

The daily addition of ammonia and Ehrlich meat extract

developed biofilm and bacterial flocks in the filter. The

nitrate linearly increased to about 500 mg-N/l, however, it

remained connnnstant after about the 85th day in all experi

mental recirculating systems. The formation of anaerobic

parts in the filter, as suggested by the above, is one possible reason for the decrease in the ammonia oxidation rate af

ter loading the organic substances. The average rate of

ammonia oxidation per square meter of effective surface

area in the filters (including filter bed walls and polyvinyl

chloride pipes) after loading the organic substance in Nos.

1, 3, 5 and 6 was 14.8, 9.1, 22.9 and 6.7mg-N/m2/h,

respectively (corresponding to 0.36, 0.22, 0.55 and 0.16 g-

N/m2/day, respectively). These values are similar to those

reported by Nijhof and Bovendeur,7) and Rijn and Rivera.6)

In this study, there was no significant linear relationship

between the ammonia oxidation rate in filters or filter media

alone and the surface area of the filter. The filters with a

net or fibriform filter media showed higher oxidation rates.

The ammonia oxidation rate in No, 5 after loading the

organic substance was 11.0 mg-N/filter/h. In the previous

paper concerning the ammonia excretion of the Japanese flounder Paralichthys olivaceus, the maximum excretion rate

of the immature flounder (189-575g body weight) after

feeding at 20•Ž was about 1 mg-N/l00 g fish/h.14) These

results indicate that the ammonia excreted from 100 kg of

immature flounder could be oxidized in the filter with

about 1001 volume of the net filter medium. Honda et al.15)

reported that the Japanese flounder grew successfully in a

closed recirculating culture system with a 200l volume of

the net filter medium (EC filter medium), and the total yield

in the system was 97.6 kg. There may be factors which

should be examined in order to extrapolate the current

results to practical fish culture tanks. However, based on

data from the experimental tank, it is proposed that the

volume of the net filter medium required to produce 100 kg

of immature Japanese flounder in a seawater recirculating

culture system ranges from 100 to 200l. This value is much

lower than that estimated in the previous study, in which

about 7 m3 of gravel media was needed for a closed culture

of I ton of carp. 16)

Acknowledgement The authors wish to thank Drs. Kazutsugu Hirayama and Atsushi Hagiwara, Faculty of Fisheries, Nagasaki University, for their valuable advice and critical reading of this manuscript. Thanks are also due to Dr. Thomas M. Losordo, Departments of Zoology and Biological and Agricultural Engineering, North Carolina State University for his

valuable suggestions.

References

1) K. Hirayama: Studies on water control by filtration through sand bed in a marine aquarium with closed circulating system-VI.

Acidification of aquarium water. Nippon Suisan Gakkaishi, 36,26-34

(1970).2) J. R. M. Forster: Studies on nitrification in marine biological filters.

Aquaculture, 4, 387-397 (1974).3) J. F. Wickins: Studies on marine biological filters. Model filters.

Water Res., 17, 1769-1780 (1983).

4) J. Bovendeur, E. H. Eding, and A. M. Henken: Design and

performance of a water recirculation system for high-density culture of the African catfish, Clarias gariepinus (Burchell 1822). Aquaculture, 63, 329-353 (1987).

5) A. Saeki: Studies on fish culture in the aquarium of closed-circulating system. Its fundamental theory and standard plan. Nippon Suisan Gakkaishi, 23, 684-695 (1985).

6) J. Rijn and G. Rivera: Aerobic and anaerobic biofiltration in an

aquaculture unit-Nitrite accumulation as a result of nitrification and

denitrification. Aquacultural Engineering, 9, 217-234 (1990).

7) M. Nijhof and J. Bovendeur: Fixed film nitrification characteristics

in seawater recirculation fish culture systems. Aquaculture, 87,

133-143 (1990).

8) F. Koroleff: Determination of ammonia, in "Methods of Seawater

Analysis" (ed. by K. Grasshoff, M. Ehrhardt, and K. Kremling),

Verlag Chemie, Weinheim, 1983, pp. 150-157.

9) K. Grasshoff: Determination of nitrite, in "Methods of Seawater

Analysis" (eds. by K. Grasshoff, M. Ehrhardt, and K. Kremling),

Verlag Chemie, Weinheim, 1983, pp. 139-142.

10) T. Almgren, D. Dyrssen, and S. Fonselius: Determination of

alkalinity and total carbonate, in "Methods of Seawater Analysis"

(ed. by K. Grasshoff, M. Ehrhardt, and K. Kremling), Verlag Chemie,

Weinheim, 1983, pp. 99-123.

11) S. Tsunogai and S. Noriki: Chemical Analysis, in "Marine

Chemistry" (ed. by M. Nishimura), Sangyo-Tosho, Tokyo, 1983, pp.

265-267 (in Japanese).

12) S. E. Siddall: Studies of closed marine culture systems. Prog.

Fish-Cult,, 36, 8-15 (1974).

13) C. E. Bower, D. T. Turner, and S. Spotte: pH maintenase in closed

seawater culture systems: Limitations of calcareous filtrants.

Aquaculture, 23, 211-217 (1981).

14) K. Kikuchi, S. Takeda, H. Honda, and M. Kiyono: Effect of feeding

on nitrogen excretion of Japanese flounder Paralichthys olivaceus.

Nippon Suisan Gakkaishi, 57, 2059-2064 (1991).

15) H. Honda, Y. Watanabe, K. Kikuchi, N. Iwata, S. Takeda, H.

Uemoto, T. Furuta, and M. Kiyono: High density rearing of

Japanese flounder, Paralichthys olivaceus with a closed seawater

recirculation system equipped with denitrification unit. Suisanzo-

shoku, 41, 19-26 (1993).

16) A. Saeki: Studies on fish culture in filtered closed-circulating aquaria

‡U -On the carp culture experiments in the system. Nippon Suisan

Gakkaishi, 31, 916-923 (1965).