R. C. SQUIRES, BTech, MIMechE
Partner, Binnie & Partners International.

Paper presented at the 82nd Summer General Meeting and Conference, held at the Hotel Piccadilly, Manchester, on Thursday, 12th May 1977.

SYNOPSIS

THIS PAPER DESCRIBES typical layouts of sea-water systems for oceanaria and the basic requirements influencing design. Arising from experience with new oceanaria at Hong Kong, Jakarta, and Port Elizabeth, the treatment of water for mammals and fish is discussed, and standards of water quality needed for keeping marine mammals are suggested.

INTRODUCTION

An oceanarium provides facilities for people to view the inhabitants of the sea. For many people, a visit to an oceanarium is the only way they can see these creatures. The livelihood of an oceanarium depends on the income derived from the public. To attract large numbers of visitors having different interests, and to satisfy them so that they return to the oceanarium many times, the exhibits must be arranged to impart a lasting impact on as many of the people as possible.

The desired impact can be generated provided the exhibit tanks are bright and clean, the shows performed by the marine mammals are slick and undertaken by healthy vigorous animals, the exhibit tanks display various different types and species, the people can get close to the animals, and a sizeable fish tank displaying a high concentration of many types of fish, including fully grown sharks and rays, is included. All these criteria are governed mainly by the quality of the water in the exhibit tanks. Poor water quality leads to dirty unattractive tanks housing sick or lethargic animals and the inability to rear the colourful, delicate fish, marine invertebrates and the sensitive sharks.

Accordingly, the water must be of the correct quality and any system used to maintain its quality must be reliable, as degradation of the water could lead to the loss of animals. An untrained dolphin costs about £5,000 to purchase and after one year's training may command a price of three times its purchase value. An untrained killer whale could cost £50,000 and would be worth several times this amount when trained. Thus, a failure of the water supply system could be catastrophic to an oceanarium.

In the early 1970s, a general lack of knowledge existed on the exact quality of water needed to keep captive pelagic animals, like dolphins, in good health and it seemed that no one could positively state the animal's tolerance levels to water treatment chemicals which were used normally; because most oceanaria at the time, simply operated on open circuit systems pumping clean sea-water through the tanks to replenish them every two hours. Chemical treatment of the water in open systems was either not necessary, or not practised because of the expense incurred by the continual wastage of dosed chemicals. However, extensive knowledge existed on the husbandry of the animals from which it was clear that dolphins, although air breathing, could suffer a wide range of viral, bacterial, and parasitic illnesses, and that infection of animals could occur from the water in which they lived and through their food.

Oceanaria which had to use filtration plant used open, slow-sand and rapid-gravity filters for fish and mammal tanks respectively. However, as exhibit sizes increased, so did the filter areas, with attendant rises in capital costs of the installations. This led to a tendency for smaller fish and mammal tanks or, where larger fish tanks were erected, to the use of a low fish load per unit volume of water and, in the case of large mammal tanks, to the use of long turnover times which produced dirty, polluted tanks. Unfortunately, this approach did not generate the required impact on the visiting public, who preferred fish tanks stocked to high density. Consequently, there was a need to investigate the factors governing water quality for keeping marine mammals and fish, and to find ways of reducing the costs of plant where closed water-circulation systems were mandatory.

The data presented in this paper has been derived from the experience gained in designing three new large oceanaria recently undertaken by the author's firm in Hong Kong, Jakarta, and Port Elizabeth, together with data arising from studies and visits in connection with nine other oceanaria and reports of experiences from elsewhere.

THE FUNDAMENTALS OF OCEANARIA WATER SYSTEMS

The sizes of exhibit tanks vary considerably but, in general, main dolphin and whale show tanks have a water capacity between 1,360 and 4,540 m3 (0.3 to 1.0 mg), with or without underwater viewing windows; subsidiary mammal tanks for displaying seals and sea lions have capacities between 100 and 700 m'; dolphin training tanks have capacities between 700 and 1,100 m3 with separate quarantine tanks having a capacity of about 90 to 225 m3. Each of these groups of tanks would normally include a main tank and four to six satellite animal-holding tanks. For displaying fish, several conventional small tanks of capacity 0.25 to 25 m3 would be used together with a large tank fitted with underwater viewing windows, having a capacity between 700 and 2,250 m3. Accordingly the total water volume in an oceanarium could be up to 9,000 m3 (2 mg) and the total circulating water flow up to 90,000 m3/d (20 mgd).

Both mammals and fish pollute their environment with toxic excreta; thus arrangements are needed to remove these pollutants, or to dilute them to concentrations which are neither harmful to the tank inhabitants, nor result in the tanks having a dirty appearance. Consequently, a flow of clean water to the exhibits is required and this can be arranged in many ways; but fundamentally these ways fall into two categories: "open" or "closed" systems, as shown diagrammatically in Fig. I and 2 respectively.

 

In an open system, all incoming water is extracted from a source and pumped through the exhibits, possibly in cascade, and discarded to waste. In a closed system, after initial filling from some source, the water is re-circulated and treated alongside the exhibit; a small proportion of the water may be discharged waste, with make-up water being drawn from the source. A closed system requires water filtration and treatment equipment, but the pumping power needed to circulate the water is generally small. With an open system, the water must be pumped to the top of the highest exhibit tanks above the source, and this often entails a relatively high lift through which all the water must be continuously pumped at the required rate of circulation.

Conventional water treatment chemicals can be used to condition water for mammal tanks as the animals are air breathing, but for fish who derive their oxygen supply from the water, chemicals cannot generally be used without risk to the fish, unless the residual chemical concentration is very low. Unfortunately, the attractive, coloured reef fish and the marine invertebrates essential to a good display are often those with a minimum residual chemical tolerance. In many cases, the level of chemical toxicity for many of these species of fish is below that which can readily be measured. It follows that the water circulation to fish and mammal tanks must be separate in closed systems; but in open systems, the effluent from a fish tank may be re-used in a mammal tank, although the effluent from a mammal tank cannot be re-used in a fish tank. The water may be either sea-water, or fresh water to which suitable chemicals are added; sodium chloride is added for marine mammals, and synthetic sea-salts (Instant Ocean) for fish. If synthetic water is used, a closed system must be adopted because of the expense of the solutions.

SEA-WATER INTAKES

Sea-water intake systems can be arranged for sub-sea-bed or Leach well-point extraction where the formation of the sea-bed or beach is suitable, or for direct, below-surface abstraction from the sea where the formation is unsuitable. Selection of the appropriate system for a given application depends on the quantity and quality of the water needed, on physical parameters such as tidal action, littoral drift, location, and finally on economics. Sub-sea-bed or well-point systems yield water requiring the least amount of primary treatment and it is often of a quality that can be directly introduced to mammal and fish tanks; marine infestation of suction pipelines is also avoided. Accordingly, whenever such intakes are economical, they should be adopted otherwise few locations exist where inshore water abstraction can be applied without primary treatment. In most cases, winds, tides, currents, land-water run-off and other pollutants produce variations in the clarity, salinity, temperature, nutrient and bacterial content of the sea-water, together with periodically high plankton, diatom and algae content. These variables could all be unacceptable for fish and although only the bacterial content might initially be critical for the marine mammals, the other parameters could soon produce unsightly algae or other blooms within the system which would in turn putrefy, with the inherent risk of infecting the animals.

The siting of a direct extraction intake has been found to justify careful studies of water quality at a number of different levels between the surface and the bottom of the sea. Often several intakes sited at different levels are required, because the concentration of nutrients, pollutants and marine organisms will vary with depth. In summer, abstraction of cooler water near the sea-bed can reduce the amount of cooling needed at the tanks; salinity variations are also less near the sea-bed. On the other hand, oxygen content usually decreases with depth. In Hong Kong, for example, surface salinity varies by 50 per cent, but bottom salinity varies by only 20 per cent; also the vertical distribution of temperature in summer varies by up to 4°C from surface to sea-bed.

The intake lines should be kept as straight as possible and provided with means for back-flushing to clear accumulated debris, such as jellyfish and seaweed. To control marine infestation, which even on smooth U.P.V.C. pipelines can rapidly cause a loss of carrying capacity, sufficient chlorine must be dosed at the remote end. Where suspended silt and sand are not a problem, a tidal forebay can be used. This can be constructed on the beach and supplied by the correct number of siphon intakes, each drawing water from an appropriate level.

DISTRIBUTION SYSTEMS

The distribution of water for open-circuit, multiple-tank systems has, in the past, normally been arranged in cascade. In this way all the water is introduced to the highest exhibit tank and flows by gravity from one tank to the next, being re-used several times until it is discarded to waste from the last exhibit. As fish tanks require water of the best quality, the highest exhibit would normally be the fish tank. In some cases, the flow from the first tank is divided and passed to several tanks at the second level, being re-collected and passed on to the remaining tanks, if any, at a lower level. Severe algae problems can be experienced in mammal tanks at the lower end of a cascade distribution system.

If a storage reservoir is used, water for mammal and fish tanks should be separated. Many aquarists recommend that water for use in fish tanks should be aged in darkness before use.

The effluent from the tanks and from the backwashing of any filters must be returned to the ocean at a point which does not result in its being drawn in again through the intake. Current and wind float-tests are often needed to locate a satisfactory site.

CIRCULATION SYSTEMS

Irrespective of the type of circulation system used, i.e. open or closed, the inlet supply and discharge from each tank must be arranged so that the incoming water sets up a circulation in the tank which continually moves between the inlet and outlet. To achieve an acceptable circulation in large tanks, multiple inlets and outlets are needed. Several rows of inlets sited at different depths are often used.

Complete circulation in mammal tanks ensures that any disinfecting agent used is mixed in the tank, and there are not any local areas where debris accumulates. For fish tanks, this ensures even distribution of oxygen and removes debris to the filters.

Circulation in large fish tanks where central decoration is used, must be a combination of vortex and downflow. Vortex flow promotes a good rotating current, and this is crucial to the successful rearing of sharks who cannot pump water through their gills, but must rely on currents to do this for them. If there is not any current, sharks are unable to rest and must move continually to obtain oxygen. All water is introduced near the surface and most is abstracted from the base of fish tanks, although a proportion of water is skimmed from the surface to remove floating material, such as fish oils, dust and scum.

In the past, water circulation in mammal tanks has also been by downward vortex flow with a high rate of surface skim to remove floating debris. However, in future tanks, a combined flow system will be tried where the major proportion of water will be introduced at the base of the tank and abstracted from the surface, thereby obtaining a better distribution of the sterilizing agent which will be introduced at the place where there is maximum demand. It will be necessary to retain a bottom outlet, abstracting a small flow to remove any debris which settles against the vertical upflow current.

A mammal tank would normally comprise a main tank with four to six holding tanks. Inlets and outlets are needed for each of these. Also, because part of the mammal training programme relies on the development of a close mammal-to-man relationship and because animal husbandry involves periodic medication, vaccination, and blood sampling, it must be possible to lower the water level in any of the holding tanks to permit the trainers to "walk" the animals in about half a metre depth of water, whilst retaining circulation in all tanks. This is normally accomplished using weir boxes with adjustable sseirs and be-pass gates positioned at the correct level. Each holding tank is fitted with removable water gatproof es for separation, when necessary, from the main tank.

The outlet flow from each tank is passed by gravity to a separate upstream wier chamber, fitted with an adjustable weir. Flows from each are collected in a common collection sump. The weir chambers are sized to permit settling of larger solids such as sea lien faeces and foreign matter thrown into the pools by the public. The chambers have hopper bottoms with individual drain valves to permit the periodic flushing out of accumulated debris.

At normal operating levels, a good fall of water over the weir, usually about, 2 to 3 m, is needed, which aerates the water, expelling free CO, and other gases. The weir box also acts as a chemical mixing chamber for the addition of pre-filtration chemicals.

Weir boxes are also used for large fish tanks, mainly for aeration but also to permit reduction and recirculation of water at low levels when the fish are receiving medication.

Water drawn from the surface of the tanks is discharged preferably direct to waste in both open and closed systems. Often a proportion is re-used in closed systems being returned from the tank by gravity to the downstream side of the weirs.

Broad-crested, surface draw-off weirs have been found to give poor performance because a relatively large head over the weir is needed to remove floating solids, which represents a large flow on tanks with long perimeters. Broad-crested weirs fitted with vee-notches along their length have proved more satisfactory.

Irrespective of the type of system used, mammal and fish quarantine tanks are needed and their circulation equipment and filter plant, if any, must be either entirely separate from the remainder of the system, or be capable of being completely separated when sick animals are being kept. If the quarantine system is normally part of the main system but is capable of being isolated, every isolation point at an interface with the main system must be fitted with double valves and a drain between them, so that the joining sections of pipe can be emptied to create a bacterial barrier.

FILTRATION PLANT

Filtration plant used for closed circuit supplies for mammal and for fish systems are quite different. For mammals, the filter has to remove as much of the suspended solids as possible and to give clarity suitable for good viewing. In many cases, underwater viewing windows are provided in mammal tanks where a visibility of 45 m underwater is required. For fish, however, the filters must not only produce acceptable clarity but must also support a bacterial population to oxidize the toxic ammonia compounds to the less toxic nitrites and nitrates. Accordingly, filter rates, volume of filtering sand, and size of filter media are extremely important for fish tank systems; whereas for marine mammal systems which use oxidizing sterilizers to accomplish the major part of the nitrification and BOD removal, filters are principally designed to give the required clarity.

Small fish tanks in the home rely on the bottom sand layer to accomplish the required filtration, abstracting water beneath it through a waffle plate and returning it to the top of the tank by air-lift pumps. However, debris from fish faeces and uneaten food soon accumulates on the sand making cleaning of the tank mandatory. Such systems cannot be used in large fish tanks because of the practical difficulties of holding the fish whilst large volumes of floor sand are cleaned. Accordingly, in large fish tanks, the tank floor is designed to permit all debris to be flushed into the filters, but as tank floor decoration is needed to produce a realistic environment, false floors of lattice construction supporting carefully positioned rocks and large stones are used giving the impression of a rocky seabed, but with sufficient openings between the stones and rocks for the debris to pass through the floor to the filters. Care must be taken in designing the underfloor to avoid debris accumulating and putrefying around any floor supports.

The Hong Kong reef tank of 2,045 ml capacity uses a multiple, hopper-bottomed floor with an outlet from each hopper, to ensure that no debris accumulates within the tank underdrain system.

MATERIALS OF CONSTRUCTION

In open circuit fish and mammal systems, the materials used for the component parts are chosen generally for minimum maintenance and most materials known to have corrosion resistance in sea-water have been successfully applied. However, with closed-circuit systems, problems have been experienced in many oceanaria which could be traced to the materials used. Most metals, epoxies, paints, and sealants are suspect, especially for fish circuits, and only 316S12 and 316516 stainless steels together with titanium have been shown to be entirely satisfactory. The expense of these materials, however, makes it necessary to reduce usage of them to a minimum, and trials were therefore made to discover the best substitute inert materials which have now been in use since 1973.

Filter shell linings of rubber or neoprene lining appear to be superior to all other linings, but success has been achieved with one chlorinated rubber paint which does not retain elements which are toxic, after drying, to fish. Fibreglass and epoxy linings are not recommended.

Valves, weirstocks, and similar, made of neoprene-covered cast iron are suitable, with bolts of 316 stainless steel.

The materials used for pumping plant present the most difficult problem, because the pump duty for declining filter rate systems must ideally produce a small variation in flow for a wide range of pumping heads; that is if the tank recirculation time is not to become too long as the filter run proceeds. Such pumps are close tolerance machines which traditionally have metallic wearing rings. One solution which has performed reasonably well, but requires considerable maintenance, is the use of all-cast-iron centrifugal machines, with stainless steel shafting and the casing and impellers coated in nylon; no wearing rings are fitted. Reasonably close tolerances between impellers and casings are achieved, but fine debris finally wears off the nylon coating at the areas of closest tolerance. The coating is easily repairable at site, but this approach, although satisfactory for the animals, is not altogether the best; machines made entirely of stainless steel probably offer a better solution.

It some sacrifice can be tolerated in the efficiency and limitation of filter runs, the Gardener - Denver type of pump can be used. This machine has a cast iron casing into which a two-part, rigid, rubber liner is inserted, with a 316 stainless steel shaft and rubber impeller. There is not therefore any metal in contact with the water being pumped. The impeller is of the open type and although not as efficient as the close tolerance machines, it can pump solids such as uneaten feed fish, thereby obviating the need for pump suction strainers with their inherent head loss. Units of this type up to 29,500 m3/d at 15 m head have been giving satisfactory maintenance-free service.

In fish tank systems where heating and cooling is required, specialized materials are needed for the heat exchangers. Glass tubed units have been tried in the U.S.A. with limited success and, at present, there seems little or no alternative to adopting titanium heat exchangers.

All mammal tanks need surfaces smooth enough to reduce the adherence of algae and must avoid crevices where bacteria can grow. Underwater viewing windows should preferably present acryllic surfaces to the water. All straight pipes should be of unplaticized P.V.C., and all fabricated bends and specials should be of U.P.V.C. with external fibreglass wrapping. Where U.P.V.C. pipes are exposed to sunlight they should be painted with an ultra-violet barrier paint.

THE TREATMENT OF WATER FOR OCEANARIA

PRIMARY TREATMENT

Primary treatment is designed to remove debris and to kill any bacteria, plankton, diatoms, or other organisms. If a high concentration of fine suspended solids is not encountered but primary treatment is necessary, the water should be filtered by rapid gravity filters operating at a filter rating of about 3.4 m/hr (70 gal/ft'/hr). This is a sufficiently low rate to permit handling of a wide variation in the concentration of marine debris with a reasonable length of filter run between backwashes. Conventional backwash and air scour rates are used for these filters. After filtration, the water should be aerated over weirs, dosed with high doses of chlorine and retained for a period of at least 30 min before passing to the oceanarium. All water used in the fish tanks should be dechlorinated by being passed through an activated carbon filter bed at a rate of 10.8 m/hr (220 gal/ft'/hr).

In areas where heavy loads of fine sand and silt are present, sedimentation is an essential stage of pretreatment. Sea-water can be successfully settled in conical, hopper-bottomed, vertical-flow, settling tanks. In one installation which supplies mammal and fish tanks with make-up water from a common source, aluminium sulphate dosing is used to assist settling when the tank is operating on a continuous flow basis to make up the water level in the mammal tanks. A filling, settling and batch discharge flow process, without aluminium sulphate dosing, is used when providing water to the fish tanks.

If the settled water is to be used in a closed system oceanarium, further pretreatment by filtration is unlikely to be necessary because the make-up water is added upstream of the exhibit filtration plant. In this case, the water should be dosed with chlorine prior to settlement so that the retention time in the tank can be used for chlorine reaction time, unless a site storage reservoir is used. Any water to be used in the fish tanks should be passed through activated carbon for removal of the halogen residual before use.

Open circuit oceanaria may require up to 100,000 m3/d of feed water, compared to 5,000 m3/d for the equivalent closed circuit system. Hence the extent of the intake and primary treatment needed to render the water suitable for the exhibits may alone determine the type of circulation system used in the oceanarium.

MAMMAL TANKS

In closed systems for mammals, it is important that the water should be disinfected and effective removal of the pollutants, which are introduced by the animals, must be achieved. In addition, inhibition of algal growth and maintenance of clarity is needed if the exhibits are to be attractive to the public. Experience has shown that suitable water for marine mammals can be obtained if treatment of the water aims to eliminate suspended solids, BOD, and to oxidize all nitrogenous matter. The total oxidized nitrogen and phosphorus levels must be kept reasonably low, and discarding daily at least 5 per cent by volume of the tank water will prevent the accumulation of these algal nutrients and any other complex components which might be harmful to the animals.

Ridgeway (1972) has published the results of studies on the production of urine and faeces from dolphins and seals, which show that a single 187 kg (410 lb) dolphin, fed on 7 kg of fish per day, produces 1.45 kg of faeces and over 4.5 litres of urine per day. However, water balance tests on the animal showed that it probably consumed only about 0.036 kg of the 5 kg of water in its feed, the balance being excreted. Thus, after allowing for the water in the faeces, the faeces weight reduces to about 0.26 kg/day and the urine to about 0.45 kg/day of which about 0.29 kg was identified as urea, containing 46 per cent nitrogen.

A healthy animal should maintain a generally stable blood-urea-nitrogen content; hence the daily pollution in the tanks must be proportional to the amount of feed. The high protein fish diet of marine mammals contains about 4 per cent nitrogen, and Murphy (1975) has shown that the BOD per day of dolphin excreta is also about 4 per cent by weight of the daily feed. If the BOD and daily amount of nitrogen present is to be removed using chlorine, a breakpoint reaction must be set up so that the amount of chlorine present in the water is always at the correct ratio with the nitrogen for breakpoint. The ratio of chlorine to nitrogen is theoretically between 5 and 12 times, for conventional sewage treatment practice. The sea water marine mammal systems installed show that about 10 times as much chlorine as nitrogen must be present in the water for the reaction to proceed.

To achieve disinfection and to inhibit algae growth, an additional amount of chlorine is needed to establish a free residual. The systems designed include a weirbox, which in addition to aeration may reduce the residual chlorine in the water; thus additional chlorine may have to be added continually to the re-cycle flow to maintain a residual in the tank. On the foregoing basis, the chlorine consumption for a given system, estimated as described, should produce a water with low concentrations of NH3-N, NO2-N, NO3-N, and BOD.

A study of the operating results obtained from three different dolphin tank systems which were installed recently, each with different water volumes and quantities of animals, shows a close correlation between the amount of chlorine actually consumed per day and the amount derived by multiplying the amount of nitrogen, or 4 per cent of the "eight of the daily feed, by ten, and adding the amount of chlorine needed to maintain the residual. Details are given in Table 1.

When analysis for BOD has been made at Port Elizabeth oceanarium, it has generally been undetectable. Water analyses have been performed daily on samples from each of the tanks at the Honk Kong oceanarium and a typical set of monthly figures are given in Fig. 3, which also shows the effects of chlorine failure. Bacterial counts for Coliforms and E Coli have also been performed three times a month and have generally shown no infection. The records show that the NH3-N residuals in the tank are generally below 0.05 mg/1 and periodic tests of nitrates show these also to be very low. It can be concluded therefore, that the pollution by the animals is effectively being removed by the chlorine injected and by the filter plant, and that the amount of pollution added by marine mammals approaches 4 per cent of the feed fish per day.

Not all the animals produce their excreta or faeces together, nor at the same rate, thus the actual chlorine demand varies. Dolphins seem to produce most of their waste products over a 12 hr period and seals probably over about 3-6 hr. In practice, the day time chlorine dose needed, in kg/hr, has been found to be about twice the night time dose.

In an open circuit system, the rate of replenishment with new water should be arranged to dilute the pollutants, calculated as above, to less than 0.10 mg/l. For normal water to animal ratios, this results in a change of tank water every 1 1/2 to 2 hr, depending on the efficiency of the tank circulation. However, in a close system, the tank water must be changed with sufficient frequency to ensure good mixing of the water in the tank and the desired clarity, together with satisfying the peak chlorine demand of the tank when the concentration of chlorine in the recycle flow is limited to about 3 mg/l.

Usually a well-defined, animal-to-water ratio exists; in the case of dolphins it is usually one dolphin per 91 m3(20,000gal). If this ratio is exceeded, it becomes extremely difficult to maintain free chlorine residuals without the combined residual rising to values which burn the animals. It has been found that the combined residuals should be maintained at about half the free ones and preferably below 0.5 mg/l. During repair work on the main dolphin tank at Port Elizabeth, five animals were kept for one month in a 91 m3 tank, with the water recycled through a filter every 30 min. Dosing chlorine at a rate of only 1.36 kg/hr compared with theoretically required 1.72 kg/hr produced combined and free residuals of 5 mg/l. Increasing the dose increases both residuals, without the combined residuals falling. The phenomenon has been reported by other oceanaria and insufficient information exists to determine the causes.

Phosphate control appears to be necessary in some mammal tanks, although dolphins have been kept for long periods in water containing up to 9 mg/1 phosphate (as PO.). Phosphates may accelerate algae growth and should be kept at low levels. Phosphate levels fluctuating up to 5 mg/I have been present in the Hong Kong tanks, whereas the maximum level in the Port Elizabeth tanks has been 1.5 mg/I; both systems use ultra-high rate filters. It is known that the feed fish have about 0.5 per cent phosphates which, if untreated, would represent daily accumulative concentrations of 0.075 mg/I and 0.19 mg/1 in the Port Elizabeth and Hong Kong tanks respectively. However, the levels seem to fluctuate. Recent experiments in Hong Kong indicate that extension of the aluminium sulphate dosing, presently used only to condition the filters, may be effective in reducing phosphate concentrations in closed systems.

If aluminium sulphate dosing is continually used, the residual aluminium level should be controlled to below 0.30 mg/l. Dolphins do not drink sea-water (Ridgeway 1972), but aluminium concentration higher than 0.30 mg/l could affect their eyes.

If insufficient primary treatment or tank chlorine residuals are used algae growth is likely to occur in outdoor pools. Where draining and cleaning is impractical, dosing of copper sulphate chelated with citric acid in the ratio 1:2 respectively to give a residual of 0.5 mg/1, has proved effective in Japan and the U.S.A.

MAMMAL FILTRATION

A test facility comprising 545 m' and a single, horizontal pressure sand filter operating at a conventional rate of about 11.7 m/hr (240 gal/ft'/hr) was set up in Hong Kong, from which it was noted that the suspended solids load on recycled mammal tanks was quite low. Consequently, ultra-high rate (23.5 to 39 m/hr (480-800 gal/ft2/hr)) pressure sand filters were investigated, only to find that although several types were available, head losses were relatively high and ranged from 1.4 to 3.5 kgf/cml (20-50 lb/in2). However, one manufacturer offered ultra-high rate filters, using a bellmouth water in-let and a high density nozzle underdrain system, which gave head losses similar to conventional filters using these as a comparison of the two types of filter was possible. Results showed that both produced water in the tanks of turbidities between 0.20 and 0.35 JTU, giving under-water viewing distances of over 45 m. Filter operating differential pressures for the ultra-high rate and conventional rate filters are similar being 0.21 kgf/cm2 clean and 0.49 kgf/cm2 dirty at 37.7 m/hr (770 gal/ft2/hr) with filter runs of 2 days. The total recycle pumping head was about 16.8 m (55 ft) with the ultra-high rate filters and about 15.2 m (50 ft) with the conventional ones. Accordingly, ultra-high rate filters have been adopted for new mammal systems installed.

Irrespective of the filter type, aluminium sulphate dosing at 10 rng'I for 1 hr after backwashing is used to assist in filtering.

CHEMICAL PLANT

The chemical plant for mammal tank systems is generally similar to that used in conventional waterworks practice, although the chemicals are mixed in sea-water.

Chlorine dosing can be by gaseous chlorine, sodium hypochlorite or electrolytic chlorination. Experience has shown that the latter method of dosing is far superior to the others, because it is the cheapest to operate and requires a minimum of associated pH correct. However, the efficiency of these units is dependent on salinity and feed water temperature; thus careful study is needed before selecting this method.

The electrolytic chlorinators in use produce chlorine at about 11.5 kWh/kg at a rate of 5.6 kg/hr and at 12.7 kW/kg/hr at 3.7 kg/hr, the kVA required being 14.0 kVA/kg/hr and I5.6 kVA/kg/hr respectively. They were manufactured in the UK, but the construction materials were altered to ensure that they conformed with those materials found acceptable for use in oceanaria. Fears that the magnesium hydroxide, formed as a by-product of the electrolysis, would interfere with the tank clarity, have proved unfounded.

RECOMMENDED WATER QUALITY FOR MARINE MAMMALS

Some countries are now introducing minimum standards of water quality permissible for the keeping and rearing of marine mammals, and it is hoped that the UK will soon pass similar legislation.

Table II suggests suitable water quality and holding conditions for marine mammals in sea-water closed-circuit systems.

THE TREATMENT OF WATER FOR FISH

The treatment of water for fish tanks relies on totally different principles to those for marine mammals, because most chemicals are toxic to fish at very low concentrations. Accordingly, treatment of fish tank water must rely on the bacteria of the nitrogen cycle in the absence of chlorine.

Most aquarists agree that filtration should take place at a rate of 3 m/hr (60 gal/ft2/hr) to accomplish acceptable results, when the tank is stocked with fish in the ratio of 1 kg fish to 1 m3 water. If these criteria were used to design a 2,000 m3 tank, the cost of filtration alone would be uneconomical.

Accordingly, a 182 m3 (40,000 gal) closed circuit fish tank was constructed using a 18.6 nil pressure sand filter operating at a rate of 7.3 m/hr ( 150 gal/ft2/hr); the recycle time of the system was set at 1 1/2 hr. This filter rate was chosen as a result of pilot tests conducted in Japan which suggested that the efficiency of the nitrifying bacteria fell rapidly at a filter rate exceeding 8.8 m/hr (180 gal/ft2/hr).

The pilot facility was steadily stocked with an increasing concentration of fish, and daily analysis of the water was made to determine the NH3-N, NO2-N, the oxygen consumption of the filter, and the turbidity of the water. Results appear in Fig. 4 which are typical of those obtained in over two years operation. The tests confirmed that the bacterial population would become established in a pressure filter, would perform adequately at the higher filter rate, and, provided that only half of the divided bed filter was washed at a time, would reseed themselves within two to three days of backwashing the filter at conventional air scour and wash water rates. The NH3-N concentration in the tanks remained generally at about 0.02 mg/1 with peaks of 0.03 to 0.05 mg/1 after backwashing. These levels of NH,-N have been low enough to permit successful rearing of invertebrates, such as anemones.

By September 1975, the fish loading was about 1.67 kg of fish per rn' which made a total load of 300 kg of fish. The fish are fed an average of 2 per cent of body Height of feed fish which amounts to 6 kg/day having a nitrogen value of about 0.22 kg/day. Theoretically 1.98 kg of oxygen is needed to completely oxidize each kilogram of nitrogen. Thus the oxygen consumption could be expected to be about 1 kg/day. The average oxygen depletion measured across the filter was about 0.30 mg/l in a flow of 2.27 m'/min which gives an actual oxygen depletion of 0.98 kg/day.

In September 1976, the tank loading was about 200 kg of fish and the oxygen depletion in the flow reduced to about 0.25 mg/1. The amount of oxygen expected on the same basis as above would be about 0.71 kg/day, and that actually measured across the filter "as about 0.82 kg/day.

Aritsune Sacki (1953) found in his tests that the amount of nitrogen (TKN) excreted he fish is about 50 mg/100g/day, of which about 50 per cent is NH3-N, and that sufficient bacteria would be present in the filter sand provided that the weight of sand present was 30 times that of the fish. If these figures are used to calculate the amount of oxygen needed, the, indicate that about 60 per cent more oxygen is being used. However, Saeki's figures referred to the actual nitrogen output and did not allow for uneaten feed fish; nor did they apply to sharks which have a notoriously high urea output.

Fry showed that the rate of oxygen consumption by fish was mainly dependent on fish activity, size, and on the concentration of dissolved oxygen in the water. He shoulded oxygen consumption figures of between 100 and 400 nil 02/kg fish/hr; thus aeration of the fish tank water is probably the most important single factor which affects the achievement of the successful operation of the system. Aeration is required both before and after filtrations, to ensure sufficient oxygen is present for the fish and the bacteria in the filter. As outdoor fish tanks tend to support algae, sufficient oxygen must be present to replace that removed by the algae at night time.

The turnover time of large fish tanks should be selected to ensure that adequate oxygen is always available to satisfy the above demands. Turnover times of about 1 to 1 ½ hr normally suffice.

Disinfection of the water for the fish tank is considered essential to avoid bacterial infection and must be accomplished without leaving a toxic residual in the water. If chlorination is used, then complete dechlorination is needed before the water is returned to the tank. Even if activated carbon dechlorination is used, a risk exists that accidental residuals may pass into the tank; this risk is unacceptable for large reef tanks due to the cost of replacing the specimens if they are accidentally destroyed. Ozone is possibly suitable but there are many drawbacks to operating such plant by inexperienced staff, and to operating in the humid atmospheres of the tropics.

Ultra-violet irradiation used in a pilot fish tank has produced excellent results. The bacteria kill was high and has remained so, provided that the units are regularly cleaned. The pilot plant used conventional units with the ultra-violet tubes mounted in quartz sleeves. However, while these units are successful for small systems, the maintenance involved prohibits their use in large systems. The units to be used to treat the 32,000 m3/day flow to the reef tank at the Hong Kong oceanarium, irradiate the water while flowing in open channels, the lamps being fitted to the roof of the channel. Access to the lamps for cleaning is made easier. The end-of-life lamp exposure rating of the ultra-violet system in use and of that being installed in the main tank system is 33,000pW,,crn'.

The control of pH for large fish tanks is generally accomplished by using dissolved calcite blocks.

Temperature control of the water is also required if the delicate multi-coloured reef fish are to be kept, because these fish are extremely sensitive to temperature variations. The water in the tanks should be kept at about 21°C ± 1.1°C (81°F ± 2°F) and this must be achieved by means of low temperature difference heat exchangers, as the sea-eater cannot safely be heated much above 38°C (100°F) without risking accidental scaling up of the heat exchanger passages with sulphate and/or carbonate scales. In some cases, winter heating and summer cooling will be needed.

CONCLUSIONS

It can be seen that the treatment of water for mammals and fish requires relatively, large plant which is expensive to purchase and operate; yet little is known of the chemical reactions which occur in the tanks. More study is needed to find out why difficulty is experienced in maintaining chlorine residuals when the animals are overcrowded, and whether or not chlorine residuals are being measured in the tanks, or whether, in reality. some combination of chlorine, bromine, chloramines and bromamines is being measured.

New analytical techniques are needed so that simple and cheap comparators or other field test instruments can be produced. These will enable measurement of many "ater quality parameters in the plant which presently require advanced knowledge, or laboratory equipment which are not retained by oceanaria.

Certainly, the quest for reduction in the capital costs of projects will create the need for filter research resulting in ultra-high rate filters operating at rates above 50 m/ hr (1.000 gal/ft2/hr) with the minimum increase in operating headlosses.

References

Ridgeway, S. 1972 Charles Thomas, Illinois, U.S.A., "Mammals of the sea".

Murphy, K. 1975 International Ozone Institute, "The use of ozone in recycled oceanarium water", Aquatic applications of Ozone.

Saeki, A. 1958 Bulletin of the Japanese Society of Scientific Fisheries, std. 23. no. "Studies on fish culture in the aquarium".