The introduction of cage or pen culture to a water body has an impact on the environment which can lead to conflict, since inland waters are often, and increasingly so, under pressure from other users and for a wide variety of purposes. It can induce the operation of negative feedback mechanisms which restrict the number of units, determine the type of species grown, and limit production. The establishment of cage and pen farming operations in a lake, reservoir or river can also have an impact outside the immediate vicinity of the site, by its demands for construction materials (Cariaso, 1983) (see Section 3.3.3 below).
Enclosures can affect water bodies both by their physical presence at a site and by the changes they can induce in the physical, chemical and biological characteristics of the water body through the method of culture (extensive/semi-intensive/intensive) and species used.
Cage and pen structures affect a water body in three principal ways: they take up space, thus potentially competing with other users; they alter flow regimes which govern the transport of oxygen, sediment, plankton and fish larvae; they have an impact on the aesthetic qualities of the site.
Enclosures can compete with lake and river fisheries for space. Stationary cages and pens, for example, are restricted to shallow areas, which are 7m or less deep, and this approximates to the littoral region of most lakes and reservoirs where rooted emergent and submerged vegetation occurs (Goldman and Horne, 1983). Such areas are important as spawning grounds for commercially important fishes such as the phytophilous cyprinids and pike (Braum, 1978), and the substrate spawning tilapias, T. zilli, T. rendalli and O. macrochir (Ruwet, 1962; Philippart and Ruwet, 1982). Inshore areas of vegetation where predators can be avoided are also important nursery grounds for fry and juveniles of many species.
In Laguna de Bay in the Philippines, pen and cage culture of milkfish and tilapias was introduced in the late 1960s (PCARRD, 1981). Since then these industries have boomed, despite the ravages of periodic typhoons and fish kills (Fig. 5). The LLDA has attempted to regulate the industry and avoid conflict with fishermen and local villagers by trying to limit production within certain areas of the lake designated as a fish pen belt by a series of laws (Republic Act No. 4850, Presidential Decree No. 813, and Resolution No. 9, 1976; Agbayani, 1983). The fish pen belt provides for other interests by leaving free a fish sanctuary area, where no fishing or pens are permitted, and by utilising a 15,000 ha area (17% of the lake surface) which is at least 200m from the shore and yet does not interfere with navigation routes (Fig. 6). Thus access to inshore areas, open waters and fish landing sites should have been protected.
However, in the last 3–4 years, there has been a rapid proliferation of fish pens outside the legal fish pen belt (Figs. 7 and 8). The use of cages to culture tilapia is not covered by existing laws (Agbayani, ibid.), and although still of relatively minor importance (100 ha; Guerrero, 1983), they are increasing. Current estimates of the area covered by cages and pens is 34–40,000 ha (38–45% of the lake).
Many of these illegal enclosures were sited in traditional fishing grounds and snail-gathering areas, and blocked the main navigation routes to the fish landing sites (see Fig. 8). In 1982 and early 1983 the widely reported conflict (theft, vandalism, killings) between the local fishermen and the fish pen owners, most of whom live outside the Laguna de Bay area, had escalated. Following public pressure, an aerial survey of the lake was carried out by the Philippine Air Force in April 1983, and when the extent of the proliferating pen industry was realised, existing regulations were enforced.
Not all impacts of enclosure structures may necessarily be negative. The attraction of fishes to free-floating and anchored objects has been widely reported from all over the world, both in freshwater and marine situations. Many of the theories proposed to explain these phenomena apply to the effects of cage and pen structures on wild fishes, and are summarised in Table 3. In Laguna de Bay, the numbers of the indigenous catfish, kanduli (Arius manilensis), had been declining for some years prior to the establishment of the cage and pen industry, due to pollution and overfishing (Santos, 1979). However, the enclosures have apparently provided shelter for the fish, thus allowing the population to recover somewhat (Guerrero, 1982a).
Snails (Lymnea and Amnicola spp.) which are intensively harvested for feed for the local duck industry (Arriola and Villaluz, 1939), are reported to have increased to high densities within the protection of the pens (Guerrero, 1982a). However, there is also evidence that these enclosed snail populations fluctuate enormously, because of unchecked growth and recruitment, and restrictions on emmigration from the area (S. Vivar, pers. comm.). Whilst some fish pen owners do permit snail gatherers to harvest their pens for a fee, others refuse, arguing that the type of fine-meshed gear in use on the lake is highly destructive and disruptive.
The flow of water through enclosures is affected by drag forces exerted by the framework and netting (Inoue, 1972; Wheaton, 1977; Milne, 1979; Wee, 1979). The reduction in flow is dependent upon a number of variables including flow rate and density of water, enclosure size and shape, mesh type (knotted/knottless, diamond/square) and material, degree of fouling, and stocking density (Milne 1970, 1979; Inoue, 1972; Wheaton, 1977, Wee, 1979; Kils, 1979). The coefficient of drag (Cd) exerted by knotted and knottless netting is related to nominal mesh size (a), and diameter of twine (d) by the following equations (Milne, 1970):-
Cd = 1 + 3.77 (d/a) + 9.37 (d/a)2 knotted net
Cd = 1 + 2.73 (d/a) + 3.12 (d/a)2 knottless net
Cd is greater for knotted than knottless mesh, and courlene and polythene have smaller Cd values than nylon or ulstron (Milne, ibid.).
Inoue (1972) noted that the current velocity inside a large (20 × 20 × 6m) cage of 5cm mesh size, stocked with fish at 1.6 kg m-3 fell to only 35% of the current speed recorded outside the cage, and he also demonstrated that when cages were located parallel to the direction of current, flow rate in successive cages fell.
Cage and pen structures, therefore, can have a considerable impact on local currents, and this has a number of implications. Sediment transportation in an aquatic system, although influenced by a number of factors, is principally determined by current flow (Smith, 1975; Gibbs, 1977). Significant reductions in flow, as can occur in some enclosure systems (see above), would cause the sedimentation of larger, denser particles in the immediate vicinity of the cages and pens. A sudden increase in the rate of sedimentation in an area would disrupt benthic communities (Brinkhurst, 1974) and accelerate filling in (ageing) of the water body, which could interfere with navigation. Siltation in the vicinity of cages and pens has been reported from Egypt, India, Malaysia, Singapore, Sri Lanka and Thailand (IDRC-SEAFDEC, 1979).
Siltation problems caused by enclosures are most likely to occur in rivers and in areas of lakes where large rivers flow in. Here the dispersion of the sediment carrying plume, which is determined by the horizontal water current speed (Csanady, 1969, 1975) could be severely disrupted.
Of more importance, however, are the effects of reduced current on the fish culture operation. The flow of water through the system governs the supply of oxygen and the removal of toxic waste metabolites from the vicinity of the fish, and in extensive and semi-intensive culture, it also controls the supply of planktonic food.
The introduction of cages and pens to a water body can transform its appearance (Fig. 4). In many countries, provision is made within conservation laws to preserve areas of outstanding natural beauty, and protect them from unsightly developments. The proposed establishment of a large floating cage fish farm in Loch Lomond, Scotland, by a private company was objected to by a number of people who thought that this would detract from the scenic value of the area, reduce tourist numbers estimated at 2 million per year, and ultimately affect local employment and incomes (Beveridge and Muir 1982). Similar objections have been voiced over cage culture developments in Hong Kong.
The introduction of cage and pen culture to inland waters can cause a number of changes to both the biotic and abiotic components of the environment. Although intensive, semi-intensive and extensive methods of culture have impacts which differ both qualitatively and quantitatively from one another and will therefore be considered separately, there are a number of factors common to all methods, and these will be examined first.
An enclosure is more of an open fish rearing system than land-based ponds, raceways or tanks, and there is a far greater degree of interaction between the caged or penned fish and the outside environment than occurs in other systems. In recycle systems, only 1–20% of the daily water requirements are replenished (Bryant et al 1980; Muir, 1982; Muir and Beveridge, in press). Incoming water passes through settlement tanks and filtration systems which effectively remove all bacteria, protozoa and plankton, and of course larger organisms such as fish. In some operations, the recycled water is treated by U.V. which kills most of the virus particles and remaining bacteria in the system (Spotte, 1979; Muir, 1982). Thus there is little opportunity for organisms to enter and influence the system from outside, and the fish are cultured in an environment where both the abiotic and biotic components are highly controlled.
In earth and concrete ponds, the fish are fully exposed to the vagaries of climate (sunlight, temperature etc.), and there is also a degree of interaction between the cultured fish and other organisms. Usually, only coarse screens and/or settlement ponds are used, which help prevent fishes (eggs, fry, adults) from entering the system (Hepher and Pruginin, 1981). However, microscopic and macroscopic organisms such as viruses, bacteria and fungi, and phytoplankton, zooplankton and insects can be carried unimpeded into the ponds in inflowing water. Birds and other vertebrates also have relatively free access to ponds and raceways unless elaborate trapping or other preventative methods are used (Meyer, 1981; Martin, 1982).
The establishment of recycle systems and ponds and raceways to grow fish is the creation of a new environment. However water usually only passes through pond and raceway systems once, and the consequences of changes to the water through fish culture is experienced where the effluent is discharged. Outflows from land-based systems of course can be treated by passing water through various settlement pond and filtration systems until acceptable standards are reached (Warrer-Hansen, 1982; Muir, 1982a). By contrast, enclosures use existing environments to grow fish. Cages and pens must thus be regarded as subcomponents of the aquatic ecosystems in which they are sited, since the enclosure and the surrounding environment are intimately related i.e. changes occurring in the water body will have an effect on the enclosure environment, and vice versa. There is little opportunity to treat wastes emanating from cages. Although various methods of waste collection and removal have been developed on an experimental basis (Tucholski et al, 1980, 1980; Tucholski and Wojno, 1980) the costs involved would prove prohibitive to the industry. These differences between land and water based systems have a number of important implications.
There are five main groups of organism which cause disease in fishes: ectoparasites and fungi, endoparasites, bacteria, viruses and organisms which produce toxins leading to fish deaths (Sarig, 1979). The occurrence of disease outbreaks in fish farming is usually asociated with bad husbandry, since the disease-causing organisms are often ubiquitous and cause few problems until the fish are stressed through inadequate dietary or environmental conditions (Wedemeyer, 1970; Snieszko, 1974; Roberts and Shepherd 1974; Shepherd, 1978). In wild fish populations, mass mortalities are rare and are also usually linked to external stress factors (Shepherd, ibid.), since the fish and the disease causing organisms are usually in a state of balance. For example, although many parasitic infections are known in wild tilapias, there is little evidence of clinical effects and thus it would seem that the presence of parasites is a normal occurrence of little significance (Roberts and Sommerville, 1982). Studies of the adult cestode Eubothrium in rainbow trout show that 1–5 parasites per fish have no effect on either nutrient absorption or fish growth (Ingham and Arne, 1973).
However, the introduction of large numbers of fish in enclosures to a system can have a dramatic effect on disease agents. Diseases from outside the enclosure site can easily be introduced by transporting fingerlings/fry from other areas in the country, or importing fish from abroad without proper precautions being taken (Avault, 1981; Mills, 1982). The danger of the spread of fish diseases in this way is widely recognised, and is currently giving cause for concern (Rosenthal, 1976; Roberts and Sommerville, 1982). In a recent survey of the ecto and endoparasite fauna of cage and wild fish communities in a Scottish loch, Sommerville and Pollock (1984 in prep.) have shown that the numbers and species of parasite present in the wild fish differ markedly from expected, and concluded that this was a result of the intensive culture of rainbow trout in the lake. Although some of the parasites may have been imported with fingerlings used in stocking, yet others may have been present in the wild fish and only reached abnormal levels due to increased densities of fish and changes to the environment subsequent to the introduction of cages.
Unfortunately, little is known about the transmission of parasites from cage to wild fish, or vice versa. However, in several cases in the U.K., cage fish have become severely infested with the cestodes Triaenophorus nodulosus and Diphyllobothrium spp. resulting in heavy mortalities, and the eventual closure of one farm (Wootten, 1979; Jarrams et al, 1980). Those infections were attributed to the wild fish populations which were subsequently found to be carrying the parasites.
Data from Matheson (1979) showed that Atlantic salmon parr raised in cages in a freshwater loch in Scotland became heavily infected with D. ditremum and D. dendriticum within two months of being introduced to the site. Surprisingly, the parasites were not isolated from the brown trout (S. trutta) in the loch, although only a few specimens were examined.
A survey of fishes in a Scottish lake site prior to the introduction of cages showed no endoparasites present, and one month later cages of rainbow trout were introduced. The stocked trout were also examined and found to be free from parasites. However, two months later mass mortalities of fish were reported, and on examination the fish were found to be heavily infected with Diphyllobothrium spp. (Sommerville, unpublished data). Phillips et al (1983) believe infestation was precipitated by inadequate use of feed, which caused the caged fish to ingest copepods, the intermediary hosts for these cestodes.
The role of increased nutrient levels often associated with intensive cage culture (see below) in promoting proliferation of parasites is not clear. However, Grimaldi et al (1973) have shown that Phycomycetes saprolegnioles infections of trout are widespread in eutrophic lakes in central Italy and northern Switzerland. Eutrophic conditions may also favour increased production of intermediary hosts (e.g. crustaceans). Recent work carried out by Soderberg et al (1983) in North America, has shown that exposure of rainbow trout to high levels of free ammonia, such as can exist in intensive culture conditions, predisposed the fish to succumb to parasitic epizootics.
Few such parasite or disease problems have been reported in cage or pen culture of fishes in the tropics, although Vass and Sachlan (1957) reported the presence of gut parasites in common carp grown under extensive conditions in a polluted Indonesian stream. These parasites were believed to have been transmitted from human faeces.
Cages and pens of fish seem to act as a magnet to a wide range of both obligate and facultative fish-eating vertebrates. The range of species reported to cause problems at cage and pen farms is listed in Table 4, and includes fish, reptiles, birds and mammals. Many of these species move into an area where a fish farm has been established, attracted by the large numbers of readily detected fish and also by the bags of commercial feed occasionally left unprotected on the cage walkways. Even comparatively rare species, such as the osprey (Pandion haliaetus) in Scotland will travel considerable distances in order to visit a fish farm. Seasonal and diurnal changes in numbers of predators have been noted (Ranson and Beveridge, 1983).
So far there has been little serious evaluation of the impact of these predators either on the environment, or on the enclosed fish. Ranson and Beveridge (1983) concluded that although herons (Ardea cinerea) and cormorants (Phalacracorax carbo) frequently attacked caged rainbow trout, these attacks were rarely successful. An examination of stomach contents of birds from the farm showed no evidence that any of the fish came from the cages, and this conclusion was supported by many hours of observation. However, 0.5% of all caged fish showed evidence of bird damage, which could lead to secondary bacterial or fungal infection.
Damage to nets by unsuccessful predators such as birds, turtles, monitor lizards and rats has been reported from several cage farms (Table 4), thus contributing to the heavy losses of fish from enclosures reported by Secretan (1979). Predation of wild fish may increase through the attraction of predators to the enclosure site. Ranson and Beveridge (1983) recorded 11 perch (Perca fluviatilis) removed from a cormorant stomach at a cage fish farm. Another serious, although as yet little studied, impact of the immigrant predator population, is their contribution to disease. In the example described in Section 3.3.1.1 above, the rapid spread of Diphyllobothrium to caged rainbow trout within two months of a farm being established may in part be due to the observed migration of large numbers of gulls (Larus sp.) into the area. Certainly both birds and mammals play important roles in the life cycles of many commercially important endoparasitic fish diseases. For example, birds act as intermediate host in the life cycle of the nematode Contracaecum, and piscivorous mammals such as the otter may act as final host for the digenean Haplorchis, both common parasites of tilapia (Roberts and Sommerville, 1982).
Caged and penned fish frequently escape through netting or mesh damaged by predators, floating objects, or rough weather (Secretan, 1979), and in this way foreign or exotic species can be introduced to an environment. In any commercial cage or pen operation it is inevitable that some fish escape. In one lake in Poland, Penczak (1982) estimated that 4 tonnes of trout escaped in one year. There are many records of the impacts of escaped or deliberately transplanted fishes on indigenous fish stocks, and these include the extermination of local fishes through predation or competition, interbreeding with native fishes and adulteration of the genetic pool, habitat destruction and the outbreak of disease epidemics (Rosenthal, 1976; Mills, 1982).
In Laguna de Bay, typhoons often cause considerable damage to fish pens (PCARRD, 1981). In 1976, 50% of the fish pens were totally destroyed, resulting in the release of millions of milkfish to the lake (Gabriel, 1979). This boosted open water fishery catches tremendously in the weeks following the disaster.
In the U.K., ferral rainbow trout which had escaped from cages were found to be breeding in feeder streams to the lake. Examination of the gut contents showed that the rainbow trout and native brown trout fry in the streams had similar diets, and therefore could be competing. Angling catch returns from the lake demonstrated that brown trout returns, which had declined to a low level many years previously, remained low after the introduction of the cages, whereas the catches of rainbow trout increased each year due to escapes (Phillips, unpublished data).
The use of chemicals and drugs in pond, tank and raceway fish farms to control disease is widespread, particularly in intensive units in North America, Europe, Israel and Southeast Asia (Bardach et al, 1972; Brown, 1977; Hepher & Pruginin, 1981; Alabaster, 1982a). In the most extensive surveys to date, carried out in Europe and the U.K., Alabaster (1982a) and Solbe (1982) found that most pond, tank and raceway farms used small amounts of chemicals (especially malachite green and formalin) from time to time to treat ectoparasitic and fungal infections. A wide range of antibiotics, such as aureomycin, furazolidene, nitrofurazone, penicillin, oxytetracycline, sulpha-merazine and terramycin are also occasionally administered to fish in their food.
However, there is very little quantitative data on the frequency or pattern of use of chemicals and drugs in cage and pen farms. Treatment is costly and difficult due to water flow through the enclosures which can rapidly dilute the chemical used, and render treatment ineffective. The addition of large quantities of chemicals to compensate can make treatment too costly. To minimise expense, many farmers enclose cages in polythene sheeting to try and reduce the flow rate, although this is highly labour-intensive. Alternatively, they transfer diseased fish to a specially modified enclosure or tank, thus minimising waste (i.e. loss to the environment) of chemicals. For these reasons, it is probable that chemicals on enclosure farms are employed much less frequently than in other systems, and the resultant additions of foreign substances to lakes etc., is small.
Intensive culture of fishes in enclosures, as discussed in Section 2 above, is at present largely restricted to lakes and reservoirs in temperate regions, where the principal farmed species are salmonids, carp and catfish. Early laboratory studies by Murphy and Lipper (1970) and Liao (1970) demonstrated that the intensive culture of fish resulted in high levels of waste production per unit live weight, compared to other livestock such as chickens, swine or cattle. As the cage industries developed and expanded in the 1970s, concern about the potential polluting effects grew.
The first studies of environmental impact of intensive cage culture were in the United States, where the Arkansas Game and Fish Commission had begun leasing areas of state-owned lakes and reservoirs to commercial catfish and trout producers in the early 1970s (Eley et al, 1972; Newton, 1980). As their lease programme expanded, a number of studies were commissioned (Eley et al, 1972; Kilambi et al, 1976; Hays, 1980). In Eastern Europe, intensive culture of common carp and trout has been practiced in lakes used as cooling ponds for heated water from power stations since the mid-1960s (VNIRO, 1977), and the water quality of several of these lakes has been monitored over a period of years (Korycka and Zdanowski, 1980). Recent studies in Poland have been concerned with the waste output of caged rainbow trout in reservoirs (Tucholski et al, 1980a, 1980b; Penczak et al, 1982).
In the U.K., two study programmes on the environmental impact of intensive cage rainbow trout farming are in progress at the Institute of Aquaculture, University of Stirling. One is a long term monitoring programme of water quality at a highly developed commercial site in a lowland reservoir, and the other concerns the study of environmental impact of intensive culture on the more typical dystrophic-type of lake that is currently being developed for cage culture in highland Scotland (Phillips et al, 1983). A desk study has also been completed on the impact of proposed cage culture developments on Loch Lomond, an important natural reservoir and Site of Special Scientific Interest in Central Scotland (Beveridge and Muir, 1982; Beveridge et al, 1982).
Several other study groups are currently studying the problems at the University of Lund, Sweden (Enell, 1982; Anon, 1983), and at the Department of Agriculture and Fisheries, Scotland, in Pitlochry (R. Harriman, pers. comm.).
In general, these studies fall into two categories. In some investigations, comparisons have been made between the environment at the cage site, and at a control site some distance from the cages, whilst in others, a study of the site prior to the introduction of cages, and during and after the period of culture have been carried out (Kilambi et al, 1976; M. Phillips, pers. comm.). The latter type of study, is preferable, but involves planning, long term commitments of manpower and resources, and of course greater capital than a short-term study.
However, irrespective of differences in methodology, species cultured and size and type of site, most studies have recorded increases in the levels of suspended solids and nutrients (alkalinity, total-P, PO4-P, NH4-N, organic N,C) and decreases in O2 in and around the enclosures (Table 5). In the sediments below cages, considerable increases in oxygen consumption and in the total-N, total-P and organic content of the muds has been recorded (Tucholski et al, 1980; Enell, 1982; Merican, 1983). (See Fig. 9).
Changes in the flora and fauna of inland waters associated with enclosure culture were first noted by Vass and Sachlan (1957), who investigated the effects of extensive carp culture on stream biota. More recent studies of intensive systems in temperate countries have noted quantitative and qualitative changes in bacteria, protozoa, plankton, benthos and fish (Table 5).
Changes in fish communities at intensive culture enclosure sites are inevitable, not only because of the high probability of fish escaping from the enclosures and the risks of disease introduction or escalation, but also because of the release of nutrients and loss of feed to the environment associated with intensive operations. Feed losses have been reported by many authors (Collins, 1971; Eley et al, 1972; Coche, 1979; Muller and Varadi, 1980; Beveridge and Muir, 1982; Penczak et al, 1982), and are dependent upon quality and type of food (wet/dry, floating/sinking), method of feeding (hand/demand feeders/automatic feeders), enclosure design (cage/pen; presence/absence of feeding ring; solid/mesh cage bottom) species, site characteristics (lotic/lentic; sheltered/exposed), and stocking density (high/low). Loss of feed to the environment is sometimes increased by the currents generated inside enclosures by feeding fish (Collins, 1971; Coche, 1979).
Wild fish have been observed in comparatively high densities in the immediate vicinity of fish cages (Collins, 1971); Eley et al, 1972; Loyocano and Smith, 1976; Hays, 1980). Using telemetry, Ross, Phillips and Beveridge (unpublished data) followed the behaviour of ferral rainbow trout in a Scottish loch where intensive cage culture was in operation, and found that during certain periods of the year at least, the fish spent comparatively long periods of time near the cages. Phillips (1982, 1983) has shown that fish can learn to come to a feeding station in a lake in response to an acoustic signal, and it may be that the feeding response of the enclosed fishes acts as a signal to the wild fish that food is available.
Growth rates, and abundance and survival of fish in some lakes and reservoirs where intensive culture is practiced, have been shown to increase (Loyacano and Smith, 1976; Kilambi et al, 1978; Hays, 1980) and although this is in part due to intake of commercial feed, it is also due to the effects of increased nutrient levels. Fish growth has been found to increase in many temperate water bodies following fertilisation (Weatherly and Nicholls, 1955; Munro, 1961). However, in other lakes the intensive culture of fish in cages has resulted in a decrease in the natural fish population (Penczak et al, 1982). All water bodies have characteristic fish communities which are dependent upon the trophic state (Vaughn et al, 1982) and changes in the trophic state will cause the fish community to change (Welch, 1980).
Negative feedback from changes in water quality on the growth and survival of caged fish have been reported from many intensive cage farms. In Lake Kasumigaura in Japan, the intensive cage carp industry has been affected by deteriorating water quality caused in part by the fish culture operation itself (Kitabatake, 1982). Interestingly, those farms using automatic feeders had significantly higher mortalities than those which practiced feeding by hand. In Scotland, off-flavours in the cultured fish, associated with the presence of high levels of blue-green algae, have been recorded from a eutrophic cage rainbow trout farm site (A. Stewart and A. Hume, pers. comm.).
Commercial extensive culture of fishes in enclosures is restricted to tropical and subtropical countries, where fish such as milkfish, tilapias and carps can be grown without recourse to the use of supplementary feeds. In the Philippines, tilapia production of up to 2 kg m-3 month-1 has been attained in this way (Table 6). The exploitation of inland waters for extensive culture follows a typical pattern: following the first and usually highly successful harvest of fish, existing entrepreneurs expand production, and other operators move in. Within a few years, there is a considerable number of both small and large operations. By the second or third year, the growth rate of the fishes has fallen, and fish farmers must either endure reduced production, or resort to the use of supplementary feeds. In both cases, economic viability is impaired and many producers may be forced to close. For those that remain, prospects usually improve (see Fig. 10).
There have been few studies which have specifically investigated the relationship between extensive cage culture and productivity. However, Henderson et al (1973), Melack (1976), Oglesby (1977, 1982) and Marten and Polovina (1982) have shown that there is a positive relationship between fishery yield and aquatic productivity, and a similar relationship seems to hold for extensive cage culture and productivity (see Section 4.4 for detailed discussion). In Seletar Reservoir, Singapore, where bighead carp have been stocked in cages since 1972 in order to combat the problems of eutrophication, there has been a steady decline in the frequency of algal blooms and plankton biomass corresponding to a decrease in fish production per unit area per unit time (Yang, in press). Aquino's study (1982) showed the growth rate of caged tilapia in Sampaloc lake in the Philippines was related both to gross primary production (Figure 11), and to visibility (i.e. related to algal biomass).
The best documented examples of the effect, of such feedback mechanisms on fish culture come from the San Pablo lakes area, Rizal Province, in the Philippines. Here five of the seven lakes are utilised for pen and cage culture of O. niloticus. At Sampaloc Lake, for example, the annual production by extensive means fell from 11.4 kg m-3 in the late 1970s to 1 kg m-3 in 1983 (Coche, 1982; Guerrero, 1983). Stocking density was reduced from 25 m-3 to 2 m-3, and the growing period to produce marketable fish (200g) increased from 4 months to 6–9 months, and many operators began to use supplementary feeds, such as rice bran, copra cake, pig manure and kitchen scraps (Aquino, 1982). Similar problems have occurred at most of the other San Pablo lakes, although to a lesser extent. For example, 150–200g fish could still be grown in Lake Calibato in 5–6 months in 1983 without resorting to the use of supplementary feeds.
Similar problems have also beset the semi-intensive enclosure culture industry. Experience from many Southeast Asian countries shows that a period of rapid and uncontrolled expansion usually follows the introduction of cages and pens to an area leading to increasingly heavier dependence on supplementary feeding. In Laguna de Bay, fish cage and pen operators are increasingly having to rely on supplementary feeds, whilst production has fallen from 0.18–0.36 kg m-3 month-1 to 0.12–0.14 kg m-3 month-1 (Mane, 1979; Lazaga and Roa, 1983).
Pens and stationary cages are commonly used in extensive and semi-intensive operations in Southeast Asia (e.g. Philippines, Indonesia, China, Thailand).
Construction can require large quantities of timber, such as bamboos (Bambusa spinosa), “anahaw” palms (Livistonia rotundifolia) and hardwoods. These materials have only a limited useful life before they must be replaced. Exposed parts usually deteriorate first through the combined action of heat, sunlight and rain, and the rigours of day to day use. Lifespan not only depends upon the climate and materials used, but also on the age and health (e.g. degree of insect damage) of the wood, and the maturation and preservation method used, if any (e.g. use of pitch) prior to installation. Details of materials and their characteristics are given in Table 7.
Despite the many advantages of using other materials (stronger, longer life etc) bamboo is still the most commonly used construction material for pens and cage frames in Southeast Asia due principally to its comparatively low price and availability. It normally lasts only 1–2 years, before being replaced (IDRC/SEAFDEC, 1979). In Laguna de Bay in the Philippines, some of the old wood is removed by local villagers for use as firewood, although most is left in the lake to decompose. Cariaso (1983) has estimated that a 1 ha pen could consume as many as 2000 bamboos and 100 “anahaw” palms, with an estimated weight of 600 (60?) tonnes (Cariaso, pers. comm.). However, the quantity of materials used per unit area decreases with increasing pen size. Provisional estimates of the wood used by the fish pens and stationary cages in Laguna de Bay are enormous, although the nature and impact of these materials on the environment is still being assessed.
As argued in Sections 3.2 and 3.3 above, water-based culture systems (pens and cages) differ from land-based systems (silos, ponds, raceways, recycle systems) in two important ways. In contrast to land-based production systems which are usually built on privately owned or rented land, cages and pens utilise lakes, reservoirs and rivers which for the most part are state owned. By definition these waters are publicly owned and should be managed for the benefit of the public. They can be used to generate hydro-power, as a supply of water for drinking, irrigation or industrial purposes, for fishing and for recreation. Many rural communities have developed around inland water bodies and depend on them for their livelihood. Thus their large-scale use for fish culture by privately owned fish farms will, unless carefully managed, lead to conflict of interests. Pen culture can cause more friction than cage culture, since pens are much larger and are usually owned by single persons or corporations, thus limiting the number of beneficiaries at a site. Because of the high investment necessary for construction outsiders are often involved, thus aggravating tensions.
Secondly, cage and pen culture systems are much more open than land based systems, and must be considered as subcomponents of the lake/river/ reservoir watershed ecosystem in which they operate. Interactions between the environment inside and the environment outside the enclosure occur with little restriction, and so changes in one part of the ecosystem inevitably have an effect on all other parts, to a greater or lesser degree.
The impacts of cage and pen culture are summarised in Figs. 12 and 13. Common to extensive, semi-intensive and intensive methods are the effects of the enclosures themselves on water flow, currents and sediment transport, and on space and aesthetics. In most situations, the most important of these impacts will be on space.
The siting of cages and pens within a water body and with respect to each other is of great importance, since the enclosed fish depend on water flow through the enclosures for food and/or O2, and to remove toxic metabolites (Schmittou, 1969; Awang Kechik et al, 1983). However, it seems that at least for extensive and semi-intensive operations, the optimum siting of enclosures within a water body is likely to maximise interference with other users, since the cages and pens should be widely dispersed (see Section 4.6 below). From the point of view of fishermen and other users, it would be best to restrict enclosures to particular areas and therefore in such multi-use water bodies there must inevitably be a compromise between parties.
Also common to all methods of cage and pen culture are the effects of enclosing large numbers of fish on local fauna - predatory birds and mammals, wild fish and especially parasites and other disease organisms. However, from both published accounts and from discussions with fish farmers and experts it must be concluded that such impacts are much less important in tropical freshwaters. For example, disease outbreaks appear to be almost unknown in warmwater cage and pen fish culture, unlike the mass mortalities which occur from time to time in temperate salmonid farms. One reason may be the paucity of data, especially from the tropics. Nevertheless, there is also some evidence to suggest that several of the more important cultured tropical fishes such as the tilapias, may be ‘tougher’ than the temperate salmonids and catfishes i.e. although tilapias may frequently harbour a range of potentially pathogenic organisms these rarely cause widespread mortalities. These apparent differences in disease resistance may in part be due to differences in culture methods. In Europe and North America intensive methods, associated with high stocking densities, are practiced in contrast to the usually less heavily stocked extensively and semi-intensively reared tropical fishes. Thus the methods practiced in temperate countries may be more stressful to the fish, resulting in suppression of the immune system, and increased susceptibility to infection (Wedemeyer, 1970; Snieszko, 1974; Roberts, 1979).
However, all evidence to date suggests that the method of culture has the greatest impact on the environment, since it directly affects nutrient concentrations, O2 levels, and concentrations of toxic metabolites (Eley et al, 1972; Penczak et al, 1982; Phillips et al, 1983) (Figs. 11 & 12). Disease organisms thrive in eutrophic conditions (Numann, 1972; Grimaldi et al, 1973; Lundborg and Lyndberg, 1977) and changes in water quality have also been shown to affect the amenity value of water for drinking purposes (Jones and Lee, 1982; Beveridge and Muir, 1982) recreation (Vaughn et al, 1982) fish production and fisheries (Henderson et al, 1973; Melack, 1976; Liang et al, 1981) and pressure from these interests may in turn curtail or restrict enclosure culture, as has happened in Laguna de Bay in the Philippines and Loch Lomond in Scotland. Negative feedback of changes in water quality on enclosure fish production have also been demonstrated (Aquino, 1982; Kitabatake, 1982).
In order to maximise fish culture potential, or calculate the relative costs and benefits of fish culture in a multiuse water body, the impact of extensive semi-intensive and intensive cage and pen culture must be quantified in water quality terms. Lack of this information has forced various agencies in both temperate and tropical countries to set development limits which, because they have been based on few data, have been viewed as somewhat arbitrary. This angers both fish farmers and opposing interests, and has resulted in several instances in their flagrant disregard.
