3. FISHERIES MANAGEMENT IN PRACTICE (Cont.)

3.3. Enhancement of fish populations and yields

3.3.1. Introductions

Current knowledge concerning introductions of fishes into each of the countries covered by the present survey was summarized by Welcomme (1988). This FAO publication lists, as far as is known, all the species introduced to each country, the waters into which they were released, the dates of their introduction and the source of supply, the purpose for which they were introduced (where this was done deliberately), and the outcome. The information contained in the report is being continuously updated in the database held at FAO headquarters, Rome. It is therefore not considered necessary to reiterate this large body of information here.

Introduced species now make a significant contribution to yields from inland capture fisheries, and especially from aquaculture, in all the Asian countries under consideration. Some introduced species have become such staple components of the fish fauna that their exotic origin has been forgotten. Perhaps the most outstanding example of this is the common carp in Southeast Asia. The native range of this species was Central Asia, China and Japan, but it is now a favoured food fish in Thailand, Myanmar, Laos, Cambodia, Vietnam, Indonesia and Bangladesh (as well as most other Asian countries). The Nile tilapia is perhaps progressing towards a similar status in some parts of the region.

Recently, concerns over the risks of possible environmental damage and consequences for indigenous species have perhaps made introductions less fashionable. Nevertheless, examples of recent highly beneficial fish introductions or transfers can still be found. One is the establishment of populations of the ice fish (also called silver fish or white fish on account of its milk-white colour) Neosalanx tangkuhkeii taihuensis (Chen) in Erhai Lake(244km²), Dianchi Lake (300km²), Fuxian Lake (212km²) and other large waters in Yunnan Province, China. This small (to 11cm and 3.5g max.) zooplankton feeder was introduced in the 1980s from its native habitat in Taihu Lake, Jiangsu Province. The species has since become one of the most valuable fishery products in the province, supporting an export industry to Japan, Hong Kong and Taiwan. The total catch of over 3,000 tonnes in 1994 had an export value of approximately US$30 million after cleaning, freezing in 1kg blocks and packaging in the eight specially built processing plants in Yunnan.

A few cases are known where species other than fishes were introduced with the aim either of increasing the food supply for fish or of controlling an undesirable organism biologically. Examples are the introductions to Thailand of water hyacinth weevil (Neochetina eichhoniae), the chrysomelid Agasicles hydrophila for control of alligator weed, and the noctuid moth Episammia pectinicornis for water lettuce (Bhukaswan and Chookajorn, 1988). In Indonesia, the cladoceran Daphnia similis was successfully introduced into Jatiluhur Reservoir in 1970 to enhance the poor zooplankton community as a food source for fish (Sarnita, 1983). However, the long-term usefulness and side-effects of these introductions are not well quantified.

3.3.2. Culture-based fisheries

The scope of activities

The governments of all eight countries included in this study operate hatcheries with the dual stated functions of providing fish fry/fingerlings for growout to market size in fish farms and for release to supplement wild populations in open waters. Where releases are of species already occurring and naturally breeding in the recipient waters it is extremely difficult to establish whether the stocking is having any significant effect on fish populations and catches. In some cases, government itself acknowledges that such activities are done mainly as a public relations exercise. For the purposes of this paper, therefore, culture-based fisheries are defined as those where all or part of the catch derives from releases of hatchery reared juveniles of species which do not normally reproduce in the recipient water body. The species most widely used in this way are the Chinese carps (silver carp, bighead, grass carp etc.) and Indian major carps (catla, rohu, mrigal etc.), but certain other species (e.g. Puntius gonionotus. which requires running water to reproduce) may fall into this category in some waters.

By this definition, culture based fisheries are insignificant in Indonesia and Cambodia, limited in Laos and Thailand, widely practised in lakes and reservoirs in Myanmar and Vietnam, and almost universally applied in suitable water bodies in China. Bangladesh has a rather different hydrological regime from those of the other countries under consideration, being dominated by the annual flood cycle to a much greater extent. The Bangladesh government has a two-pronged strategy for attaining large increases in freshwater fish output. One is the development of aquaculture, but the second is enhancement of production from the floodplains by stocking with hatchery-reared juveniles, especially carps. To test the technical and economic practicability of this approach, two large internationally-funded pilot stocking programmes are under way.

Difficulties in evaluating the results of stocking

Culture-based fisheries in some small, privately owned or leased water bodies are worked by individuals, companies or cooperatives (examples are given below). However, the vast majority of stocking programmes in Asia are undertaken by governments.

The viability of fish release-recapture activities clearly depends on the value of fish caught, minus fishing and marketing costs, in relation to the costs of stocking. But proper evaluation of both the biological and economic results of government-funded activities is made difficult for the following major reasons:

1. Catch statistics are almost always incomplete. As pointed out above, inland fishing is done by large numbers of small, part-time and subsistence fishermen. Enumerating the catches made by these people is almost impossible. Even where control of the fishery is officially vested in a single authority, e.g. a state-owned Lake Management Company, many unofficial boats and gears can be seen operating. Thus official figures for freshwater fish yields are generally underestimates.

2. State-owned hatcheries rearing juveniles for release are normally given a production target each year by government. When the target is achieved, government will often increase the quota for the following year, frequently without allocating a proportionate increase in pond area or other resources. Once released, the fish have effectively disappeared, and it is not even within the remit of many state hatcheries to attempt to assess their fate. Since it is almost impossible to accurately audit the releases, it is widely acknowledged that actual numbers liberated frequently fall well below those stated in official reports.

   This phenomenon of “paper fish” is not necessarily improved when production of fry/fingerlings is contracted out to private hatcheries. There, the temptation towards corrupt dealings between the producers and government officials responsible for placing contracts and auditing releases is obvious.

Unfortunately both the above problems work synergistically in making the returns from culture-based fisheries appear poorer than they actually are. In the present period of universal privatization and economic rationalization it is difficult for governments to maintain stocking activities when the benefits cannot be properly quantified. Even where it can be demonstrated that the value of returns exceeds stocking expenses, much of the financial benefit is reaped by fishermen who make no contribution towards costs. Many of these are quite poor people, who badly need the income and food which fishing provides. Introduction of cost recovery systems, e.g. by levying licence fees or taxes on fish at landing points, is subject to the same difficulties of implementation as described for the enforcement of regulations.

Documentation of the results of stocking programmes

Attempts are made by many governments to assess the impact of their fish stocking programmes, with varying degrees of success. Data available from the countries and provinces included in this study are presented in Tables 7–29. The results all refer to fisheries in lakes or man-made reservoirs except in Bangladesh, where floodplain areas were also stocked. Fishery yields from similar unstocked water bodies are given for comparison.

Bangladesh

There are currently three major internationally-assisted projects aimed at demonstrating the viability of culture-based inland fisheries enhancement in Bangladesh. These are:

1. The baor (oxbow lakes) fisheries development project, funded by DANIDA and IFAD.

2. The World Bank third fisheries project, co-funded by ODA and UNDP.

3. The Asian Development Bank second aquaculture development project.

Oxbow lakes. As the name suggests, the oxbow lakes project aims to increase fish production from this specific type of perennial water body. The lakes used range in size from 20 to 200 ha, and water depth varies from 3–4m maximum in the wet season to 1m in the dry season. When the lakes were first taken into use some clearing of weeds was done, particularly removal of water hyacinth, but no fertilization or fish feeding is carried out. Fingerlings of silver carp, catla, rohu, mrigal, grass carp and common carp can be stocked at any time of year, and 95% of the fish are recaptured within one year. Fishing is done all year round except when the water is very muddy. Most fish are taken by seining.

In Phase 1, from 1978 to 1986, six oxbow lakes totalling 1040–1090 ha in area were used, and are still in production today. Management is organized and stocking paid for by the Department of Fisheries, with the income from fishing being divided in the proportion 60% to the government and 40% to the fishermen who are employed on a catch share basis. Recent release/recapture results for three years are summarized in Table 7.

Table 7. Summary of results from stocking of six oxbow lakes in southwest Bangladesh, 1991–1994.

Source: Status of Baor Fisheries Development Oxbow Lakes Small Scale Fishermen Project, Phase I and II (1991–94). Project Implementation Unit, Danida Technical Assistance, January 1995.

Phase 2 of the project started in 1988 and is on-going. New oxbow lakes have been leased from government each year since then, until in 1995 there were 22 in operation with a total area of 1334 ha. The management system is completely different from that used in the first phase. Now, cooperative groups of fishermen (called lake fishing teams) are given exclusive rights to use the baors. Rights are guaranteed by the Ministry of Fisheries and Livestock and Ministry of Land under tenure of a renewable annual licence issued by the Department of Fisheries. The fishermen themselves organize stocking of fingerlings, harvesting and marketing. They pay all expenses, using credit administered through a national NGO (BRAC), and receive all benefits from sale of fish.

Provisional results from stocking/recapture operations in nine oxbow lakes involved in Phase 2 are shown in Table 8.

Table 8. Summary of provisional results from stocking of oxbow lakes in Bangladesh (project phase II), 1991–1994.

* Number of oxbow lakes used increased year-by-year. The figures given are means from 9 waters used each year from 1991.

** Gross income. In practice about half of this was deposited in the central fund to cover repayments of credit and expenses for fingerlings, lease fees etc.

Source: As Table 7.

Of total catches, only 13% was “miscellaneous” (= non-stocked) species. Stocked fish recovery rates averaged 9% by number, 300% by weight and 290% by value. The best recovery rates were obtained with rohu, followed by grass carp, then silver carp, mrigal, catla and common carp. The net incomes of participating fishermen have been greatly increased by the project, and many NGOs are now adopting the same approach in other oxbow lakes. Clearly the results of stocking juvenile fish into such small, easily manageable bodies of water are good, and it is anticipated that all similar waters in Bangladesh will be managed as culture-based fisheries in the future. However, the number of oxbow lakes available in the country is very limited.

Third Fisheries. The third fisheries project aims to assess the results of stocking much larger areas of water more typical of the vast floodplains of Bangladesh. The smallest beel (= floodplain depression) used was 1,500 ha in area and the largest floodplain stocked over 20,000 ha.

The approach used was to stock fingerlings early in the rainy season, between June and the beginning of August. Most fingerlings were bought from private hatcheries. Payment has to date been covered by the project, but partial recovery of costs is now being sought via introduction of a licence fee system, with a sliding scale depending on type of gear used. The main fish species stocked were rohu, catla, mrigal, silver carp, common carp, grass carp and Puntius gonionotus.

Fishing continues all year round, but catches peak during the flood recession period when fish are concentrated in residual water bodies, as is normal in floodplain fisheries. Similarly the gears used vary with the rise and fall of water. A wide variety of gears is used on the floodplains of Bangladesh, including seines, hooks (single and long-lines), push nets, lift nets, cast nets, spears, traps, gillnets, bamboo fences (used like seine nets), and hand fishing. Often a high percentage of the catch is taken in katha (brush parks used as fish aggregating devices) and kua (fish pits dug to concentrate fish as the water recedes). The project uses several methods of catch assessment, including household monitoring, gear dependent monitoring, fish landing centre monitoring and special surveys of katha, kua and large lift nets.

The results available from the third fisheries project's stocking programme to date are summarized in Tables 9 to 14.

Table 9. Bangladesh Third Fisheries Project. summary of fish stocking and harvesting results from BSKB floodplain.

Source: Third Fisheries Project, Third Annual Report (July 1993–June 1994. Bangladesh Centre for Advanced Studies, November 1994, and Internal Report to the UK Overseas Development Administration, 1995.

Table 10. Bangladesh Third Fisheries Project. Summary of fish stocking and harvesting results from Haiti floodplain.

Source: As Table 9.

Table 11. Bangladesh Third Fisheries Project. Summary of fish stocking and harvesting results, chanda floodplain.

Source: As Table 9.

Table 12. Bangladesh Third Fisheries Project. Summary of fish stocking and harvesting results, Garalia Beel (approx. 1500 ha).

Source: As Table 9.

Table 13. Bangladesh Third Fisheries Project. Summary of fish stocking and harvesting results, Hilna Beel (approx. 2000 ha).

Source: As Table 9.

Table 14. Bangladesh Third Fisheries Project. Estimated total production of stocked fish by species in three floodplains, July 1993 to June 1994.

Source: As Table 9.

It should be noted that some species, especially common carp and Puntius, may have reproduced naturally in the recipient waters. In addition, some wild Indian major carps may have entered where the floodplains have connections with rivers. Occasional anomalous results, e.g. impossibly high returns of grass carp, are probably attributable to mis-identification of species by field workers. These minor problems are to be expected in such a large project, which has yielded one of the biggest and most valuable databases on the results of fish stocking available anywhere in the world.

As the summary of results in Table 14 shows, recapture rates were generally highest for rohu, followed by common carp, catla and mrigal. Returns of silver carp were disappointing, perhaps because they migrated out of the stocked areas. Total yields of stocked species ranged between 5 and 13 times the weight of fish released.

Biologically and economically, the third fisheries stocking programmes can be considered largely successful. Recapture rates have been high enough to give positive economic rates of return in most waters (except Haiti, where activity has now been suspended). Calculated ERR ranged from 14% in BSKB floodplain to 78% in Garalia Beel. As might be expected, returns were generally better in the smaller, more manageable waters.

Thus the results indicate that it would be technically possible to increase fish yields from the floodplains of Bangladesh by widespread stocking. However, as is commonly the case in fisheries development projects, the most serious problems in further implementation and extension of activities are more social than technical. The main ones are:

1. Who will pay? As discussed above, the implementation of cost recovery mechanisms through collection of licence fees etc. is extremely difficult. The third fisheries project is aiming at 20% cost recovery in 1995, but management doubts that full recovery will ever be possible. Since it is unlikely that government can contribute more than a small fraction of what would be necessary to cover the costs of widespread floodplain stocking with fingerlings, many specialists believe this approach has no viable long-term future.

2. Who benefits? As also mentioned above, control of fisheries in Bangladesh tends to be monopolized by wealthy elites. The unfortunate side effect of increasing the value of the fisheries by stocking has been to encourage the powerful to seize control of the resource even more vigorously, and to exclude the access of the poor even more than before. For example, historically the owners of kua generally had no objection to subsistence fishermen operating on the floodplain during the wet season. Rights of ownership were exercised only when the flood waters had receded to the extent where discrete water bodies became discernible. Where stocking has been done, however, there is a tendency for the kua owners to discourage all fishing on the floodplain even in the wet season, on the grounds that the fish will eventually be concentrated in their kua, and are therefore their property. Thus the social effects of development efforts can be the opposite of what was intended.

   Whilst it may be possible, if difficult, to organize some form of community management to ensure a more equitable distribution of benefits in small water bodies, this is almost certainly impossible in large areas.

Second ADB aquaculture development project. The second ADB project has taken a different approach to floodplain stocking from that tested in the third fisheries programme. Instead of stocking expensive hatchery-reared fingerlings, ADB has sought to use natural beels (floodplain depressions) as nurseries for producing fingerlings from larvae stocked at only a few days old. The resulting fingerlings subsequently spread out onto the floodplain when the rainy season comes, or alternatively they can be caught and transported to release sites.

Selected beels were chemically de-weeded, treated with rotenone to remove predatory fishes, then limed and fertilized both organically and inorganically during the dry season (February–March). Pesticides were then used to reduce insects and other predators which could destroy carp larvae. Larvae of silver carp, common carp, catla, mrigal and rohu were then stocked at a target rate of l.25kg/ha of 3–5 day old fish (numbering about 360,000/kg). Beels were fertilized further and guarded to protect the young fish. Subsequent sampling monitored the growth and survival of fry and fingerlings until their dispersal following the onset of rain. The normal nursing period was 60–90 days. Fishing activity began 2–3 months later.

The number of beels used has increased since the project started in 1991. In 1994, 27 beels totalling 450ha were used, and in 1995 21 with a total area of 500ha. Catches from the floodplains were evaluated by five methods:

-Fish landing monitoring (and subsistence fishing survey)
-Fishermen household survey
-Fish market survey
-Fish consumption survey
-Lessees' survey

The yields resulting from releases were estimated from the increase in catches of young-of-the-year over the previous year for each species. Those species not occurring naturally in Bangladesh, especially silver carp, were also used as “markers” for successful stocking. The gears used by fishermen, and the seasonal pattern of catch, were similar to those described above for the third fisheries project.

The results of the “beel-nursery” approach are not yet complete, but some preliminary indications are given in Table 15.

Table 15. ADB Second Aquaculture Development Project. Provisional total fish stockings and recaptures (all waters combined) 1994–January 1995.

Source: Bangladesh Second Aquaculture Development Project. An interim evaluation of 1994 floodplain stocking. Working Document 17, March 1995.

During the 1994–5 season, around 25% of hatchlings survived to reach fingerling size. By January 1995, the recapture rate of marketable fish had reached an estimated 6% of fingerlings released. This was expected to rise to around 10% by the end of the fishing season in February–March. Common carp, rohu and catla showed better growth and recapture rates than mrigal and silver carp.

Despite some loss of control caused by unexpected flash flooding and consequent premature escape of fry, preliminary indications are that the beel nursery approach can provide fingerlings cheaper than their purchase price from hatcheries, and that recapture rates are high enough to give a positive economic rate of return. However, the social and economic problems indicated above for the third fisheries project apply equally here, and the second ADB strategy has come in for additional criticism on environmental grounds for its use of chemicals, particularly the killing of native fish species with rotenone during beel preparation.

During years when insufficient time was available to prepare and stock beel nurseries, the second ADB project instead stocked fingerlings into Hail Haor. a floodplain with maximum area of over 14,000 ha (minimum dry season area 2,834 ha) in north-east Bangladesh. Fish catches were estimated mainly from surveys of landing places and markets. The results from this experiment are presented in Tables 16 and 17, and are directly comparable with those of the third fisheries project. Similar conclusions can be drawn.

Table 16. Bangladesh ADB Second Aquaculture Development Project. Fingerling stocking results from Hail Haor, 1991–2.

Source: Bangladesh second aquaculture development project, working document No. 16, July 1992.

Table 17. ADB Second Aquaculture Development Project. Preliminary results of fingerling stocking programme in Hail Haor, 1994–5.

Source: Bangladesh second aquaculture development project. Working document no. 17: An interim evaluation of 1994 floodplain stocking.

China

China is generally considered as the Asian nation in which culture-based fisheries are most advanced and widely applied. Most suitable lakes and reservoirs are routinely stocked to some degree by provincial fisheries departments or state-owned lake management companies.

Decisions about which species should be stocked and in what numbers are made on a relatively scientific basis by comparison with most other countries. In China, it is usual for fisheries officers to conduct physical, chemical and biological surveys on waters, the outcome of which is to classify the lake or reservoir on the basis of its trophic status. Stocking policy is then formulated, loosely depending on the resources of plankton, benthos etc. available to support fish populations. An example of the guidelines used is given in Table 18.

Reality, of course, does not always correlate well with the theory, and in practice fish releases in China and in all other countries included in this study generally depend more on the instinct of the officials responsible, and above all on the availability of fry/fingerlings, than on strict scientific principles. Similarly the evaluation of the results of stocking are always less than perfect, due to the general problems already discussed above.

Thus some stocking of non-reproducing species is done in many lakes and reservoirs in Yunnan and Guangxi, but accurate records of releases and returns are scarce. Statistics on total fish yields, including both naturally breeding and stocked species, from the major lakes of Yunnan are given in Table 19.

Table 18. Recommended numbers of silver and bighead carp fingerlings (over 1.3cm) for stocking in Chinese lakes/reservoirs.

Source: Chinese Academy of Fisheries Sciences, Beijing.

Table 19. Fisheries yield of lakes in Yunnan Province, China, 1994.

* Excludes Crustacea and molluscs. In some cases catches of these add a further 25–30% to total yield of aquatic products.

Source: Calculated from figures provided by the Department of Agriculture, Yunnan Provincial Government.

Expressed as yield per unit area, there is a clear inverse correlation between fish production and lake area and depth, i.e. small shallow waters are more productive than large, deep ones.

Many of the lakes were stocked with some hatchery-reared fingerlings. Available records of releases and yields of these are collated in Table 20.

Table 20. Fish stocking and yield in lakes in Yunnan Province, China, 1994.

Source: Calculated from data provided by the Department of Agriculture, Yunnan Provincial Government.

Clearly yields are not strictly comparable with releases the same year, but result from stocking in the previous year or years. However, numbers and species released were said to be similar in successive years.

Generally, yields of silver and bighead carp appear good, with the weight of fish recaptured exceeding that of fingerlings released by a ratio ranging from three times in Ji Lu Lake up to over 80 times in Erhai Lake. Returns of grass carp were similarly valuable. Yields of common carp look even more impressive, ranging up to 300 times the weight of released fingerlings in Chenghai Lake, but here it must be assumed that the fish were also breeding naturally. Wuchang fish (Megalobrama amblycephala), which require running water to reproduce, and were therefore presumably not breeding in the lake, gave a return to Erhai fishermen of over three times their stocked biomass.

Indonesia

Stocking of non-reproducing species is not generally done in Indonesia, mainly because the species usually used for this purpose, i.e. the Chinese and Indian carps, are virtually unsaleable in Indonesian markets.

For purposes of comparison with other countries, figures on fish yields from some Indonesian reservoirs are given in Table 21. Only a few inland waters are leased; most are open access.

Table 21. Fish yields from some Indonesian reservoirs.

Source: from Balyut, 1983.

No clear correlation can be discerned between fish yield and reservoir area and/or depth. This does not mean that there is no relationship between biological fish production and lake morphology, but more likely reflects differences in fishing effort between the waters, for which no figures are available.

Laos

Some releases of hatchery reared juvenile Chinese carps and tilapias are made by government into reservoirs in Laos, but no records of returns are available.

Myanmar

The Department of Fisheries of Myanmar operates a policy of stocking suitable water bodies throughout the country. Most waters are leased, but the terms of contract require the leaseholder to buy fingerlings from government to a minimum value of 30% of the lease price-In addition, some lessees run their own hatcheries. As elsewhere, the species most used for stocking are the Chinese and Indian major carps and common carp. To encourage leaseholders to improve the fishery, the period of lease was increased in 1991 from one year to three.

No attempt is made to assess recapture results from large lakes or reservoirs, but Table 22 shows returns from some “inns” (small lakes, generally connected to a river, the water level and area of which rises and falls with the flood cycle; similar to the beels of Bangladesh). There are about 3,500 inns in Myanmar, with the greatest concentrations in the floodplains of the Irrawaddy River in Mandalay Division and the delta area. Some inns are perennial, others flooded only seasonally.

Table 22. Average annual fisheries production and numbers of juveniles¹ stocked in some “inns” of Mandalay Division, Myanmar, 1991–94.

^1. The main species stocked were rohu, silver carp, common carp, bighead and grass carp.

Source: Department of Fisheries.

The fishing methods used in inns vary according to whether they are permanently or only seasonally flooded. In perennial waters, gillnets are most used for capture of free swimming species, and longlines for bottom species, but many other gears including seines, cast nets, traps, lift nets etc., and khone-cha (brush parks, similar to the Bangladeshi katha, but in Myanmar usually surrounded by bamboo screens instead of nylon netting for harvest) can also be seen. In seasonal inns special large bamboo traps are placed on outlet channels to catch out-migrating fish during the recession. Sometimes men splashing in the water with bamboo poles drive the fish into the traps. Fish pits are also used.

As the fish yields shown in Table 22 suggest, inns are often managed quite intensively. Some receive heavy inputs of fertilizer and supplementary feeding. For example, Tet Thay Inn is a shallow (2–5m deep) perennial lake, which the leaseholder has divided into workable sections with nylon net fences (formerly bamboo screens were used). Larvae of Chinese, Indian and common carp are bought in and reared to fingerling size in ponds adjacent to the inn. Fingerlings are stocked regularly throughout the year, and harvesting is done twice daily (early morning and evening), when small wholesalers come to the waterside to buy. The inn receives a continuous input of municipal sewage from Mandalay City via an open and rather malodorous canal. In addition, the fish receive supplementary feeding with cottonseed and peanut cake, and rice bran. Fishermen are hired for a fee plus a share in the catch of non-stocked, carnivorous species. The main method of fishing for Chinese and Indian carps is to drive the fish into set surface gillnets with small boats. In each boat one man paddles towards the nets, while another bangs on the bottom of the boat with a stick.

Thailand

Large lake and reservoir fisheries in Thailand are generally open access. Therefore all stocking activities must be done by government. Licence fees are sometimes levied on gears in areas where fisheries officers are present.

Regular stocking of fish into reservoirs was started in 1978 under the “nationwide fish stocking programme”, and it is said that over 700 million fish and shellfish were stocked in 5.6 million hectares of water between 1978 and 1987 (Bhukaswan, 1989). However, most of these were naturally reproducing species, including common carp, O . mossambicus and O. niloticus, Trichogaster pectoralis etc. so the impact of the programme on yields was difficult to assess. Recorded returns of Chinese and Indian major carps were apparently poor. In 1990, policy changed to centre on conservation of indigenous species, and subsequent stocking in large water bodies has since concentrated on these. Current annual releases total over 100 million juveniles.

Total fish yields from some large Thai reservoirs are shown in Table 23.

Table 23. Average fish production from major reservoirs in Thailand.

* HE-hydro-electric, MT = multipurpose, I = irrigation.

Sources: Bhukasawan (1979), Bhukasawan and Pholprasith (1977), Leenanond and Chookajorn (1993), Pupipet et al. (1989), Chantsavong et al. (1988).

Again, some inverse correlation between fish output and water body size and depth is apparent. Yields of Thai reservoirs generally appear low by comparison with those in southern China, despite the more favourable climate prevalent in Thailand. However, as with all fisheries statistics, care should be taken in drawing firm conclusions due to the unreliability of official figures in all countries.

For some years the Thai government has supported a programme of construction of “village ponds”; small multi-purpose reservoirs normally less than 20 ha in area, predominantly in the dry north and north-eastern parts of the country. The ponds are managed by the local community, via a fisheries committee. The Department of Fisheries can provide training courses for committee members. In the first year of operation of a community reservoir, the Department will provide fish seed for stocking free of charge. In the subsequent three years, 25%, 50% and 75% of seed cost must be paid by the committee, until from year five onwards total costs are paid.

Some waters receive fertilization and supplementary feeding in addition to fish stocking. Revenue to pay for fisheries management is generated through sale of tickets on fishing “open days”, and the surplus profits are officially earmarked for use in other infrastructure improvement projects for the benefit of the community, e.g. road and school building. Most waters have only one open day each year.

The species most commonly stocked in such small water bodies are mostly non-indigenous, including Chinese carps (silver, bighead and grass carp), Indian major carps (rohu and mrigal), and Nile tilapia. Thai silver barb is also stocked. Stocking is normally done once per year between May and August. Stocked species typically make up over 90% of total catch, with the balance comprising indigenous snakeheads, catfishes and minor species.

Average yields of village ponds in the various regions of Thailand are shown in Table 24.

Table 24. Fish production of village ponds in Thailand, 1992.

Source: Fishery Economics Division, Department of Fisheries (1994) .

Fisheries in several large “swamp” areas in Thailand are also partially managed by the provincial fisheries departments. Management measures include weed clearing and dredging, as already mentioned above, and stocking of juvenile fish reared in purpose-built hatcheries. Species stocked include tilapias, Chinese and Indian major carps, common carp, Thai silver barb and Macrobrachium. Average annual fish yields from four of the largest Thai swamps are shown in Table 25.

Table 25. Average fish production from major “swamps” in Thailand.

Sources: 1. Duangsawat et al. (1994).

2. Phayao Freshwater Fisheries Station (1987).

3. Nakorn Sawan Freshwater Fisheries Station (1987).

4. National Inland Fisheries Institute (1980).

Vietnam

Stocking with non-reproducing fish-species is commonly done in lakes and reservoirs all over Vietnam, either paid for by the Ministry of Fisheries or the autonomous water management companies which are responsible for many large waters. The usual Chinese and Indian carps are used.

Official figures for fish production in reservoirs in many provinces of Vietnam are summarized in Table 26, and species stocking structure and yields for eight individual waters listed in Table 27.

The inverse relationship between water body size and fish productivity per unit area is shown in Table 28.

Growth of stocked exotic carps, especially Chinese carps, in Vietnamese reservoirs is frequently very fast (Table 29), and these species are well accepted on local markets.

The relatively low yields per unit area recorded for many Vietnamese reservoirs is partly a reflection of the large proportion of the catch being taken unofficially (and therefore unreported). Under the new economic policy, many water management companies are now in financial difficulty. Some have been forced to suspend their fisheries activities, and instead function only as “landlords” selling licences or sub-leases to individual fishermen. Consequently the scale of government-financed fish stocking activity is in decline, but private fishermen lack the resources and organization to take over this work.

Table 26. Productivity of fisheries in Vietnamese reservoirs, 1990–94.

* Figures rounded to the nearest kg.

Source: Ministry of Fisheries, Hanoi.

Table 27. Fish stocking structure and productivity of some reservoirs in Vietnam.

¹ S.C.= silver carp, B.H.= bighead, M.C.= mud carp, G.c.= grass carp.

Source: Research Institute for Aquaculture No. 1, Hanoi.

Table 28. Fish production of reservoirs of various sizes in Vietnam.

Source: Research Institute for Aquaculture No.l, Hanoi, 1995.

Table 29. Growth of fish stocked in two reservoirs in Northern Vietnam (Mean fish weight in g).

Source: Research Institute for Aquaculture No.l, Hanoi.

3.3.3. Cage and pen aquaculture

Bangladesh. Cage and pen culture is almost non-existent in Bangladesh, where only one commercial farm producing tilapias in net cages sited in a river close to Dhaka is known. Development projects have demonstrated the technical viability of, for example, growing O. niloticus in small cages moored in oxbow lakes. Ownership of the cages and fish was vested in landless people and groups of women. However, once profitability was achieved the owners or leaseholders generally re-took possession of the waterbodies, evicting the target beneficiaries. This question of access to water resources is central to all fisheries development programmes in overcrowded Bangladesh.

Technically, the prospects for introduction of commercial cage culture in Bangladesh are present, but limited by the shortage of materials suitable for use as fish feeds. Sites are also limited, but can be found in the perennial Kaptai Lake and some oxbow lakes, in large canals in the south west of the country, and possibly in many seasonal lakes and ponds associated with the extensive floodplain.

Cambodia. The use of floating cages to grow fish to market size developed from the older practice of employing cages for storage of captured fish and their transport to market. This practice is still followed today, especially for moving fish from the Great Lake down the Tonle Sap River to the capital, Phnom Penh.

Cage culture developed rapidly in the 1950s and 1960s, but underwent serious decline in the subsequent period of internal troubles in the country. A revival in the 198 0s probably brought production back to prewar levels by 1990.

Recent estimates of fish production in cages in the various provinces of Cambodia are given in Table 30.

Cage culture is done in the Great Lake itself and in a number of locations in the Mekong River system, including the Tonle Sap River (which joins the Great Lake to the Mekong), the Mekong River proper, and the Bassac River (the second major branch of the Mekong system below Phnom Penh). Pen culture is relatively rare in Cambodia, but it is carried out in shallow water at the margin of the Great Lake in a few places. Occasionally fish are cultured in pens during the dry season, then transferred to floating cages when the water level rises with the onset of the rains (Mekong Committee, 1992).

Cage farms over 50m² officially require a licence from the provincial agriculture department, but no rental fee is charged.

Table 30. Fish production from floating cages in Cambodia, by region, 1993.

Pangasius spp. (predominantly Pangasius sutchi) was the major fish cultured in cages, with a total annual production of about 4,800 tonnes. Second in importance were the snakeheads (mostly Channa micropeltes), which contributed about 1,345 tonnes in 1993.

Source: NACA country paper, delivered at the final workshop of project TCP/RAS/2253, RAPA, Bangkok, February 1994.

Almost all fish cages in Cambodia are made either of split bamboo or timber slats, both attached to a wooden frame. Cages are usually rectangular, but some are constructed in the shape of a boat to facilitate towing to transport fingerlings to growing sites or harvestable fish to market. Sometimes timber cages are virtually solid wood on their long sides, with water permeable meshes only at the ends. These must be orientated correctly in rivers so that water flows through them. Most commonly, flotation is provided by bundles of bamboo attached to the long sides of the rectangular cages, but metal drums are also sometimes used. Cage size is very variable, but growing cages are usually less than 200m in area and not more than 3m deep. However, boat-shaped transport cages tend to be larger, commonly over 30m long. Bamboo rots fairly quickly in water, and bamboo cages must be replaced every two years. Timber lasts much longer; up to five years or more, depending on quality.

Most fish cages are owner-operated. The owner's family normally lives permanently in a timber house constructed on top of his cage (or one of his cages if several are owned). A trap-door in the floor gives access for feeding the fish.

Cages are moored together in large groups in suitable locations, so that self-contained floating villages are formed. In the Tonle Sap River just upstream of Phnom Penh, for example, there are three separate floating villages, each one a distinct community made up of a single ethnic group: one Cambodian Buddhist, one Moslem, and one of Vietnamese people.

The major fish species produced in cages in Cambodia are the Pangasius catfishes, and snakeheads. Most fry/fingerlings are captured from the wild. Pangasius sutchi. the most commonly cultured species, is mainly caught as large fingerlings (averaging approx. 12–15cm, or 6–10 fish/kg) in the Tonle Sap River or Great Lake. Snakeheads are taken somewhat smaller, down to 2–3 cm, but large fish up to 3/kg are also stocked, depending on availability. Sometimes bamboo cages are at first lined with a nylon net to prevent escape of fish when small fry are used. Lesser numbers of other species, including other Pangasius spp. , sand goby (Oxyeleotris marmorata) and clariid catfishes are also cultured in cages in Cambodia, either in monoculture when sufficient fry are available or as an extra “side crop” in polyculture with P. sutchi or snakeheads.

The preferred feed for the predominantly carnivorous species cultured in Cambodia is trash fish. Many cage farmers are also fishermen, selling the large fish they catch and retaining the small ones as feed. The species most commonly used as fodder in the Tonle Sap area are Chanda spp. (chanda ranga, C. siamensis etc.) and Cirrhinus jullieni. Alternatively trash fish can be purchased. During the peak season of December–January small fish are so abundant that they can be bought for less than US$0.1/kg, whilst at other times they may cost four times this. Trash fish are usually fed whole. During periods of year when small fish are not freely available, especially from July to October, rice bran and other poorer quality feeds are used either alone or mixed into balls with some trash fish.

Obviously the time taken to grow-out fish to market size depends partly on size of fingerlings stocked. For example, snakeheads stocked at a large size of 3 fish per kg will attain a marketable weight of l-l.5kg in 5–6 months, whilst 2–3cm fry require around 10 months to reach this size. Pangasius sutchi stocked as 100g fingerlings similarly need 8–10 months growth to harvest at l–.5kg. Stocking densities and survival rates of fish in cages no doubt vary widely. Some farmers interviewed in the Tonle Sap River just outside the Great Lake in Kampong Chhnang Province reported they stocked up to 5 tonnes of fingerlings in 6×30×2.5m deep cages, i.e. about 11 kg/m³ , and harvested up to 30 tonnes, i.e. about 67 kg/m³. Others working further downstream, close to Phnom Penh, stocked smaller 10×6×2.5m cages with 1 tonne of seed (=6.7kg/m³) and harvested up to 4 tonnes/cage (=26.7 kg/m³). These figures fall within the normal range used for other species in cage culture all over the world. Similarly survival rates reported from fingerling to market size varied from 60 to 95%.

Most Cambodian fish farmers sell their crop to middlemen who come to the cage villages to buy, but some organize their own transport down-river to market. Sometimes the middlemen provide credit. There is some export of market fish to Vietnam and Thailand, and of fingerlings (especially Pangasius spp.) to Vietnam.

China. Cage culture is a relatively recent introduction to Yunnan and Guangxi, having started only in the 1980s. Total output of fish from cages in 1994 was estimated at 12,000 tonnes in Guangxi and a little over 2,000 tonnes in Yunnan. Some cages are sited in rivers, but the majority are in lakes or reservoirs (predominantly natural lakes in Yunnan, man-made reservoirs in Guangxi).

Cage construction is different from that described above for Cambodia. In Guangxi and Yunnan, most cages are based on nylon or polypropylene netting bags, usually suspended from timber, bamboo or steel floating collars. Occasionally steel or bamboo mesh is used either instead of nylon or outside a nylon net where it is believed there is a risk of predators breaking nets. Bouyancy is most commonly provided by empty steel oil or chemical drums or bundles of bamboo, but sometimes specially made polystyrene or plastic floats are used. Typically cages are square or rectangular and small, with sides up to 8m and depth not more than 4m. Where water level is regulated to a standard depth, cages are sometimes placed near the shoreline supported by bamboo poles stuck into the bottom mud. This system is used in parts of Erhai Lake in Yunnan, where over 1,000 tonnes of fish is produced annually from just over 10 ha of cages, but it is not common elsewhere.

In a few locations huts are provided on top of floating cage groups to provide storage space and especially shelter for a watchman, but more usually cages are moored within sight of the on-shore homes of the owners. Permanent family dwellings are very seldom built on fish cages in Yunnan and Guangxi.

The vast majority of fish cages are family owned, but the provincial governments and lake management units responsible for administration of some large waters run their own cage farms as demonstration and training facilities. There are fairly good government-run agriculture and fisheries extension services operating in Yunnan and Guangxi, but the spread of cage culture has been patchy, with many remote regions left unserviced. However, in some locations local government has achieved good success in encouraging the adoption of cage culture as a business in poor rural communities. Credit is made available through the Agricultural Development Bank or other local lending institution.

The species cultured in cages in south China are Chinese carps (grass carp and occasionally silver carp), Nile tilapia and common carp. Silver carp are kept only in eutrophic waters where phytoplankton is abundant and the fish do not have to be fed. Grass carp are widely favoured for culture because they command a good market price, but can be fed on cheap, locally available foodstuffs (grasses, waterweeds and other vegetation collected free of charge). Common carp and tilapia are fed pelleted diets manufactured in the many animal feed mills to be found in all provinces of China. Monosex tilapia are available from some hatcheries in Guangxi.

Due to the predominantly higher elevation, the waters in Yunnan are on average colder than those of Guangxi, and fish growth therefore tends to be slower. Examples of fish growth rates routinely attained in the region are:

Grass carp stocked in cages at Gonghe village, Wuming County, Guangxi, at the rate of 40–50 15cm fingerlings per m² in March reach marketable size of l–2kg in 5–6 months, with a survival rate of 90%.

Two-month old tilapia fry stocked in cages in Chen Bi He Reservoir, Baise District, Guangxi, reach a saleable weight of 500g in 4–5 months at the ambient water temperature of 18–24 C. Stocking rate is around 15,000 fry per 4×6×2m cage.

In Yang Zhong Hai Lake, Gongming District, Yunnan, common carp take 1.5 years to grow from “fingerling” size of 200g to 1kg plus at temperatures of 16–26 C. Fingerlings are stocked at 3–3, 500 fish per 5×5×2m cage, and food conversion on the pelleted diet averages 2.8:1.

In Erhai Lake, Yunnan, grass carp and common carp fingerlings stocked at 13–15kg/m of cage require two years growth to attain market sizes of above 1.5kg and 0.5kg respectively. Maximum fish density at harvest is around 50kg/m³.

Occasionally fish are stocked in cages in polyculture. For example, in Dong Feng Reservoir, Yuxi District, Yunnan, approx. 3% of bighead and grass carp are stocked in common carp cages.

Some experimental trials are being conducted with high-value species, especially catfishes, but commercial production of carnivorous fishes is not yet practised in this region of China with the exception of a few trout in cold waters close to Kunming City, Yunnan Province. Indeed, traditional fish farming systems all over the country concentrate on species low in the food chain, which give the highest yields from the available natural and supplementary food resources.

Pen culture is practised in this part of China, but on a lesser scale than cage methods. The biggest area of pens is in the l,300ha, 2m deep “grass lake”, adjacent to the 30,000ha Dianchi Lake, close to Kunming city. Most of the grass lake is divided up by nylon net pens, which are managed in much the same way as fish ponds. However, the future of this venture is now threatened because engineering work has closed the connection between the grass lake and the main body of Dianchi. Since the grass lake receives much of the sewage from Kunming City, environmental conditions in the water are now deteriorating rapidly to the extent that fish kills can be expected.

Other, more limited areas of pen culture are present in lakes and reservoirs in both Guangxi and Yunnan. Examples can also be found where whole arms of a lake up to several hectares in area are fenced off from the main water body and used as big fish ponds. In Chen Bi He, Guangxi, for example, this technique is used despite the very large seasonal fluctuations in water level experienced in this 4,000 ha hydro-electric and irrigation reservoir. When water level is low, the poles supporting the nylon netting fence can be seen sticking 20m out of the water. Additional panels of netting are attached to the poles as water level rises, and removed again as it falls. As with cages, ownership of net pens normally rests with local families, and to encourage these activities government does not usually make any charge for use of the water.

Indonesia. The market for freshwater fish in Indonesia is unlike those of the other countries included in this study. Indonesian consumers strongly favour common carp over other species, and demand for carp is predominantly for small fish in the 200–300g range. The cage culture industry in this country therefore concentrates heavily on production of small common carp, though Nile tilapia has also achieved good market penetration in recent years. Indeed. even pond fish farms in Indonesia focus on growing common carp in monoculture; a very unusual practice for Asia. On a very much smaller scale, a few other species are produced in cages, including giant gouramy, sand goby, Pangasius and clariid catfishes and snakeheads. The latter, carnivorous fishes are mostly produced in Kalimantan, where there is some concern over the capture of wild fry of other fish species for use as feeds in cages.

Very small cages (approx. 2–5m²) made from bamboo or timber slats have been used to grow common carp in canals and streams in Indonesia for many years, since at least 1940. The cages are closed on all sides and sit on the bottom. In this system fish are fed on domestic kitchen wastes, which are put into the cage through a trap-door on the top. The practice continues, but its scale has actually declined due to problems caused by the obstruction of water flow by the cages.

Nowadays, most cage aquaculture in Indonesia is done in nylon or polypropylene net cages suspended from timber or bamboo walkways on top of oildrum floats, as used in many other countries. This type of cage was introduced during the 1970s, but did not really “take off” until the late 1980s. It is used predominantly in lakes and reservoirs. Average cage size is 50–100m² , and depth of net not more than 4m. Cage groups are usually provided with temporary shelters for watchmen, but not permanent family dwellings.

The main areas for cage culture are in West, and to a lesser extent Central Java, close to markets in the main centres of population. There is also some export of cage reared fish, especially of monosex red tilapia. A single foreign-owned company in Central Java produces several tonnes of this fish per day for export as frozen fillets to the United States. However, the biggest concentrations of fish cages are in the three reservoirs on the Citarum River in West Java. From up to downstream, these are the Saguling (about 5,600ha, completed 1985), Cirata (around 6,700ha, completed 1988) and Jatiluhur (8,300ha, 1967) Reservoirs. Between them, these reservoirs account for the majority of cage fish production in the country.

In some sections of these waters cages are very densely crowded. The resulting enrichment, together with some special hydrological conditions causing a turnover of deoxygenated water, has led to periodic mass mortalities of caged fish in Cirata. This phenomenon has in turn resulted in a trend towards change in culture practice in the reservoir away from common carp into tilapia. This is simply because carp are normally sold alive, and the difference in price between live and dead carp in Javanese markets is very large. Tilapia, on the other hand, are easily saleable dead as well as at small size.

Official government policy in these reservoirs was to encourage cage ownership by small-scale fish farmers, especially families displaced from their land by inundation following construction of the dams. To this end, regulations specify that only local people can be given licences to deploy cages, and that each permit holder can have a maximum of four cages. In practice, however, wealthy individuals from Jakarta and other cities now dominate production, some of them owning over 60 cages worked by hired labour.

Feeding of common carp and tilapia is with commercially made, dry pelleted diets. Several Indonesian and foreign companies market these locally, at a mid-1995 delivered price of US$0.32–0.36/kg. Food conversion ratio is 1.5–1.7:1. 25–50g common carp fingerlings reach 250g ready for market in 3–4 months, allowing three crops to be produced per year. Tilapia can be sold even smaller, from 100g up.

Laos. There is only a little cage culture practised in Laos. In Nam Houm Reservoir, the government of Vientiane Prefecture and the army operate about 150 cages, most of them 5×5×2.5m deep. All the cages use nylon nets imported from Thailand, some supported by bundles of bamboo and others by drum floats.

The species grown are mostly bighead and silver carp, but some tilapia are also produced. None of the fish are fed, but they grow well on the natural supply of plankton. 2cm Chinese carp fry stocked in March reach a saleable size of 600g in six months. Each cage produces around 500 fish, and survival during growout is said to be 90%.

There is significant potential for application of cage technology for these and other species in currently unutilized reservoirs in Laos. Preliminary, promising trials have been conducted in the largest: the 37,000ha Nam Ngum.

Myanmar. There is currently very little cage culture known in Myanmar, except on an experimental basis, but the country has substantial potential for application of this technology in its many natural and man-made lakes and reservoirs. The division of stocked “inns” into workable sections with netting barriers described above in section 3.3.2 may be regarded as a type of pen culture which is fairly widespread in this country.

Thailand. By comparison with pond and rice field aquaculture, freshwater fish production from cages, and pens in Thailand is very small, comprising well under one percent of the total aquaculture yield. Official Department of Fisheries statistics show cage (including pen) output as 1,139 tonnes in 1992, out of a total aquaculture production of 142,105 tonnes.

Net pens are deployed in only a few reservoirs and lakes, including the largest lake, Songkla, in which rohu and Nile tilapia reach marketable size after 14 months feeding on aquatic plants (Tansakul, 1990).

The use of cages is concentrated in rivers in Central Thailand. There, bamboo or wooden cages are most used, similar to the Cambodian models described above except that they do not usually carry permanent family dwellings. Goby and snakeheads are the species most commonly grown in Thai cages, also using the “Cambodian” technology. Pangasius is produced only in small quantity nowadays due to the low price resulting from the high output of these fishes from pond farms in Thailand. Some goby is exported.

There are a few experimental cage units sited in reservoirs in North and Northeast Thailand, where nylon net cages supported by steel drums are used primarily to grow Nile tilapia fed on commercially made pelleted diets.

Vietnam. In 1993, it is said that about 6,000 tonnes of freshwater fish was produced from around 10,000 floating cages in Vietnam.

Cage culture was introduced to south Vietnam from Cambodia in the late 1960s by ethnic Vietnamese people who moved down the Mekong River to the delta area. This migration of Vietnamese refugees from Cambodia continues today.

As a consequence, the technologies and species used in cage culture in the south of Vietnam are largely the same as those described above for Cambodia. The main species grown are Pangasius catfishes (in this case predominantly P. pangasius) and snakeheads, but goby, common carp and a few other species are also cultured. Occasionally the high density of cages has resulted in outbreaks of disease and high fish losses, and in these cases some farmers have gone over to common carp production, believing that this species is less prone to disease problems. A few fish farmers also grow crocodiles in floating cages made of steel mesh. The cages have a sloping timber floor partly in and partly out of the water, giving the crocodiles the option of immersing themselves or climbing out onto dry land. Both the crocodiles and the cultured carnivorous fishes are fed largely on trash fish. Some farmers catch their own fodder fish using seine nets in nearby waters, but supplies can also be bought in either fresh or dried. Some is imported from Cambodia, as are many of the Pangasius fingerlings. Additional feeds for caged fish are rice bran and water plants, which are usually mixed into a paste with trash fish before use.

The main sites for cages are in rivers, where “cage villages” have formed in locations where water conditions are suitable and access to infrastructure on land is good. Many cage farmers are clearly doing quite well. Their houses are built and furnished to a good standard, and television aerials are common on their roofs. However, newly returning refugees often have very few resources of their own, and are in dire need of assistance in the form of loans to enable them to enter the cage culture business.

Several large new reservoirs for hydro-electric production and irrigation were completed in the south of the country within the past 10 years, e.g. Tri An (32,400ha), Dau Tieng (27,000ha) and Thac Mo (10,600ha). Similar cage designs to those described above are now beginning to be used to grow the same fish species in the reservoirs. In this case it is common for the owners to live in houses on-shore close to the cage site rather than on the cages themselves. Sometimes fingerlings are reared in small cages in nearby canals.

Cage culture is less common in central and northern Vietnam, but the practice is beginning especially in reservoirs, which are most numerous in the mountainous north of the country. Here bamboo or occasionally nylon net cages are used predominantly to grow grass carp, which are fed on vegetation collected from the adjacent land. Cassava leaves are a frequently used feed source. Many of the people who have adopted this technology were displaced from their homes when their farmland was inundated following dam construction. Cage culture provides them with an alternative income to the environmentally-destructive slash and burn farming on the steep hillsides around the reservoirs, but they also need initial assistance in the form of expertise and credits.