Previous Page Table of Contents Next Page


The diversity and the density of the reservoirs in Sri Lanka represent a laboratory ‘in situ’ for limnologists and fisheries biologists, with great potential for developing and testing basic limnological and fishery biological models and concepts, as well as combinations between basic concepts in these two areas of interest. Sri Lankan reservoir studies have clarified some biological and other factors responsible for the success of especially O.mossambicus, and in this way have provided some useful information for the future developments in reservoir fisheries elsewhere in the tropics. The reservoir capture and culture fisheries, of Sri Lanka are good examples of rapid developments in the inland fisheries sector in the tropics. These developments have been responsible for providing an animal protein source to a very poor section of the population, and also for creating job opportunities.

6.1 Success of Oreochromis mossambicus

In Sri Lanka, O.mossambicus has been most successful in stimulating the development of a major fishery and its continued sustenance. The question that has to be posed is what are the reasons for its success in Sri Lanka.

Fernando and Indrasena (1969) mentioned several reasons for its success in major perennial reservoirs in Sri Lanka: lack of lacustrine species in the indigenous fauna; high predatory pressure; and restriction of breeding sites preventing the building up of large populations, which in turn prevents stunting.

The readily available blue-green algae as a food source could also enhance the success of O. mossambicus in the major perennial reservoirs (De Silva and Fernando, 1980), although a more extensive survey has shown that blue-green algae are not the major food item (Maitipe and De Silva, 1985). The ability of O.mossambicus to change from detritivory to zooplanktivory and its ability to efficiently digest the wide array of ingested materials (De Silva, 1985b; De Silva et al., 1984) enhance the growth and condition of this fish. In contrast to the situation reported from Lake Valencia, Venezuela (Bowen, 1979), the abundance of good quality of digestible food in the relatively shallow major perennial reservoirs is likely to favour good growth of the species in Sri Lanka.

Apart from the adaptability of its feeding habits O.mossambicus can also shift its life cycle pattern from an altricial to a precocial pattern (Balon, 1981; Noakes and Balon, 1982; Arthington and Milton, 1986). The reservoir populations in Sri Lanka show a precocial pattern and moreover the species has also been shown to have the ability to regulate its fecundity in relation to the fishing pressure (De Silva, 1986).

In the major perennial reservoirs of Sri Lanka there is no convincing evidence of it becoming stunted (De Silva, 1985 and Fig.18, Table 21), and in the author's view it will continue to flourish and sustain the all important reservoir fishery for years to come.

Apart from Sri Lanka, success of this species is being reported from the Sepik River floodplains in Papua New Guinea (Coates, 1985). It is possible that its lack of success and the tendency to stunt in smaller water bodies - in effect an erronneous comparison - has led in some situations to the opinion that tilapias are undesirable. Evidence from India is one of the examples Sreenivasan, 1988). There is no direct evidence elsewhere from Asia to indicate that O. mossambicus has become a nuisance in major reservoirs (Sarnita, 1987).


Fig.18 Mean landing size of O.mossambicus from different reservoirs during 1982-1986.

6.2 Future of Oreochromis mossambicus

With the increasing tendency to stock major perennial reservoirs of Sri Lanka with O.niloticus some possible long-term effects on the fishery must be considered.

The present yields of O.mossambicus have been shown to be a result of its adaptability, including its high reproductive potential and precocial life-style, which in the case of the Sri Lankan perennial reservoirs is advantageous. The continued introduction of O. niloticus may result in hybridization between these two species, this being evident to a small extent in two major reservoirs. The O.mossambicus × O.niloticus hybrids or hybrids of the reverse cross are predominant in males (Hickling, 1971). The hybrid is likely to have a higher rate of growth but, it is not certain whether the intricate adaptability seen in O.mossambicus would be found in the hybrid; would the hybrid be equally viable and maintain a high reproductive rate and a precocial life style which is needed to sustain the fishery? Would the population balance be upset by a predominance of male hybrids after a few years? The impact of the stocked O.mossambicus and on tilapia fishery should be closely monitored to clarify the possible interaction of these two species. It should be noted that the hybrid found in the Sri Lankan reservoirs is different from the mutant hybrid, ‘red tilapia’, which is gaining importance in intensive culture in parts of Asia. Apart from the isolated instance reported by Chandrasoma (1986) there has been no reports of complete replacement of O.mossambicus by O.niloticus either in Sri Lanka or elsehwere. In Sri Lanka the reservoir dried up completely and on refilling it was stocked heavily with O. niliticus which continues to dominate the yield. Such complete drying up of major reservoirs in Sri Lanka is of rare occurrence and one cannot expect the process to be repeated. Hence it is not a strict replacement of O.mossambicus by a more desirable species.

Extensive hybridization in large water bodies has been reported from Lake Itasy, Madagascar, between O.macrochir and O.niloticus (Daget and Moreau, 1981). In this instance the advantages were very short lived and necessarily result in an obligatory benefit for will not necessarily result in an obligatory benefit for fish production or for the preservation of the indigenous fauna.

Declining viability and long term detrimental effects on water quality from tilapia species replacement, which were expected to be favourable, have also been reported. In Laka Kinneret, Israel, stocking of Sarotherodon aureus is reported to have undesirable impact on water quality as well as on S.galilaeus, the native cichlid species of the Lake (Gophen et al., 1983). According to Gophen et al. (1983) yields of the most commercially valuable S.galilaeus have declined inspite of intensive stocking and the introduced S.aureus competes with the former for the same food source which resulted in the disruption of the trophic dynamics of the lake.

6.3 Use of irrigation canals for fish production

It has been estimated that by the year 2000 two thirds of the total stream flow in the tropics will be regulated (Szestay, 1972), and in the South East Asian tropical belt reservoirs will cover 20×106 ha (Fernando, 1980). This vast reservoir storage is primarily meant for irrigational purposes, and it will involve an extensive and an intricate canal systems. Estimates of the extent of irrigation canals in most countries are not available. Sudan for example is reputed to have 18,500 km of minor canals with an approximate width of 4-8 m (Coates, 1984).

The potential use of these irrigation canals for fish production have received little attention worldwide. Irrigation canals in China, with the flow velocity between 0.1-0.3 m sec-1 are utilized for semi-intensive culture (Tapiador et al., 1977). Beveridge (1984) considered in detail the suitability of lentic water for cage- and pen-fish culture. In Indonesia irrigation canals are used for extensive culture by partitioning stretches of canals with stakes. Irrigation canals are also known to increase the potential spread of human diseases, particularly parasitic diseases involving secondary molluscan hosts, especially in the African continent (Petr, 1978) and the use of fish in the management of ecological problems has received some attention (Coates, 1984).

In Sri Lanka most of the canals are unlined and the major canals retain water throughout the year, or through most of the year. These were identified as potentally suitable habitats for crayfish by De Silva (1985), and he recommended that suitable species be identified for introduction. If crayfish cannot be introduced, a significant portion of the irrigation canals should be used for fish production.

Plate IX.

Plate IX. Irrigation canals: note the overhanging weeds and grass on the banks.

In Sri Lanka the major irrigation canals, for most of their length are 2–4 m wide and 1.0 to 3.0 m deep and retain around 0.5 to 1.0 m of water. The banks of the canals are covered in vegetation, which is covered or uncovered with the fluctuating water level. It could be possible to portion out stretches of canals with wooden stakes and prevent the movement of fish out of such area while permitting a free water flow. Depending on the nature of the bank, average speed of water flow, nature of the bottom sediment, the extent of usage of the canal for domestic purposes, and most of all depending on the extent of care and feed that the person involved in the fish and crayfish culture operation is able to provide, suitable species could be introduced into selected lengths of such canals. The ready availability of grasses on the bunds and near the canals should be taken into consideration in the selection of species.

As the irrigation canals in Sri Lanka are not utilized for transportation, portioning out of canal stretches should not pose a problem. It is surprising that the use of this resource has hitherto attracted only marginal attention in the Indo-Pacific region as a whole. A concerted effort in this direction, including research and experimental work which could perhaps be organized on a regional basis, is likely to pay high dividends. If efforts of rehabilitating of paddy-cum-fish culture and that needed for utilization of irrigation canals are compared, the latter is likely to be more favourable; the latter involves less conflicts between users and does not call for changes in some of the conservative agricultural traditions of most countries of the region.

6.4 Predictive modelling and its applicability to Sri Lankan reservoirs

The importance of predictive modelling using physico-chemical and biological parameters as a tool in bringing about effective management in natural lakes has been aptly demonstrated over the last two decades. The simplest and the most widely used index- the morphoedaphic index (MEI)- for predicting fish yields was introduced by Ryder (1965). As pointed out by Ryder (1982) it has been criticized both as a methodology and a concept. Since the introduction of the MEI concept a number of other parameters such as the gross primary production (Melack, 1976), phytoplankton standing crop (Oglesby, 1977), total phosphorus concentration and macrobenthos biomass/ mean depth (Hanson and Leggett, 1982) have been utilized to develop predictive models of fish yields and or biomass of lacustrine water bodies. Relationships between fish yields and lake area (Youngs and Heimbuch, 1982), and total dissolved solids (TDS) with temperature included as a parameter (Schlesinger and Regier, 1982) have also been derived. In the tropical region the MEI (Henderson and Welcomme, 1974) and primary production (Melack, 1976) have been successfully utilized to predict fish yields of groups of water bodies. The applicability of available predictive models for African lakes and reservoirs (pre-and post- impoundment) were reviewed by Marshall (1984) and Oglesby (1985) pointed out the need to predict potential yields and compare those with current harvests as a first critical step to management of the lacustrine fisheries in the tropics.

With respect to tropical Asia so far only three attempts have been made to develop predictive models. Melack (1976) used data on primary production and fish yields available for small water bodies in Tamil Nadu, India, while Wijeyaratne and Costa (1981) and Wijeyaratne and Amarasinghe (1984) attempted to develop a predictive model based on the MEI concept.

For the major perennial reservoirs of Sri Lanka Wijeyaratne and Costa (1981) found the potential yield (C) to be related to the MEI in the following manner:

Log C = log 19.0677+0.7050 log MEI

The validity of this derivation was questioned by De Silva (1985a) on the grounds that (a) the data base was limited, (b) that the authors rejected some of the data on the assumption that they were unrealistic, although all data were from the same source, and (c) that the basic assumptions that in the five reservoirs the volume fluctuations are minimal was incorrect. Subsequently Wijeyaratne and Amarasinghe (1984) found the maximum sustainable yield (MSY), computed from the simplified version of the Schaefer model for catch statistics, to be related to the MEI based on a more broader data base. The relationship is defined by the following equation:

Log- MSY = 0.9005 log- MEI + 1.922

The MEI concept is not applicable to Sri Lankan reservoirs, because reservoir characteristics do not fulfill the assumptions made in the derivation of the concept (Ryder, 1982). The relative shallowness, the high draw-down and flushing rates of the Sri Lankan reservoirs do not permit them to act as basin releasing nutrients for biological processes. Moreover, most of the major reservoirs in Sri Lanka do not act in isolation but as components of an complex system. In such a context the productivity of a reservoir may be influenced by factors other than those that are common to a system, such as the specific inputs into each reservoirs, its own catchment characteristics and perhaps some others.

Cochrane and Robarts (1986) have shown that in culturally eutrophicated waters the ionic composition is disturbed and the relative contributions of nitrogen and phosphorus increase and the conductivity therefore ceases to be a valid indicator of nutrient concentrations. Most major reservoirs of Sri Lanka are culturally eutrophicated and this could be another reason for the lack of a relationship between MEI and fish yield.

Possible relationships (linear, semi-log, loglog) between MEI and reservoir area to the total yield and/or the catch per unit effort were explored for data given in Table 31. Statistically valid relationships obtained from this computation are summarised in Table 31 and are also shown in Fig.19.

For the dominant O.mossambicus fishery of the major reservoirs, the total yield and catch per unit effort CPUE are related significantly to reservoir surface area, the most significant relationships being a simple linear one (De Silva, 1985a). The lake area has been shown to be a powerful predictor of fish yield for African lakes (Youngs and Heimbuch, 1982). It is suggested that this model be utilized at present until further refinements/new models are developed using other parameters. At present data are not available to explore the possible relationships between yield and the primary productivity, chlorophyll content, and total phosphorus. The large array of Sri Lankan reservoirs with diverse hydrology and limnology provides an ideal opportunity to test such modelling, which may provide further insight into the use of empirical indices for assessment of fish production.

The yield statistics for the seasonal tanks are accurate, and where other parameters are available, modelling of yield in relation to total dissolved solids, total phosphorus, reservoir area, primary productivity, chlorophyll content and suspended organic matter and perhaps some other parameters could provide evidence for adjustment of stocking densities and species composition. This element of research needs to be enhanced.

The seasonal tanks unlike the major reservoirs are small isolated water bodies. The allochthonous nutrient input from cattle and human activities are likely to have a significant influence on their nutrient load. The extent of these influences is highly variable from tank to tank, primarily being dependent on the number of direct users and the size of the adjacent village population. These factors will undoubtedly complicate modelling but they will also provide a greater challenge to the inquisitive scientific mind.


Fig.19 Relationships between yield or CPUE and area.

Table 31

Statistically valid relationships between yield (t annum-1) and CPUE (t boat-1annum-1) and other parameters for major perennial reservoirs in Sri Lanka. The data base is given in Table 18. (a-SE of slope ; b-SE of intercept; r=correlation coefficient).
Total Yield(TY)
1nTY=0.801 1nA-0.1490.2151.6100.67<0.01
CPUE (CE)/Area(A)    

6.5Development programmes

The reservoir capture fishery in Sri Lanka is barely three decades old and the reservoir culture fishery is only six years old. The accent on the reservoir fisheries is a post - 1980 phenomenon. Prior to that in Sri Lanka the accent was on the marine sector and the inland sector was relatively neglected, with the cheap source of protein remaining underutilized.

All major fisheries development programmes in Sri Lanka have had a substantial foreign funded component. All development programmes have concentrated on upgrading of existing and/or building of new infrastructural facilities, mainly hatcheries. They have also involved training of middle and lower technical grades. However, none of the development programmes has envisaged a research component. The shortage of baseline information is still one of the major constraints for obtaining a better yield in the seasonal tanks, and an experimental programme for determining optimal stocking densities for different types of seasonal tanks is much needed.

The total cost of the Aquaculture Development Programme funded by the Asian Development Bank is U.S. $ 21.65 million, spread over a period of six years. It is believed that an allocation of 1% of the project cost to research could result in developing management strategies for obtaining better yields, evaluating the suitability of seasonal tanks for culture purposes and developing suitable post-harvest technology.

A number of other programmes considered for the future, and to be funded by foreign governments or international organizations, concentrate on the development and/or building of infrastructural facilities, with little attention being given to considerations for reservoir fisheries. It has been pointed out that for the seasonal tank aquaculture Programme a minimum fingerling requirement would be about 20 million by 1990, and that producing the fry requirements poses no problems. However, not sufficient attention, in my view, was given to the fingerling production. Construction of numerous new hatcheries throughout the island is likely to increase the demand for experienced man-power possibly resulting in an overall decrease in fingerling production as such demand cannot be met within a short time. Consolidation of the already available facilities, perhaps with improvements to those in which constraints to further development are minimal, may be currently a better approach, than building of new hatcheries.

6.6 Freshwater fish as a protein source

The per capita fish consumption in Sri Lanka is estimated as 21.8 kg annum-1, and fish account for 56.3% of the total protein animal intake (Josupeit, 1981). As such Sri Lanka ranks sixth amongst Asian countries in the importance of fish in the diet. Subasinghe (1982) estimated that 25% of the total population is dependent on freshwater fish for their animal protein requirements; this population component is essentially being restricted to the dry zone and constitutes the farming community. The projected animal food requirements for Sri Lanka for the period 1981–2001 (Table 32) shows that by year 2000 the freshwater fish production must be increased by nearly 2.33 times the present level of 27,000 to 30,000 t annum-1, to meet the projected requirements.

These figures indicate the increasing importance of freshwater fish as a protein source since 1958 when freshwater fish consumption was almost zero.

Is it possible for inland fisheries to fulfill these requirements? As discussed earlier, the freshwater fish production will mainly depend on
- the existing capture fishery in major perennial reservoirs,
- introduction of a subsidiary gill-net fishery for minor cyprinids in major perennial reservoirs, which is estimated to yield 15–30% of the yield of the existing fishery,
- the seasonal tank aquaculture programme, and
- intensive aquaculture, such as cage and pen culture in major perennial reservoirs.

Table 32

Projected animal food requirements for Sri Lanka (from Subasinghe, 1982; *-assuming 25% of the population is dependent on freshwater fish).
Mid-year population (x104)15.3718.3821.29
Fish & fishery product requirements
(60% of the total in 103 tons)251.5304.9357.5
Freshwater fish requirements   
Total animal protein requirements   

On the basis of the predicted yields (Table 33) there would be a shortfall of nearly 5500 tons in the freshwater fish requirements by year 2000. However, it is expected that the yields from both the existing capture fishery and the seasonal tank aquaculture programme could be considerably increased with the introduction of proper scientific managerial measures over the next 15 years or so and with the harnessing of untapped fish resources in the perennial reservoirs. The increasing cost of production of seafish is likely to increase the demand for freshwater fish. Sri Lanka is also likely to take up extensive forms of aquaculture, on a relatively large scale, to meet the shortfall in freshwater fish supplies.

6.7 Research needs

In the earlier sections it has been stressed that effective managerial measures for optimization of the yields from different reservoir resources should be based on a sound scientific research. Sri Lankan reservoirs provide a diversity of water bodies for testing basic limnological hypotheses and empirical models on relationships between fish yields and a diversity of physico-chemical and biological parameters. The need for research of these aspects has been emphasised earlier. The envisaged irrigation expansion should be preceded by detailed pre-impoundment studies of rivers. Such an opportunity was lost in the case of the Mahaweli Basin Development (De Silva, 1985) and also to lose future opportunities for detailed preimpoundment environmental studies would be regrettable.

Table 33

The expected yields from reservoir resources on the basis of available figures at present.
Water ResourceMethodYield(mt)
Perennial reservoirsCapture fishery24,863
(87854 ha)283 kg ha-1yr-1 
 (- Table 19) 
–do–Capture fishery for minor 
 cyprinids -low estimate3,729
 -high estimate7,458
Seasonal tanksExtensive culture20,000
-20,000 ha byit ha-1annum 
Intensive practicesExtimated5,000
in perennial reservoirs  
 Total (low)53,592

There is an urgent need to commence a survey on fish as food. Sri Lanka is often cited as a classic example of the successful use of DDT for control of malaria in the dry zone (Domros, 1976; Perera, 1984; Weeraratne, S., 1983) Over the 1970–1980 period there was a 3.5 fold increase in the pesticide importation (Weeraratne, S., 1983) into Sri Lanka, most of which is utilized for rice paddies in the dry zone. Because of the intricacy of the irrigation systems, particularly in the dry zone region of the island, a significant percent -age of these pesticides are likely to be carried into reservoirs, and some will accumulate through the food chain, in fish. It is suggested that quantification of pesticide levels in the different food fish species along ‘chains’ of reservoirs should be given priority.

Due to the increasing importance of the reservoir fisheries, there is a need to intensify research on the dominant, economically important fish species. With the exception of the biology of O. mossambicus that of other species is little understood. Basic biological research should complement research on aspects of management, for the latter to be implemented property. In South East Asia research on lesser known species of reservoir fauna is still poor or nonexistant. In some instances some of these species, either singly or collectively, could provide a subsidiary fishery.

Previous Page Top of Page Next Page