PREDATOR-PREY RELATIONSHIPS, FISHERIES PRODUCTIVITY AND FISH POPULATION DYNAMICS IN LAKE KARIBA - A REVIEW

by

C. Machena
Lake Kariba Fisheries Research Institute
P.O. Box 75
Kariba
Zimbabwe

ABSTRACT

The tigerfish, (Hydrocynus vittatus), is the major predator in the Lake Kariba system. Its prey preferences have been changing since the dam was built and are consistent with the changes in the abundance of prey species. The introduction and successful establishment of Limnothrissa miodon has played an important role both in increasing the size of the of the tigerfish population and in extending its habitat into more pelagic areas. Other predators are present in Lake Kariba but there is little data available on them.

Fish yields in the lake were initially high following high nutrient loads from flooded vegetation. However, there was a rapid decline in yields because the large outflow caused a rapid depletion of the nutrients.

Estimations of biomass and yield have been made by a number of authors using a variety of methods. Mortality coefficients of a number of species are now available.

1. INTRODUCTION

When the damming of the Kariba Gorge was completed in December 1958, Lake Kariba became the largest man-made lake in Africa though it has since been surpassed by Lake Volta.

The formation of the lake created development prospects in the infertile Sebungwe Valley. Although the production of hydroelectric power provided little benefit to the local inhabitants, the development of a fishery has been very helpful to them. Minshull (1973) analysed the early development of the fishery and reported disappointing fish yields after an initially encouraging production. The high initial yields were correlated with a high nutrient load from flooded vegetation (Coche, 1974) but the nutrient levels quickly decreased because of the large lake outflow.

The introduction of the pelagic sardine, Limnothrissa miodon created an economically viable industry on the Zimbabwean side. Concommitant with the development of the fishery has been the development of a programme of fisheries and limnological research on the lake.

This paper is a review of the research conducted on predator-prey relationships, population dynamics and fisheries productivity on Lake Kariba. The work conducted prior to 1973 is reported in Balon and Coche (1974).

2. PREDATOR-PREY RELATIONSHIPS

2.1 Hydrocynus vittatus

Donnelly (1969, 1971), Kenmuir (1973) and Mitchell (1976) concurred that the tigerfish, Hydrocynus vittatus, is the major predator in the Lake Kariba system. It is an extremely active and efficient predator which preys upon fish up to 40% of its own length (Bell-Cross, 1976). Donnelly (1969) suggested that the behaviour patterns of the prey species is dictated by the predatory behaviour of H. vittatus.

Kenmuir (1973, 1975) conducted an intensive study on the ecology of H. vittatus in the lake between May 1969 and May 1972. Three salient features emerge from his study:

1. H. vittatus is predominantly a predator.
2. There has been a change over the years in the food preferences of H. vittatus which is consistent with the change in the abundance of the prey species.

3. The fry and fingerlings of H. vittatus occupy a different habitat and have different food preferences from the adult.

In the first period of Kenmuir's (1973) study, (April, 1969 to March, 1970), of 2,144 tigerfish examined, 34.1% and 20.6% of the diet consisted of cichlids and Alestes lateralis respectively (Table 1). Bowmaker (1973), working largely between 1967 and 1969, found that A. lateralis and Micralestes acutidens comprised 45.1% and unidentifiable fish remains comprised 44.7% of the food. In the second period of Kenmuir's study (April, 1970 to March, 1971) the importance of cichlids in the diet had diminished to 15.3% and the major food item had become Limnothrissa miodon which comprised 41.1% of the diet (Table 1) (Kenmuir, 1973). Thus, within the two year study period, there had been a major shift in diet from A. lateralis and cichlids to L. miodon. This shift coincides with the introduction of L. miodon into the Lake between 1967 and 1968 (Bell-Cross and Bell-Cross, 1971) and its successful establishment by 1970 (Balon, 1971; Begg, 1974).

It appears that H. vittatus is an opportunistic feeder which preys upon the most abundant potential prey species. Donnelly (1971) showed that during the years 1964 to 1971, H. vittatus fed upon the prey which was most abundant and available at any one time (Table 2). This seems to be a common pattern amongst predatory fish as was shown in Lake Erie, where Knight et al (1984) demonstrated that walleyes and yellow perch feed on the most abundant appropriate-sized prey. In Lake Kariba, M. acutidens was initially abundant (Matthes, 1968) but became scarce after the filling period was over. The earliest record of A. lateralis, which seems to have replaced M. acutidens in the lake, dates back to 1965 (Donnelly, 1971). The A. lateralis population expanded and reached its peak in 1967/68. This was followed in 1970 by the rapid expansion of the L. miodon population (Mitchell, 1976).

L. miodon is a preferred prey because of its small size, soft fins and high turnover rate which maintains high population levels. Equally important is the fact that it occupies open water and is thus vulnerable to predators.

Jackson (1961), Donnelly (1971) and Kenmuir (1975) showed that H. vittatus becomes predominantly ichythiophagus at an approximate length of 40mm whilst below that length it is basically entomophagous. Kenmuir's (1975) data is summarised in Table 3.

Whereas fingerlings below 24mm in length feed largely on zooplankton (Cladocera and Copepoda), those between 24 and 40mm in length feed on both zooplankton and insects. Amongst the Cladocera, the adults and nauplii of Diaphanosoma sp. are the most important prey comprising 34% of the diet. Of the Copepoda, Tropodiaptomus sp. and Mesocyclops sp. are the most important and comprise 52.8% of the diet. Ephemeroptera nymphs and Hemiptera and chironomid larvae are the most important insect prey comprising 68.4%, 17.5% and 12.2% respectively of the total insect component of the diet.

H. vittatus is largely an open water species with a natural preference for regions fairly near to the shore with a water depth of less than 15m. It is rarely found amongst aquatic vegetation but usually cruises well above the substrate (Langerman, 1984). In Lake Kariba, the introduction of the pelagic L. miodon has extended the habitat of H. vittatus to deeper, offshore water where it cruises near to the surface. The prey fishes which inhabit the shallow inshore waters are usually found near vegetation where they can seek refuge, a habit which Donnelly (1969) suggests has been influenced by H. vittatus.

2.2. Limnothrissa miodon

L. miodon is largely planktivorous, with the cladoceran Bosmina longirostris and the copepod Mesocyclops leuckartii accounting for over 80% of the food intake (Table 4) (Cochran, 1984). Begg (1974) found B. longirostris to be the single most important food item, accounting for 80% of the food intake of L. miodon. Other food organisms are taken and Mitchell (1976) found that in the inshore areas, as much as 55.5% of the food of L. miodon is of terrestrial origin (insects etc.).

2.3. Other fish species

There are a number of other predatory species in Lake Kariba, but these tend to be generalised predators as distinct from the more specialized H. vittatus and L. miodon. The feeding habits of these fish, which include Mormyrops deliciosus, Mormyrus longirostris, Marcusenius macrolepidotus, Alestes imberi, A. lateralis, Micralestes acutidens, Hippopotamyrus discorhynchus, Heterobranchus longifilis, Clarias gariepinus, Eutropius depressirostris, Synodontis zambezensis, S. nebulosus, Haplochromis darlingi, H. codringtoni and Pseudocrenilabrus pilander, are described by Bowmaker (1973) and Mitchell (1976). The following prey items are taken by the above species but their relative importance in the diet varies from species to species:- Caridina and Cyclaestheria (crustacea), Povilla adusta nymphs, anisopteran nymphs, tricopteran larvae, chironomid and Chaoborus larvae, Diaphanosoma and Bosmina (zooplankton) and fish etc. For example, the mormyrids prey heavily upon Caridina and Povilla nymphs whilst molluscs (Bulinus, Biomphalaria, Curbicula etc.) are only important as prey items for S. zambezensis, C. lazera and H. codringtoni.

2.4. Piscivorous birds

There has been very little work on any bird species associated with the lake. The only major work is that of Birkhead (1978) on the reed cormorant (Phalacrocorax africanus) and the darter (Anhinga rufa). Both species have a similar diet with cichlids representing over 90% by numbers and 70% by weight (Table 5). It has been estimated that approximately 6.4 tonnes of fish are taken per year from a 5km stretch of shoreline by 4.59 tonnes of cormorants and 1.82 tonnes of darters. Junor (1972), working in lake Kyle (Zimbabwe), estimated the mean daily consumption for various cormorants to be 10–14% of their body weight. Both P. africanus and A. rufa feed in the littoral areas of the lake which have a mean depth of 2m. The size of the prey taken is related to size of predator (Table 6). Table 5 shows that a wide variety of prey is taken and there seems to be little ecological separation with respect to prey type and feeding areas. Birkhead (1978) attributes this to an abundant food supply, particularly at a 2m depth where Coke (1968) says the maximum density of fish occurs.

The other piscivorous birds in the lake are Haliaeetus vocifer (fish eagle), Ardea goliath (goliath heron), Ceryle rudis (pied kingfisher) and Milvus migrans parasitus (yellow-billed kite). There seems to be no quantitative data on these species. The above species are primarily inshore predators though C. rudis and Chlidonias leucopterus (whitewinged black tern) also prey upon the largely pelagic L. miodon (Junor, 1972; Begg, 1973).

3. FISHERIES PRODUCTIVITY

3.1. Biomass estimations

Estimations of the fish biomass of the inshore areas of Lake Kariba were carried out by Balon (1973), Mitchell (1976) and Langerman (1984). Balon and Mitchell (op. cit) worked in shallow areas where they were able to block off coves with nets and rotenone all the fish within the enclosures. Balon sampled deeper areas than Mitchell. This approach limited the programme to shallow coves and bays which could be completely netted off. Langerman conducted his sampling by using an explosive grid and extrapolated the weight of fish recovered from each quadrat to one hectare. This approach was more flexible and permitted sampling in deep water and on rocky shores that are not easily accessible. Whereas Balon's work was more extensive, covering large areas of the lake, Mitchell limited his sampling to the eastern Basin while Langerman restricted his work to an unfished area in the middle of the lake. These differences in approach may explain the differences in the results. The biomass estimates from these workers are summarised in Table 7.

Mitchell and Langerman's ichythiomass values are similar and this is probably due to the similarity of the habitats sampled (Marshall, 1984). Balon's higher value could reflect a wider range of habitat sampled. The importance of habitat type is shown in Table 8. Gentle slopes with soft mud and silt have rich vegetation and these support high fish biomasses. In contrast rocky shores have little vegetation and therefore less fish (Marshall, in prep).

The higher values of H. vittatus from Langerman is a reflection of the method used. Langerman was able to sample in deeper water where the other methods would not be effective. The high values of M. deliciosus, H. discorhyncus and C. gariepinus from both Balon and Mitchell reflect the presence of extensive mats of Salvinia molesta, (which are favoured by mormyrids (Joubert, 1975)), at the time of sampling. Mitchell found that under permanent S. molesta mats, M. deliciosus was the dominant species comprising 42% of the standing stock. Mitchell suggested that these mormyrids are nocturnal and take refuge under the mats during the day.

Balon extrapolated the biomass estimates to obtain a standing stock for the lake. His figure of 17,814 tonness is based on the assumption that 25% of the 0–20m zone is occupied by fish.

3.2. Fish yields

Inshore fishing in the lake began on the Zambian side in 1960 (Marshall, 1979) and on the Zimbabwean side in 1962. Both artisanal fishing by local villages and commercial fishing developed (Minshull, 1973). Catches from both sides of the lake were initially good but declined from a peak of 5,200 tonnes in 1963 (from both the Zambian and Zimbabwean shores) to 1,200 tonnes (from the Zimbabwean shore) in 1970 (Minshull, 1973) (Table 9). There seem to be no catch data from the Zambian side since 1968. Inshore catches on the Zimbabwean side appear to have stabilised at about 1,000 tonnes a year.

The initial high yields were the result of the abundant nutrients available from flooded vegetation but the nutrient level soon declined, as the lake's rapid volume-replacement time of about 4 years meant that nutrients were quickly flushed out. This is reflected in the change in conductivity of the water which reached a peak of 111 u mhos in 1959 but soon dropped and stabilised at 80 u mhos (Coche, 1974).

Annual sardine (L. miodon) catches in Kariba have increased from 66 tonnes in 1973 when fishing started (Marshall et al, 1982) to 11,000 tonnes in 1981 (Marshall, 1981) (Table 10). The increase up to 1980 reflects an increase in fishing effort as more permits were issued. Since 1980, no further permits have been issued and the catch fluctuations from 1980 onwards may therefore reflect fluctations in stock size. L. miodon has a short life cycle and therefore responds rapidly to fluctations in the environment, particularly to food supply. There appears to be a correlation between rainfall in the catchment area and sardine yields, as the major source of nutrients in the lake is from river inflow.

3.3. Yield from prediction models

Machena and Fair (in press) used Melack's (1976) regression model : Log FY = 0.113 GP + 0.91 where FY is fish yield (kg ha^-1 yr^-1) and GP is gross photosynthesis (g O₂m^-2 d^-1 to predict fish yields in Lake Kariba. The predicted figure of 27 kg ha^-1 is very close to the known yield of 28 kg ha^-1 (Table 11).

Marshall (1985) gives predicted yields for the lake (Table 11). Marshall's (1985) value of 29 kg ha^-1 predicted from the morphoedaphic index of Lake Kariba, differs from the value he gave earlier of 23.2 kg ha^-1 (Marshall, 1982) but no explanation for this is given.

It seems that the PO₄ prediction (Table 11) is well below the known yields and may thus not be very useful. These empirical models are derived from known yields and thus their accuracy is influenced by the degree of fishing on the lakes selected, and on the completeness of fishery data from those lakes. Ryder et al (1974) point out that for yield to be used as an index, lakes must be subjected to moderately intensive to intensive fishing pressure.

In spite of its limitations, Rigler (1982), Jones and Hoyer (1982) and Marshall (1984) argue that empirical modelling is important in that it motivates the search for better correlations and the collection of new data.

4. POPULATION DYNAMICS

Balon (1974) provides mortality coefficients for a number of inshore fish species from Lake Kariba (Table 12). Species such as Alestes lateralis, A. imberi, Micralestes acutidens and Pharyngochromis darlingi have short life spans and instantaneous rates of total mortality that are very high. These high mortality rates are related to the high density of these species. Hippopotamyrus discorhynchus is an exception as it has a relatively long life span of 4–6 years but has a high instantaneous mortality rate. This is probably due also to the high population density of the species (Balon, 1974).

Langerman (1984) calculated the instantaneous mortality coefficient for Hydrocynus vittatus in the unexploited Mwenda estuary and in the commercially exploited Sanyati basin. Langerman's coefficient of 0.58 in the Mwenda estuary (Table 13) is similar to Balon's coefficient of 0.52. In the Sanyati basin however, Langerman demonstrated an increase in mortality coefficient from 0.65 in 1965 to 0.73 in 1981 (Table 13), and this is a strong indication of increasing fishing pressure. Langerman's very high mortality coefficients for H. vittatus from the pelagic fishery (Table 14) seem to support the effect of increasing fishing pressure, as in the pelagic zone H. vittatus is caught in the Limnothrissa miodon fishing nets.

Marshall (unpubl.) provides the mortality coefficients for the pelagic L. miodon in the Bumi estuary and Sanyati Basin (Table 15). These mortality rates are very high, perhaps corresponding to the rapid life cycle (less than 3 years; Balon, 1974) and also the high density of the species.

5. CONCLUSION

Research into the biology of Lake Kariba's fish stocks is continuing. Current studies place emphasis on the evaluation of the population dynamics of the major commercial species - the cichlids Serranochromis codringtoni, S. macrocephalus, Oreochromis mortimeri, O. macrochir and Tilapia rendalli and the sardine Limnothrissa miodon. There is a need to work out population dynamics parameters for these species and make definite recommendations that will ensure sustainable yields.

6. REFERENCES

Balon, E.K. 1971 First catches of Lake Tanganyika Clupeids (Kapenta-Limnothrissa miodon) in Lake Kariba. Fish.Res.Bull. Zambia, 5:175–85

Balon, E.K., 1973 Results of fish population size assessments in Lake Kariba coves (Zambia) a decade after their creation. Geophys.Monogr., 17:149–58

Balon, E.K., 1974 Total production, available production and yield of major fish taxa from Lake Kariba. Monogr.Biol., 24:428–45

Balon, E.K. and A.G. Coche. 1974 Lake Kariba: a man-made tropical ecosystem in Central Africa. Monogr.Biol., 24:767p.

Begg, G.W. 1973 The feeding habits of the whitewinged black tern on Lake Kariba Ostrich, 44:149–53

Begg, G.W., 1974 Investigations into the biology and status of the Tanganyika Sardine Limnothrissa miodon in Lake Kariba Fisheries Research Institute Project No. 17 (mimeo)

Begg, G.W., 1976 The relationship between the diurnal movements of some of the zooplankton and the sardine Limnothrissa miodon in Lake Kariba. Limnol.Oceanogr., 21: 529–39

Bell-Cross, G. and B. Bell-Cross, 1971 Introduction of Limnothrissa miodon and Limnocaridina tanganicae from Lake Tanganyika into Lake Kariba. Fish.Res.Bull.Zambia, 5:207–14

Bell-Cross, G., 1976 The Fishes of Rhodesia. Salisbury, National Museums and Monuments of Rhodesia.

Birkhead, M.E., 1978 Some aspects of the feeding ecology of the Reed Cormorant and Darter on Lake Kariba, Rhodesia. Ostrich, 49:1–7

Bowmaker, A.P.M., 1973 An hydrobiological study of the Mwenda river and its mouth. Ph.D. thesis, University of Witwatersrand, Johannesburg

Coche, A.G., 1974 Limnological study of a tropical reservoir Monogr.Biol., 24:1–247

Cochrane, K.L., 1984 The influence of food availability, breeding seasons and growth rate on commercial catches of Limnothrissa miodon in Lake Kariba, Zambia - Zimbabwe. J.Fish.Biol. 24:623–36

Coke, M., 1968 Depth distribution of fish on a bush-cleared area of Lake Kariba, Central Africa. Trans.Am.Fish.Soc., 97:460-5

Donnelly, B.G., 1969 A preliminary survey of Tilapia nurseries on Lake Kariba during 1967/1968. Hydrobiologia, 34:195–206

Donnelly, B.G., 1971 The fish population changes on Lake Kariba between 1960 and 1068. Part II. Characidae and Atharinidae. Lake Kariba Fisheries Research Institute. (mimeo)

Jackson, .P.B.N., 1961 Ichthyology. The fish of the Middle Zambezi. Kariba Stud., 1:1–36

Jones, J.R. and M.V. Hoyer, 1982 Sportfish harvest predicted by summer chlorophyll a concentration in Midwestern Lakes and Reservoirs. Trans.Am.Fish.Soc., 111:176–79

Joubert, C.S.W., 1975 The food and feeding habits of Mormyrops deliciosus (Leach) 1818 and Mormyrus longirostris Peters 1952 (Pisces : Mormyridae) in Lake Kariba. Kariba Stud., 5:68–85

Junor, F.J.R., 1972 Offshore fishing by the pied kingfisher (Cyrule rudis) in Lake Kariba. Ostrich 43:185

Kenmuir, D.H.S., 1973 The ecology of the Tigerfish, Hydrocynus vittatus Castelnau, in Lake Kariba. Occas.Pap.Natl. Mus.Rhod., 5:115–70

Kenmuir, D.H.S., 1975 The diet of fingerling tigerfish, Hydrocynus vittatus Castlenau, in Lake Kariba, Rhodesia. Arnoldia,Rhod., 9:1–8

Knight, R.L., F.J. Margraf and R.F. Carline, 1984 Piscivory by Walleyes and Yellow Perch in Western Lake Erie. Trans.Am.Fish Soc., 113:677–93

Langerman, J.D., 1984 Optimum harvest strategies for tigerfish in Lake Kariba Zimbabwe. M.Sc. thesis. University of Witwatersrand. Johannesburg.

Machena, C., 1983 A study of the primary productivity of the Sanyati Basin of Lake Kariba and the application of a regression model to estimate fish yields in the lake. M.Sc. thesis. University of Zimbabwe.

Machena, C. and P. Fair, Comparison of fish yields from in press prediction models between Lakes Tanganyika and Kariba. Hydrobiologia

Marshall, B.E., 1979 Fish populations and the fisheries potential of Lake Kariba. S.A.J.Sci., 75:485–8

Marshall, B.E., 1982 Fish production potential of Lake Kariba. Zimbabwe Agric.J., 79:173–7

Marshall, B.E., 1984 Predicting ecology and fish yields in African reservoirs from preimpoundment physico-chemical studies. CIFA.Tech.Pap., 12:1–21

Marshall, B.E., 1985 Limnology and fisheries management in Zimbabwe. Verh.Internat.Verein.Limnol., 22:2671–5

Marshall, B.E. and Junor, F.J.R. 1981 The decline of Salvinia molesta on Lake Kariba, Zimbabwe. Hydrobiogia 83:477–84

Marshall, B.E. and Shambare, L. 1984 Fisheries Statistics, Lake Kariba. L. Kariba Fisheries Research Institute Project Report, 48:1–21 (mimeo)

Marshall, F.J.R. Junor and J.D. Langerman, 1982 Fisheries and fish production on the Zimbabwian side of lake Kariba. Kariba Stud., 10:177–231

Matthes, H., 1968 The food and feeding habits of the tigerfish, Hydrocynus vittatus (Cast.1861) in Lake Kariba. Beaufortia, 15:143–53

Melack, J.M., 1976 Primart productivity and fish yields in tropical lakes. Trans.Am.Fish.Soc., 105:575–80

Minshull, J.L., 1973 The establishment of a fishery among the resettled Batonka people at Lake Kariba, Rhodesia. In Eleventh Congress, International Commission for Large Dams. pp.413–9

Mitchell, S.A., 1976 The marginal fish fauna of Lake Kariba. Kariba Stud., 8:109–61

Rigler, F.H., 1982 The relation between fisheries management and limnology. Trans.Am.Fish.Soc., 111:121–9

Ryder, R.A., S.R. Kerr, K.H. Loftus, and H.A. Ryder, 1984 The morphoedaphic index, a fish yield estimator - review and evaluation. J.Fish.Res.Bd.Can., 31:663

Table 1. Result of stomach content anyalyses of Hydrocynus vittatus from Lake Kariba from April 1969 to March 1971 (After Kenmuir, 1973)

Table 2. Changes in the diet of H. vittatus in Lake Kariba between 1964 and 1971 (Donnelly, 1971)

Table 3. Results of stomach content analysis of 186 H. vittatus fingerlings (Kenmuir, 1975)

* Other items included sandgrains, small pieces of vegetation, Microcystis and a shrimp, Caridina nilotica.

Table 4. Total numbers and percentage occurrence of organisms found in 100 L. miodon stomachs examined each month from June 1975 to June 1976 (Cochrane, 1984)

Table 5. Diet of Reed Cormorant and Darter: percentage of fish species by numbers and by weight (Birkhead, 1978)

Table 6. Analysis of gut contents; number, mass and length of fish caught by Reed Cormorant and Darter (Birkhead, 1978)

Table 7. Standing stock of major fish species (ie. >1%) in Lake Kariba (from Marshall, 1984)

Table 8. Effect of vegetation on fish abundance in Lake Kariba

* 1. Langerman (1984)
2. Mitchell (1976)
3. Balon (1973)

Table 9. Inshore fish yields in Lake Kariba (Minshull, 1973)

Table 10. Sardine (L. miodon) catches on the Zimbabwean side of Lake Kariba

Table 11. Predicted and known yields in Lake Kariba

Table 12. Instantaneous mortality coefficients for some inshore species (Balon, 1974)

Table 13. Mortality coefficient for H. vittatus (Langerman, 1984)

Table 14. Estimates of mortality rates for tigerfish from the pelagic fishery (Langerman, 1984)

Table 15. Preliminary mortality coefficient for L. miodon (Marshall, unpubl.)