4. FISHERY RESOURCES OF NIGERIAN INLAND WATERS (Contd.)

Table 21. Species composition, density and ichthyomass of fish captured from three stations with rotenone at Bakolori Reservoir, Sokoto State, Nigeria in May 1982.

* Classified as carnivorous species.
Percentage weight of carnivores = 0.3%
Additional species from gillnet and commercial catches = Synodontis membranaceus, Malapterurus electricus, Monnyrus rume and Physailia pellucida.

A gillnet survey was conducted with great difficulty on account of the uncleared bush. A total of 1,260 m² of net was set along the shore and a similar fleet set at the surface of the open water. Because of the problem of nets entangled in the bush, the gillnet catches did not reflect the relative abundance of species in the rotenone survey. Only nine species were recorded out of the sixteen recorded in the rotenone samples. Day fishing was attempted using the gillnets because of the relatively low transparency of the water (0.2 m). It was confirmed that the lake was highly turbid throughout the year. The ratio of day to night catches was 1:4. The estimated catch per 1,000 m² of gillnet per night was about 261 gm. This was extremely low compared with the other reservoirs discussed earlier.

A statistical frame and catch assessment survey of the artisanal fisheries was also conducted in Bakolori Reservoir. The frame survey revealed a total of about 80 professional fishermen and 313 others, located in 16 fishing villages out of a total of 27 villages identified along the shore of the lake. Thirty boats and 82 gourds were also counted. Only two out of the 30 boats were operated with outboard engines for passenger transport services. gillnet were recorded in only six villages compared with hooks in thirteen villages. The low concentration of gillnet (usually the most popular gear in other man-made lakes) was attributed to the relatively poor catches and the short life span of nets in areas with uncleared bush. This was confirmed by the experience of the survey team.

The mean catch per boat from a random selection of artisanal fishermen was 4.3 kg per night. This was low when compared with Jebba, Shiroro and Tiga Reservoirs, as expected on the basis of the low standing crop estimate for the lake. The only species recorded in the artisanal landings were Alestes spp., Clarias spp., tilapias, Synodontis spp. and Labeo spp. Alestes dominated the catch, followed by tilapias.

With a surface area of 80 km², only about 80 boats and 160 fishermen should have been registered to fish in the reservoir. At its current low production level the estimated 80 professional fishermen and 313 others then operating in the reservoir could overfish the stock.

The potential conductivity of Bakolori reservoir was estimated to be about 323.8 μS cm^-1 using the mean percentage increase between the maximum conductivity observed for The River Sokoto (Holden and Green, 1960) and that expected in the reservoir (Table 11). However, with a flushing rate of 2:1 (i.e. the reservoir empties itself twice in one year), as estimated from the discharge in the main irrigation canal of (30 m³/sec) and the storage capacity of 450 million m³, the conductivity is likely to be lower than expected with about a 12% increase above that recorded for the river (i.e. about 287 μS cm^-1). Thus with the MEI of 48 matched against the regression line of Henderson and Welcomme (1974) the expected catch could be approximated at 50 kg/ha.

The above estimate of potential yield is far above the observed standing crop of 32 kg/ha suggesting that the present fish production in Bakolori Lake is much below its potential. Assuming that the expected potential catch of 50 kg/ha is about one third of the production, then the fish production at Bakolori could be estimated at 150 kg/ha with a total production of 1,200 mt for the 8,000 ha reservoir. The expected yield at one third of the production is estimated at 400 tonnes a year.

The actual range of conductivity observed in Bakolori Lake (Adeniji, 1980) was between 58 and 70 μS cm^-1 during the rainy season. Adeniji cites FAO (1969) as confirming the low conductivity of the surface waters of the Sokoto River in the area of Bakolori Lake with figures ranging between 20 and 60 μS cm^-1 during the rainy season but rising to about 200 μS cm^-1 during the dry season. This suggests that the highest conductivity (256 μS cm^-1) observed by Holden and Green (1960) was recorded during the dry season. Adeniji (op.cit.) using a derived total dissolved solids (from conductivity records) of 47 to 58 ppm for the lakes estimates the potential fish production at about 25 to 40 kg/ha, or 200 to 300 mt/yr for the lake. This estimate was rather lower than ours which is not unexpected since our figures were based on the upper limits of dry season conductivity whereas Adeniji's estimates were based on the lower conductivities of the rainy season.

Other factors, such as low Secchi disc transparency (high turbidity) and low shoreline development (a relatively regular shoreline with few inlets) could mitigate against the productivity of Bakolori Reservoir. The relationship of Secchi disc transparency to depth of euphotic zone (light penetration) (Table 22) shows that turbidity affects light penetration, and therefore the rate of photosynthesis, resulting in decreased primary productivity. With an observed maximum transparency of 0.5 metres in May, light penetration was at its lowest in Bakolori reservoir and hence primary productivity would be expected to be relatively low. The observed pelagic primary production (Adeniji, 1980) was 0.125 mg C/1/day at 0.1 m and also at 0.3 m depth. No production was observed beyond a depth of 0.3 metres. Karlman (1973) compares the range of primary production in Kainji Lake with those of other tropical lakes. The data show that Kainji, with a range of 0.0–1.8 g C/m²/day is about the lowest, compared with Lake Volta with a range of 0.8–5.3 g C/m²/day, and Lake Nasser with a range of 3.0–7.2 g C/m²/day.

The turbidity in Bakolori was similar to that observed in Kainji Lake during the “White Flood” around September. The colloidal turbidity in Lake Kainji is attributed to the kaoline clay particles in the local soils which form the major component of the colloidal particles (Henderson, 1973). The settling time of the particles in Kainji Lake was observed to be very low (less than 50% per month) and it is believed that under the turbulent conditions of the lake it must be negligible. Whereas Kainji Lake is blessed with the arrival of the “Black Flood” from the upper Niger which flushes away the white flood and restores its blue colour and higher transparency, the turbidity in Bakolori was relatively permanent because of the turbulence of the water. In Kainji Lake, the colloidal particles in the water are observed to hinder the transmission of blue and green light and rather favour the transmission of red light. The persistence of such turbidity is bound to have a negative effect on the productivity of Bakolori Reservoir.

The poor shoreline development, with a relatively small littoral area, is bound to limit spawning and nursery areas of most of the species. The shore development factor is defined as the ratio of the actual shoreline to the shoreline of a perfectly circular lake of the same area, and is thus a measure of the shoreline extension produced by bays and other indentations of the lake. Kainji Lake with a perimeter of 720 km and surface area of 1,280 km² has a shore development factor at high water of 5.65. Natural lakes with idealized circular form have values around 2. With a surface area of 80 km² and a shoreline (perimeter) length of 107 km the shore development factor for Bakolori Reservoir is 3.4. This is much lower than that of Kainji, as can be observed from the regular shape of the shoreline on the map. Whereas Tiga and Kainji reservoirs are blessed with abundant shore vegetation during most periods of the year, on the evidence of the sparsely distributed vegetation cover along the drawdown area observed during the survey in May 1982, Bakolori reservooir was unlikely to develop a stable shoreline vegetation.

Table 22. The ratio of maximum depth of Secchi disc transparency (metres) to
maximum depth of euphotic zone (metres) in some African reservoirs.

N.B. * = Data extrapolated from mean ratio
Source: (Ita et. al., 1982)

GORONYO RESERVOIR (pre-impoundment survey)

The Goronyo dam was still under construction at the time of this investigation. However, since the main objective of the survey was to predict the productivity of the proposed Goronyo Reservoir, with an estimated surface area of 200 km² at maximum water elevation, and since the Goronyo Reservoir was to be colonized by species from the Rima River, a closer look at the species of fish in the temporary pools of the river was necessary.

The Rima is a seasonal river and flows only during the rainy season. At the time of the investigation (May 1982) only stagnant pools were observed along the channel. Fishing activities were minimal with only a few local fishermen seen using clapnets and dragnets. The dragnets, measuring over 100 metres in length and 6 metres deep, were made of nylon netting with a 1.0 inch stretched mesh. In order to predict the future species composition and successional changes in the proposed reservoir three stations were blocked and sampled with fish toxicant along the temporary pools of water. No portion of the river channel containing water was more than 20 metres wide within the 5 km stretch covered by the survey. Some pools were however, over one kilometre long. The pools were devoid of aquatic macrophytes but rich in filamentous algae, possibly resulting from fertilization by cattle dung littered along the edges. The bottom of the pools was muddy and rich in organic matter. In some of the pools recently disturbed by cattle the water was turbid but the particles appeared to settle and in some undisturbed pools presented the blue green appearance of an algal bloom. Some pools had a maximum depth of more than 2 metres and the channel was winding, with steep cliffs at irregular intervals along the shore. The opposite shore to most of the cliffs was always gently sloping and cattle were able to reach the water to drink without difficulty.

The relative composition of the species sampled from the temporary pools of the Rima River is compared with those of Kainji, Tiga and Bakolori Reservoirs in Table 23. The family Mochokidae ranked first in terms of percentage weight, mostly dominated by Synodontis gambiensis of relatively small sizes (mean weight 20 g). The mormyrids (mostly Petrocephalus bane) dominated in terms of percentage number (mean weight only 4 g). Only one Gnathonemus senegalensis and one Mormyrops deliciosus (mean weights 20 g and 3 g respectively) were recorded, making up a total of 3 mormyrid species. The Mochokidae were second in importance in terms of percentage number followed by the Schilbeidae with 26% and 22% respectively. The Characidae were fourth in importance in terms of number and were dominated by Alestes baremose and A. leuciscus in order of importance. The cichlids came fifth in terms of number with Oreochromis niloticus, S. galilaeus and Tilapia zillii represented in the proportion of 1.5:1:1 respectively. The cyprinids although represented by Labeo senegalensis, L. coubie and Barbus spp. were not important in terms of either number or weight. These three species had mean weights of 15,2 and 1 g respectively. The total estimated density and ichthyomass of fish per hectare were 16,911 and 141 kg respectively. An estimate of the ratio forage to carnivorous fish was 7. This was considered to be good for such exposed pools of water without cover for the forage species.

Unlike the Rima River pools, in Kainji, Tiga and Bakolori reservoirs (Table 24) the family Cichlidae ranked highest in terms of both ichthyomass and density of fish per hectare. The dominant species of cichlids in the three lakes were Sarotherodon galilaeus and Oreochrom is niloticus, popularly called tilapia. This agrees with the pattern observed in most African reservoirs where the cichlids are known to dominate. In Kainji Lake however, the cichlids were not the dominant species during the first few years after impoundment. Factors associated with unstable shore vegetation cover were responsible for the deviation of Kainji Lake from the pattern observed in other African reservoirs. Figure 7 shows a significant correlation between standing crop (ichthyomass) and density of fish in Kainji Lake.

Three other families, the Bagridae, Characidae and Cyprinidae (Table 24) were also dominant in all three reservoirs. While the tilapias dominated among the cichlids, Chrysichthys auratus tended to dominate among the bagrids and Alestes baremose, A. leusiscus and A. nurse among the characids. Among the cyprinids, Labeo senegalensis, L. coubie and Barbus species tended to dominate. There is thus a close similarity between the three reservoirs (Kainji, Tiga and Bakolori) in terms of dominant species.

Table 23. Comparison of standing stock (kg) and density of fish species and families captured in rotenone samples in Kainji Lake between 1975 and 1976 (From Ita. 1982) with those captured in Tiga (1976), Bakolori (1982) and Rima River (1982).

Seven out of the nine dominant species grew to reasonable sizes of 200 g and above in Kainji Lake (Table 25). However, in spite of the relatively small sizes of Alestes nurse, Alestes leuciscus and Barbus species, these species are of great economic importance in both Tiga and Bakolori. These small fishes, together with juveniles of A. baremose, are fried whole and sold in small heaps at relatively cheap prices to the local population. Some fishermen sun dry the excess of these small fishes as a way of preservation. This is similar to the method used in processing clupeids from Kainji Lake.

Species composition, abundance and succession in some African man-made lakes and their application to Goronyo Reservoir

The species composition of a new reservoir results directly from colonization by species from the inflowing rivers. However, because of the drastic changes in environmental conditions compared with those of the original rivers, species dominance and succession change significantly with time in man-made lakes.

A comparison of the dominant fish families in the River Niger before impoundment with those of Kainji Lake one year after impoundment revealed a drastic decline in the population of mormyrids and a dramatic increase in the population of citharinids. The mochokids, though still second in terms of number, declined. A more elaborate comparison of family dominance in some African rivers and subsequent successional changes in their respective man-made lakes (Petr, 1975) showed that in most cases, the dominant families in the river are replaced in the lake by other families which were of lesser importance in the river. In Kainji Lake however, the Citharinidae retained their dominant position during the first year of the lake's formation but were later replaced by the Characidae, which dominated the catch for about 7 years, before the emergence of the cichlids as the dominant family in the reservoir (Fig. 6).

Although the Cichlidae, (notably the tilapias) became dominant after some years in most African reservoirs, they initially required between 2 and 7 years to become established. They were not significant in either the commercial or experimental catches from Kainji Lake until 5 years after the lake's formation. In Tiga and Volta Lakes, they became the dominant group only 2 years after impoundment.

Species density and stocking rate in some man-made lakes and their implication for Goronyo Reservoir

Available evidence from large African man-made lakes constructed across perennial rivers, and smaller ones across seasonal rivers, indicates that the latter are characterized by a paucity of fish families and species at lower densities. In Kainji, Tiga and Bakolori some years after the lake's formation the following species were the most abundant within the dominant families:

The Cichlidae and Cyprinidae are primary consumers feeding on algae and decayed organic matter on the bottom of the lake. The Bagridae (Chrysichthys) and Characidae are secondary consumers feeding on aquatic insect larvae, pupae and nymphs. The Alestes spp. are opportunistic feeders with a range of food items including both animal and plant materials.

Apart from adapting to the available food in the lakes, the species also adapt to the lacustrine environment for their breeding pattern. From the above observations it becomes pertinent that, if a new man-made like such as Goronyo Reservoir, with a relatively low species density, is to be stocked, the species should be selected from the dominant species listed above depending on the availability of fingerlings. The selection of species for stocking should preferably follow the order of dominance also listed, namely: Cichlidae (tilapias), Cyprinidae (labeos), Bagridae (Chrysichthys) and Characidae (Alestes).

Table 24. Comparison of percentage density of fish families in Kainji, Tiga and Bakolori reservoirs (7,2 and 3.5 years respectively after impoundment) and the Rima River before impoundment.

NB. The percentage numbers of the four major dominant lacustrine fish families in each lake are underlined.

Stocking of a new reservoir can be effected in two ways:

1. by collection of the required species from the wild in other bodies of water and introducing them into the new reservoir,

2. by natural and induced breeding of the required species in hatcheries, growing them in nurseries to reasonable sizes for a couple of months and then transferring them to the new reservoir.

Table 25. Dominant fish species in Kainji, Tiga and Bakolori reservoirs with their
respective maximum weights as observed in Kainji Reservoir.

Among the species listed, only Oreochromis and Sarotherodon (tilapias) have been intensively studied with regards to natural breeding in confined habitats. Techniques have been developed in Nigeria for induced breeding of Chrysichthys nigrodigitatus and the common carp (Cyprinus carpio) a cyprinid with similar feeding habits to Labeo. Chrysichthys nigrodigitatus has not been very succesful in reservoirs even though it would be a better substitute for Chrysichthys auratus, being much bigger and with a faster growth rate. The common carp has not yet been introduced into any reservoir in Africa even though it has been found to be very successful in Asian reservoirs (Bhukaswan, 1980).

Fish fingerling collection from the wild, although possible, is labout intensive and more expensive to accomplish, but in the absence of hatchery produced fingerlings it is much preferable to nothing if increased production of a new reservoir with low fish density is to be accomplished.

Although stocking was recommended for Tiga Lake (Ita, 1979a) because no organisation was made sufficiently responsible for its management, nothing was done until stocking was initiated by the National Institute for Freshwater Fisheries Research during the late eighties.

Stocking Density in Bakolori/Goronyo Reservoirs

Stocking density is important if the stocked species is expected to make an impact on the future fisheries of the lake. Bhukaswan (1980) reported that some reservoirs in Thailand stocked at a low density of about 1–8 fish per hectare did not contribute significantly to the total catches in the lakes. It is important that stocking densities be estimated. The difference between potential yield estimates and the observed density per hectare could serve as guidelines for determining the stocking density. For example the potential yield of Bakolori Lake was estimated at 50 kg/ha and the estimated density per hectare was 3,708 fish/ha. Assuming that the lake would be exploited at one third of its estimated production of 150 kg/ha the stocking density at 150 kg/ha could be extrapolated from Figure 7. which shows the relationship between ichthyomass and density of fish sampled with rotenone at 21 stations in Kainji Lake. The density at 150 kg/ha approximates to 5,000 fish/ha.

Fig. 6. Percentage composition (by numbers in families) of catches in gillnets at Kainji Lake (1969–1977).

The observed density/ha during the survey period was 3,708 fish/ha showing a deficit of about 1,300 fish/ha. Depending on the species it is intended to stock, the deficit could be shared in accordance with their observed percentage dominance in the natural habitat. Thus for instance if only Sarotherodon is to be stocked the 1,300 fish could be shared between the two dominant species of tilapia (Oreochromis niloticus and S.galilaeus) in the ratio of 1.3: 1 or 1:0.8, based on their relative percentage composition. A new ratio could be worked out if other species are to be stocked.

In addition to estimates such as those above, the habitat of the stocked species should be taken into consideration. In Kainji Lake, for instance, the tilapias are confined to the littoral zone down to a depth of 7 metres. Bakolori reservoir with a mean depth of 6 metres, may not present any problem for the tilapias but in deeper lakes the problem of overstocking could arise and the surface area of that portion of the lake likely to favour the growth of tilapia should be estimated under such circumstances. Pelagic species like Alestes and clupeids could be stocked in relation to the total surface area of the reservoir. Bottom feeders could be stocked in relation to the total surface area of the bottom having oxygen available during thermal stratification of the lake. For instance, in Tiga Lake, it was observed that gillnets set below a depth of 15 metres caught no fish on account of oxygen stratification. If Tiga Lake were to be stocked with a bottom feeder such as the common carp, the depth of oxygenation should be taken into consideration.

Finally, it is relevant to state that, because of the relatively high cost of stocking a reservoir, it is important to ensure that fishing is controlled in order to prevent over-exploitation of juveniles of the introduced species. Minimum mesh size regulations should be enforced by the organisation responsible for the management of the lake.

Fig. 7. Relationship between the ichthyomass and density of fish sampled by rotenone in blocked areas of Kainji Lake in 1976.