Three major barrages have been installed in Zambia for electrical power generation: Lake Kariba, Kafue Gorge and Itezhitezhi Reservoir. Lake Kariba is not within the scope of the present study; this lake was dealt with by Kapetsky and Petr (1984).
There are large dams across the Kafue River at either end of the Kafue Flats floodplain. The Itezhitezhi dam forms a reservoir of 360 km2, while the Kafue Gorge dam controls, to some extent, the water level on the Kafue Flats (Vanden Bossche and Bernacsek, 1990a). Their geographical positions are presented in Figure 8.1.
The Kafue has its source near Ndola at an altitude of 1 513 above sea level. Its total length is approximately 945 km and it drains about 154 200 km2. The major tributaries are the Lufupa and Mushingoshi Rivers. The Kafue River discharges into the Luswishi. The high water period is from January to May.
The two dams across the river, the Kafue Gorge dam and the Itezhitezhi dam, were closed in 1972 and 1977 respectively. The latter was constructed upstream of the Kafue Gorge dam in order to regulate the inflow of the Kafue Flats, which are located between the two dams.
Regulation of the Kafue River, above (Itezhitezhi dam) and below (Kafue Gorge dam) the flats, alters the natural flooding regime by decreasing and regularizing the maximum area flooded, decreasing the flood duration, delaying the flood, increasing the area permanently flooded during the dry season and reducing the amplitude of water level fluctuation by raising the minimum level (Vanden Bossche and Bernacsek, 1990a).
Construction of the Kafue Gorge and Itezhitezhi dams required multi-disciplinary studies of the Kafue River system with emphasis on the Kafue Flats. The FAO carried out a multipurpose survey of the Kafue River Basin. The Universities of Michigan and Idaho studied the possible effects of water regimes on fish and fishery. Also biologists of the Zambian Department of Wildlife, Fisheries and National Parks conducted numerous studies on the Kafue fishery (Muncy, 1973).
According to Lagler et al. (1971) fishermen utilized the floodplain during the entire cycle of flooding and drying. At maximum flood level nearly all fishermen fished with gillnets from camps along the floodplain margin. Few villages on the river banks or mounds were sufficiently elevated to be habitable during this high water period. As the flood waters receded and the natural levels of the main river channel emerged and dried, fishermen and their families migrated from the villages at the floodplain periphery to settle along the river and adjacent lagoons in temporary camps, still fishing with gillnets. As soon as suitable riverine and lagoon beaches were exposed by receding water, seining (locally termed draw netting) commenced and continued until these beaches were again inundated by rising water. Fishermen then migrated back to the floodplain margin, from where they continued to fish as indicated above. The local fishing population, including helpers and families, approximated 5 000.
In 1970 there were some 1 200 canoe-owning fishermen on the Kafue Flats. In addition there were many persons who owned nets and shared the use of canoes.
The principal fishing gear consists of gillnets and draw nets, both of which in 1970 were required to have a stretched mesh of not less than 7.6 cm (3 inches). Other types of gear, such as weirs, hook and line, and spears contributed only modestly to the catch.
Lagler et al. (1971) estimated the ichthyomass for the Kafue Flats in 1970 for both high- and low-water situations. Their results are summarized in Table 8.1. The estimate of total ichthyomass on the Flats at low water in 1970 was 38 000 t less than at high water; the difference was due to both natural and fishing mortality. Lagler calculated that 8% of the difference could be accounted for by fishing mortality. The remaining 92% was assumed to be natural mortality.
Relative abundance for 19 fish species was estimated as well and data are summarized in Table 8.2. Reduction in surface area of the Flats due to drying up was of the order of 1 500 km2. As a result fish are concentrated and become vulnerable to predatory fish and birds.
The altitude of the Kafue Flats, which have a total surface area of 14 000 km2, is 1 021 m above sea level at their upstream end and 976.6 m at Full Supply Level at the Kafue Gorge dam. The dam's height is 45 m. The maximum surface area during the wet season is 4 340 km2 and from 600–1 600 km2 in the dry season. During the wet season the mean depth is 2 m. The volume of the Kafue Gorge Reservoir is of the order of 0.8 km3. Its maximum length is 425 km while its width is 9.7 km. The Kafue is both the in and outflowing river. The catchment area is estimated at 45 325 km2. The mean annual fluctuation in level is 3.3 m (Vanden Bossche and Bernacsek, 1990a).
The surface temperature varies between 17 and 33°C (with a mean value of 24°C), the conductivity range is 140–382 μS/cm, the concentration of dissolved solids ranges from 68–220 mg/l and the pH ranges from 6.7 to 8.2. The ionic composition ranged as follows (in mg/l): Na, 3.5–8.5; K, 1.17–3.13; Ca, 15.8–45.3; Mg, 4.1–18.0; HCO3 + CO3, 73.2–268.4 and Cl, 0–3.55 (Vanden Bossche and Bernacsek, 1990a).
Magadza (1977) reported the existence of a serious weed problem. The weeds came from the Kafue River Flats where the increase in water level as a result of the Kafue Gorge dam had caused extensive growth of emergent macrophytes, particularly Vossia cuspidata. Another weed that showed explosive growth was Aeschynomene elaphroxylon, mainly in sheltered bays along the lake. Magadza feared ecological changes that could have a negative impact on the Kafue fishery.
Other ecological changes were reported by Sheppe (1985), who stated that a dam at the lower end of the Flats had permanently inundated parts of the floodplain, and a dam at the upper end had reduced the seasonal flooding, so that much of the floodplain was permanently dry. When Sheppe visited the Flats in 1983, he found that the effect of the dams was obscured by the effects of drought, hunting, and human use.
Everett (1971) presented detailed information on the location of the fish camps on the Kafue Flats. The overall total of fish camps was 162, of which 16 were permanent, 48 semi-permanent and 98 temporary. A total of 1 262 professional fishermen was recorded. Two aerial surveys were performed, one in April and the other in September 1970 resulting in the following numbers of boats:
| April | September | |
| Dugout canoes | 824 | 1 012 |
| Banana boats | 69 | 59 |
| Other boats | 5 | 6 |
Gillnets used generally had 3- or 4-inch mesh (stretched). Drifting gillnets are set at night; longlines are rare and hook and line fishing is common near population centres. Reed hides, from which people spear fish, were common around Nijunba, 206 km upstream of the Kafue Gorge. Beach seines or draw nets varied in length from 122 to 328 m. A total of 550 nets were in use on the Flats in the dry season. Around 2 000 gillnets were used throughout the year.
A list of commercial fish species was composed by Everett (1971) and is presented in Table 8.3.
The number of fishermen increased slightly to 1 280 in 1972 and dropped to 1 034 in 1973. They operated 1 037 and 1 021 boats in 1972 and 1973 respectively. The number of villages or camps dropped enormously, from 162 in 1970 to 134 in 1972. An exodus of fishermen, probably to Lake Mweru Wantipa, in 1974 and 1975 was reported; however, an influx of fishermen from Lake Kariba in 1976 and 1977 brought the numbers back to the 1973 level (Muyanga and Chipungu, 1978).
Munene (1983) reported that fishermen adapted themselves to changes in the floodplain, moving to higher grounds when it rains and to riverside fishing camps during the dry season. As a result fishermen are scattered along 240 km of the coastline when fish are abundant, and small-scale traders find it difficult to collect the fish. On the other hand at low water the traders have easy access, but then catches are poor. Due to the demand for fish the fishermen then resort to an illegal fishing method, ‘kutumpula’, which involves driving the fish into nets by beating the water surface with wooden plungers.
8.2.5.1 Annual yield
Table 8.4 summarizes fish production figures for the period 1957–1969 as presented by Lagler et al. (1971), while Table 8.5 shows data for the period 1966 to 1982. It may be observed that data from the different sources for the overlapping period (1966–1969) differ considerably. Production figures exceeded 10 000 t in 1958, 1966 and 1979.
8.2.5.2 Catch rates
Annual catch rates for the Kafue Flats are summarized in Table 8.5. Available data show a gradual decline in catch per boat per year from 1970 to 1974 (Muyanga and Chipungu, 1978). This reduction may have caused the departure of a number of fishermen to Lake Mweru Wantipa. From 1975 the catch rates increased again.
8.2.5.3 Fishing effort
No data available.
8.2.5.4 Mesh selectivity
No data available.
8.2.5.5 Species composition
Muyanga and Chipungu (1978) noticed changes in relative abundance of a number of species of commercial interest. Catches of Sarotherodon macrochir, Tilapia rendalli and T. sparrmanii declined as compared to the pre-impoundment situation, while some predators like Clarias gariepinus, Marcusenius macrolepidotus, Labeo molybdinus and Hepsetus odoë significantly increased. A year earlier, however, only one predator had increased in abundance (Dudley and Scully, 1980).
8.2.5.6 Fish population dynamics
Fish catches from multimesh gillnetting operations were studied from January 1964 to June 1965 by Carey and Bell-Cross (1967). They recorded sexual activity for males and females for commercial fish species occurring in the Kafue River. Results are summarized in Table 8.6.
Chapman et al. (1971) studied food habits, fecundity, and growth of several important fish species. The most significant result was that extent and duration of flooding significantly affects growth. Higher, longer floods increase growth of Tilapia. The age of T. andersonii was successfully determined and it appeared that the size of the flood accounted for most of the variation in annual growth.
In the 1960s, seine netting in the Kafue River floodplain area was an illegal fishing method. Carey and Bell-Cross (1967) observed that it was impossible to enforce effective control on fishermen's activities in those days. However, they considered that the use of seine nets was not very harmful to breeding fish populations, because seine netting was seldom carried out during peak breeding periods.
Chapman et al. (1971) recommended that the water regime should be regulated to permit at least five weeks of drying on as much of the floodplain as possible, and that those areas of the Flats which are not heavily grazed by cattle and wildlife should be burned before the rainy season, to reduce biological oxygen demand during the next flood. The reservoir should be allowed to rise soon after the Flats are burned, to minimize nutrient losses, and should rise slowly enough to allow aquatic grasses to keep pace.
According to Munene (1983) the kutumpula method, i.e. driving the fish into nets by beating the water surface, is prohibited.
Handlos (1982) summarized the various views on water level regulation in the Flats. Regulation of water from the Itezhitezhi dam would be subject to ecological considerations. For instance a ‘freshet’ or discharge in March of a great amount of water would simulate the natural flooding in dry years. On the other hand reduced flow before and after the freshet would result in a minimum ecological impact. Prolonged, but not continuous, flooding would be essential for the success of the indigenous fishing industry; any other measure would likely have a negative impact on the fishery. Handlos quoted White who suggested that the Kafue Gorge Reservoir be drained annually in September to combat the weed problem. The impact on the fishery of such a measure is evident. A solution would be to raise the level of the Itezhitezhi dam to allow for a larger freshet (as was also suggested by de Groot and Marchand (1982) in the context of ‘polder’ development).
Dudley (1974), in fact, demonstrated by his Tilapia growth studies that higher flooding regimes caused by the Kafue Gorge dam will increase growth and survival of Tilapia.
(See Section 8.2.1.)
(See Section 8.2.2.)
The Itezhitezhi Reservoir is fed by the Kafue River, which is also the outflowing river. At Full Supply Level the reservoir's altitude is 1 029.5 m above sea level, while its surface area is 360 km2 (Vanden Bossche and Bernacsek, 1990a).
No data available.
8.3.5.1 Annual yield
Research activities resulted in a production level between 500 and 1 000 t. The standing fish biomass would not exceed 3 000 t (Mubamba, 1992).
8.3.5.2 Catch rates
No data available.
8.3.5.3 Fishing effort
No data available.
8.3.5.4 Mesh selectivity
No data available.
8.3.5.5 Species composition
No data available.
8.3.5.6 Small clupeids
Mubamba (1992) reported the attempts to introduce Limnothrissa miodon from Lake Tanganyika into the Itezhitezhi Reservoir. The first release took place 16 March 1992. The fish were flown from Mpulungu to Itezhitezhi. The second release was on 20 March 1992. During the second shipment mortality was reduced significantly. Mubamba expected that another 20 000 fry would be necessary to fully stock the reservoir.
Mubamba (1992) reported the introduction of L. miodon into Itezhitezhi Reservoir. He expected that the annual fish production from the reservoir would increase from the current levels of 500–1 000 t to about 5 000–8 000 t/yr.
