2.1 The environment
Lake Kariba (277 km long; 5 364 km2; 160 km3; 29 m mean depth and 120 m max. depth) is located on the Zambezi River between latitudes 16E28to 18E04S and longitudes 26E42to 29E03E. It was the largest man-made reservoir in the world at the time of construction. Today it is the second largest reservoir in Africa by volume. The catchment area covers 663 817 km2 extending over parts of Angola, Zambia, Namibia, Botswana and Zimbabwe. The dam wall (128 x 580 m) was completed in 1960 and the filling phase lasted from December 1958 to September 1963 when the water reached the mean operation level at 485 m above mean sea level. The lake is naturally divided into five basins (Figure 1) and is almost equally shared by the two riparian countries Zambia and Zimbabwe with 45 percent and 55 percent respectively. The impact of the artisanal inshore fishery on the two sides of the lake can be considered not to affect each other because of the deep channel in between the two shores along most of the lake.
Lake levels fluctuate annually from one to five metres (mean = 2.9 m) as a function of inflowing floods between December and June and continuous drawdown through the turbines and, before 1981, spillage through the sluice gates. Since 1982 the lake levels have declined due to a series of droughts and the lowest levels recorded was in December 1992 and January 1997 at 476 m (Figure 2). Since 1997 the lake levels have risen rapidly, and in April 2000 the sluice gates were opened again for the first time in 19 years.
FIGURE 1. Map of Lake Kariba showing the five natural basins (B1..B5), the designated inshore fishing grounds on the Zimbabwean side (C1..C7), the sampling strata in Zambia (S1..S4), the selected experimental fishing stations in Zambia around Sinazongwe (open circles), and the experimental fishing station (Lakeside) in Zimbabwe near Kariba town (open square).
FIGURE 2. Mean monthly and mean annual lake levels (m a.m.s.l.) in Lake Kariba from 1962 to 2000. Between 1981 and 2000 no surplus spillage through the dam wall floodgates have been performed.
The limnological and various biological characteristics of Lake Kariba are well described elsewhere (see e.g. Coche, 1968; Balon and Coche, 1974; Marshall, Junor and Langerman, 1982; Marshall, 1984; Machena, 1988 and Moreau, 1997 for comprehensive reviews). The most salient features are a warm (mean surface temperature of 26E), oligotrophic, monomictic lake with overturn occurring in June-July. Stratification begins around September with a thermocline around 15 m depth which gradually moves down to around 35 m at the time of the turnover. The oxygen concentration in the hypolimnion declines steadily during the stratification period and often reaches a stage of deoxygenation. The volume to inflow (c. 50-70 km3 per year) ratio is low giving a mean water retention time of only 2.6 years. The Zambezi River provides about 80 percent of the inflow and there is a natural gradient in the lake ecosystem from east to west. Basin 1 and 2 have more riverine characteristics gradually attaining a more lacustrine environment to Basin 5 near the dam wall.
The waters from the Zambezi River are very clear and low in plant nutrients. Phytoplankton blooms occur just after turnover and locally at the onset of floods from tributaries, but most of the year primary production may be very low. Zooplankton abundance shows variation in response to phytoplankton biomass (Begg, 1976; Marshall, 1997). Initially after filling, the lake was eutrophic due to the vast amount of dissolved nutrients through inundation of vegetation and land. From 1963 the lake became slightly alkaline with a marked decrease in phosphate and nitrates (Thornton, 1980). During the early eutrophic years the floating fern, Salvinia molesta, colonized large areas of the lake and retained vast amounts of plant nutrients (Mitchell, 1973). From the 1970s it has gradually decreased and been replaced by increasing amounts of rooted macrophytes down to a depth of around 10 m (Machena, 1989).
2.2 The fish fauna
The natural limnological gradient in the lake is also reflected in the fish communities which are dominated by potamodromous species in the east (cyprinidae and distichodontidae) and by more sedentary cichlids in the western basins (Begg, 1974). The fish community and biology of the fish species is well described in a number of reviews and research works (Balon and Coche, 1974; Mitchell, 1976a; Marshall, Junor and Langerman, 1982; Kenmuir, 1984; Marshall, 1984; Karenge, 1992; Machena, Kolding and Sanyanga, 1993, Musando, 1996; Karenge and Kolding, 1995a). Several of these discuss the observed biological succession in detail and only a brief update will follow. Most notable is that the number of fish species in Lake Kariba seems to be steadily increasing.
Jubb (1967) listed 55 species occurring in the Middle and Lower Zambezi River system, that is the portion of the drainage system from below the Victoria Falls to the mouth of the Zambezi River into the Indian ocean. In a pre-impoundment gillnet survey (Jackson 1961b), 28 fish species were caught in the Zambezi river at the location of the new lake (but very small mesh sizes were not used). Shortly after filling in 1964, Harding (1966) reported 33 species. In the early 1970s, Balon (1974a) recorded 39 species, later updated to 43 species by Marshall (1984). Since then five more species, the garpike Hepsetus odoe (Sanyanga and Feresu, 1994), the tilapias Tilapia sparmanii and Oreochromis niloticus, the chiselmouth Varicorhinus nasutus, and the largemouth bass Micropterus salmoides (Karenge and Kolding, 1995b; Anon., 1995) have been caught in the lake by the Lake Kariba Fisheries Research Institute. Bell-Cross and Minshull (1988) list 62 species in Lake Kariba area, defined by the Victoria Falls upriver and all south bank tributaries. Several of these have not been recorded in the lake itself.
2.3 Invasions, introductions and disappearances
The presence of non-introduced Upper Zambezi fish (i.e. above the Victoria Falls) in Lake Kariba a decade after creation, such as the cichlids Sargochromis giardi, Sargochromis carlottae, Oreochromis andersonii, and Pseudocrenilabrus philander, the mormyrid Marcusenius macrolepidotus, the cyprinid Labeo cylindricus, the schilbeid Schilbe intermedius (mystus), and the small barbs Barbus poechii, B. paludinosus and B. unitaeniatus, caused some scientific debate when Balon (1974c) suggested that they survived the drop down the mighty Victoria Falls (e.g. Jubb, 1976a, 1976b, 1977). There are, however, indications that the so called upper Zambezi invaders may also have been present in the middle Zambezi before inundation but were missed in the sampling programs. They may also have been accidentally introduced or could have reached the lake through the Victoria Falls power station overflow (Jubb, 1976a; Kenmuir, 1984; Bell-Cross and Minshull, 1988). Whatever the origin, it seems that the new lacustrine environment give these species enhanced conditions to become established.
Presently 50 different fish species have been observed in the lake, five of which are introduced (Songore and Kolding, 2003). However, seven species have been reported just once (Leptoglanis rotundiceps, Serranochromis angusticeps, Hepsetus odoe, Barbus radiatus, Labeo lunatus, Varicorhinus nasutus, and the introduced Micropterus salmoides). Thus it can be is questioned whether they are stray specimens or have established viable populations in the lake. Two of the introduced species, Tilapia rendalli and Serranochromis robustus, may as well have invaded the lake naturally (e.g. Kenmuir, 1984). In fact T. rendalli was already caught by Jackson (1961b) during the pre-impoundment survey. Thus only three truly exotic species have established in the lake. One is the cichlid Oreochromis macrochir, which was stocked into the lake in 1959-62 from Chilanga, Zambia. It was never caught during Balons (1974a, 1974b) intensive sampling programme and was believed extinct, but started appearing in the Lakeside sampling programme in 1974. It has since then been caught every year in the experimental surveys at Lakeside but only in very small numbers (on average about ten specimens per year out of 5-10 000 total sampled). In the Zambian experimental gillnet surveys it has been recorded only three times: in 1985, 1992 and 1996. The other true exotic is the small pelagic clupeid Limnothrissa miodon introduced from Lake Tanganyika in 1967-1969 (Bell-Cross and Bell-Cross, 1971). This introduction is a well-known success story (see section below) and there are no indications that this by far largest single stock in the lake with an annual total production rate around 125-150 000 tonnes, has had adverse effect on the other species (Marshall, 1991; Karenge and Kolding, 1995b). The third true exotic is the Nile tilapia (Oreochromis niloticus) which since 1993 has been caught in ever increasing quantities in gillnets by Lake kariba Fisheries Research Institute (LKFRI). This species was never introduced deliberately but is cultured at several farms along the lakeshore, which drain directly into the lake.
Some species have disappeared or have become rare (Kenmiur, 1984). The rheophilic species, Chiloglanis neumanii, Opsaridium zambezense (and possibly also Leptoglanis rotundiceps) are now confined to the tributaries or the two more lotic western basins (Balon, 1974a, 1974b). Species diversity could be greater in the eastern effluent part of the lake than is assumed today (Begg, 1974), as Balon (1974a, 1974b) recorded 39 species in the Zinazongwe area (Basin 3) in a poisoning sampling programme. In a similar study in the Sanyati basin (Basin 5) only 27 species were recorded (Mitchell, 1976a). It was feared that the potamodromous mottled eel, Anguilla bengalensis, eventually would disappear from the lake, since the elvers coming up from the sea would not be able to mount the 128 m dam wall (Jubb, 1967). Marshall, Junor and Langerma (1982) considered the situation even more problematic with the construction of the formidable Cahora Bassa dam wall (160 m high) about 400 km down-river. However, some elvers still seem able to do so, although in low numbers, and eels have been recorded in the lake at all times (Marshall, Junor and Langerman, 1982). For instance, two mottled eel of around 60 cm TL were captured in March 1993 near the dam wall. If these have not ascended the two dam walls on their way up the Zambezi river, they were by then more than 30 years old (Anon., 1993).
2.4 Management and regulations:
The lake was primarily constructed to generate hydro-electricity but additional expectations were also derived from different estimates of the potential fish yields. Before inundation, an area of some 950 km2 (about 18 percent of the lake area) was bush cleared at various places within the 20 m contour in order to establish inshore fishing grounds. Between 1959 and 1962, 26 tonnes of cichlid fingerlings (Tilapia rendalli and Oreochromis macrochir) were stocked into the lake to boost the fishery (Coche, 1971). From the very beginning the essential issue in the management of the fishery was the question Who was to fish? On this question the Zambian and Zimbabwean authorities fundamentally disagreed, which eventually led to divergent policies that still mark the fisheries in the two countries today (Bourdillion, Cheater and Murphree, 1985). In Zimbabwe the authorities divided the shorelines into 14 areas, separated in spheres of white and black interest. Along the Native Area shoreline, black fishers, who were encouraged to become professionals, would exploit the inshore fishery. Therefore the 26 camps in which they were located were strictly for fishing: permanent settlement and shoreline agriculture were prohibited. The native areas were shared with white-owned concessionaires, who in addition to their own fishing concession areas, would also purchase the fish from the black fishers. The area demarcations and concessions allocated for fishing were changed in 1972 into eight larger areas and again in 1976 into the present seven areas (Figure 1), (Marshall, Junor and Langerman, 1982; Karenge and Games, 1995), mainly to conserve the stocks and to meet the demands for more recreational waters from the burgeoning tourist industry. Whilst the total area available for inshore fishing was reduced by these changes, the area allocated to local fishers was increased slightly. At present about 470 km2 (63 percent) of the fishable water on the Zimbawean side is available to the inshore fishery (Marshall, Junor and Langerman, 1982) and the number of fishing villages has grown to about 40.
From the very beginning, the Zambian authorities took a different view. The interest of the local population was paramount and there was no racially based segmentation. The whole Zambian shoreline was designated as Native trust land and could not be utilized without consent of the local people. Furthermore, the Zambian authorities initiated infrastructure and institutions for developing the fishery. A large resettlement compensation was paid to the Gwembe Rural Council, mainly used for development of the fishery. In addition, a Fisheries Training Centre was build in Sinazongwe with harbour, boat building, and ice plant facilities. Up to 1994 there were about 270 fishing villages on the Zambian shoreline, but these were amalgamated into 67 villages in 1995 (Jul-Larsen, 2003). In Zimbabwe, preparations for the initial fishery were much more modest and little money was allocated for the fishery development. Since its beginning until today any kind of economic investment in the inshore fishery in Zimbabwe has been virtually absent (Marshall, Junor and Langerman, 1982; Bourdillon, Cheater and Murphree, 1985), in contrast with a relatively strong management regime and enforcement capacity.
In Zimbabwe the inshore fishery is controlled by the State through the Department of National Parks and Wildlife Management (DNPWM) who limits access, closes areas to fishing and restricts fishing gear methods. A permit system is in place whereby the DNPWM informs the two riparian local authorities, the Nyaminyami (west) and Binga (east) District Councils (separated by the Sengwe river in Basin 3), of the limits placed on the number of fishing permits that may be issued for the particular year. The Councils then issue these fishing permits to the individual fishers or fishing co-operatives. The distribution of permits per individual fisher can differ according to local authorities. For example for the greater part of the 1980s individual fishers in Binga were allocated 2.5 nets each whilst fishers in Nyaminyami were allowed five nets each. Fishing permits for fishers operation off state lands are issued directly from DNPWM. According to officially set limits there should be a maximum of 2 530 nets on the Zimbabwian side of the lake belonging to 771 fishers, independent fishers and cooperative members inclusive. It is however difficult to establish how these effort limits were arrived at since it proved impossible to obtain documentation on the method used (Songore, 2000). The official limit, however, has never been reviewed and the actual recorded number of nets and fishers, although fluctuating, for most of the time has been below these values (Figure 4).
The fishing patterns are very simple because only one type of gear (gillnets) are used throughout the whole Zimbabwean inshore fishery. Fishing is not permitted using nets with less than a four-inch (102 mm stretched) mesh size. In addition explosives, chemicals, poisons, intoxicating substances, scoop nets, jigging and fish driving may not be used to catch fish. Fishing is also not permitted along parts of the shoreline (about 20 percent, Figure 1) belonging to the DNPWM, notably all the Chete Safari Area, most of the Matusadona National Park and parts of the Charara Safari Area. Other restrictions are in place for river mouths, large population centres, harbours, and river estuaries. Fish net manufacturing is not permitted for persons who do not hold a valid manufacturer and dealers license and fishing nets can only be sold to holders of valid fishing permits.
Two institutions dealing with the fishery resort under the DNPWM: The Lake Kariba Fisheries Research Institute conducts research and collects data on the fish stocks, the fishers and their catches, while the management branch of the DNPWM is tasked with policing the fishery.
In Zambia access to the fishery is free in principle and fishers can fish anywhere. When the lake was created the management regulations were similar to the other fisheries in the country, but these were seen as interim to be revised after obtaining further information (Malasha, 2003). Initially, therefore, the legal minimum mesh size was set to four inches and a closed season was to be observed from 16 December to 16 March. In 1962 the Department of Game and Fisheries in Northern Rhodesia tried to abolish the mesh size regulation based on the results of experimental fishing. However, the authorities in Southern Rhodesia rejected this on the grounds that the Lake Kariba Fisheries Research Institute had not yet conducted sufficient experiments to support this move. From the independence of Zambia in 1964 to 1986 no mesh restrictions for gillnet existed and beach seines were allowed. After 1986 the minimum mesh size for gillnets was set at three inches (76 mm) and beach seining was prohibited. In practice, however, there has been little enforcement due to lack of resources (Musando, 1996). Also the kutumpula method (fish driving) is very popular among fishers although it is officially banned by the Fisheries Act (Scholz, 1993).
 Before 1964 called
Northern and Southern Rhodesia under the Central African Federation of Rhodesia