As part of the preparatory phase of ALCOM's pilot project on the Utilization of Small Water Bodies in Botswana, a socio-economic survey was carried out among communities living around seven selected dams in south-eastern Botswana. The main aim of the survey was to determine fish consumption patterns and attitudes as well as the type of fishermen and the level of their current activity.
Results show that approximately 20% of the households interviewed (343) have eaten fresh fish, albeit very rarely. The most common reason for not eating fresh fish was that it was not available for sale. The main source of fresh fish was a fellow-member of one's own household catching fish in the dam. Tilapia is the preferred species, although barbel (Claris gariepinus) is also eaten.
Households said there were no cultural taboos concerning the consumption of fish, and this was confirmed by the large proportion of households that eat tinned fish. Just over 80% of the total sample eat tinned fish, usually pilchards in chili sauce. Other types of fish such as frozen or salted fish are eaten by a negligible number of households, because they are not easily available.
The results indicate that most local people presently fishing are mainly children or youth, predominantly male, who fish with hook and line during summer only. Most were at school or unemployed and all were interested in improving their catches.
The survey repudiated early assumptions that people in eastern Botswana do not fish and suggested that availability and knowledge concerning how to cook fish are the main constraints preventing people from eating more fish. It was also concluded that fishing is still considered as a part-time activity undertaken by people who have some spare time.
An inventory of reservoirs in eastern and southeastern Botswana was compiled from secondary data and completed in May 1994 by ALCOM. Data was input on 331 dams, of which three were completely silted, 13 have no name and four belong to WUC instead of WDS. Data registered for the reservoirs are: name, district, location, year constructed, capacity and (estimated) surface area at full supply level, maximum and average depth and status of reservoir.
Figures for most reservoirs mention only the storage capacity, which varies from 2,000 to 2,800,000 m3. Only about 55 reservoirs are well documented on physical characteristics. If we count only the WDS dams, only 42 out of 324 are noted to have a storage capacity exceeding 70,000 m3.
A very rough estimate of the surface area was calculated by assuming that:
Based on these estimates, the average surface area, including that of the WUC reservoirs, varies from 0.2 to 1900 ha. They can be divided into four categories:
Dams that exceed 100 ha in size (which account for only 1% of the dams area in Botswana but take up 71% of the total surface area)
Dams between 20 ha and 100 ha in size (which account for 4% of the dams)
Dams between 5 ha and 20 ha (which account for 14% of the dams)
Dams less than 5 ha in size (which account for 79% of the dams by number, but only about 9% of the surface area).
The total estimated surface area of WDS reservoirs is 1390 ha. The four WUC reservoirs totalled an area of 3255 ha when full (this figure will have to be confirmed by checking real area data). Therefore the total estimated area of reservoirs in Botswana is 4645 ha.
Officers at the WDS report that the reservoirs built by them after 1988 should be perennial and give a safe yield in about 20 years. The smaller reservoirs built before 1988 do not have such a good water holding capacity; perhaps as many as 50% of them dry out every year. The reservoirs with a maximum depth of 2 m or more should have water throughout the year, if not over-utilized for irrigation or cattle watering.
Information on how the dam groups are registered and what kind of a constitution they have is almost non-existent. To evaluate the potential fish production from all the reservoirs, a follow-up to the preliminary inventory is essential, to update it and add new information on number, size, water capacity, time with water, status of reservoir, potential yield, management etc. In July 1992, collaboration was initiated with WDS and the Irrigation Section (IS) about a more complete inventory of the reservoirs, and a detailed questionnaire was drafted. A simplified version of the questionnaire was sent out to all the regions/districts by the Land Utilization Officer. This version covered the status of the reservoir rather than its use. The original version was later sent out in the Gaborone region by the regional irrigation officer. Responses to the two questionnaires sent out during December 1992 have not yet been received.
Initially seven reservoirs were selected (1990) for the socio-economic survey in rural areas of south-eastern Botswana, with three dams in Kweneng District and four in Southern District. An eighth reservoir was selected (1990) in Gaborone because of its easy access. It is a collection of barrow pits filled with seepage water. And in 1992, three additional reservoirs were selected; two of these are among the biggest dams in Botswana (Gaborone and Bokaa dams). These 11 reservoirs have been used for test fishing. They were selected because they were considered representative of the different dams that can be found in Eastern Botswana.
The reservoirs have been built for the primary purpose of supplying domestic water, providing water for cattle and also for horticulture. They usually hold water for the whole year, and are therefore classed as perennial. Because of the extreme drought during 1992 four of the eight reservoirs had no water left by the end of September 1992 but all of them were again filled during the beginning of December 1992.
The reservoirs represent the socio-economics of Botswana settlements, which fall under three main categories: “village”, “lands” and “cattlepost”:
The “village,” where we find schools and health clinics, is the most permanent of the settlements. People are registered and also stay during non-farming season;
The “lands” refer to the compound in the farming area to which people move to plow and sow during the rainy season and stay there until the harvest is completed; and
The “cattlepost” is the area where cattle are kept for grazing. Family members stay only for a part of the year and are usually boys herding cattle.
Multimesh monofilament bottom set gill nets consisting of 14 randomly distributed panels of various mesh sizes, 6.25 to 75 mm bar height, were used. This method of test-fishing was developed by the Institute for Freshwater Research, Drottningholm, Sweden. For a more detailed description of this method see Appendix 1 and the report of Furst et Fjalling (1991). In conjunction with each test-fishing exercise, water quality parameters were measured.
One drawback of this method is that information is not available on the catch of each panel. This information would have provided data on mesh selectivity for different species, resulting in recommendations on mesh size to the Fisheries Department.
If the Drottningholm method is used on one reservoir for a full one-year cycle, it becomes a useful tool for decisions on fishing curbs during certain times of the year, i.e. during the breeding season. Annual changes in any reservoir can be monitored by repeating the previous year's test-fishing exercise under climatic and water conditions as similar as possible to the year before.
The mending of the monofilament net is also very difficult, especially if scientific accuracy should be maintained with exact mesh sizes. The twine has in some instances been shown to be too weak and fish caught easily damage the nets, both in the water and while cleaning the nets. Another problem sometimes experienced is when big specimens get caught in a panel; the whole panel gets entangled, making it very difficult to catch other specimens.
Test fishing was carried out from 1990 to 1992 to assess the stocks of representative reservoirs. A total of 17 species of fish were caught on 11 dams selected. Freshwater crab and turtle were also caught. For identifying the species, the books “The Fishes of Zimbabwe” (Bell-Cross and Minshull, 1988) and “Freshwater Fishes of Southern Africa” (Bruton, Jackson and Skelton, 1982) were used. What is reported as Oreochromis andersonii might in fact be a hybrid between O.andersonii and Oreochromis mossambicus.
During our test fishing, breams, especially O. andersonii, were found in all reservoirs except one. In the smaller very shallow reservoirs, such as Thaone dam, all species can be found in all the areas where nets are set.
Mmakgodumo and Gampudi dams have one and three species respectively. These reservoirs are situated in the same area. Both have been stocked with the species found and no additional species were caught. In Semarule reservoir, all species found in the reservoir, except Labeo molybdinus, were caught at each test-fishing exercise. Labeo was caught only in November 1990, May, June, and September 1991.
The biggest variety of species was found in Gaborone ponds and in Gaborone dam, where 15 species were found. Mmakgodumo dam showed the least species diversity, with only O. andersonii caught. The two species Marcusenius macrolepidotus and Cyprinus Carpio were caught only in Gaborone dam.
The catch per unit effort (CPUE) from the Drottningholm multimesh nets for the 11 different reservoirs is presented in Fig 1–2. The catch was dominated in mass by O. andersonii in four reservoirs, by Clarias gariepinus in five and by small Barbus spp. in two reservoirs. Table 3 gives details of CPUE by species for four reservoirs which were test-fished in 1992.
The results from test fishing in May 1992 in Semarule did not at all show similarities with the May 1991 fishing, because of the drought and the low water levels. The CPUE, both in weight and numbers, was much higher in May 1992, probably because the density of fish per surface area increased quite dramatically with the decrease in water level. In Mamokhasi dam, which was the only dam with two depth zones, it was shown that the catch was much lower in depth zone 3 to 6 meters compared to the depth zone 0–3 meters.
The length-frequency distribution for O. andersonii in Mmakgodumo reservoir is shown in Fig 3 and the length/weight correlation in Fig 4. The specimen caught during the test fishing had a relatively low maximum size. The length frequency distributions for O. andersonii for the different fishing exercises in Semarule dam are shown Fig 5. The length/weight correlation for O. andersonii in this dam is presented in Fig 6.
The temperature (at 0.5 m depth) varied between 13.5 and 27.5 °C. The temperature curves for Semarule reservoir for 0.5, 1.5 & 2.5 m depth, for the period October 1990 – October 1991, are shown in Fig 7.
The pH, measured from the surface water, was found to be basic in all tested reservoirs. pH as well as transparency showed seasonal change. Recorded pH and transparency for Semarule are presented in Fig 8.
No oxygen or conductivity could be measured during the test-fishing exercises October 1990 – November 1991, as no equipment was available. In December 1991, a conductivity meter was made available. The one test done in Semarule reservoir showed a conductivity of 205 μmS/cm.
Water quality parameters are presented in Table 4. The dissolved oxygen (DO) level was measured only once in each reservoir. In the deeper reservoir, Mamokhasi, it was 5 mg/l at 3 m depth. The DO in Semarule in May 1992 was very high, 11 mg/l at 1 m depth.
With the aim of identifying the right equipment to sample fish populations, ALCOM compared two kinds of nets. The Drottningholm method introduced by ALCOM used monofilament multimesh gillnets. These nets are scientifically well calibrated but are not available locally. They tear easily and are not easy to mend, they are also expensive. Therefore, ALCOM decided to conduct trials to find out if a cheaper method or a combination of methods can be used.
Both Drottningholm test nets and monomesh multifilament nets of five mesh sizes (bar height 31 mm 62mm) were used during the same exercise on three occasions (Semarule, Bokaa dams and Gaborone ponds), to enable comparison of results from the different methods.
The two methods showed some differences in species caught but mainly for the small species such as Barbus spp., Mesobola brevianalis and Pseudochrenilabrus philander, more of which are caught with the Drottningholm nets of mesh size less than 31 mm bar height. But the multifilament nets are still a good method of providing an idea of the the relative abundance of fish, and they can be bought and mounted locally. It will be necessary to mount 5–7 different panels of mesh size from 20 to 65 mm bar height, and if it is not possible to get the smaller mesh size locally, the method can be combined with a beach seine to get the small fish.
The fact that the population of Botswana are by tradition non-fish eaters and non-fishermen, made it difficult to apply some of the Rapid Appraisal techniques on the fish-related issues the team was focusing on. The three-site settlement system of Botswana also caused problems for the team, as a number of households were not inhabited at the time of the appraisal.
In most aspects the results of Sen's 1989 socio-economic survey (Sen, 1990) were confirmed during a RRA exercise undertaken in 1991 by Westerlund at Semarule dam. The findings indicate that fishing activities and consumption of fresh fish are very low because of people's fear of water and the fact that many people have never eaten fish. Hook and line fishing on the reservoirs is sparse, and people eat fish very rarely. Some people would be interested in fishing intensively if training in more efficient methods than hook and line could be provided; they would like to catch more and eat more fish.
The main activities of people in the dam area are farming and livestock herding, which take up considerable time. In Botswana many people also move away from the farming areas (lands) after the harvest is done. This means that either people are too busy with farming or that they are not living by the reservoir when time would allow them to fish. In a year such as 1992 when no rain fell, many farmers didn't even bother to move out to the lands area (where most reservoirs are situated).
Increased fisheries could be promoted as a full-time activity for unemployed persons, and also as a part-time complementary activity for the other groups.
Botswana people like fish, and it is abundant in reservoirs. Local people prefer catfish, while residents of the bigger villages and cities prefer bream and Labeo, but no one has shown any interest in the smaller Barbus species. Management should therefore concentrate on the species already in demand. The small Barbus species seem to be an abundant unexploited resource from ALCOM's test fishing experience. If there is any further fishery exploitation, this fish species should probably not be caught by small-mesh size gillnet, as cleaning it is very time-consuming. These small fish are also easily spoiled; therefore post-harvest processing of this fish species (sun-drying, salting) is needed before it is sold.
At a later stage, the Fisheries Department can consider introducing new or improved fishing gears for the small barbus.
There is a need for the Fisheries Department and ALCOM to organize fishing through dam groups. Training is needed for people on how to fish, how to cook fish, how to exploit a reservoir.
