7. CONCLUSIONS

All fish communities harbour species or segments of populations that are resilient to fishing to different degrees. For Mweru-Luapula these can be categorized as (Coulter, unpublished; see also Chapter 5 in Jul-Larsen et al., 2003):

1. Susceptible to fishing were the spawning runs of the Labeo altivelis (Mpumbu), a species that has disappeared from the fishery^[13]. Equally susceptible are large, old, specimen of long-lived species such as all catfish species (Clarias spp., and the bagrids Auchenoglanis occidentalis and Chrysichthys sharpiii) and the predatory tigerfish Hydrocynus vittatus.

2. Resilient to fishing are the tilapias - the mainstay of many African fisheries as well as Mweru - notably the detritus feeder Oreochromis mweruensis (Pale). Stocks of O. mweruensis appeared to be depleted in the 1970s, but returned, resulting in huge catches of two year old sized fish two years after the high increase in water levels in 1998; The predators Serranochromis spp. and the mollusc and insect eater Tylochromis mylodon are relatively resilient to fishing as well.

3. Highly resilient to fishing are the species with high population turnover rates: Microthrissa moeruensis (chisense) with a P/B ratio^[14] of around 5 (Zwieten et al., 1996), and small Alestes macrophthalmus. A. macrophthalmus can grow to large sizes, but matures already at small sizes. The standing stock of these species varies tremendously as can be seen in Figure 10.

As of 1997 the fish community of lake Mweru has shifted towards smaller sizes, in particular with the onset of the pelagic light fishery, that now dominate the catch. In fact now that the more resilient part of the fishery is being exploited, total production has increased tremendously and as a result also varies more (Figure 1). Fishing therefore has become more uncertain for the individual fisher, and the total outcome will be more susceptible to boom and bust periods. Whereas the average size of fish landed by the fishery becomes smaller, due to the size selection of the dominant mesh sizes used, it is less clear what happens with the length of individual species in the stocks. If, that is, part of the available (larger) sizes of the fish population is released from fishing pressure, as result of the interplay between size selective fishing mortality and recruitment variability. Considering the large period examined over which a dramatic increase in fishing effort has taken place, average length of fish caught has not changed very fast. Average length thus is only a good indicator of change caused by fishing effort for larger species with generally slow growth rates. In other words, for the mainstay of the fishery of lake Mweru, the cichlids, and smaller target species, length does not seem to be a good and timely indicator of change (contrary to Welcomme, 1999).

Based on this analysis it can be concluded that smaller mesh sizes could be less harmful and more productive for a multispecies fishery such as that in lake Mweru. The danger that small mesh sizes will lead to growth over-fishing is diminished, as a result of pulses in floods that lead to very strong year classes of which a considerable proportion still can outgrow the size-selective fishery. However, extended periods of low pulses such as between 1986 to 1998 may then be particularly dangerous in affecting this type of resilience. With the recent high water levels an increase in biomass of larger sized specimen, also of the predators Hydrocynus and Alestes, can be expected.

Lake Mweru has a multigear fishery whereby the dominant part of the catch of gillnets and traps is formed by species feeding on low trophic levels such as the microphytovore and detrivore Oreochromis mweruensis (around 20 percent of weight of the gillnet catch), while at present the largest catch comes from the zooplankton feeder Microthrissa moeruensis. Intermediate levels such as the molluscivore Tylochromis mylodon (6 percent of the weight in gillnet catch), the omnivore Clarias gariepinus (10 percent) and the predatory species such as Hydrocynus vittatus (22 percent) and Serranochromis spp. (10 percent) still form a substantial part of the catch. Theoretically a fishery that utilizes available stocks at all trophic levels in proportion to their biomass turnover could be a resilient fishery not affecting the structure of the fish community (Jul-Larsen et al., 2003). As changes in the fishing methods will be slower than changes in the stocks and as effort keeps on increasing, it will depend on the flood pulse driven dynamics whether a longer term pressure, in particular on the more vulnerable parts of the fish community, will result in structural changes in the ecosystem. The fishery of Lake Chilwa (Zwieten and Njaya, 2003) presents a case where catastrophic natural events leads to a simplified fish community that can withstand very high fishing mortalities, while changes in total catch reflect recruitment variability. After the tremendous increase in effort in the 1980s and 1990s, the gillnet fishery in Mweru, now with a few dominant mesh sizes, selectively fishes out fish between 10 and 20 cm, while the pelagic fishery targets even smaller sizes. Structural changes in the fish community have taken place as a result of the increase in effort, and fishing patterns have moved in a direction comparable to Lake Chilwa. At present Mweru-Luapula can be described as an open exit fishery where a highly dynamic population of fishers makes use of a variable but resilient resource base.