The abundance of Nile shrimp, Caridina nilotica (Roux) in Lake Victoria has fluctuated significantly over time, from the periods of 1900–1950s, 1980s–1990s and from 2000s to present day. To elucidate its current abundance, its contribution to the Rastrineobola argentea fishery catches, and its importance to the composition of the diet of Nile perch (Lates niloticus L.), and to the bottom trawl catches, were investigated. From August 2006 to January 2007, the mean C. nilotica CPUE was 1.68 ± 1.20 kg boat−1 day−1 at Kijiweni and Igombe, and its 0.37% contribution to the R. argentea total catches was insignificant. Of the 230 Nile perch stomachs containing prey items, C. nilotica comprised 54% by volume of the diet of fish <50 cm TL; while haplochromines contributed 41% to the diet of fish >50 cm TL. Mean catch rates of C. nilotica in bottom trawls in 8 months between 2005 and 2008 ranged between 0 and 2.45 ± 2.50 kg hr−1. Low catches of the C. nilotica in the R. argentea fishery, and the dietary shift of Nile perch of >50 cm TL to once again include haplochromines, may indicate a decrease in C. nilotica's abundance in the lake, and vice versa. Overfishing and the selectivity of the fishery to take only large Nile perch for fish filleting factories, have resulted in reduced stocks and dominance of juveniles in the perch's populations. The highly reduced Nile perch stocks are currently leading to an apparent reversal of the 1980s regime with a shift to a new cichlid-dominated and C. nilotica low abundance state. Increase in predation on juvenile C. nilotica by the recovering haplochromines, and juvenile perches, as well as environmental degradation, especially eutrophication and pollution, along with the effects of global warming impacts, account for the observed decline in C. nilotica.
Over the past 75 years, Lake Victoria has undergone major ecological disruptions as a result of the introduction of alien species, overfishing, and eutrophication (Ogutu-Ohwayo, 1990; Wanink et al., 2001; Witte et al., 2007). While increased nutrient loadings resulted in higher algal biomass, shifting from green algae and diatoms to blue-green algae (Muggide, 1993), the higher primary production also resulted in higher organic sedimentation to the hypolimnion and this significantly increased the extent and duration of hypoxia in water layers below 20 m (Wanink et al., 2001). Most likely the negative effect of Nile perch, (Lates niloticusL.), an introduced top predator in the Lake Victoria ecosystem (Fryer, 1960), was not the only cause of the decline of the haplochromines; eutrophication and fisheries also played a role (Ogutu-Ohwayo, 1990; Katunzi et al., 2003; Witte et al., 2007). Yet, no attempt has been made to link the three processes to understand feedbacks in food webs, resource use patterns or trade.
In 1950s, Lake Victoria fish catches were based on a wide range of mainly cichlid species which predominated its demersal fish stocks, accounting for 80% of the demersal fish mass (Graham, 1929). During this period, the abundance of atyid shrimp Caridina nilotica (Roux) was very low (Lehman et al., 1996). In the 1980s, the lake's food web changed from a complex fauna with many species, to a simplified food web dominated by three fish, Nile perch (Lates niloticus L.), Nile tilapia (Oreochromis niloticus), dagaa (Rastineobola argentea) (Witte et al., 1992; Medard and Ngupula, 2007) and C. nilotica (Witte et al., 1992; Kaufman and Ochumba, 1993). This regime shift in the ecosystem drastically changed the resource base of the fishery with far reaching consequences for local livelihoods (Katunzi et al., 2003; Witte et al., 2007).
Currently, the lake catches are dominated by Nile perch (Katunzi et al., 2006; Budeba and Cowx, 2007a). Large Nile perch, which are required by fish filleting factories, have been declining (Mkumbo and Mlaponi, 2007). This selective harvesting of Nile perch at the size at which it becomes piscivorous, could be leading to the re-appearance of haplochromines, in particular the zooplanktivorous species (Katunzi et al., 2003; Machumu et al., 2008) which are main predators of C. nilotica (Goudswaard et al., 2006; Machumu et al., 2008). Dominance of juveniles in the perch populations could be impacting C. nilotica as well as the resurging haplochromines. Furthermore, juvenile perches also feed on larvae and juveniles of C. nilotica as haplochromines did (Ogutu-Ohwayo, 1990; Goudswaard et al., 2006; Machumu et al., 2008). This could possibly lead to an apparent reversal of the regime shift which took place around the 1980s, more than 30 years after Nile perch introduction.
Since the C. nilotica boom of the 1980s was mainly due to the predation relief caused by the disappearing haplochromines (Goudswaard et al., 2006; Budeba and Cowx, 2007b), our study assumes that their recent recovery negatively influences the abundance of C. nilotica. We also assume that the selectivity of fishery in Lake Victoria has ultimately reversed the perches population structure, from the populations dominated by huge adults in the 1980s to the current populations dominated by juveniles.
In this study we intend to answer the following questions: (a) what is the current status (abundance) of C. nilotica after haplochromines reappearance? (b) Is the selectivity of fishery changing the population structure of the perch? And, (c) What is the contribution of C. nilotica to the diet of perches?
It is the purpose of this study to discuss the main drivers of these changes; i.e. increasing eutrophication, selectivity of fishery or overfishing.
Materials and Methods
Data on C. nilotica and dagaa were collected tri-monthly from the dagaa artisanal fishermen between August 2006 and January 2007 at Igombe and Kijiweni landing sites. Random selections of fishermen's boats were sampled to provide information on total catch weight (g), species composition, fishing effort and gear type. A sub sample of the total catch was taken, weighed, labelled, and then preserved in 4% formalin for analysis in the laboratory. Resulting statistics were used to assess the percentage contribution of C. nilotica in the dagaa fishery catches and to examine both spatial and temporal differences in catch returns, with the six month survey spanning two dry months (August–September) and four wet months (October–January). Quarterly bottom trawl surveys undertaken routinely by TAFIRI in the Tanzanian side of Lake Victoria waters were used to provide data on haplochromines and C. nilotica abundances between March 2005 and August 2008.
Calculation of Catch per Unit of Effort (CPUE) and Total catches estimation
Average catch rate (CPUE) was calculated from the sum of the total catch (in kg) of all boats sampled (Equation 1):
In order to estimate the total catch site−1 day−1, the CPUE (kg boat−1 day−1) of each species was multiplied by the number of boats that went fishing on that day (Budeba and Cowx, 2007b). CPUEs for C. nilotica for the two stations along with the monthly CPUEs of dagaa, C. nilotica, and haplochromines were established.
The total catches during the sampling periods were estimated as follows:
C = CPBD × FD × proportion of fishing boats × total number of boats recorded/1000 where C = catches within that period (August 2006 to January 2007; 6 months), CBD = mean Catch per Boat per Day (kg boat−1 day−1), FD = number of Fishing Days in a month (21 days average) (Budeba and Cowx, 2007b). Friedman test (for repeated measures) was used for the test of spatial and temporal variations in catch rates of C. nilotica while one sample t- test statistic for comparison.
Nile perch diet composition was determined from samples collected during a bottom trawl survey conducted between 3rd and 19th November 2006 on the entire Tanzanian sector of L. Victoria. Details on the catch techniques and analysis process are given by Budeba and Cowx, 2007a.
The abundance of C. nilotica in the R. argentea fishery catches
At Igombe, C. nilotica's average CPUE was 2.68 ± 0.87 (kg boat−1 day−1, N = 158) while that at Kijiweni was 0.69 ± 0.41 (N = 257). It contributed about 0.33% of the total dagaa fishery catches at Igombe but roughly doubled (0.63%) at Kijiweni. Based on the combined data from Igombe and Kijiweni, the mean CPUE for C. nilotica in the dagaa fishery was 1.68 ± 1.20 kg boat−1 day−1, compared to mean CPUEs of 174.92 ± 145.52 and 270.82 ± 120.86 kg boat−1 day−1 for dagaa and haplochromines respectively. The C. nilotica contributed 0.37% as a bycatch of the dagaa fishery catches, haplochromines formed a major catch (59.41%) followed by dagaa (38.38%) (Table 1).
Both C. nilotica, dagaa and haplochromines exhibited significant spatial and temporal variations (Friedman test, p < 0.001). On average, Igombe had higher catches of all the species than Kijiweni, so too the months of September, November, and October which constituted the wet season during the course of this study (Figure 1).
The catch rate of C. nilotica we report (1.68 ± 1.20 (kg boat−1 day−1), 0.37% of the total catches) differed significantly (one sample t- test, p < 0.001) with that reported by Budeba and Cowx, 2007a (9.60 ± 6.10 (kg boat−1 day−1)) at Kijiweni and Igombe.
The abundance of C. nilotica in the bottom trawls catches
Table 2 shows the mean catch rates of haplochromines and C. nilotica from bottom trawl surveys. The bottom trawl mean catch rates of C. nilotica conducted in Lake Victoria between 1979 to 1984 indicated zero catches (Goudswaard et al., 2006), implying that C. nilotica was previously very low in abundance, but increased in subsequent years.
The diet of the former and reappearing haplochromines
The diet of the recovering detritivorous haplochromines showed a remarkable increase in diversity, demonstrating versatility in feeding behaviour. For the period between 1991 and 2008, C. nilotica contributed significantly to the diet of the detritivorous and zooplanktivorous haplochromines contrary to the situation in the period between 1977 and 1982 (Table 3).
Size structure and the diet of Nile perch
Of the total 629 gutted Nile perch for dietary analysis, only 230 had prey items in their stomachs, and of those 92 were full, 9 were three quarter filled, 54 were half filled, 25 were quarter filled, and 50 were less than quarter filled. C. nilotica dominated in the diet of Nile perch by contributing 54% by volume to the diet of fish < 50 cm total length (TL) and haplochromines contributed 41% to the diet of fish > 50 cm TL. Juvenile Nile perch between 1 and 2 cm TL fed on zooplankton before switching to C. nilotica at a length of 3 cm TL. All Nile perch below 20 cm TL fed predominantly on C. nilotica, but switched to piscivory predominantly on haplochromines at 60 cm TL.
The current abundance of C. nilotica was very low, and contributed insignificantly to the R. argentea fishery catches in contrast to previous findings of the C. niloticas' significant contribution in this regard. Budeba and Cowx, (2007b) report a mean catch rate of 8.97 ± 11.83 and a contribution of about 7.6% of C. nilotica in the total R. argentea(dagaa) catches in the Tanzanian waters, data they collected between 2000 and 2002. They reported a mean catch rate of 9.60 ± 6.10 (kg C. nilotica boat− 1 day−1) in the Igombe and Kijiweni in contrast to a mere 1.68 ± 1.20 (kg boat−1 day−1) of our study. We attribute the current low abundance of C. nilotica in the lake to three interdependent factors.
Firstly, is the reappearance of the detritivorous and macrofauna feeding haplochromine cichlids, which are known to effectively feed on larvae and juveniles of C. nilotica (Goudswaard et al., 2006; Machumu et al., 2008). In that period, haplochromine cichlids constituted more than 83% of the lake ichthyomass and were believed to have been driven towards extinction by Nile perch in the late 1980s (Witte et al., 1992, 2007). Our study reflects their astounding recovery, manifest in their contribution as an important by-catch in R. argentea fishery at Igombe landing site and their reappearance in bottom trawl catches. The reappearing zooplanktivorous haplochromines and dietary switch to C. nilotica of the detritivorous and benthivorous haplochromines significantly impact the C. nilotica abundance.
Secondly, is the increased predation from juveniles of Nile Perch and the increase in abundance dagaa. Currently, Nile perch stocks are declining and there is increased dominance of juveniles in its populations (Katunzi et al., 2006), a result of overfishing and selectively harvesting of large perches required by filleting factories. Contrasting with large perches which are mainly abundant in the offshore waters (Mkumbo, 2002), juvenile perches prefer most inshore waters (Katunzi et al., 2006) the same as larvae and juveniles of C. nilotica (Budeba and Cowx, 2007b). As Nile perch tendency is to feed on the abundant food resource (Mkumbo, 2002), then the current juvenile perches impact on the C. nilotica is significant as that of haplochromines. Haplochromines are implicated of limiting the abundance of C. nilotica during 1900–1950s because they predated mostly on its larvae and juveniles (Witte et al., 1992, 2007; Goudswaard et al., 2006) contrasting the Nile perch which by that time its populations were dominated by offshore large adult perches (Ogari and Datzie, 1988; Mkumbo, 2002; Katunzi et al., 2006). And, when Nile perch significantly reduced the haplochromines (during 1980s–1990s), the resulting predation relief allowed the C. nilotica to recruit strongly to the reproductive stage, thus causing its boom (Witte et al., 1992; Goudswaard et al., 2006).
Furthermore, dagaa which also predates on C. nilotica's larvae and juveniles, currently are increasing in abundance (Wanink, 1998; Wanink et al., 2002; GW Ngupula, pers.obs). Its predation impacts coupled with its competition (i.e. for food resources, habitat, etc) with the same, also contribute to the further decrease in C. nilotica's abundance.
The third factor which negatively influences the abundance of C. nilotica is the currently observed eutrophication and pollution, as well the global warming impacts affecting the health of Lake Victoria ecosystem. Eutrophication and pollution in the lake have resulted into increased environmental stress like hypoxia and change in food quality, among many others. Majority of the bottom debris in the past was mainly made up by decaying diatoms (in particular Aulacoseira spp.) (Mugidde, 1993) whereby at post Nile perch era, it is mainly consisting of decaying less nutritious species like that of Nitzschia acicularis (Bacillariophyta), and Microcystis flos aquae(cyanophyta) from the abundant algal blooms in the lake, the fact which have several implications to the detritus feeders like C. nilotica(Hart et al., 2003).
Ngupula and Kayanda (accepted) observation on the abundance patterns of benthic macroinvertebrates in the near shore, intermediate, and offshore waters of Lake Victoria, observed a shift, from oligochetes and insects dominated community in 1980s to molluscs dominated community in 2008.They implicated eutrophication and pollution as the main driving factor towards that shift. The changed Lake Victoria environment negative affects the C. nilotica physiology, thus impacting its reproduction efficiency and making it highly vulnerable to its predators (Hart et al., 2003)
Furthermore, the current global warming impacts also influence the climate of the Lake Victoria basin impacting on its mixing and stratification patterns. The lake basin climate is characterised by two distinct seasons: a dry season with lake mixing in July–September and a wet season with two rainfall maxima, short rains period with thermal re-establishment and moderate stratification in October–December and long rains with stratification in March–May (Ngupula and Kayanda, accepted). The increasing temperature and incomplete mixing of the lake might as well impact the C. nilotica physiology.
The current low abundance of C. nilotica in the lake is explained by the increase in predation on its juveniles by the recovering haplochromines and by dagaa and the juvenile perches as well. Fishing pressure, selectivity of fishery and eutrophication are the main driving factors for this regime shift. Since C. nilotica are so important in young Nile perch and haplochromine diets, future recruitment of Nile perch can be adversely affected by its further decline, and therefore we call upon for a continued research and monitoring of the diet of Nile perch and abundance of C. nilotica
We express our sincere thanks to the European Union for funding part of this research through the Implementation of Fisheries Management Plan Project (IFMP). Thanks also go to the TAFIRI – Mwanza staffs, and crew members of R/V Lake Victoria Explorer for their valued support during data collection and comments on the early versions of this paper. The anonymous reviewers are acknowledged for their constructive criticisms.