Spatiotemporal variation in fishing patterns and fishing pressure in Lake Victoria (East Africa) in relation to balanced harvest

Highlights

• We analyse fishing patterns and fishing pressure in Lake Victoria (East Africa) in relation to balanced harvest.

• Historical fisheries followed a more balanced fishing pattern than current fisheries.

• Big adult demersal fisheries are overexploited; small pelagic fisheries are underexploited.

• An indication of fishing-induced phenotypic changes in Nile tilapia, but not Nile perch.

Abstract

Balanced harvest (BH) refers to applying moderate fishing pressure across a broad range of species, trophic levels (TL), stocks, or sizes in an ecosystem in proportion to productivity (gross production per biomass unit) or production (total cumulated biomass over a given period) instead of exerting pressure on particular taxa or sizes. Both modelling and empirical studies have shown that BH can lead to higher fish yield than selective fishing, with minimal changes to ecosystem structure and function. This concept has stimulated considerable debate, and one question that needs to be addressed is how close or far fisheries are from being balanced. Here, we investigated whether fishing on Lake Victoria (East Africa), the world’s second largest inland fishery, is consistent with BH, and whether there is any link between past and present fishing patterns and life history traits of major commercially-exploited species: Nile perch (Lates niloticus) and Nile tilapia (Oreochromis niloticus). We found exploitation rates to be relatively higher in high TL groups than low TL groups; however, the overall fishing pattern has been consistent with BH until recently (from 2000 onwards). Exploitation rates (E) above 50% of annual production (“overexploitation”) were observed for large Nile perch, Nile tilapia, and other less productive demersal groups (for example, catfishes, squeakers, and lungfish). Unexpectedly, there was no evidence of significant reduction in body size of Nile perch associated with this fishing pattern, which may be attributed to limited compliance of minimum size regulations. On the other hand, we found a significant reduction in body size of Nile tilapia, which tended to be associated with high fishing pressure. We conclude that the fishery is currently inefficiently utilized in terms of food energy value, whereby groups with highest production (including small Nile perch) are underexploited (E < 10%). However, moving towards BH in the case of Lake Victoria, and generally in many inland fisheries, will require a paradigm shift both in mindset and the law.

Introduction

Ecosystem approach to fisheries (EAF) has become a popular tool in fisheries management in recent years (Garcia et al., 2003, Patrick and Link, 2015). The fundamental objectives of the EAF, which are also in line with the objectives of the Convention on Biological Diversity (CBD), are to maintain ecosystem structure and functioning and to avoid overfishing (Garcia et al., 2003, Zhou et al., 2019). Balanced harvest (BH), a fishing strategy that considers moderate fishing pressure across a broad range of species, trophic levels (TLs), stocks, or sizes in an ecosystem in proportion to productivity (i.e., the gross production per biomass unit of an ecological group during a given period (P/B) (Garcia et al., 2012)) or natural production (i.e., the total cumulated new biomass produced from an ecological group during a given period (Heath et al., 2017)) is viewed as a holistic fishing strategy that can achieve the objectives of CBD in general and EAF in particular (Zhou et al., 2019). Studies have shown that BH can minimize the adverse fishing impact on the relative size and species composition of aquatic ecosystems, and increase or maintain sustainable yield from such ecosystems (Jacobsen et al., 2013, Kolding et al., 2016c, Law et al., 2014, Zhou et al., 2015). This new concept has attracted wide interest and discussion in the management of fisheries (Burgess et al., 2015, Froese et al., 2016, Kolding et al., 2016b, Reid et al., 2016, Zhou et al., 2019), and one question that needs to be addressed is how close or far fisheries are from being balanced.

Previous investigations into the distribution of fisheries exploitation patterns across stocks and species show exploitation (both in marine and inland fisheries) to be skewed towards high TL species, with significantly lower fishing pressure on low TL (highly productive) groups (Kolding et al., 2016a, Kolding et al., 2019). This fishing pattern is associated with inefficient utilization of the fishery in terms of food energy value (Kolding et al., 2016a). The reasons for this fishing pattern, however, tend to vary depending on the nature of the ecosystem and objectives of the fisheries. In large scale marine fisheries, for example, this selective exploitation pattern is largely driven by the market preference for large fishes (Charles et al., 2015, Sethi et al., 2010). In small scale inland fisheries, however, size-oriented fishery regulations (minimum mesh sizes, gear types, and minimum landing sizes), which are prevalent on most inland fisheries (Jul-Larsen et al., 2003), are likely to the biggest drivers. Otherwise, small scale inland fisheries are expected to be balanced as food provisioning and nutritional security (i.e., from maximizing biomass catch of small and low-value fishes) are as crucial as profit maximization from large (high-value) (Kolding et al., 2019). What is still unknown for the inland fisheries is whether this unbalanced fishing pattern observed in 13 African lakes (Kolding et al., 2019), for instance, has persisted over time, and whether there are any spatial differences within the same ecosystems that could be explained by differences in application of regulations or compliance.

Lake Victoria (Fig. 1) is the world’s second-largest lake and the largest tropical lake with a surface area of 68,800 km². It is also one the largest inland fisheries with annual fish landings approximately equal to one million tons (Taabu-Munyaho et al., 2016). Over the past five decades, the lake has undergone considerable changes, both in limnology and fisheries, aided by new species introductions, intensive fishing, eutrophication and habitat degradation, invasive weeds, and climate variability (Hecky et al., 2010, Kolding et al., 2008b, Mkumbo and Marshall, 2015, Taabu-Munyaho et al., 2016, van Zwieten et al., 2015). However, the sustainability of the fisheries amid increasing fishing pressure has been a central focus of discussion by scientists and managers, albeit no consensus being reached to date. Some authors have suggested that the levels of exploitation are not sustainable, and that fisheries are exhibiting signs of overfishing as seen from the reduction in catch rates (catch per unit effort (CPUE)) and size structure of main commercial fish species, i.e., Nile perch (Lates niloticus) and Nile tilapia (Oreochromis niloticus) (Matsuishi et al., 2006, Mkumbo and Marshall, 2015, Njiru et al., 2006, Pitcher and Bundy, 1995). Indeed, after the initial increase during 1980–1990, mainly due to the sudden Nile perch explosion (Njiru et al., 2006, van Zwieten et al., 2015), the overall CPUE plummeted. The catch per boat decreased from 35–45 tons per in the early 1990 s to less than 15 tons per year after 2000, while the catch per fisher decreased by 2–3 folds during the same period (Kolding et al., 2019). Similarly, the proportion of large Nile perch (i.e., individual fish greater than 50 cm total length) in the catches has decreased from 80% in the year 2000 to less than 20% by the year 2011 (Mkumbo and Marshall, 2015). Yet, other scientists are strongly opposed to the view that this decline is a manifestation of overfishing, arguing that the overall landings have not plummeted with the increase in fishing effort, and that the reduction in CPUE and biomass of primary commercial fisheries is a normal response to fishing, while the decrease in size structure is due to selective fishing on adults, possibly compensated by increased production and earlier maturation (Kolding et al., 2014, Kolding et al., 2019, Kolding et al., 2008b).

These disagreements among scientists may largely be due to the mixed nature of the fishery and differences in interpretation of data, but also the confounding decreased catch rates (CPUE) with decreased total catches (Jul-Larsen et al., 2003). The former makes it difficult to assess fishing effort on individual fish species and to estimate species-specific catch rates precisely. In Lake Victoria, the CPUE is mainly approximated using indirect methods, for example, from the combined catches and either the total number of fishers or fishing crafts (see, for instance Kolding et al., 2014, Kolding et al., 2019). While this approach may help to give a general picture of the fishery, it has shortcomings arising from lack of separation of catches and fishing effort. For example, Fig. 2 shows that the catches of most species, especially the long-lived demersal and bentho-pelagic species (Table 1), have declined markedly in the past; only Nile perch has remained stable since the 1990 s (with small annual fluctuations), while Silver cyprinid (Rastrineobola argentea), locally called dagaa, has increased over time. Those opposed to overfishing mainly argue based on data in Fig. 3, which always shows that the overall catches have been increasing with increasing effort, a sign that the fishery is productive and healthy (Kolding et al., 2014, Kolding et al., 2008b). Yet, only one species (dagaa) is solely responsible for the observed continual increase in catches. If we continue to rely on CPUE estimated from combined catches, the high productivity of dagaa (which can reproduce its biomass up to 4 times per year (Natugonza et al., 2016)) may mask the effect of increasing effort on other species, especially in the event of localized or species-specific overfishing. Consequently, a different method is needed to assess the status of the fisheries at a finer scale (preferably at species level). One potential approach is by examining how fishing pressure is distributed across the broader trophic spectrum in relation to the biological productivity or production of an individual species or TL (i.e., balanced harvest), similar to the one used in Kolding et al. (2016a) and Heath et al. (2017). This approach would help in identifying species that are likely overexploited relative to biological production or productivity, and those that are least exploited.

In addition, the findings reported in Mkumbo and Marshall (2015), regarding the decrease in the proportion of large Nile perch, require further scrutiny. A 60% decrease in the proportion of large fish in a period of 10 years could be an example of the fastest fishery-induced evolution (FIE) known from wild fisheries populations to date. FIE can manifest through human-induced phenotypic change in natural populations, including slower growth, higher reproductive investment, and maturation at younger ages and smaller sizes (Audzijonyte et al., 2013a, Enberg et al., 2012, Heino et al., 2015). Fishing targeting large and fast growing individuals have been linked to FIE (Bundy et al., 2005, Law, 2007). Although FIE is generally a slow process (Andersen and Brander, 2009), there are examples from modelling and empirical studies which suggest that in some populations, it can be faster than expected (e.g., Audzijonyte et al., 2013a). In marine fisheries, examples include the decline of age-specific maturation length for North Atlantic cod by about 10 cm in 7 years (Olsen et al., 2004), the decrease in weight-at-age for Atlantic silversides of ca. 40% in four generations (Conover and Munch, 2002), and harvest-induced evolutionary changes in wild fish stocks in the range of 20–30% over 13–125 years (Jørgensen et al., 2007). FIE has also been suggested in tropical inland fisheries. In Lake Victoria, for example, FIE has been suggested for dagaa, involving a reduction in length at 50% maturity from 47 mm to 36 mm between 1990 and 2010 (Sharpe et al., 2012) and a 50% reduction mean length between mid-1970 s and 2015 (Mangeni-Sande et al., 2018). Given the selective nature of fishing in Lake Victoria, driven by size-oriented regulations, especially for the high value commercial fisheries (Njiru et al., 2010), the disappearance of large Nile perch in catches (Mkumbo and Marshall, 2015) might be an indicator of FIE. This phenomenon needs additional investigation, preferably using fisheries-independent data and metrics that are less likely to be influenced by changes in selectivity and recruitment (e.g., mean adult length).

The specific objectives of this paper, therefore, are twofold. First, we aim to explore whether fishing on Lake Victoria (in space or time) is consistent with the BH strategy. Because of the size-oriented fishery regulations, e.g., the minimum mesh size and minimum size of landed fish (Njiru et al., 2010) and the emphasis by governments on maximizing revenue from high-value fisheries, e.g., through the processing of big Nile perch for export (Johnson and Bakaaki, 2016, LVFO, 2016a), one would expect higher-TL species to be exploited harder than lower-TL groups, resulting in a fishing pattern that is inconsistent with the BH strategy. More recently, however, especially after 2000, food provisioning and nutritional security (i.e., from maximizing biomass catch of small and ’low-value’, but highly productive fishes, e.g., dagaa and juvenile Nile perch) have become as crucial as profit maximization from the export of large (high-value) Nile perch (Kolding et al., 2019). Accordingly, one would expect current fisheries to be closer to balance than fisheries in the past (for example, before 2000). Second, we aim to examine selected life-history aspects of major commercially exploited fisheries in Lake Victoria (Nile perch and Nile tilapia) to ascertain the likelihood of fishery-induced phenotypic change associated with fishing. Because of the emphasis on big-sized fisheries in Lake Victoria fisheries management (LVFO, 2016a), one would expect a decrease in body size and size at maturation of these fisheries in line with what has been reported for dagaa (e.g., Mangeni-Sande et al., 2018; Sharpe et al., 2012), although phenotypic changes in dagaa and Nile tilapia could also be environmentally-driven (Kolding et al., 2008b, Wanink and Witte, 2000).

To test the two hypotheses for the first objective, we use a framework proposed by Kolding et al. (2016a), showing the relationship between yield and production by trophic level (TL). We supplement this analysis by exploitation indicators, e.g., the trophic balance index (TBI) and mean exploitation (Bundy et al., 2005), showing the variability of exploitation across ecosystem-specific TLs. Our interpretation of results follows that a species/trophic group is overexploited if exploitation rate (E, i.e., the ratio of yield to production) is above 50% (Alverson and Pereyra, 2001) and sustainable when E is less than 40% (Patterson, 1992, Pikitch et al., 2012). The data required to perform these analyses is yield (normally annual total catch) and total annual production across all exploited species in an ecosystem, either by TL or size. These data are readily available in standardized formats in Ecopath models by species or functional groups and TL (Christensen and Pauly, 1992). Whereas size/age is part of the Ecopath model formulation, through the multi-stanza input (Christensen and Walters, 2004), none of the Ecopath models for Lake Victoria separates the species or functional groups by stanzas (Musinguzi et al., 2017). To address the second objective, therefore, we examine length frequency distributions of Nile perch and Nile tilapia using data from trawl surveys, which has been collected almost on annual basis between 1997 and 2017 from the Ugandan side of Lake Victoria by the National Fisheries Resources Research Institute (NaFIRRI). Unlike Mkumbo and Marshall (2015), we choose to use fishery-independent data as it is less influenced by changes in gear selectivity observed in commercial fisheries. Also, since the average size can also be greatly diminished by recruitment, which is shown to have increased recently (van Zwieten et al., 2015), we compare mean adult length over time instead of average length (see below). We also examine changes in length at first maturity mainly using information from vast literature sources.