Discussion

Species composition, rareness, and distribution

The total species count recorded was comparable with the FishBase overall estimate of 278 described fish species in Uganda (Froese & Pauly 2019). Also, the species accumulation curve approached the asymptote, suggesting that most of the species were considered in the analysis (Gotelli & Colwell 2001). Large lakes had higher species richness and more rare species compared with rivers and small lakes. Differences in species richness and rarity among water bodies may be attributed to geomorphological, abiotic, and biotic factors (Brown et al. 2007). Also, isolated inland water bodies favor rapid allopatric speciation and adaptive radiation (Basiita et al. 2018), which exposes the species to different evolutionary pressures. For example, along the Victoria Nile, Murchison Falls hinders fish migration to Lake Kyoga from Albert (Basiita et al. 2018). The falls along River Semiliki prevents fish passage to Lake Albert from Edward (Acere & Mwene-Beyanga 1990). Similarly, between Lake Victoria and Kyoga have been separated by the Owen and Bujagali falls (Basiita et al. 2018), and a sandbar between Lake Nabugabo and Victoria (Stager et al. 2005). These biogeographical barriers may have led to allopatric speciation; for instance, Lake Nabugabo which was once connected to Lake Victoria had five endemic species (Ogutu-Ohwayo 1993; van Alphen et al. 2004). At the species level, L. niloticus from lakes Kyoga and Victoria were genetically different (Basiita et al., 2018) possibly due to manmade barriers between the two lakes, which has impeded gene flow (Basiita et al. 2018).
High species richness and rarity are expected in habitats of long geological age, geographically isolated, and where species are prone to speciation (Strayer 2013). Also, according to the biogeographical principle, large areas have more species (Rosenzweig 1995). In this study, the high species richness and rareness observed in Lake Victoria is not surprising, given its surface area, explosive speciation and adaptive radiation of haplochromine cichlids because of hybridization, allopatric, microallopatric and sympatric mechanisms (van Alphen et al. 2004; Meier et al. 2017). A similar trend of high species richness with increased lake area was observed (Amarasinghe & Welcomme 2002). In large water bodies, species can shift to favorable habitats, for example, Squalius lucumonis and Telestes muticellus that migrated upstream in the Mediterranean Rivers due to climate warming (Carosi et al. 2019). The level of rarity of recently described species could not be determined in the available data. For example, the distributions of H. akika in Lake George (Lippitsch 2003) andH. katonga in River Katonga (Schraml & Tichy 2010) are not well understood except where the specimen was obtained. Also, haplochromine cichlids are mostly lumped as haplochromines (Marshall 2018), which may have accounted for the absence of certain species in particular water bodies.
For rivers, habitat degradations and manmade obstructions mostly damming have affected fish species migration, feeding, and recruitment patterns (FAO 2001). For instance, only 70 of the 177 large rivers in the world are free from damning (WWF, 2020). In Uganda, along the Upper Victoria Nile, three dams were constructed. These obstructions affected species richness and gene flow along the river (Basiita et al., 2018). For example, the stocks of H. simotes and L. victorianus have been affected by dams along the Upper Victoria Nile (Sayer et al. 2018). Similar effects of damming have been reported along Yangtze and Mekong (Dugan et al. 2010; Yi et al. 2010).

Conservation Priority Index (CPIw) for inland water bodies

Biological metrics including indices, indicators, or targets have been used by managers and decision-makers to prioritize areas for conservation (Tognelli et al. 2019; Linke et al. 2011). These metrics may include prioritizing areas with threatened species (Kirkpatrick 1983), threatened species affected by climate change (Tognelli et al. 2019), or fixed percentages of an area (Linke et al. 2011). However, some indices such as species richness do not incorporate the component of complementarity of the areas in question, thus, highly-ranking areas with the same species (Kirkpatrick, 1983). Shannon Weaver diversity index does not distinguish habitats with the same species evenness and richness (Omayio & Mzungu, 2019). Most priority indices are computed for terrestrial ecosystems or specific for a particular region, thus, cannot be extrapolated to other systems (Brum et al. 2017). For example, the Forest Conservation Priority index (FCPI) (de Mello et al. (2016) used the area and shape of the forest without considering the species conservation status. The Cave Conservation Priority Index (CCPi) (Souza Silva et al. 2014) did not consider the species or their conservation status but species richness, distribution, and impact weights. In the species conservation importance index, the species conservation status and rarity were considered (Halmy & Salem, 2015), but the index applied to terrestrial plants but not an aquatic ecosystem. Thus, these priority indices cannot be used to rank water bodies for site-based priority conservation.
In most instances, biodiversity measures such as species richness are usually higher for large water bodies, which would mean they are prioritized for conservation. Indeed, this study also showed higher species richness for large lakes compared to small ones. However, the novel conservation priority index (CPIw) was significantly higher for small lakes compared with large lakes. This observation is consistent with expectation, given the low ecological substitutability for the species and higher levels of exposure to human-induced threats in small water bodies compared to large systems. For a system such as Lake Victoria, with a vast habitat heterogeneity, fish species can easily seek refugia in other habitats (Seehausen et al. 1997; Chapman et al. 2003), which may not be possible in a small water body. The index showed that, for example, Lake Gigate with a surface area of 1.7 km2, but with 4 critically engendered haplochromine cichlids (H. latifasciatus, H. obesus, H. parvidens, and H. argenteus ) could be prioritized for conservation ahead of a large system with higher species richness. The likelihood of the species getting extinct in Lake Gigate is eminent if a similar magnitude of stress is applied to both lakes. However, the index does not imply that other water bodies with low CPI should not be monitored, but it may allow a conservation manager or decision makers to rank water bodies for urgent intervention, especially if the resources are limiting. Because conservation interventions should address social needs (Linke et al., 2011), large waterbodies, which are usually productive would be difficult to fully conserve. The index should, therefore, be adopted as a rapid metric measure to rank water bodies to enable prioritizing them for conservation. Further, if the size and species in habitats in the waterbody are known, the index can be downscaled to a habitat level. However, index could not be applied on rivers because most of them are dammed or obstructed, creating distinctive habitats along the river.

Supporting Information

The map large lakes in Uganda (Appendix S1), Waterbody species richness and IUCN status (Appendix S2), Rare species (Appendix S3), Ordination plot for waterbodies (Appendix S4). Ordination plot for species (Appendix S5), Simper analysis plot (Appendix S6), Waterbody Conservation priority index values (Appendix S7)