Fishes of the Loeme River System, Republic of the Congo, Lower Guinea, west-central Africa

Freshwater systems in central Africa remain poorly studied for many reasons, mostly pertaining to the inaccessibility of the region, access to funding for research, problematic political situations and disease outbreaks (World Bank, 2020). Mamonekene et al. (2018) note that it is often through investigations conducted to document social and environmental impacts of major infrastructure projects that substantial discoveries and understanding of these systems occur, which drives home the importance of scientifically sound survey approaches to document freshwater ecology prior to development (Larsen, 2016). The Loémé River is the second last major drainage in the Lower Guinea ichthyofaunal province before the Congo basin, and the catchment has been earmarked for various major mineral developments. The Loémé runs for 160 km from the headwaters in Dimonika Biosphere Reserve to the coast near Pointe Noire with a drainage catchment area of 3250 km2 (Stiassny et al., 2007). Despite being a major River system situated near a well-developed city (Pointe-Noire), only four prior ichthyological explorations of the Loémé River feature in published work (Daget, 1961 which is reported in Daget & Stauch, 1968; Mamonekene & Teugels, 1993; Stiassny et al., 2007) which report approximately 17% of the species known from the Lower Guinea ichthyofaunal province (Stiassny et al., 2007). Although the number of prior studies conducted is low, it is still generally better known than other systems in the country (Mamonekene et al., 2018). The objective of this paper is to represent the results of a survey conducted in the middle Loémé basin in 2014 where fish data were collected to provide baseline information for future monitoring of the study area. The paper also assesses the success of the rapid field approach applied in this study in terms of sampling representivity, and suitability of the field approach for future studies. Finally, we incorporate data from historical surveys of the upper and lower Loémé basin to provide a synthesis of material that may be used as a baseline for future studies. The general study area lies in the Ogooué-Nyanga-Kouilou-Niari Freshwater Ecoregion in the Lower Guinea ichthyofaunal province (FEOW, 2020). More locally, the study area includes the Tchivouba River which drains several smaller tributaries and has its confluence with the middle Loémé River (the main river of the basin) at Tchivouba village. The Loémé River catchment drains north-east to south-west through a series of large lakes (lakes Bindi, Loufoualeba and Cayo) into the Atlantic Ocean. The Cayo-Loufoualeba wetland complex is a Ramsar site located 40 km downstream of the Loémé and Tchivouba confluence (Pritchard et al., 2010). Sampling was undertaken in two field-survey periods at the end of wet-season (19 days, April/May 2014), and the peak dry-season (16 days, June 2014). Twenty-six sites were sampled (Figure 1; Table S1) based on vegetation/habitat types that were observed in the study area to make inferences on species habitat associations. Five main aquatic habitats were identified and included in the sampling regime: Rivers (R), Forest Streams (FS), Tall Swamp Forest (TS), Mitragyna stipulosa Swamp Forest (MS) and Lakes (L) (Table 1). Distichodontidae (3 of 5) Nothobranchiidae (3 of 5) Protopteriidae (1 of 1) Mormyridae (4 of 5) Anabantidae (2 of 3) Amphiliidae (1 of 5) Fish sampling efforts were site-specific and based on habitat type and accessibility. Typically, 1.5 days were allocated per site, applying both passive (fyke nets, monofilament gill nets, hook and line) and active (electrofishing, cast and scoop nets) methods (Table 2). This was complemented by observation of local fishermen's catch. Fishes were collected according to the guidelines for the use of fishes in research (AFS/AIFRB/ASIH, 2003) and taxonomic nomenclature follows Eschmeyer (2020). Species collected were fixed using 10% buffered formalin and were later transferred to ethanol. All collected vouchers were deposited at the Ichthyology Department of the American Museum of Natural History (AMNH) in New York, USA. Data are accessible online at the AMNH (2020) vertebrate zoology database. All fishes were collected and exported with permission of the Congolese Direction Générale de l'Economie Forestière, Direction de la Faune et des Aires Protégées (Permit No. CG1125635, on file at AMNH). The collection efficiency of the 2014 study was evaluated using a species accumulation curve plotted in Primer 7 (Clarke & Gorley, 2015). Estimators included Chao 1 (Incidence-based Coverage Estimator; Chao, 1987) and Jacknife1 (First-order jackknife estimator of species richness; Burnham & Overton, 1978, 1979) which were used to compare estimates of species richness against observed species values (Sobs). A sampling target of 90% of expected species based on the estimators was considered as representative sampling (Moreno & Halffter, 2000). Ichthyological survey results of this study are presented per aquatic habitat type sampled (Table 1) and include a checklist of species identified in this study and previous studies in the Loémé basin. A list of all sites and species of the Loémé basin is provided in Tables S1 and S2 for information (Daget, 1961; Daget & Stauch, 1968; Mamonekene & Teugels, 1993; Stiassny et al., 2007). In the current study, a total of 41 species of fishes representing 16 families were sampled across the five habitat types (Table 1). A large diversity was noted in the Characiformes, Perciformes and Siluriformes orders, which, cumulatively, made up 67.4% of fishes sampled in the Middle Loémé, which corresponds with what is expected in the region (Brooks et al., 2011). All species sampled in the present study were primary freshwater species. The number of taxa per habitat type for the present study was highest in Rivers (28 - Loémé River), Forest Streams (26) and Lakes (22) habitat type. Tall Swamp Forests had the lowest fish diversity (12) of the habitat types sampled. Our study highlighted new records for 14 species in the Loémé catchment, where Kamden-Toham and Teugels (1998) contributed 17 new records on the initial studies of Daget and Stauch (1968) and Daget (1961) (Table S2). A notable new record for the Loémé catchment from our study was Chrysichthys thysi which is known from the Congo, Ogooué and Nyanga Basins (Stiassny et al., 2007; AMNH, 2020). Another species of interest collected in the area included a putatively new species of Plataplochilus, which was sampled in Tall Swamp Forest in the Tchivouba catchment. The type locality of Plataplochilus loemensis is close to the study area in the Lombo Stream, a tributary of the Loémé approximately 30 km south of the Tchivouba Stream (Eschmeyer, 2020). However, P. loemensis was not sampled in the present survey and is distinctly different from the Plataplochilus sp. sampled in the study area by several morphological features, most notably body depth. In general, the taxonomy of Plataplochilus is poorly resolved (Cutler, 2019). Undescribed Plataplochilus species have been sampled in the Mboungou Badouma region, Sébé and Ogooué river drainage (Cutler et al., 2019) and Wonga-Wongué Presidential Reserve (Agnèse et al., 2018) in Gabon. It is therefore likely that more species of Plataplochilus exist in the study area than what have previously been described. The study area showed a moderate-high comparative species richness in relation to basins just north of the study area considering the size and heterogeneity of the systems, and position in the catchment (Walsh et al., 2014). Rivers exhibited a notable number of fishes that were sampled in this habitat only, with approximately 19% of fishes exclusive to this habitat (Table 1). Small cichlid species were restricted to Forest Stream habitat, and Lakes showed a lower number of habitat-specific species, where the remainder of the habitat types only contained generalists. The Amphiliid catfish occurred at a lower than expected frequency of occurrence in the study area based on the rheophilic habitat availability, with only one specimen of Phractura brevicauda sampled over the entire study period. Generally, Amphiliids are a sensitive, rheophilic group dependent on flowing water for survival (Lujan & Conway, 2015). The low abundance of this group from the current area indicates a potential decline in ecosystem integrity (likely a decrease in water quality), as rheophilic habitat and species are usually the first to be impacted when watercourses are disturbed by anthropogenic activities such as agriculture, damming and water abstraction (Welcomme et al., 2006). Larger-bodied members of the Cyprinidae were also conspicuously missing from the study area in the present study but were well-represented in earlier studies of the basin (Daget & Stauch, 1968; Mamonekene & Teugels, 1993). This family usually undertakes large-scale reproductive migrations during the onset of the rainy season. Many of these fish move from the main channel up into smaller tributaries where they reproduce, leaving their offspring to feed in the forest habitat, while returning downstream (Kamden-Toham & Teugels, 1998). This aspect of the ecology of large cyprinids (Pinder et al., 2019) highlights the importance of the habitat continuity between larger river systems and smaller streams. The low frequency of occurrence and lower than expected diversity in the area may also indicate ecosystem degradation due to urbanisation, mining and industrial developments, and potential species over utilisation due to overfishing (van Treeck et al., 2019) as no signs of major habitat fragmentation (e.g. dams or weirs) were observed in the study area. Fishermen were observed collecting Oreochromis niloticus (Nile Tilapia) in the area's lakes, and specimens were sampled in lakes Bindi and Tchikoutou in the present study. The species naturally occurs in the Nile basin and the rivers Niger, Benue, Volta, Gambia and Senegal in West Africa (Vreven et al., 2007). Records of the fish from Lower Guinea have been established by museum vouchers from Cameroon and Gabon; however, no collection records were available from the Republic of the Congo until the present study. Teugels et al. (1992) did observe O. niloticus in Lake Nanga in the Kouilou River basin. It is suspected that the species escaped from an aquaculture farm after its introduction from Sudan in 1953, and it is likely that it was introduced to various systems since (Stiassny et al., 2007). A few intermediary specimens of Tilapia were also collected in the lakes, and it is suspected that Tilapia guineensis and T. tholloni may also be hybridising in these habitats. Overall, the subtropical coastal rivers associated with the Lower Guinea ichthyofaunal province are rich in freshwater species (Brooks et al., 2011) and species richness is likely driven by the refuge offered by the forested West Coast Equatorial region during Pleistocene climatic fluctuations (Lévêque, 1997). In a review of West African riverine biodiversity, Hugueny and Paugy (1995) highlighted a strong correlation between species richness, watershed area and river discharge volume. Based on these relationships, the fish fauna of the Lower Guinea ichthyological province is disproportionately rich in relation to the sizes of the various watersheds (Teugels et al., 1992). The high fish biodiversity in the region is hypothesised to be a result of the relative stability of the hydrological regime in these rivers and the large number of microhabitats created by the forest element itself (Teugels et al., 1992). When combining the data from this survey of the middle Loémé River, with data from previous studies of the upper and lower Loémé system (Daget & Stauch, 1968; Mamonekene & Teugels, 1993; Stiassny et al., 2007) over the past 60 years, a total of 101 species in 27 families and 12 orders were recorded from the upper Loémé tributaries (Loukenéné and Loukoula rivers) to the Loémé estuary. The expected number of species for the Middle Loémé study area, according to accumulation curves, was 47 species based on Chao 1, and 48 species based on Jacknife1. The observed species richness (41 species) represented 91.3% of the species estimated by Chao 1 and 89.4% by Jacknife1, indicating the satisfactory efficiency (Moreno & Halffter, 2000) of the sampling effort applied for the collection for the seasons investigated and the section of the catchment considered (Figure 2). Based on species accumulation estimates from models applied in this study, the present effort is considered as representative of work requirements to investigate a small to medium scale study area (230 km2) and the authors suggest repeating this effort intensity as a minimum configuration for such investigations (Moreno & Halffter, 2000). We thank COMINCO Resources Limited for granting access to the data from this study. Thank you to Nicole Burri (Eawag, Switzerland) who provided technical support for this study. Invaluable taxonomic input was provided by Melanie L.J. Stiassny (AMNH), Jouke van der Zee (Royal Museum for Central Africa) and Anton Lamboj (University of Vienna) who are thankfully acknowledged. Thank you to Barbara Brown and Lowell Flanders (AMNH) for accessioning and cataloguing of the collection. Finally, thanks to mama Melanie and papa Phillip from Tchivouba village for their assistance in the field. None declared. GW, MNJ and VNBK made the collections and identified specimens. GW, JG, AJH and MPA wrote the text with contributions from VM, MNJ and VNBK. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions. Table S1 Table S2 Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.