Next Article in Journal
Lag Synchronization of Complex Lorenz System with Applications to Communication
Previous Article in Journal
Identity Authentication over Noisy Channels
Previous Article in Special Issue
Modeling Soil Moisture Profiles in Irrigated Fields by the Principle of Maximum Entropy
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Broad Niche Overlap between Invasive Nile Tilapia Oreochromis niloticus and Indigenous Congenerics in Southern Africa: Should We be Concerned?

1
South African National Biodiversity Institute, Pretoria National Botanical Gardens, 2 Cussonia Avenue, Brummeria, Private Bag x101, Pretoria, 0184, South Africa
2
Department of Ichthyology and Fisheries Science, Rhodes University, PO Box 94, Grahamstown, 6140, South Africa
3
DST-NRF Centre of Excellence for Invasion Biology (CIB), Department of Zoology and Entomology, University of Pretoria, Private Bag X20, Hatfield, 0028, South Africa
*
Author to whom correspondence should be addressed.
Entropy 2015, 17(7), 4959-4973; https://doi.org/10.3390/e17074959
Submission received: 26 February 2015 / Revised: 7 July 2015 / Accepted: 8 July 2015 / Published: 14 July 2015
(This article belongs to the Special Issue Entropy in Hydrology)

Abstract

:
This study developed niche models for the native ranges of Oreochromis andersonii, O. mortimeri, and O. mossambicus, and assessed how much of their range is climatically suitable for the establishment of O. niloticus, and then reviewed the conservation implications for indigenous congenerics as a result of overlap with O. niloticus based on documented congeneric interactions. The predicted potential geographical range of O. niloticus reveals a broad climatic suitability over most of southern Africa and overlaps with all the endemic congenerics. This is of major conservation concern because six of the eight river systems predicted to be suitable for O. niloticus have already been invaded and now support established populations. Oreochromis niloticus has been implicated in reducing the abundance of indigenous species through competitive exclusion and hybridisation. Despite these well-documented adverse ecological effects, O. niloticus remains one of the most widely cultured and propagated fish species in aquaculture and stock enhancements in the southern Africa sub-region. Aquaculture is perceived as a means of protein security, poverty alleviation, and economic development and, as such, any future decisions on its introduction will be based on the trade-off between socio-economic benefits and potential adverse ecological effects.

1. Introduction

Nile tilapia, Oreochromis niloticus (Linnaeus 1758), is an endemic African freshwater cichlid that is native to the Nile River basin, southwestern Middle East, and the Niger, Benue, Volta, and Senegal rivers, as well as lakes Chad, Tanganyika, Albert, Edward, and Kivu [1,2]. Owing to its hardy nature, and its wide range of trophic and ecological adaptations, O. niloticus has been widely introduced worldwide for aquaculture, augmentation of capture fisheries, and sport fishing [1,3]. It is currently among the most widely distributed invasive fish worldwide and has established feral populations in most tropical and sub-tropical environments to which it has gained access [35].
Within sub-Saharan Africa, O. niloticus was initially introduced into Lake Victoria in the 1950s and its distribution has since expanded to include most river systems in eastern and southern Africa [6]. It was imported into Zambia in 1982 for aquaculture purposes and appeared in the Kafue River in the mid-1990s after escaping from nearby fish farms [7]. Subsequent aquaculture introductions occurred in Lake Kariba. It is now widely distributed and common in most sub-catchments of the Middle Zambezi, Nata (Makgadikgadi/Okavango), Runde-Save, Buzi, and Limpopo river systems [812]. The advent of O. niloticus into novel river systems is a cause for concern for the conservation of indigenous congenerics that are at an extirpation risk through hybridization and competition arising from habitat and trophic overlaps. Despite its widespread distribution within sub-Saharan Africa, several river systems are still free of O. niloticus but remain vulnerable. These areas currently act as refuges for the conservation of indigenous congenerics and it is important to identify such areas to prevent further O. niloticus introductions. These areas include the Cunene, Okavango, and the Upper Zambezi rivers where threespot tilapia (O. andersonii) and greenhead tilapia (O. macrochir) are endemic [1,6]. The localised distribution of endemic tilapiines within southern Africa predisposes them to invasion as the introduction of O. niloticus into a given catchment is likely to have a significant impact on species with limited natural ranges. Examples include: (1) Kariba tilapia (O. mortimeri) that is naturally confined to the middle Zambezi and Luangwa river systems; (2) black tilapia (O. placidus) that is confined to the east coastal plains from the Lower Zambezi to the Mkuze rivers in KwaZulu-Natal Province, South Africa; and (3) Mozambique tilapia (O. mossambicus) that occurs from the lower Zambezi River system to the Bushman’s River system of Eastern Cape Province, South Africa, spreading far inland within the Limpopo River Basin, but south of the Phongolo River system, and is confined to the closed estuaries and coastal reaches of rivers [6].
This study is an extension of a recent study that applied ecological niche models to predict the potential invasive range of O. niloticus and revealed broad invasive potential over most river systems in southern Africa [13]. The present study attempts to explicitly quantify the niche overlap between O. niloticus with indigenous congenerics in southern Africa and highlights conservation concerns in areas where there is broad niche overlap. This was achieved by developing niche models for the native range of indigenous Oreochromis species in southern Africa and assessing how much of their range is climatically suitable for the establishment of O. niloticus. We specifically addressed the following two research goals: (1) what is the degree of niche overlap between the projected invasive range of O. niloticus and the natural ranges of endemic congenerics in river systems in southern Africa? and (2) we investigated the degree of niche overlap between the projected invasive range of O. niloticus and the natural ranges of endemic congenerics in river systems in southern Africa, and then reviewed the conservation implications for indigenous congenerics as a result of overlap with O. niloticus based on documented congeneric interactions.

2. Methods

2.1. Environmental Data Sources

The environmental variables dataset was composed of bioclimatic variables that have been widely used in ecological niche modelling of species [14]. These variables represent annual trends (mean annual temperature and annual precipitation), seasonality (annual range in temperature and precipitation), and either extreme or limiting environmental factors (temperature of the coldest and warmest months and precipitation of the wet and dry quarters). Variable selection for model training followed [13] and only used six out of 19 bioclimatic variables that represented the availability of water and energy. This allowed for comparative analysis of niche models between the two studies and the six variables have been shown to produce better a model performance than any other variable combinations [13].

2.2. Specimen Data Sources

Georeferenced occurrence data for four Oreochromis species (O. andersonii, O. mortimeri, O. mossambicus, and O. niloticus) that are known to occur in southern Africa were obtained from various sources, including museum specimen records and biodiversity databases such as FishBase [15], Global Biodiversity Information Facility [16], and the published literature and fish survey data from various fisheries departments in southern African countries that included Botswana, Mozambique, Zambia, and Zimbabwe.

2.3. Model Building

Modelling was based on the maximum entropy algorithm that was implemented in the Maxent modelling package (Maxent Version 3.3.3k) that utilizes associations between environmental variables and known species’ occurrence localities to predict potential areas where a given species is likely to establish [17]. The Maxent method has been shown to perform better than other correlative methods that use presence and background data [18]. For all models, the algorithm’s parameters were set to a maximum number of 500 iterations, a regularization multiplier of 1, convergence threshold of 0.00001, test percentage = 0, and only hinge features selected. Hinge features allow for simpler and more concise approximations of the true species response to the environment variables [19,20], thus preventing over-fitting of the model without significantly increasing the complexity of the models and, hence, improving model performance [17]. In addition, clamping was selected to minimize predictions to regions of environmental space outside the limits encountered during training because extrapolation may overinflate the degree to which species niches are estimated to overlap [2123]. The logistic output format was used to indicate the probability of a species presence at a default prevalence of 0.5 [20,24]. Values range from 0, indicating low probability, to 1.0, indicating greatest probability of a species presence in a given area.
Maxent makes use of presence and background localities to project potential species distribution models. The extent of the background is known to influence model performance, where a broad background can cause overestimates and a constrained background can cause underestimates [25,26]. In this study, the extent of the native range for each of the three Oreochromis species endemic to southern Africa was limited to hydrological basins where each species was known to occur as defined by previous studies [1,6]. This was achieved by overlaying the hydrological basins layer over known occurrence records for each species. A given hydrological zone was included as part of the background if it contained an occurrence point using ArcGIS® v10. (2011; ESRI, Redlands, CA, USA). The native range model for each of species was then calibrated with 10,000 pseudo-absence points drawn at random from the defined native range. Ten niche models were then constructed for each species within its native range and, in each model, all native occurrence records were partitioned using the statistical software package R (R Development Core Team 2008) into a calibration set (training set = 70% of all records) and a testing set (validation set = 30%) using k-fold partitioning [17]. Average model performance was obtained by repeating the process for 10 iterations. A consensus map was then created as an average of the 10 native range projection maps. The predicted distribution of the O. niloticus across African river systems was obtained from niche models constructed in a previous study [13]. These models were constructed using both native and introduced range occurrence records while the background selection was delimited using the buffering methods from a previous study [25]. In each model, the calibration set consisted of a subset of occurrence records from the native range and a subset of known introductions elsewhere in Africa. The projected models were then evaluated using a subset of records from the introduced range only. A consensus map was then created to show the average of the 10 introduced range projection maps.

2.4. Niche Overlap and Similarity

Niche similarity between niche models of O. niloticus and its southern African congenerics was estimated using ordination techniques [27] in the ecospat package in R (R Development Core Team 2008). Ordination techniques for quantifying niche overlap are recommended as a better alternative to other methods that use geographical projections derived from species distribution models, e.g., see [28], because the former allows for random points for similarity tests to be selected in environmental space instead of geographical space, thereby allowing for correction of bias associated with geographical dimension [27]. Niche overlap was measured using the Schoener’s index of niche breadth (D) [28]. Indices may range from 0 (indicating that niche models are completely different) to 1 (indicating that niche models are identical). The significance of the D values was then evaluated using null models of niche similarity. Ecospat quantifies niche similarity using several ordination techniques (27), but in this study we only used the principal component analysis (PCA-env) that was shown to consistently out-perform other ordination metrics of niche overlap [27]. For niche similarity, we tested the hypothesis that ecological niche models drawn from a partially or entirely non-overlapping distribution of O. niloticus and that of Oreochromis congenerics in river systems in southern African are more different from one another than expected by random chance. This test was conducted by calibrating the PCA-env with introduced records of O. niloticus but trained on a background randomly drawn from both the known introduction range of O. niloticus within southern Africa and the native range of each of its southern African congenerics (see model building section for native range delimitation). The same process was repeated by running a PCA-env based on known occurrence records for each of the southern African congenerics and trained on a randomly drawn background from its native range and from the hydrobasins with known occurrence records of O. niloticus within southern Africa. This process was repeated in either direction (native ↔ introduced) to generate 100 pseudo-replicate datasets for each species comparison with O. niloticus. The observed measures of niche similarity (D) from the two original populations (O. niloticus and each of its southern African congenerics) were then compared with percentiles of these null distributions. The hypothesis that niche similarity (or divergence) is different from that expected by chance between O. niloticus in its introduced range and that of each southern African congeneric based on the availability of habitats was rejected when the empirically observed values of D were either lower or higher than values obtained from pseudo-replicate datasets, resulting in a Type 1 error of 0.01.

2.5. Model Evaluation

The performance niche models of each Oreochromis species were evaluated using the Maximum test area under the curve (AUC) [29] and the Boyce index [30,31]. AUC defines the discrimination ability (between presence and background) of the models where values range from 0 (indicating random distribution) to 1.0 (indicating perfect prediction), with values > 0.5 being considered to indicate that the model discriminates better than random [29]. All AUC model performance measures were calculated in Maxent and predictions with an AUC value greater than 0.8 were considered to be acceptable [32,33]. Although AUC has been widely used as measure of model performance, it is not necessarily an appropriate measure for presence-only model evaluation [34]. Therefore, we used the Boyce index to further evaluate our model outputs. The Boyce index validation method involved partitioning the habitat suitability scores from each model output into 10 classes of equal intervals. For each class, we calculated the predicted and expected frequencies of pixels. Predicted frequency is the number of occurrence points predicted by the model to fall into habitat suitability class i divided by the total number of occurrence points used to build the model. The expected frequency is the number of grid cells belonging to habitat suitability class i divided by the overall number of cells in the whole study area. The predicted-to-expected (P/E) ratio was then calculated for each class and Spearman rank correlation was used to test if the P/E ratio increased with habitat suitability. The predicted-to-expected (P/E) ratio values may range from –1 to 1, where negative values indicate incorrect models that predict unsuitable habitats where presences are more frequent, values closer to zero indicate random predictions, and positive values indicate model predictions that are consistent with presence distribution in the evaluation dataset [30,31].

3. Results

3.1. Native Range Predictions

The native ranges of all three tilapia species (O. andersonii, O. mortimeri, and O. mossambicus) that are endemic to southern Africa were correctly predicted with good model performance (AUC > 0.83, Boyce index p < 0.001) for all four species (Table 1, Figure 1). There was a wide variation in the relative contribution of each variable to model performance for each respective tilapiine species (Table 2). The most important predictive factor for O. andersonii was maximum temperature in the warmest month, while for O. mortimeri it was annual precipitation. For O. mossambicus, it was annual mean temperature.

3.2. Introduced Range

The predicted potential geographical range of the O. niloticus reveals a broad climatic suitability over most of central and southern Africa. Its potential distributional range covers the entire basins of the Zambezi and Limpopo rivers and the continent’s coastal rivers along the Indian Ocean. The models, however, predicted low suitability for most of the Orange River basins and west-flowing rivers in the south Atlantic coast of Africa. There was no significant differences (P > 0.05) between training (0.92, SD = 0.002) and test datasets (0.91, SD = 0.021), and the most important predictor variables in the model performance were annual precipitation, annual mean temperature, and minimum temperature of the coldest month (Table 2).

3.3. Niche Overlap and Similarity

The predicted potential geographical range of O. niloticus within southern Africa shows significant pairwise niche overlaps with the native ranges of all three endemic congenerics. Niche overlap was most pronounced between O. niloticus and O. mossambicus (D = 0.59), followed by O. niloticus and O. andersonii (D = 0.55), and O. niloticus and O. mortimeri (D = 0.12). The observed overlap values between O. niloticus in its introduced range and that of each southern African congeneric’s native range was greater than 99% of the simulated values, indicating that for each species comparison, the species occupy environments that are more similar to each other than expected by chance (Table 3).

4. Discussion

The predicted potential range of O. niloticus reveals a broad climatic suitability over most of southern Africa. Eight out of 11 of the river basins in the sub-region are predicted as being climatically suitable for O. niloticus and it has already been confirmed to be present in six of them. This is a concern as 55% of the river basins in the region already have established O. niloticus populations and there is potential for further invasion. In addition, the predicted geographical range of O. niloticus overlaps with the complete native ranges of endemic congenerics in the sub-region. From previous evidence, these indigenous congenerics are likely to become extirpated from the river systems through either competitive exclusion and/or hybridisation [5]. Of additional concern are those areas that have been free of exotic species but are now vulnerable due to the introduction of fish mainly for aquaculture and sport fishing.

4.1. Upper Zambezi and Okavango

The first area of concern is the upper Zambezi and Okavango river systems. The upper Zambezi River is an ecologically diverse “reservoir river” with varied and extensive habitats [35]. As a result, the area has the highest fish diversity within the Zambezian River system [36]. The Upper Zambezi and the Okavango river systems are regarded as pristine areas where minimal fish introductions have occurred, but the ardent promotion of small-scale aquaculture as a means of poverty alleviation and livelihoods has put such river systems at serious invasion risk [37]. In Zambia, aquaculture projects rearing O. niloticus have been keenly promoted within the Zambezi River system, and the inevitable fish escaping from such facilities has led to the establishment of feral populations in river systems such as the Kafue River [7]. The distributional range of O. niloticus has also spread to include tributaries of the Upper Kapombo River and it is highly likely that it will reach a further spread within the upper Zambezi River, where O. andersonii and O. macrochir are at risk of being extirpated [37]. In the Kafue River, the Oreochromis population may already comprise a hybrid swarm that consists of fish with an admixture of O. niloticus, O. andersonii, and O. macrochir ancestry [38].
It is of concern that beside the introgression of O. niloticus alleles into the indigenous Oreochromis congenerics, the resultant hybrids are non-sterile and are able to backcross with each other and with O. macrochir and O. andersonii, facilitating gene mixing between the indigenous congenerics that do not otherwise frequently hybridise in sympatry. This facilitated hybridisation will likely have a major impact on the conservation of indigenous congenerics as they are likely to be extirpated from the river system and replaced by an admixture of introgressed hybrids. Other areas of concern include pristine regions that have experienced little or no fish introductions due to civil strife. These include Angola and the Democratic Republic of Congo (DRC). With the end of these conflicts, it is anticipated that increased development in the form of aquaculture projects may expose novel river systems to a serious risk of invasion [10]. In Angola, potential river catchments for aquaculture development, such as the Cuito, Cubango, and Cunene rivers, were also predicted as potentially suitable systems for the establishment of Nile tilapia.

4.2. Middle Zambezi

Further downstream, within the middle Zambezi River, O. niloticus is already established and appears to be displacing the indigenous Kariba tilapia, O. mortimeri. In Lake Kariba, O. niloticus appeared in the mid-1990s after escaping from in situ cage-culture fish farms and has become abundant at the expense of Kariba tilapia that has declined significantly in abundance [3941]. As a result, Kariba tilapia is now listed as Critically Endangered (CR) on the International Union for Conservation of Nature (IUCN) Red List of threatened species [42]. In Zimbabwe, O. niloticus has been extensively propagated by farmers and anglers for recreational and sport fishing. It is now prevalent in most catchments where it has supplanted indigenous tilapiines in most medium- to small-sized dams [43]. For example, in Lake Chivero, a medium-sized reservoir on the Upper Manyame River, a sub-catchment of the middle Zambezi, O. niloticus has displaced greenhead tilapia and is now the dominant commercial species harvested [44].
Further south in the Zimbabwean Lowveld, O. niloticus is known to attain a body weight of approximately 6 kg in Inyankuni, Mayfair, and other smaller dams. The ability of O. niloticus to attain such a large body size could partially explain its success in displacing indigenous congenerics, such as O. mortimeri, in Lake Kariba [41]. Oreochromis niloticus is an aggressive competitor and large males often out-compete other congenerics for limited spawning and nursing grounds [45]. Although the hybridisation between O. niloticus and indigenous Oreochromis congenerics has been illustrated in the upper Zambezi, a comparative analysis using the same approach has unfortunately not been done for the Middle Zambezi. Such a study could have, for example, used experimental gill-net data that has been collected at fairly regular intervals by the Lake Kariba Fisheries Research Station for the past 45 years. These data would have likely shown the genetic diversity of indigenous Oreochromis species before and after the introduction of O. niloticus.

4.3. Lower Zambezi

The third area of concern comprises the Lower Zambezi River and other eastern river systems such as the Buzi, Save-Runde, and the Limpopo River systems, where O. niloticus is now established and spreading. It poses a threat to the other indigenous tilapia species such as O. mossambicus, which is also listed as Near-Threatened (NT) on the IUCN Red List of threatened species as a result of habitat and trophic overlaps with O. niloticus [46,47]. Feral populations of O. niloticus are already established in the Runde-Save, Buzi, and Limpopo river systems where adverse ecological impacts such as reduced abundance of indigenous species have already been documented [11]. In addition, similar to the Kafue River, the Oreochromis cichlid population in the Limpopo River may already be comprised of a hybrid swarm that consists of fish with an admixture of O. niloticus, O. andersonii, and O. mossambicus [48,49]. The presence of the O. andersonii mitochondrial haplotypes in the Limpopo River system indicates another invasion event, also linked to aquaculture. Oreochromis andersonii was introduced in the early 1970s into Shashe Dam in Botswana for aquaculture [50] and it appears to have spread further downstream unnoticed because it is morphologically similar to O. mossambicus and the two species are difficult to distinguish from each other in sympatry. The impacts of invasive species are most insidious when they affect the genetic integrity of indigenous congenerics [51].
Despite the introduction of O. niloticus in the Limpopo River basin, several rivers within the system are still free of invasion but remain vulnerable. While O. niloticus has established feral populations along the main arm of the Limpopo River and the immediate reaches of its associated tributaries east of the Shashe/Limpopo rivers’ confluence, it is yet to establish in most river systems in the upper bushveld sub-catchment [52]. Its spread may have been retarded by limited natural dispersal pathways and physiological limiting environments such as low water temperatures in upstream rivers [52]. Similarly, another comparative study [48] further demonstrated the limited expansion of alien Oreochromis species in the lower Limpopo despite multiple introductions in upper sub-catchments, as well as the importance of headwater regions as “refugia” for unintrogressed O. mossambicus populations that are crucial for conservation and management of the species.

4.4. Lake Malawi

The fourth area of concern is the Lake Malawi system which is comprised of Lake Malawi, its affluents, and the Shire River down to Kapachira Falls. Oreochromis niloticus is now present in the Lake Malawi catchment, where it was reported from rivers in Tanzania that drain into Lake Malawi [53]. It was introduced from Morogoro in central Tanzania between 1998 and 2010 as part of initiatives to develop aquaculture. The presence of O. niloticus in the Lake Malawi catchment obviously raises concerns about its impact should it reach the lake, which seems likely. Its presence will largely impact the endemic chambo (O. shiranus, O. squamipinnis, O. lidole, and O. karongae) and haplochromine complexes in the lake [54]. This has been noted in another Rift Valley lake, Lake Victoria. It is a comparatively similar system to the Lake Malawi system and the introduced O. niloticus displaced the indigenous O. variabilis and O. esculentus [55]. However, the indigenous tilapias in Lake Victoria were severely depleted by overfishing before O. niloticus was introduced and the commonly held hypothesis that O. niloticus displaced the indigenous species might not be entirely supported. Oreochromis niloticus increased slowly in the lake and did not dominate the ichthyofauna until Nile perch Lates niloticus was introduced and they more or less eliminated the endemic haplochromines. Much the same happened in Lake Kyoga and it seems that competition with haplochromines might have prevented O. niloticus from becoming established [56]. If this is correct, then the impact of O. niloticus on Lake Malawi might not be too severe, especially if they can maintain their haplochromine stocks.

5. Conclusions

The major conservation concern is that the predicted potential geographical range of O. niloticus within southern Africa overlaps with the complete native ranges of all endemic congenerics. In river systems in which O. niloticus is already established within the region, it has been implicated in reducing the abundance of indigenous species through competitive exclusion and hybridisation. Despite these well-documented adverse ecological effects, O. niloticus remains one of the most widely cultured and propagated fish species in aquaculture and stock enhancements in the sub-region. Aquaculture is perceived as a means of protein security, poverty alleviation, and economic development and, as such, decisions on exotic fish introductions are usually based on the trade-off between socio-economic benefits and potential adverse ecological effects. This is further compounded by the fact that in most invaded systems, the establishment of O. niloticus has not led to a decrease in overall yields, but rather a replacement of indigenous species and, in some cases, pronounced fisheries production and poverty alleviation by creating alternative aquaculture and fisheries livelihoods.

Acknowledgments

This study was funded and undertaken under the auspices of the South African Working for Water (WfW) Programme through the DST-NRF Centre of Excellence for Invasion Biology (CIB), University of Pretoria, South Africa. It also benefited from partial funding from the South African National Biodiversity Institute (SANBI).

Author Contributions

All authors conceived the idea, designed the study, analyzed the data, and drafted the manuscript. All authors have read and approved the final manuscript.

Conflicts of Interest

The authors declare no conflict of interest

References

  1. Trewavas, E. Tilapiine Fishes of the Genera Sarotherodon, Oreochromis and Danakilia; British Natural History Museum: London, UK, 1983. [Google Scholar]
  2. Daget, J.; Gosse, J.P.; Teugels, G.G. Thys van den Audenaerde, D.F.E. In Check-List of the Freshwater Fishes of Africa (CLOFFA); ISNB: Bruxelles, Belgium; MRAC: Tervuren, Belgium; ORSTOM: Paris, France, 1986; Volume 4. [Google Scholar]
  3. Welcomme, R.L. International Introductions of Inland Aquatic Species; FAO Fisheries Technical Paper; FAO: Rome, Italy, 1988. [Google Scholar]
  4. Costa-Pierce, B.A. Rapid evolution of an established feral population (Oreochromis spp.): The need to incorporate invasion science into regulatory structures. Biol. Invasions. 2003, 5, 71–84. [Google Scholar]
  5. Canonico, G.C.; Arthington, A.; McCrary, J.K.; Thieme, M.L. The effects of introduced tilapias on native biodiversity. Aquatic Conserv: Mar. Freshw. Ecosyst. 2005, 15, 463–483. [Google Scholar]
  6. Skelton, P.H. A Complete Guide to the Freshwater Fishes of Southern Africa; Struik Publishers: Cape Town, South Africa, 2001; pp. 324–332. [Google Scholar]
  7. Schwank, E.J. The introduced Oreochromis niloticus is spreading on the Kafue floodplain, Zambia. Hydrobiologia 1995, 315, 143–147. [Google Scholar]
  8. Van der Waal, B.C.W.; Bills, R. Oreochromis niloticus in the Limpopo System. Ichthos 1997, 52, 14–16. [Google Scholar]
  9. Van der Waal, B.C.W.; Bills, R. Oreochromis niloticus (Teleostei: Cichlidae) now in the Limpopo River System. South. Afr. J. Sci. 2000, 96, 47–48. [Google Scholar]
  10. Wise, R.M.; van Wilgen, B.W.; Hill, M.P.; Schulthess, F.; Tweddle, D.; Chabi-Olay, A.; Zimmermann, H.G. The Economic Impact and Appropriate Management of Selected Invasive Alien Species on the African Continent; Prepared for: Global Invasive Species Programme, CSIR Report Number: CSIR/NRE/RBSD/ER/2007/0044/C; CSIR: Pretoria, South Africa, 2007; p. 43. [Google Scholar]
  11. Weyl, O.L.F. Rapid invasion of a subtropical lake fishery in central Mozambique by Nile tilapia Oreochromis niloticus (Pisces: Cichlidae). Aquatic Conserv: Mar. Freshw. Ecosyst. 2008, 18, 839–851. [Google Scholar]
  12. Zengeya, T.A.; Marshall, B.E. The inshore fish community of Lake Kariba half a century after its creation: what happened to the Upper Zambezi invasion? Afr. J. Aquat. Sci. 2008, 33, 99–102. [Google Scholar]
  13. Zengeya, T.A.; Robertson, M.P.; Booth, A.J.; Chimimba, C.T. Ecological niche modeling of the invasive potential of Nile tilapia Oreochromis niloticus in African river systems: Concerns and implications for the conservation of indigenous congenerics. Biol. Invasions. 2013, 15, 1507–1521. [Google Scholar]
  14. Hijmans, R.J.; Cameron, S.E.; Parra, J.L.; Jones, P.G.; Jarvis, A. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 2005, 25, 1965–1978. [Google Scholar]
  15. FishBase. Available online: www.fishbase.org assessed on 10 July 2015.
  16. Global Biodiversity Information (GBIF). Available online: http://www.gbif.org assessed on 10 July 2015.
  17. Phillips, S.J.; Anderson, R.P.; Schapire, R.E. Maximum entropy modelling of species geographic distributions. Ecol. Model. 2006, 190, 231–259. [Google Scholar]
  18. Elith, J.; Graham, C.H.; Anderson, R.P.; Dudík, M.; Ferrier, S.; Guisan, A.; Hijmans, R.J.; Huettmann, F.; Leathwick, J.R.; Lehmann, A.; et al. Novel methods improve prediction of species’ distributions from occurrence data. Ecography 2006, 29, 129–151. [Google Scholar]
  19. Phillips, S.J.; Dudik, M. Modelling of species distributions with Maxent: New extensions and a comprehensive evaluation. Ecography 2008, 31, 161–175. [Google Scholar]
  20. Elith, J.; Phillips, S.J.; Hastie, T.; Dudík, M.; Chee, Y.E.; Yates, C.J. A statistical explanation of Maxent for ecologists. Divers. Distrib. 2011, 17, 43–57. [Google Scholar]
  21. Elith, J.; Kearney, M.; Phillips, S. The art of modelling range-shifting species. Methods Ecol. Evol. 2010, 1, 330–342. [Google Scholar]
  22. Zurell, D.; Elith, J.; Schröder, B. Predicting to new environments: tools for visualising model behaviour and impacts on mapped distributions. Divers. Distrib. 2012, 18, 628–634. [Google Scholar]
  23. Owens, H.L; Campbell, L.P.; Dornak, L.L.; Saupe, E.E.; Barve, N.; Soberón, J.; Ingenloff, K.; Lira-Noriega, A.; Hensz, C.M.; Myers, C.E.; Peterson, A.T. Constraints on interpretation of ecological niche models by limited environmental ranges on calibration areas. Ecol. Model. 2013, 263, 10–18. [Google Scholar]
  24. Merow, C.; Smith, M.J.; Silander, J.A. A piratical guide to MaxEnt for modelling species’ distributions: What it does, and why inputs and setting matter. Ecography 2013, 36, 1058–1069. [Google Scholar]
  25. Van Der Wal, J.; Shoo, L.P.; Graham, C.; Williams, S.E. Selecting pseudo-absence data for presence-only distribution modeling: How far should you stray from what you know? Ecol. Model. 2009, 220, 589–594. [Google Scholar]
  26. Anderson, R.P.; Raza, A. The effect of the extent of the study region on GIS models of species geographic distributions and estimates of niche evolution: preliminary tests with montane rodents (genus Nephelomys) in Venezuela. J. Biogeogr. 2010, 37, 1378–1393. [Google Scholar]
  27. Broennimann, O.; Fitzpatrick, M.C.; Pearman, P.B.; Petitpierre, B.; Pellissier, L.; Yoccoz, N.G.; Thuiller, W.; Fortin, M.J.; Randin, C.R.; Zimmermann, N.E.; et al. Measuring ecological niche overlap from occurrence and spatial environmental data. Global Ecol. Biogeogr. 2012, 21, 481–497. [Google Scholar]
  28. Warren, D.L.; Glor, R.E.; Turelli, M. Environmental niche equivalency versus conservatism: Quantitative approaches to niche evolution. Evolution 2008, 62, 2868–2883. [Google Scholar]
  29. Mantel, S.; Williams, H.C.; Ormerod, S.J. Evaluating presence-absence models in ecology: the need to account for prevalence. J. Appl. Ecol. 2001, 38, 921–931. [Google Scholar]
  30. Boyce, M.S.; Vernier, P.R.; Nielsen, S.E.; Schmiegelow, F.K.A. Evaluating resource selection functions. Ecol. Model. 2002, 157, 281–300. [Google Scholar]
  31. Hirzel, H.M.; Le Lay, G.; Helfer, V.; Randin, C.; Guisan, A. Evaluating the ability of habitat suitability to predict species presences. Ecol. Model. 2006, 199, 142–152. [Google Scholar]
  32. Swets, K.A. Measuring the accuracy of diagnostic systems. Science 1988, 240, 1285–1293. [Google Scholar]
  33. Fielding, A.H.; Bell, J.F. A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ. Conserv. 1997, 24, 38–49. [Google Scholar]
  34. Lobo, M.L.; Jiménez-Valverde, A.; Real, R. AUC: A misleading measure of the performance of predictive distribution models. Global Ecol. Biogeogr. 2007, 17, 145–151. [Google Scholar]
  35. Jackson, P.B.N. Fish of the Zambezi system. In The Ecology of River Systems; Davies, B.R., Walker, K.F., Eds.; W. Junk: Dordrecht, The Netherlands, 1986; pp. 269–288. [Google Scholar]
  36. Bills, R.; Marshall, B.E. Fishes of the Four Corner Area. In In Biodiversity of the Four Corners Area; Timberlake, J.R., Childes, S.L., Eds.; Occasional Publications in Biodiversity No 15; Biodiversity Foundation for Africa; Bulawayo/Zambezi Society: Harare, Zimbabwe, 2004; Technical Reviews; Volume 2, pp. 353–371. [Google Scholar]
  37. Tweddle, D. Overview of the Zambezi River System: It’s history, fish fauna, fisheries and conservation. Aquat. Ecosyst. Health. Manag. 2010, 13, 224–240. [Google Scholar]
  38. Deines, A.M.; Bbole, I.; Katongo, C.; Feder, J.L.; Lodge, D.M. Hybridisation between native Oreochromis species and introduced O. niloticus in the Kafue River, Zambia. Afr. J. Aquat. Sci. 2014, 39, 23–34. [Google Scholar]
  39. Chifamba, P.C. Status of Oreochromis niloticus in Lake Kariba, Zimbabwe, following escape from fish farms. In Stocking and Introduction of Fish; Cowx, I.G., Ed.; Fishing News Books: Oxford, UK, 1998; pp. 267–273. [Google Scholar]
  40. Marshall, B.E. A checklist of the fishes of Lake Kariba (Zimbabwe-Zambia). Arnoldia Zimbabwe. 2006, 10, 257–274. [Google Scholar]
  41. Chifamba, P.C.; Videler, J.J. Growth rates of alien Oreochromis niloticus and indigenous Oreochromis mortimeri in Lake Kariba, Zimbabwe. Afr. J. Aquat. Sci. 2014, 39, 167–176. [Google Scholar]
  42. Marshall, B.; Tweddle, D. Oreochromis mortimeri. Available online: http://www.iucnredlist.org/details/63337/0 accessed on 10 July 2015.
  43. Marshall, B.E. Freshwater fishes of the Zambezi basin. In Biodiversity of the Zambezi Basin Wetlands; Timberlake, J., Ed.; Occasional Publications in Biodiversity No 8. Biodiversity Foundation for Africa; Bulawayo/Zambezi Society: Harare, Zimbabwe, 2000; pp. 393–459. [Google Scholar]
  44. Zengeya, T.A.; Marshall, B.E. Trophic interrelationships amongst cichlid fishes in a tropical African reservoir (Lake Chivero, Zimbabwe). Hydrobiologia 2007, 592, 175–182. [Google Scholar]
  45. Lowe-McConnell, R.H. The role of tilapias in ecosystems. In Tilapias: Biology and Exploitation; Beveridge, M.C.M., McAndrew, B.J., Eds.; Kluwer: Dordrecht, The Netherlands, 2000; pp. 129–162. [Google Scholar]
  46. Cambray, J.; Swartz, E. Oreochromis mossambicus. 2007. Available online: http://www.iucnredlist.org/details/63338/0 accessed on 9 February 2010.
  47. Zengeya, T.A.; Booth, A.J.; Bastos, A.D.S.; Chimimba, C.T. Trophic interrelationships between the exotic Nile tilapia, Oreochromis niloticus and indigenous tilapiine cichlids in a subtropical African river system (Limpopo River, South Africa). Environ. Biol. Fishes. 2011, 92, 479–489. [Google Scholar]
  48. Firmat, C.; Alibert, P.; Losseau, M.; Baroiller, J.F.; Schliewen, U.K. Successive invasion-mediated interspecific hybridizations and population structure in the endangered cichlid Oreochromis mossambicus. PLoS ONE 2013, 8, e63880. [Google Scholar]
  49. D’Amato, M.E.; Esterhuyse, M.M.; van der Waal, B.W.C.; Brink, D.; Volckaert, A.M. Hybridisation and phylogeography of the Mozambique tilapia Oreochromis mossambicus in southern Africa evidenced by mitochondrial and microsatellite DNA genotyping. Conserv. Genet. 2007, 8, 475–488. [Google Scholar]
  50. De Moor, I.J.; Bruton, M.N. Atlas of Alien and Translocated Indigenous Aquatic Animals in Southern Africa; South African National Scientific Programmes Report No 144; Foundation for Research and Development, Council for Scientific and Industrial Research: Pretoria, South Africa, 1988; pp. 202–204. [Google Scholar]
  51. Lockwood, J.L.; Hoopes, M.F.; Marchetti, M.P. Invasion Ecology; Blackwell Publishing: Oxford, UK, 2007; pp. 185–205. [Google Scholar]
  52. Zengeya, T.A.; Robertson, M.P.; Booth, A.J.; Chimimba, C.T. A qualitative ecological risk assessment of the invasive Nile tilapia, Oreochromis niloticus in a sub-tropical African river system (Limpopo River, South Africa). Aquat. Conserv. Mar. Freshw. Ecosyst. 2013, 23, 51–64. [Google Scholar]
  53. Genner, M.J.; Connell, E.; Shechonge, A.; Smith, A.; Swanstrom, J.; Mzighani, S.; Mwijage, A.; Ngatunga, B.P.; Turner, G.F. Nile tilapia invades the Lake Malawi catchment. Afr. J. Aquat. Sci. 2013, 38, 85–90. [Google Scholar]
  54. Weyl, O.L.F.; Ribbink, A.J.; Tweddle, D. Lake Malawi: Fishes, fisheries, biodiversity, health and habitat. Aquat. Ecosyst. Health. Manag. 2010, 13, 241–254. [Google Scholar]
  55. Ogutu-Ohwayo, R.; Hecky, R.E. Fish introductions in Africa and some of their implications. Can. J. Fish. Aquat. Sci. 1991, 48, 8–12. [Google Scholar]
  56. Ogutu-Ohwayo, R.; Odongkara, K.; Okello, W.; Mbabazi, D.; Wandera, S.B.; Ndawula, L.M.; Natugonza, V. Variations and changes in habitat, productivity, composition of aquatic biota and fisheries of the Kyoga lake system: Lessons for management. Afr. J. Aquat. Sci. 2013, 38, 1–14. [Google Scholar]
Figure 1. The projected distributional range for O. mossambicus, O. mortimeri, and O. andersonii within their native ranges, and O. niloticus within its potential invasive range in river systems in southern Africa. Each map represents an average of 10 replicates for each species created using the k-fold partition method. Potential distribution is indicated by shaded areas, with red and blue indicating high and low probabilities of suitable conditions, respectively.
Figure 1. The projected distributional range for O. mossambicus, O. mortimeri, and O. andersonii within their native ranges, and O. niloticus within its potential invasive range in river systems in southern Africa. Each map represents an average of 10 replicates for each species created using the k-fold partition method. Potential distribution is indicated by shaded areas, with red and blue indicating high and low probabilities of suitable conditions, respectively.
Entropy 17 04959f1
Table 1. Average model performance evaluated using the maximum test AUC and Boyce index for O. andersonii, O. mortimeri, and O. mossambicus within their native ranges, and O. niloticus within its potential invasive range in river systems in southern Africa.
Table 1. Average model performance evaluated using the maximum test AUC and Boyce index for O. andersonii, O. mortimeri, and O. mossambicus within their native ranges, and O. niloticus within its potential invasive range in river systems in southern Africa.
SpeciesTest AUCpValue Boyce index (Spearman’s rho)
O. andersonii0.89 ± 0.035< 0.001
O. mortimeri0.93 ± 0.005< 0.001
O. mossambicus0.83 ± 0.016< 0.001
O. niloticus0.91 ± 0.021< 0.001
Table 2. Environmental variables and their relative contribution (%) to build ecological niche models for O. andersonii, O. mortimeri, and O. mossambicus within their native ranges, and O. niloticus within its potential invasive range in river systems in southern Africa.
Table 2. Environmental variables and their relative contribution (%) to build ecological niche models for O. andersonii, O. mortimeri, and O. mossambicus within their native ranges, and O. niloticus within its potential invasive range in river systems in southern Africa.
Bioclimatic variableO. andersoniiO. mortimeriO. mossambicusO. niloticus
BIO1 = Annual mean temperature17.80.140.723.0
BIO5 = Maximum temperature warmest month37.521.51.10.9
BIO6 = Minimum temperature coldest month9.10.429.121.8
BIO12 = Annual Precipitation10.669.61.143.7
BIO13 = Precipitation of Wettest Month12.53.28.46.7
BIO14 = Precipitation of Driest Month12.55.219.64.0
Table 3. Niche overlap (Schoener’s index of niche breadth, D) and niche similarity tests for O. andersonii, O. mortimeri, and O. mossambicus within their native ranges and with the potential invasive range of O. niloticus in river systems in southern Africa.
Table 3. Niche overlap (Schoener’s index of niche breadth, D) and niche similarity tests for O. andersonii, O. mortimeri, and O. mossambicus within their native ranges and with the potential invasive range of O. niloticus in river systems in southern Africa.
OverlapO. niloticus compared to
Similarity
invasive→nativenative→invasive
O. andersonii0.55>0.01>0.01
O. mortimeri0.12>0.01>0.01
O. mossambicus0.59>0.01>0.01

Share and Cite

MDPI and ACS Style

Zengeya, T.A.; Booth, A.J.; Chimimba, C.T. Broad Niche Overlap between Invasive Nile Tilapia Oreochromis niloticus and Indigenous Congenerics in Southern Africa: Should We be Concerned? Entropy 2015, 17, 4959-4973. https://doi.org/10.3390/e17074959

AMA Style

Zengeya TA, Booth AJ, Chimimba CT. Broad Niche Overlap between Invasive Nile Tilapia Oreochromis niloticus and Indigenous Congenerics in Southern Africa: Should We be Concerned? Entropy. 2015; 17(7):4959-4973. https://doi.org/10.3390/e17074959

Chicago/Turabian Style

Zengeya, Tsungai A., Anthony J. Booth, and Christian T. Chimimba. 2015. "Broad Niche Overlap between Invasive Nile Tilapia Oreochromis niloticus and Indigenous Congenerics in Southern Africa: Should We be Concerned?" Entropy 17, no. 7: 4959-4973. https://doi.org/10.3390/e17074959

Article Metrics

Back to TopTop