University of Groningen The biology and impacts of Oreochromis niloticus and Limnothrissa miodon introduced in Lake Kariba Chifamba, Chiyedza Portia

(1)

The biology and impacts of Oreochromis niloticus and Limnothrissa miodon introduced in

Lake Kariba

Chifamba, Chiyedza Portia

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Chifamba, C. P. (2019). The biology and impacts of Oreochromis niloticus and Limnothrissa miodon introduced in Lake Kariba. Rijksuniversiteit Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

528794-L-bw-Chifamba 528794-L-bw-Chifamba 528794-L-bw-Chifamba 528794-L-bw-Chifamba Processed on: 6-2-2019 Processed on: 6-2-2019 Processed on: 6-2-2019

Processed on: 6-2-2019 PDF page: 139PDF page: 139PDF page: 139PDF page: 139

Synthesis

Portia C. Chifamba

(3)

(4)

141

8

Introduced fish species contribute more to the fisheries (capture and aquaculture) in Zimbabwe than indigenous fish species. The most important Nile tilapia (Oreochromis

niloticus) and the Lake Tanganyika sardine Limnothrissa miodon are the subject of

this thesis. Limnothrissa miodon contributes more than 90% to the fish catches in Lake Kariba, and the fisheries output from Lake Kariba contributes about 90% of fishery output in Zimbabwe. Even though the catches of O. niloticus are much less than those of L. miodon, O. niloticus is also of commercial importance in the artisanal fishery of Lake Kariba, and it is the prime aquaculture fish in Kariba. Knowledge on these two species is needed to secure and enhance socio-economic benefits through informed management decisions.

As explained in this thesis, both species were introduced (one deliberately and the other incidentally) into the lake after it formed by damming part of the upper course of the Zambezi river in 1958. In addition to applied aspects, the fish intro-ductions in Lake Kariba present an opportunity to learn more about the reaction of an ecosystem in the case of a planned introduction into a vacant/open niche and an introduction into a system already occupied by a congeneric that shares an ecological niche. Therefore, Lake Kariba was the focus of the current study.

The creation of Lake Kariba in 1958 transformed a portion of the Zambezi River from a riverine (lotic) to a lake (lacustrine) habitat. The lake environment improved the survival of the cichlids, which prefer still water and reduced the abundance of fish species that favour flowing water, restricting their distribution to the more riverine upper section of the lake (Begg 1974). The lake came with a deep water and pelagic habitat that the native fish species could not utilize. A survey of the lake by Coke (1968) showed that the native fish were restricted to the inshore area (< 15m deep). Hence, about 70% of the lake, being deeper than 17 m, was unutilized (Begg 1970). Introduction of a pelagic fish species was therefore considered necessary to convert the pelagic plankton production into fish that could be harvested to bring socio-economic benefits.

A pre-introduction study of the biology of two candidate clupeid planktivorous pelagic species from Lake Tanganyika, Stolothrissa tanganicae and L. miodon, was carried out by Matthes (1965-66) in order to select the best species. Limnothrissa

miodon was chosen because of its inshore breeding habits, and it was introduced into

Lake Kariba in 1967 and 1968. Fry were captured from near Mpulungu (Zambia) and, after accounting for mortalities, about 362 400 L. miodon fry were released in the Sinazongwe (Zambia) area of Lake Kariba (Bell-Cross & Bell-Cross 1971). Limnothrissa miodon was considered established by 1969 (Kenmuir 1971). The species was also introduced in a natural lake, Lake Kivu (DR Congo / Rwanda; 1958-1960) and the Itezhi-Tezhi Reservoir (Zambia; 1992) and became established. Like in Lake Kariba, these introductions were deliberate and intended to fill a vacant pelagic

(5)

142

niche, and to increase fish production. Escapees of L. miodon from Lake Kariba reached and populated the next Zambezi dam located approximately 250 km down-stream of Kariba, Lake Cahora Bassa (Mozambique; 1976-1982). Cahora Bassa was formed in 1975 but L. miodon was not yet found in that year (Jackson & Rogers 1976). High densities were observed in the first hydro-acoustic survey in 1983 (Lindem 1983) but before 1994 no commercial fishing took place (Tweddle 2010). In all these waterbodies there is a fishery for L. miodon.

Unlike L. miodon, the introduction of O. niloticus was incidental and unplanned presumed from escapees from fish farms on the shores of the Lake Kariba (Chifamba 1998). The species was first noticed in the Experimental Lakeside Gill-netting catches in 1993. Between 1993 and 2006, O. niloticus spread from areas close to the fish farming unit to the whole of the Sanyati basin and beyond (Chifamba 2006; Zengeya & Marshall 2008). Aquaculture is not the only reason O. niloticus was introduced into some parts of the country. A reservoir in Zimbabwe which was stocked O. niloticus in 1992 by recreational anglers, resulted in its introduction in downstream Lake Chicamba, Mozambique in 1996 (Weyl 2008). This introduction of O. niloticus and that of L.

miodon into Lake Cahora Bassa, demonstrate that once introduced in a river system,

fish can spread to the rest of the river basin.

The suitability of the new environment to O. niloticus and L. miodon was evaluated through the study of their growth. Fish size and trends in the catches of O.

niloticus, since its introduction, were compared to those of O. mortimeri. In addition,

the study examined various aspects of competitive advantage of the exotic O.

niloticus over the displaced native O. mortimeri, a species endemic to the middle

Zambezi River. The aspects compared were growth, diet, aggression, size at maturity and reproductive effort.

In summary, in this thesis I explore the following questions (Figure 8.1):

1. How suitable is the new ecosystem for the two species introduced in Lake Kariba, and is their growth similar to that in their native habitat (Chapters 3 and 7)? 2. Did the introductions cause a competitive displacement of native riverine fish

species or did the new species fill an open niche (Chapters 2 and 6)?

3. What are the key aspects of the biology of the two introduced species, with consequences for their competitive interaction, in reproduction (Chapter 2), growth (Chapter 3), diet (Chapter 4) and aggression level (Chapter 5)? 4. Is the current fishing intensity on the introduced species in Lake Kariba

pro-ducing sustainable economic benefits or is the lake overfished and negatively affected by the changing climate (Chapters 6 and 7)?

5. What are the general lessons that can be drawn from this system for the management of river systems and new lakes and dams in Zimbabwe?

(6)

143

8

Figure 8.1 Schematic overview of the main questions addressed in this thesis, focussing on the introduction of Limnothrissa miodon and Oreochromis niloticus, the spatial and depth segregation of the inshore and pelagic fishery, and some factors affecting the catches of the pelagic fishery of Lake Kariba.

8.1 How suitable is the new ecosystem for the two introduced species?

Trends in the catches of the two introduced fish species O. niloticus (Chapter 2) and

L. miodon (Chapter 6) in Lake Kariba indicate that both have established populations

that support fisheries. To become established, they adapted to unique environmental conditions in the new ecosystem. Life-history traits of these introduced species are assumed to reflect the suitability of their new environment, with small size and slow growth indicating poor conditions compared to other waterbodies where the fish occur naturally or were introduced.

Growth parameters of O. niloticus, estimated for the first time in Lake Kariba (Chapter 3), provide insight into the suitability of the new ecosystem. The growth rate k (year-1_{) of O. niloticus in Lake Kariba was (mean ± SE): 0.25 ± 0.14, 0.29 ± 0.25 and}

(7)

144

0.29 ± 0.32 in 1997, 2003 and 2010, respectively. The asymptotic size (L∞) was: 32.4 ±

3.4, 44.6 ± 0.25 and 37.8 ± 0.32 cm standard length (SL), and the maximum age was: 10, 6 and 8 years in 1997, 2003 and 2010, respectively. Compared to other water bodies, the maximum size (converted to approximate (≈) SL for other studies where necessary), growth rate and maximum age from Lake Kariba are typical (Table 8.1). Estimated maximum size (L∞ = 21.3 to 53.9 cm) and growth rate (k = 0.07 to 0.56) of O.

niloticus varies among the water bodies in Table 8.1.

Oreochromis niloticus grows to a relatively large size in Lake Kariba, though the

largest asymptotic size (L∞ = 59.3 cm) was found in Lake Victoria (Getabu 1992).

The largest specimen, caught by the gill net fishery in Lake Kariba, is smaller (48 cm) than the largest from Lake Turkana (≈ 53.4 cm) (Lowe-McConnell 1958). Compared to other waterbodies, growth is slow in Lake Kariba and similar only to one value from Lake Victoria (k = 0.25) but higher than in the Lagoon Coatetelco in Mexico (k = 0.07) (Getabu 1992; Gómez-Márquez 1998). Therefore, O. niloticus in Lake Kariba grows slowly to a large size, a strategy also reflected in the size at first maturity (Chapter 2).

The size at first maturity in Lake Kariba was 17.91 and 18.28 cm in the data sets from 1993 – 2002 and 2003 – 2012, respectively, and the age at maturity was 2 years (Chapter 2). Size at first maturity in Lake Kariba lies between the values from other lakes (9.8 – 29 cm) (Table 8.2). Lake Victoria has the highest size at maturity, ≈ 26 and ≈ 29 cm for females and males, respectively, even though lower values were recorded in more recent years (Njiru et al. 2006b, 2018; Yongo et al. 2018). Com-pared to Lake Kariba, much smaller sizes at maturity were found in two crater lakes, Lake Nyamusingiri and Lake Kyasanduka in western Uganda (9.8 – 12.3 cm) and in eight reservoirs of varying sizes in Côte ďIvoire (9.1 – 14.7 cm) (Bwanika et al. 2004; Duponchelle & Panfili 1998). The smallest mature specimens of O. niloticus were 8.0 cm in Lake Kariba and ≈ 6.6 cm in Mississippi coastal catchments (USA) (Peterson et al. 2004). Furthermore, the age at maturity varies from between 5.6 and 10 months in the reservoirs in Côte ďIvoire, 2 years in Lake Kariba, and 2.6 years in Tabaru River & Yonaguni-jima Island in Japan (Duponchelle & Panfili 1998; Ishikawa 2013). Using studies where the size and age at maturity is given; it appears that early maturity occurs at a small size. Therefore, O. niloticus in Lake Kariba grows slowly to a large size, maturing later than some of the other populations, and at a relatively large size. This indicates that the environmental conditions in Lake Kariba are favourable, though not the best.

The level of primary production, type of food available and the size of the waterbody may explain the variation in growth and maturity among waterbodies. Primary produc-tion sets limits to energy in an ecosystem hence the amount available to the fish for growth. Primary production in terms of carbon (C) fixation is 0.1 to 1.7 g C2_h-1_{in Lake}

(8)

145

8

Table 8.1 Growth parameters L∞ (cm standard length) and k (year‐1_{) of Oreochromis} niloticus in African lakes, a lagoon in Mexico and Japanese rivers, obtained from the von

Bertalanffy growth model. Maximum age Amax (year), sex, aging method (LF = length

frequency analysis) and year of analysis (YoA) are also given. Values marked by * have been converted from total length to standard length using the equation: standard length = 0.838 total length – 0.2408 (R2_{= 0.99; n = 25) from Chapter 2.}

Country Lake/River L∞ k Amax Sex Method Reference / YoA Zimbabwe Kariba (i) 32.4 0.25 10 ♀♂ scale this thesis / 1997 Zimbabwe Kariba (i) 44.6 0.29 6 ♀♂ scale this thesis / 2003 Zimbabwe Kariba (i) 37.8 0.29 8 ♀♂ scale this thesis / 2010

Egypt High Dam 36.9 0.55 7 ♀ scale Yamaguchi et al. (1990)

Egypt High Dam 42.8 0.38 7 ♂ scale Yamaguchi et al. (1990)

Ethiopia Tana 36.3 0.5 ♀ otolith Wudneh (1998)

Ethiopia Tana 34.7 0.5 ♀ otolith Wudneh (1998)

Uganda Nabugabo (i) 32.7 0.52 7 ♀ otolith Bwanika et al. (2007) Uganda Nabugabo (i) 39.9 0.45 8 ♀ otolith Bwanika et al. (2007)

Uganda Albert 39.0 0.50 10 ♀♂ LF Moreau et al. (1986)

Kenya Victoria (i) 53.9* 0.25 12 ♀♂ LF Getabu (1992)

/ 1985‐1986 Kenya Victoria (i) 52.6* 0.35 8.6 ♀♂ LF Njiru et al. (2008a)

/ 2004‐2006 Kenya Victoria (i) 38.5* 0.69 ♀♂ LF Yongo & Outa (2016)

/ 2014‐2015

Kenya Turkana 21.3* 0.44 ♀♂ LF Moreau et al. (1995)

Ethiopia Awassa 21.8 0.57 5.1 ♀♂ otolith Admassu & Casselman (2000) Japan Tabaru R/ Yonaguni‐ jima Isl (i) 32.6 0.37 8 ♀ otolith Ishikawa et al. (2013) Japan Tabaru R/ Yonaguni‐ jima Isl (i) 28.2 0.42 13 ♀ otolith Ishikawa et al. (2013) Mexico Lagoon Coatetelco (i) 29.19 0.07 ♀♂ scale Gómez‐Márquez (1998) (i) = O. niloticus was introduced

(9)

146

Table 8.2 Standard length (SL), age at first maturity (Am) and sex of Oreochromis niloticus

in Lake Kariba and some other waterbodies in Africa, USA, Mexico and Japan. Values marked by * were originally in total length and were converted using the equation: standard length = 0.838 total length – 0.2408 (R2_{= 0.99; n = 25) from Chapter 2.}

Country Lake/River SL (cm) Am (y) Sex Reference

Zimbabwe Kariba (i) 17.91 2 ♀♂ This thesis; 1993‐2002

Zimbabwe Kariba (i) 18.28 2 ♀♂ This thesis; 2003 ‐2012

Kenya Turkana 22.4* ♀ Stewart (1988)

Kenya Turkana 24.1* ♂ Stewart (1988)

Ethiopia Tana 18.1 ♀ Wudneh (1998)

Ethopia Tana 20.7 ♂ Wudneh (1998)

Kenya Victoria (i) 26* ♀ Njiru et al. (2006b)

Kenya Victoria (i) 29* ♂ Njiru et al. (2006b)

Kenya Victoria (i) 19.9* ♀ Njiru et al. (2008a)

Kenya Victoria (i) 20.7* ♂ Njiru et al. (2008a)

Kenya Victoria (i) 21.5* ♀ Yongo et al. (2018)

Kenya Victoria (i) 25.7* ♂ Yongo et al. (2018)

Uganda Albert 22.2* 2 ♀♂ Lowe‐McConnell (1958)

Uganda George 22.2* 2 ♀♂ Lowe‐McConnell (1958)

Uganda George 16.9* ♀♂ Gwahaba (1973)

Uganda Nyamusingiri & Kyasanduka

9.8 ‐12.3* ♀♂ Bwanika et al. (2004)

Sudan Jebel Aulia Dam 23 ♀♂ Babiker (1984)

Côte ď’Ivoire Small Reservoirs (i) 9.1 ‐ 14.7 0.5‐0.8 ♀ Duponchelle & Panfili (1998) USA Mississippi coastal (i) 9.23* ♀♂ Peterson et al. (2004) Mexico Emiliano Zapata (i) 12.4* 1 ♀ Peña‐Mendoza et al. (2005) Mexico Emiliano Zapata (i) 12.5* 1 ♂ Peña‐Mendoza et al. (2005) Japan Tabaru R/ Yonaguni‐ jima Isl (i) 19.45 2.5 ♂ Ishikawa et al. (2013) Japan Tabaru R/ Yonaguni‐ jima Isl (i) 17.44 2.6 ♀ Ishikawa et al. (2013) (i) = O. niloticus was introduced

(10)

147

8

difference may explain the comparatively lower values of growth, maximum size and maturation of O. niloticus in Lake Kariba compared to Lake Victoria. In addition, differences in the types of food consumed by O. niloticus in locality may also explain some of the differences in growth. The diet of O. niloticus in Kariba consists of benthic pennate diatoms, which contribute 83% to stomach content by volume (Chapter 4). Most stomachs contain detritus (ca 80%) and sand (ca 70%). This diet may be poorer than the omnivorous diet, including a substantial proportion of increasingly available chironomid larvae, to which O. niloticus in Lake Victoria switched during approxi-mately 1985 – 1995 (Getabu 1994; Balirwa 1998; Wanink 1998). For example, O.

niloticus in Lake Nabugabo with a relatively energy-rich omnivorous diet, grew

faster than in Lake Wamala, where like in Lake Kariba, they fed on phytoplankton (Bwanika et al. 2007). Oreochromis niloticus of over 4 years of age from Lake Nabugabo were, on average, 10 cm larger than those from Lake Wamala. Lastly, the large size of O. niloticus in Lake Kariba supports the observation that small water-bodies tend to have small fish that mature early compared to larger waterwater-bodies (Lowe-McConnell 1958; Duponchelle & Panfili 1998). This may be attributed to more stable conditions in a large lake than a smaller one, in terms of seasonal and annual variation in the amount of water and its physical and chemical properties. Growth and maturation of O. niloticus are therefore constrained by the existing environmental conditions.

The general trend is that Oreochromis niloticus in impoverished conditions miniaturizes by maturing early at a small size and reaching a small maximum size. Therefore, the life-history strategy of O. niloticus switches between one of growth and one of maximizing reproduction, depending on existing environmental con-ditions. Judging from the growth parameters and the size and age at maturity, O.

niloticus has adopted a growth strategy in Lake Kariba in response to favorable

conditions. This knowledge is important when introducing this fish in a new reservoir, because its growth rate and its maximum size would vary, depending on the new environmental conditions.

Growth rate (k), asymptotic size (L∞) and age at maturity of Limnothrissa miodon

(Chapter 7) also provide insights on the suitability of Lake Kariba to this introduced fish. Estimates of L∞ (± SE) from the von Bertalanffy model were 12.4 ± 0.9 and 25.2

± 7.3 cm TL in 1993 – 1994 and 1996, respectively. The estimates for the 2013 data (L∞ = 51 ± 43.2 cm TL; k = 0.01 ± 0.01) were unrealistic, had large standard errors,

and could not be used for comparisons. The Gompertz model fitted the data better and provides a more realistic magnitude of the asymptotic length for 2013. Estimates of

L∞ (± SE) from this model were 18.0 ± 1.8, 9.6 ± 0.2 and 15.2 ± 1.4 cm TL in 1993 –

1994, 1996 and 2012 – 2013, respectively. The estimates of L∞ from the Gompertz

model are generally lower than those from the von Bertalanffy model. The Gompertz and the von Bertalanffy growth models are dissimilar in such a way that the former

(11)

148

provides the exponential decrease in the relative growth rate, and the latter the relative growth parameter k. Since the von Bertalanffy growth model was used in all the other studies in Table 8.1, I shall use it for comparisons.

Growth rate (k) of L. miodon varies within a lake, among years, and between lakes. However, the variation within the two reservoirs, both for k and L∞, is much larger

than that in the two natural lakes. A large part of the variation in Lake Kariba can probably be attributed to the use of two different methods for growth assessment: length frequency analysis and otolith reading. However, the lack of sufficient growth estimates from otolith reading in the other two lakes prevents the drawing of sound conclusions on the relative suitability of Lake Kariba for L. miodon in terms of growth rate. Except for 1993 – 1994, estimates for Lake Kariba have a lower asymptotic size compared to Lakes Tanganyika and Kivu (Table 8.3). The two reservoirs, Lake Cahora Bassa and Lake Kariba, have the two smallest asymptotic sizes recorded, suggesting that conditions in these reservoirs differ from those in the natural lakes. Ndebele-Murisa et al. (2010) reported differences in plankton composition and a lower range of phytoplankton production in Lake Kariba (0.10 – 1.76 g C m2_h-1_{) than}

in Lakes Tanganyika (0.16 – 4.30 g C m2 _h-1_{) and Kivu (0.85 – 2.20 g C m}2_h-1_{), which}

may explain the differences in the growth and maximum size found between these lakes.

Differences in the Lake Kariba estimates of growth parameters from length frequency analysis and otolith reading may reflect some underlying differences in the fish population. Absence of large fish in the samples used for the length frequency analysis has led to low estimates of L∞, and to the conclusion that the L.

miodon population in Lake Kariba was stunted compared to the population in Lake

Tanganyika (Cochrane 1984; Marshall 1987a). The lack of large fish in the samples may result from most fish dying at a small size, some ecological bottleneck, or a failure to adequately sample the littoral area where the very large fish live. Sampling at a depth between 20 and 25 m, close to the shoreline, yielded some large specimens (Chapter 6). In agreement with my findings, the largest sizes (110 – 150 mm) were caught using gill nets near the lake margins of Lake Kivu (de Iongh et al.1983). There seems to be a gap in our knowledge with respect to the spatial distribution of the fish with age, or the diet switch required at a particular size, which is evident from the cannibalistic tendency reported for large L. miodon in the littoral areas in Lake Kariba (Chapter 6) and Lake Kivu (de Iongh et al.1983). An-other different aspect in Lake Kariba is the intensive fishing pressure on the medium sized fish, which may reduce the numbers that reach a large size. These are potential areas of investigation in Lake Kariba.

Overfishing may have caused the reduction in the size at maturity of L. miodon in Lake Kariba (Chapters 6 and 7) hence size at maturity may not be regulated by environmental conditions only. Size at first maturity of L. miodon in Kariba in 1970 –1972 was close to those in Lake Tanganyika and Lake Kivu, regulated

(12)

149

8

Table 8.3 Growth parameters L∞ (cm total length) and k (month‐1_{) of Limnothrissa miodon}

in Lakes Kariba, Tanganyika, Kivu and Cahora Bassa, obtained from the von Bertalanffy growth model. Aging method (LF = length frequency analysis) and year of analysis (YoA) are also given. Total length, where necessary, estimated from *standard length or **fork length.

Country Lake YoA L∞ k Meth Reference

Zimbabwe Kariba 1993 25.2±7.3 0.04±0.02 otolith This thesis, chapter 7 Zimbabwe Kariba 1996 12.4±0.9 0.10±0.01 otolith This thesis, chapter 7

Zimbabwe Kariba 1988 13.5 0.079 otolith Chifamba (1992)

Zimbabwe Kariba 1976‐77 8.1 0.145 LF Cochrane (1984)

Zimbabwe Kariba 1981‐83 7.42 0.254 LF Marshall (1987a)

Zaire Tanganyika 1987‐89 18.2* 0.096 LF Mulimbwa &

Shirakihara (1994) Zambia Tanganyika 1990 20.6* 0.072 otolith Kimura (1995)

Burundi Tanganyika 1980 16.1 0.096 LF Moreau et al. (1991)

Rwanda Kivu 1980 16.3 ** 0.1 LF Spliethoff et al. (1983)

Rwanda Kivu 1980 16.7 0.108 LF Mannini (1990)

Mozambique Cahora Bassa 1993 7 0.45 LF Gliwicz (1984)

environmental conditions only. Size at first maturity of L. miodon in Kariba in 1970 – 1972 was close to those in Lake Tanganyika and Lake Kivu, though in more recent this was not the case (Chapter 7). In 2013, age and size at maturity were 8.02 and 7.90 months, and 3.43 and 3.63 cm for females and males, respectively. In 1970 – 1972, the size at maturity in Lake Kariba was 5.2 to 5.6 cm for females and 7.1 to 7.3 cm for males (Woodward 1974). Over-exploitation may have reduced the size at maturity, as observed in some fisheries (Wanink 1998; Hutchings 2005; Olsen et al. 2005; Mollet et al. 2007; Morita & Fukuwaka 2007). Early estimates of maturity in Lake Kariba were comparable to those in Lake Tanganyika (7.5 and 6.4 cm for females and males, respectively) and Lake Kivu (6.24 and 6.06 cm for females and males, respectively) (Ellis 1971; Spliethoff et al. 1983). The more recent estimate is closer to that obtained from Lake Cahora Bassa in 2003/4 (≈ 4.3 cm for females, reported as 3.9 cm FL) (Mafuca et al. 2011). Similarity between these two reservoirs in age at maturity, also observed in growth rate and L∞, may reflect similar

(13)

150

Tanganyika and Lake Kivu. Conditions in Lake Kariba in the early years without the fishery were more favourable and comparable to other lakes.

Both introduced species, O. niloticus and L. miodon, have displayed flexible life-history traits in Lake Kariba. The calculated growth parameters indicate favourable, though not optimal conditions for both species. However, their establishment impacted the native species differently, depending on whether the introduction was meant to fill a vacant or an already occupied niche. This is the next question I shall consider.

8.2 Did the introductions cause a competitive displacement of native riverine fish species or did the new species fill an open niche?

Catches of Oreochromis mortimeri, a native and endemic fish species of the middle Zambezi River, in the newly created Lake Kariba were initially high (1960s to 1980s) (Chapter 2). Subsequently, catches declined as nutrients from inundated vegetation were lost, and the ecosystem matured and stabilized. In the 1990s, O. niloticus, introduced into the lake through escapees from an aquaculture farm, displaced O.

mortimeri in the Sanyati basin (Figure 1.1). Oreochromis mortimeri almost

dis-appeared from the catches in less than a decade after the first appearance of O.

niloticus in 1993. I found a significant negative relationship in the catches of these

two fish species from 1993 to 2012 (Chapter 2). Between 1993 and 2003 – 2005, the ratio of O. mortimeri to O. niloticus changed from 1:0.1 to 1:27 (Chifamba 2006). Thus, the unplanned introduction of O. niloticus into Lake Kariba resulted in dis-placement of the native O. mortimeri, and the introduced species became dominant in the catches.

Oreochomis niloticus and O. mortimeri share an ecological niche and as such,

have similar diets (Chapter 4) and reproductive strategies (Bell-Cross & Minshull 1988; Trewavas 1983). Reproduction involves nest building in a breeding arena by the males, to attract females to lay eggs which the male then fertilizes and the female subsequently broods in the mouth. Competition between the two species could be predicted, and perhaps also the displacement of the native species. There are ex-amples of displacement of native species by O. niloticus in Africa and other parts of the world. Oreochromis niloticus displaced the native planktivores Oreochromis

esculentus and Oreochromis variabilis in Lake Victoria (Balirwa 1992) and three

native cyprinids in Lake Luhondo, Rwanda (De Vos et al. 1990). In Lake Chivero (Zimbabwe), O. niloticus replaced another introduced planktivorous species,

Oreo-chromis macrochir, and accounts for about 95% of the catches from that lake

thereafter (Marshall 1999; Tiki 2011). In Lake Chivero, both O. niloticus and O.

macrochir fed mostly on blue-green algae (> 50%) and their diets, in all size classes,

overlapped almost completely (Zengeya & Marshall 2008). The dietary overlap coefficient (Pianka 1973) for O. niloticus and O. macrochir in Lake Chivero was

(14)

151

8

0.98, comparable to the value of 0.95 that was found for O. niloticus and O.

morti-meri in Lake Kariba (Chapter 4).

Outside Africa, the introduction of O. niloticus in Gargalheiras Reservoir in Brazil, caused the decline of another introduced species, Plagioscion

squamo-sissimus and two indigenous fishes, Prochilodus brevis and Hoplias malabaricus

(Attayde et al. 2011). At the same time, the proportion of introduced species in-creased from 74% to 96% of the total fish catch. Like O. niloticus, Prochilodus

brevis feeds mostly on sediments and microalgae that include Chlorophyceae,

Bacillariophyceae and Cyanophyceae (Da Silva et al. 2010). Generally, O. niloticus tends to displace species with which it shares a niche. Hence, to avoid displacement of native species, introductions should be into a vacant niche as in the case of the introduction of L. miodon into Lakes Kariba and Kivu.

Limnothrissa miodon inhabits the once vacant pelagic niche of Lake Kariba

(Chapter 6). Juveniles are found in the marginal area only, while adults use both the pelagic and shallow areas where breeding may occur. The presence of L. miodon had a positive impact on tigerfish (Hydrocynus vittatus) by providing an alternative prey to this native predator. Hydrocynus vittatus showed a dietary switch to L. miodon and began to inhabit the open water in pursuit of its prey (Cochrane 1976; Marshall 1987b, 1991). From April 1969 to March 1970, only 1.5% of the stomach contents of H. vittatus consisted of sardines (L. miodon) whereas from April 1970 to March 1971, the amount had risen to 41.4% and remained high (Kenmuir 1973; Mhlanga 2003). Hydrocynus vittatus together with small amounts of some inshore fish species, are caught in the deep water as bycatch in the L. miodon fishery (Chapter 6).

Limnothrissa miodon is the dominant fish in the pelagic area.

Whilst the pelagic area was a vacant niche when L. miodon was introduced into Lake Kariba, the inshore area that became subsequently occupied by the juveniles, was not. The impact of this introduction on the inshore fish species that compete with the juvenile and adult sardines in the inshore area, is unknown. Inshore inter-actions of the native and the introduced fish species have never been investigated. When considering future fish introductions, all life stages of the exotic species must be included both in the pre- and the post-introduction stage, for impact assessment.

A possible negative ecological impact of the introduction of L. miodon was on a zooplanktivore, Brycinus lateralis, a species from the Upper Zambezi River that was already present in Lake Kariba when L. miodon was introduced and occupying the inshore zone where juvenile L. miodon stay. Brycynus lateralis was believed to be capable of utilizing the pelagic area. In support of this view, early catches of L.

miodon from the open water contained 20.5% B. lateralis (Woodward 1974).

Expansion of B. lateralis into the pelagic area may have been prohibited by competition from L. miodon (Marshall 1991). Since both species are not native to

(15)

152

the Middle Zambezi River, and B. lateralis is still found in Lake Kariba, the introduction of L. miodon may not have had serious ecological consequences.

The introduction of Limnothrissa miodon and Oreochromis niloticus into lake Kariba demonstrates the positive consequences of a planned introduction into an open niche, and the negative consequences when the introduction is into a filled niche. An introduction into an open niche has little ecological impact and can be beneficial through an increase in fish production. In contrast, the (unintended) introduction of O. niloticus led to the disappearance of O. mortimeri, a congeneric species with equal productivity. The negative impact of O. niloticus on O. mortimeri may have resulted from competitive advantage of the introduced species, that I will consider in my next question.

8.3 What are the key aspects of the biology of the two introduced species, with consequences for their competitive interaction, in growth, diet and aggression level?

High reproductive effort (Chapter 2), growth rate (Chapter 3 – Chifamba & Videler 2014), diet overlap (Chapter 4) and aggression (Chapter 5 – Chifamba & Mauru 2017) could give competitive advantage to Oreochromis niloticus over the native O. mortimeri, that enabled the invader to displace O. mortimeri in Lake Kariba.

In Chapter 4, the diets of O. niloticus and O. mortimeri are compared and found to be similar. Schoener’s and Pianka’s overlap indices are 0.75 and 0.95, respectively (Pianka 1973; Schoener 1974). The overlap in the diet between the two fish species is biologically significant according to the 0.60 limit suggested by Galat & Vucinich (1983). In addition, there was no significant difference in the protein content of the ingested material in the stomach, as well as in the relative digestion efficiency of the proteins, a further indication of similarity in their utilization of the food resource. As the digestion efficiencies of O. mortimeri and O. niloticus did not differ, this factor may not have given O. niloticus competitive advantage. In contrast to this finding,

O. niloticus in Lake Chivero has been reported to have competitive advantage from

a higher digestion efficiency compared to O. macrochir, the species it displaced from that lake (Marufu & Chifamba 2013). This suggests that the diet overlap of O.

niloticus and O. mortimeri in Lake Kariba was a comparable potential competition

parameter, that would favour the most fit.

Even though it has been suggested that a diet of detritus may not be limited in the environment, its quality has been shown to vary in space. Bowen (1979) found that juvenile Sarotherodon mossambicus that feed in the shallow part of Lake Sibaya (South Africa) have access to detritus with a higher protein level than the adults, which inhabit and feed in deep water. Such a heterogeneity in the distribution of food resources may lead to competition between O. niloticus and O. mortimeri. Under

(16)

153

8

conditions of heterogenous distribution and inadequate resources, access may be enhanced by the ability to win a contest and dominate.

Laboratory experiments showed that O. niloticus is less aggressive than O. mortimeri and that a larger fish is more aggressive than a smaller one (Chapter 5 – Chifamba & Mauru 2017). Oreochromis mortimeri attacked O. niloticus first in most of the encounters, irrespective of whether it was the bigger or the smaller of the pair. In those encounters where O. niloticus was smaller, O. niloticus made the first bite two times (8.3%). When bigger, O. niloticus made the first bite eight times (40.0%). Over a 30-min encounter, O. mortimeri was dominant and delivered significantly more bites (7.79 േ 2.31 bites) than O. niloticus (4.53 േ 1.53 bites). Oreochromis niloticus tended to be less aggressive, it attacked first only when it was much bigger than the opponent, while O. mortimeri attacked even when it was smaller. Considering aggression alone, it was unexpected to find higher aggression in the displaced fish. However, considering the association of large body size with higher aggression, O. niloticus would be at an advantage, because it grows faster than O. mortimeri (Chapter 3 – Chifamba & Videler 2014). At any particular age, therefore, O. niloticus would be generally larger than O. mortimeri, hence would have a size advantage in a contest. Mean length at age of O. niloticus was higher than that of O. mortimeri during 1997 – 2000 (Chapter 3 – Chifamba & Videler 2014), the period in which species replace-ment occurred. Oreochromis niloticus grew to a larger size (1997 – 2000: 32 ± 3.4 cm; 2003 – 2005: 44 ± 17.7 cm) compared to O. mortimeri (1997 – 2000: 30.2 ± 4.2 cm; 2003 – 2005: 3.6 ± 14.6 cm). Being larger at any particular size during the displace-ment years, gave O. niloticus competitive advantage over O. mortimeri.

Dominance status has a competitive advantage because it is associated with growth rate, reproductive advantage and metabolic rate. Ten out of twelve dominant steel-head trout (Salmo gairdneri) grew faster than their paired subordinates, given the same amount of food (Abbott & Dill 1989). Tiira et al. (2009) found that dominance status affected growth in brown trout, Salmo trutta populations, with individuals having the lowest ranks growing less compared to those with a higher rank. There is also a significant association between some agonistic profiles and metabolism (Alva-renga & Volpato 1995). Oreochromis mortimeri has competitive advantage through a high aggression level, which is counteracted by the size advantage of O. niloticus in a contest for resources.

Competition between O. niloticus and O. mortimeri for other resources, such as nesting and nursery space, is inevitable because of the similarity in their reproductive behaviour. Breeding males in African cichlids, including O. niloticus and O. mortimeri, make and fiercely defend their nest in an arena. The best nests are optimally posi-tioned for visits by females, and they are occupied by the dominant males (Philippart & Ruwet 1982). Compared to their subordinates, dominant male Oreochromis

(17)

154

niloticus have the highest gonadosomatic index and higher levels of gonadotropin

hormones that trigger spermatogenesis, whereas the subordinates show reduced gene expression of key factors for steroid production (Pfennig et al. 2012). Reproductive success is higher in dominant fish because they occupy the best nesting sites, which attracts potential mates (Philippart & Ruwet 1982; Seppänen et al. 2009). Hence, the larger O. niloticus males (L∞ = 32.2 and 44.6 cm SL in 1997 – 2000 and 2003 – 2005,

respectively) would be at a competitive advantage for mating and reproduction in comparison with the smaller O. mortimeri (L∞ = 30.2 and 36.8 cm SL in 1997 – 2000

and 2003 – 2005, respectively) (Chapter 3 – Chifamba & Videler 2014).

Oreochromis niloticus has a higher reproductive effort than O. mortimeri, lending

it a comparatively greater capacity to increase its population (Chapter 2). The pro-portion of mature fish with ripe gonads in the catch is always higher for O. niloticus compared to O. mortimeri. In addition, O. niloticus has a slightly lower length at maturity (17.2 cm) than O. mortimeri (19.6 cm). Furthermore, the breeding of O. mortimeri seems to be associated with rainfall more than that of O. niloticus, which is mostly dependent on temperature. This implies that the breeding season of O. niloticus is comparably longer than that of O. mortimeri, which seems to depend more on rain-fall as a signal for gonad maturation. Higher reproductive effort of O. niloticus could have also given this species competitive advantage.

One important factor that was not considered in this study because of prohibitive costs, is hybridization. Congeneric species such as Oreochromis niloticus and O.

mortimeri have a high chance of hybridization, and hybrids within this genus have

been observed in nature (Gregg et al. 1998; D’Amato et al. 2007). In the Limpopo River, introduced O. niloticus hybridized with the native O. mossambicus (D’Amato

et al. 2007). Gregg et al. (1998) reported that O. mortimeri in Lake Kariba has O. macrochir genes, indicating hybridization following the introduction of the latter

species into Lake Kariba. The authors argue that there were no recognizable O.

macrochir in the lake because their genes may have been swamped by those of the

more numerous O. mortimeri. Due to back crosses or introgressive hybridization, hybrids of O. niloticus with O. macrochir in Lake Itasy (Madagascar) and Lac Ihema (Rwanda), and with Oreochromis variabilis and possibly Oreochromis esculentus in Lake Victoria, resemble O. niloticus (Trewavas 1983; Kudhongania & Chitamwebwa 1995; Lowe-McConnell 2000). Progeny of introgressive hybridization will have phenotypic resemblance of one of the parents. It is possible that O. mortimeri has formed hybrids with O. niloticus in Lake Kariba, but these are not visually identi-fiable. Suspected hybrids, with intermediate characteristics between O. niloticus and

O. mortimeri were caught during sampling. This suggests that the introduction of O. niloticus may have caused a decline in genetic diversity, through hybridization and

(18)

155

8

competition. A genetic study to determine hybridization is essential to establish if this has indeed occurred.

Each of the factors discussed here could be associated with a small competitive advantage for Oreochromis niloticus. These advantages act simultaneously, and their effect would be cumulative and synergetic over generations. Hence, combined effects of all the many traits of O. niloticus give it a competitive advantage over the native

O. mortimeri. The next question is whether the introduction was beneficial to the

capture fishery.

8.4 Did it improve the fishery or just replaced native species that could have been fished otherwise?

Since Oreochromis niloticus appeared in the Lakeside experimental gill net fishing programme catches in 1993, the catches of this species increased. At the same time, those of O. mortimeri declined until the species disappeared in 2006, except for an occasional catch (Chapter 2). About 50% of the variation in the catches of O. mortimeri could be explained by the variation in O. niloticus catches. The latter rose to a peak in 2000, thereafter remaining stable with a mean catch of 2.8 kg per set. Combined catches of these two species show the replacement of the O. mortimeri by O. niloticus, with the trends in the catches strongly fluctuating between years but without a clear trend in the total catch during the period of displacement from 1992 to 2012 (Chapter 2). The total amount of fish harvested from Lake Kariba did not improve with the introduction of O. niloticus, an outcome that is also found elsewhere.

In Gargalheiras Reservoir, Brazil, introduction of O. niloticus did not improve the overall catches. Following the introduction of O. niloticus, the catch per unit effort of introduced Plagioscion squamosissimus (Heckel) and the indigenous species

Pro-chilodus brevis, (Steindachner) and Hoplias malabaricus (Bloch) declined, but the

overall catches did not (Attayde et al. 2011). In addition, the number of fisherman and their per capita income did not change. Therefore, in terms of increasing fish production the introduction of O. niloticus may have been unnecessary in this reservoir, as well as in Lake Kariba.

The introduction of Limnothrissa miodon in Lake Kariba, in contrast to that of

O. niloticus, resulted in a highly productive pelagic fishery, landing 24 000 tons at

its peak (Chapter 6). In terms of energy flow in the ecosystem, the new species must utilize a niche that is not already occupied by another fish caught in the fishery. Unlike O. niloticus, that feeds on the same food as O. mortimeri, Limnothrissa

miodon utilizes pelagic plankton, that inshore fish are unable to use, converting it

(19)

156

I conclude that the introduction of Oreochromis niloticus into Lake Kariba was ecologically unnecessary, and a mistake in terms of conservation of biodiversity, because it displaced the native endemic Oreochromis mortimeri.

8.5 Is the current fishing in Lake Kariba economically and ecologically sustainable?

Fishing offers social and economic benefits in the form of food, revenue, income and livelihood, thus contributes to food security. These benefits are derived from the fishing operation itself and the associated activities such as marketing of fish, and the manufacture of the fishing necessities (fishing nets, vessels, and equipment). Economic and social pressures drive exploitation, because the venture must be worthwhile for fishing to continue. Maximum economic benefits are derived when optimum fishing effort is applied, and the fish population can sustain itself. Environ-mental factors are key to the natural maintenance of the fished population through its effect on growth, reproduction and mortality. Hence, unfavourable conditions can hamper the ability of the population to sustain itself. A balance between factors that diminish the fish population (fishing and natural mortality) and those that increase the population (reproduction and growth) is needed for sustainable fishing. The sustainability of the fisheries of the introduced fish in Lake Kariba will be discussed in this section.

Economic impact of the L. miodon and O. niloticus fishery

The impact of introduced fish species, particularly those introduced for the purpose of improving fish production or for sport fishery, should be judged in an economic as well as an ecological sense. Ecological and economic impacts can be profound. A classic example is the disappearance of the native and diverse flock of haplochromine cichlids following the introduction of Nile perch, Lates niloticus in Lake Victoria (Witte et al. 1995). However, in social and economic terms the fishery for L. niloticus had significant positive outcomes, acknowledged and assessed by Reynolds & Greboval (1988) and Njiru et al. (2005). In Lake Victoria, the artisanal fishery was replaced by a multimillion-dollar export industry and the government derived benefit from licensing of fish processing and fishmeal production. Many new jobs were created in the processing and fisheries-related industries.

Similarly, the introduction of L. miodon in Lake Kariba brought into existence a semi-industrial fishery for this pelagic species, which strongly increased the total fish production. The fishery for L. miodon is more productive than the artisanal fishery and contributed about 95% of the total landings (16 387 tons) in 1996 in Zimbabwe (unpublished data from Lake Kariba Fisheries Research Institute, LKFRI). The trends in effort and catches in Lake Kariba are described in Chapter 6. Annual

(20)

157

8

commercial L. miodon catches landed in Zimbabwe reached a maximum of about 22 000 tons in 1990 but decreased to about 2 500 tons in 2011 (unpublished data from LKFRI), representing a huge loss in food production. At its peak, the Zambian and Zimbabwean L. miodon fishery at Lake Kariba generated about USD 37.7 million per year, against USD 20 million in 2013 (unpublished report of the Programme for the Implementation of a Regional Fisheries Strategy for the Eastern and Southern Africa - Indian Ocean region).

In contrast to the intended introduction of L. miodon, the introduction of O.

niloticus seems to have had no effect on the total landings, as it was almost directly

replacing the native O. mortimeri (Chapter 2). However, the fishery for the introduced

O. niloticus is important in many lakes. For example, in Lake Chivero (Zimbabwe)

the species contributes about 95% to the total catches (Tiki 2011). The yield of introduced O. niloticus by the artisanal fishery on Lake Chicamba (Mozambique) was 5.2 tons per month (Weyl 2008). An increase in mesh size may have occurred in 2003, to enable the fisherman to catch the larger O. niloticus. This suggestion is based on size selectivity of gill nets and the size distribution of the catches landed by the artisanal fishery (Chapter 2). The size of fish caught by the artisanal fishermen and the mesh size in use need to be monitored and managed to ensure the sustain-ability of the O. niloticus fishery in Lake Kariba.

In economic terms, the use of O. niloticus in Kariba for aquaculture was a huge success in terms of boosting fish production and economic activity. Lake Harvest Aquaculture produce about 9 000 tons of fish per year in Kariba (Zimbabwe) and 2 000 tons in Siavonga (Zambia). That production surpasses the gill-net fishery catches (about 5 000 tons per year, all species in Zimbabwe). When assessing the impact of introductions, the economic aspects should be taken into consideration.

Sustainability of the fisheries

The estimates of Maximum Sustainable Yield (MSY) and fishing effort were used to make inference on the sustainability of the fishery. Using the Schaefer and Fox models, MSY was estimated from combined catch and effort data from Zimbabwe and Zambia for L. miodon. MSY estimated for the period 1974 to 2011, was 23 185 (Schaefer) and 22 355 (Fox) tons per annum (Chapter 6). The effort (Fmsy) used to

catch the MSY was 10 867 and 11 226 boat nights for the Schaefer and Fox models, respectively. These estimates are lower than those made by Pitcher (1995) who used production from a baseline lake, where the species forms a successful fishery as a ‘predictor of ecology’. Pitcher’s estimate of the yield of L. miodon in Lake Kariba is 27 000 to 36 000 tons per year. The maximum fish landing of 30 000 tons of sardines from Lake Kariba is higher than my estimates of MSY, which indicates that over-fishing is the main factor that caused the decline in the catches of L. miodon.

(21)

158

Overfishing of L. miodon occurred in spite of the recommendation by the fisheries managers within the Zambia/Zimbabwe SADC Fisheries Project, to limit the number of sardine-fishing rigs/vessels to 230 and 270 for Zambia and Zimbabwe, respectively. The Protocol on Economic and Technical Cooperation stating how the two countries were going to manage the fishery, was signed in 1999. But both countries did not adhere to the effort limits. There were 460 rigs in Zimbabwe and 962 in Zambia in 2013, excluding an unknown number of illegal fishing vessels (Paulet 2014). Lake Kariba is a transboundary resource, bringing its own political complexity into the management of the fishery, which has to do with laws, policies and practices. Effort increases in Zimbabwe arose largely from attempts to address the race imbalance in the allocation of fishing permits which, before independence, were reserved for whites. One of the problems in Zambia was the ineffective management of the number of fishing vessels. Rises in fishing effort despite the knowledge of the limits to the L.

miodon fishery, highlights the importance of including social and economic factors

in the management of a natural resource.

The current management objectives are to reduce the fishing effort by limiting the number of fishing vessels and fishing nights, in order to fish within the limits of MSY. These objectives should broaden and include variation in MSY. MSY may not be the most appropriate management objective, because it varies annually according to environmental conditions and other factors such as fishing, which may affect life-history parameters and replacement rate of the population. Growth of the sardines and population size of the juveniles and adults need to be monitored. Fishing effort, measured as boat nights, also changes. Fishing vessel and gear improvement increase the fishing power of the fishing units (Chifamba 1995). More recent developments include mobile phones that allow the fishermen to communicate with each other more intensively, even during fishing. Effective fishing effort rises, hence a need to adjust for vessel improvement or perhaps, the use of catch quotas to regulate the fishery and prevent overfishing.

The size at maturity of L. miodon in Lake Kariba decreased from 5.2 – 5.6 cm and 7.1 – 7.3 cm in 1970, to 3.43 and 3.63 cm in 2013 for females and males, respectively (Chapter 7). A high adult mortality (such as in fishing) tends to reduce size and age at maturity. Reznick & Ghalambor (2005) proved this in a study of guppies (Poecilia

reticulata) from Trinidad, which live in either a high- or a low-predation environment,

and in experiments where mortality rates were manipulated in nature. The decrease in the size at maturity of L. miodon in Lake Kariba is consistent with fishery-induced evolutionary change. According to life-history, fitness in fish is optimized by an increase in reproductive effort and a decrease in the size and age at maturity, in response to high adult mortality (Stearns & Koella 1986). Fishery-induced evolution of fish can make it hard for the fishery to recover, even when the fishing effort is reduced.

(22)

159

8

Environmental effects on fish manifest on the reported negative relationship between air temperature and catches of L. miodon in Lake Kariba (Chapter 6). The mean maximum air temperature in Kariba rose from 29 to 32 0_{C between 1974 and 2008;}

a manifestation of global climate change. In the absence of continuous data on water temperature, it can be assumed that the water temperature rose together with air temperature (Chifamba 2000). One effect of temperature increase is the change in the phytoplankton, from a community dominated by Chlorophytes to one dominated by Cyanophytes (Sibanda 2003; Butterwick et al. 2005). Due to their toxicity and morphology, Cyanophytes are a poorer diet for the zooplankton on which L. miodon feeds, which may have led to the reduction of L. miodon catches (Magadza 2011). Catches of L. miodon are negatively correlated with mean maximum temperature in Kariba (Chifamba 2000).

Another effect of high temperature is an increased stability of the thermocline, as is the case in Lake Tanganyika. This reduces the incidence of circulation that brings nutrients to the euphotic zone which would have been lost to the hypolimnion (Verburg

et al. 2003). This cycling provides 95% of the phosphates and 97% of the silicon

used for algae production in Lake Tanganyika (Verschuren 2003). Consequently, the stability of the thermocline in Lake Tanganyika is estimated to have reduced primary production by about 20% and fish yield by 30% (O'Reilly et al. 2003). Stability of the thermocline may have reduced the primary production in Lake Kariba in a similar way and, in turn, the sardine production. New production estimates are needed to come with more current estimates of population parameters and fishable portion.

Changes in productivity caused by environmental variation could be the cause of the variation in growth of L. miodon that was found in Lake Kariba (Chapter 7). Growth in 2013 was slower than in 1993 and 1996, thus increasing the time needed to grow to a fishable size and thereby reducing the rate of biomass regeneration (Figure 8.2). Environmental conditions may determine the growth and size of the sardine population in Lake Kariba. This must be borne in mind when setting catch limits. Multiple regression analysis shows that fishing effort, lake level and temperature are correlated with catches and catch per unit effort of L. miodon (Chifamba 2000). The environment and fishing thus simultaneously affect fish abundance. Therefore, to manage the L. miodon fishery sustainably, there is a need to reduce the fishing effort, assess the productivity of the lake considering the environmental variables, and determine the desired fishing effort to achieve sustainability.

Sustainability of the fishery is also challenged in the case of Oreochromis niloticus and O. mortimeri, judging from the reduction in age groups in the catch. In 1997, the oldest O. niloticus and O. mortimeri were 10 and 13 years old, respectively, but in 2003 the maximum ages had reduced to 5 and 6 years, respectively (Chapter 3 – Chifamba & Videler 2014). In Lake Victoria, O. niloticus reduced its size at maturity

(23)

160

Figure 8.2 Effect of annual variation in the growth rate of Limnothrissa miodon in Lake

Kariba on the age at maturity, marked by arrows from the maturity line drawn from the size at maturity (3.43 cm) for females in 2013.

from 35 cm TL (early 90s) to 21 (males) and 22.7 (females) cm in 1998 – 2000, a sign of overexploitation (Njiru et al. 2008a). Fishing reduces the chances of the fish to survive many years, hence many exploited fish species have few age groups of young fish (Trippel 1995). Reduction in the number of age groups in the catch reduces sustainability and resilience of the fished population. Recruitment failure or other calamities are felt more strongly in a population with fewer age groups. Hence, there is a need to monitor the age of Oreochromis niloticus and O. mortimeri, and to manage their fisheries sustainably.

8.6 What are the general lessons that can be drawn from this system for the management of river systems and new lakes and dams in Zimbabwe?

The main lessons from this thesis are that:

1) dams create a new ecosystem that can be filled by exotics, and destroy the river ecosystem to which the native species are adapted;

2) reservoirs create opportunity for aquaculture, which can result in the use of exotic species that can escape into the river system, with attached costs and benefits;

Age (months) 0 2 4 6 8 To ta l l en gt h (c m ) 0 2 4 6 8 2013 1996 1993 1988 Size at maturity line

(24)

161

8

3) general conditions that promote the establishment of introduced species, could guide other introductions to minimize negative impacts; and

4) natural resources need to be utilized sustainably, considering environmental variation.

Dams create and destroy ecosystems

The construction of Lake Kariba created an ecosystem that allowed the cichlid species to thrive, and a pelagic niche that could be filled by introducing a pelagic fish. New dams can offer the same opportunities as Lake Kariba. In relatively shallow lakes, the open water area may be shallow enough to be utilized by the native fish species. Deep reservoirs such as the planned Batoka dam, upstream of Lake Kariba, will create a habitat in which pelagic species such as L. miodon are likely to thrive.

The Kariba dam is a barrier to fish migration that eliminated the connectivity of the Zambezi River almost completely. River-adapted fish were lost when the lake was formed. One species now absent in the catches from Lake Kariba, is the cata-dromous African longfin eel (Anguilla mossambica). This eel is endemic to the Malagasy area in the south western Indian Ocean and migrates into rivers on the eastern coast of Africa, Madagascar and other West Indian Ocean islands. The dam prevents migration of glass eels into the rivers and silver eels back to the ocean. Some fish species in Lake Kariba, such as Hydrocynus vittatus and Labeo altivelis, are potamodromous: they migrate upstream to breed during the rainy season. Con-struction of dams has to be kept to a minimum, and those not in use or heavily silted to be useful decommissioned, to allow free movement of fish and river continuity. Fish migration passages need to be built into every dam and weir to reduce the impact of these structures on fish migration and recruitment. Unfortunately, the loss of fish species does not raise much public concern because they are not as visible to most people as terrestrial animals. Visibility needs to be raised on the plight of fish in the face of human activities.

Dams create opportunity for aquaculture, which can result in exotic species introduction

Availability of water limits aquaculture in Zimbabwe, and water storage is critical for electricity generation, agriculture and drinking-water supply for cities. Hence, a strong driver to construct reservoirs of different sizes in this country without natural lakes. Aquaculture development is strongly encouraged and backed by government policy because of its potential role to increase food production and food security. The most favoured aquaculture species are exotics, which include Oreochromis

niloticus. Introduction of this species into other dams is likely to have negative

bio-diversity consequences, similarly to what occurred in Lakes Kariba and Chivero. In economic terms, the dilemma is that O. niloticus farming in Kariba is a huge success

(25)

162

with regard to boosting the fish production and the economy, thus fuelling the drive to repeat this success. The use of exotics in aquaculture is a global concern. About 30% of the fish introductions worldwide were for aquaculture, and only 17% for improvement of wild populations (Gozlan et al. 2010). Therefore, limiting fish intro-ductions will be difficult because of their benefits to humans. However, the economic drive needs to be balanced with the environmental impacts and economic consequen-ces of the introductions, that are often ignored.

The threatened biodiversity may require costly measures to protect, and costly measures also may be needed to reduce or remove introduced fish species. Mechanical removal of exotics is most effective in small water bodies. This is also done to limit the exotic populations in large areas, with varying success. For example, mechanical removal of Channel catfish (Ictalurus punctatus) and common carp (Cyprinus carpio) from the San Juan River (Utah, USA) has reduced the population, but cannot eradicate the exotic species despite the money and time invested since the 1990s (Pennock et

al. 2018). Traps and oral toxicants were developed to reduce C. carpio menace (Stuart et al. 2006; Poole et al. 2018). Bajer et al. (2011) describe how winter aggregates of

common carp are detected using radio and acoustic telemetry, and subsequently removed using seine nets with high efficiency. These costs are often ignored. For Lake Kariba, research is needed to investigate possible hybridization of O. niloticus and O. mortimeri. To salvage O. mortimeri from extinction would require a substantial investment of time and money. For example, measures are needed to seek and protect any refugia populations.

The Convention on Biological Diversity recognizes that the main cause of bio-diversity loss in freshwater is species introduction, and that there is need to take preventative action. Introduction of exotics should be done only after proper assess-ment of their impact. The need to increase aquaculture in Zimbabwe should be viewed together with the need to prevent further introductions of O. niloticus or other exotics in new areas. This study found that introducing a fish into and occupied niche does not necessary increase fish production. There are tools that can be developed or adapted to screen potential candidates for introduction, such as the protocol described by Copp et al. (2005). Risk identification and risk evaluation assessment is needed prior to all introductions, and the law on introduction of exotics in Zimbabwe has to be adhered to.

Factors that determine the success of introductions

The fish introductions into Lake Kariba present an opportunity to evaluate the environmental factors and biological traits of the species, that determine the proba-bility of their establishment. The case of O. niloticus demonstrates various aspects of an introduced fish that may be associated with competitive advantage. Such data