Habitat Degradation and Subsequent Fishery Collapse in Lakes Naivasha and Baringo, Kenya

Home , Lake Baringo, Lake Bogoria, Lake Naivasha, Oreochromis leucostictus

Habitat degradation and subsequent fishery collapse in Lakes Naivasha and Baringo, Kenya.

Item Type Journal Contribution

Authors Hickley, P.; Muchiri, M.; Boar, R.; Britton, R.; Adams, C.; Gichuru, N.; Harper, D.

Download date 30/09/2021 12:47:36

Link to Item http://hdl.handle.net/1834/7302 Vol. 4 No 4, 503-517 2004 Habitat degradation and subsequent fishery collapse UNESCO IHP EIFAC FAO in Lakes Naivasha and Baringo, Kenya

Ecohydrology and physical fish habitat modifications in lakes

Phil Hickley1, Mucai Muchiri2, Rosalind Boar3, Robert Britton1, Chris Adams1, Nicholas Gichuru4, David Harper5 1National Fisheries Technical Team, Environment Agency, Arthur Drive, Hoo Farm Industrial Estate, Kidderminster, DY11 7RA, UK, e-mails: [email protected], robert.britton@environment- agency.gov.uk, [email protected] 2Department of Fisheries, Moi University, P.O. Box 3900, Eldoret, Kenya, e-mail: [email protected] 3Centre for Ecology, Evolution and Conservation, School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, UK, e-mail: [email protected] 4Kenya Marine & Fisheries Research Institute, P.O. Box 81651, Mombasa, Kenya, e-mail: [email protected] 5Department of Biology, University of Leicester, Leicester LE1 7RH, UK, e-mail: [email protected]

Abstract Lakes Naivasha and Baringo in the eastern Rift Valley of Kenya are shallow, freshwater lakes that are subject to major fluctuations in water level and suffer from habitat degradation as a consequence of riparian activity. Lake Naivasha is approximately 160 km2, is bordered by Cyperus papyrus and its aquatic macrophytes are in a state of flux. The most significant riparian activity is the large scale production of flowers for the European market. Lake Baringo is approximately 140 km2 and lies in a semi-arid region. Its most noticeable feature is its extreme turbidity which is mainly due to excessive soil erosion resulting from deforestation and overgrazing. This turbidity has led to near extinction of submerged macrophytes and a lake bed virtually devoid of benthic fauna. Fishing pressure has added to the environmental stresses being endured by the fish populations and commercial catches have been detrimentally affected. Accordingly, periods of fishery closure are now imposed upon both lakes. Limited remedial action is feasible and some local stakeholders are attempting to introduce mitigation measures. For Lake Naivasha there is an agreed riparian owners' management plan which tackles issues such as water usage and protection of the C. papyrus margin. For Lake Baringo there is a Rehabilitation of Arid Environments initiative which promotes such activities as restoration of riparian vegetation and appropriate stock management. Key words: Rift Valley lakes, ecohydrology, fish, macrophytes, turbidity, soil erosion. 504 P. Hickley et al.

1. Introduction brackish or saline. Bathymetric maps and associated information are given in Figs 1 and 2. Ecohydrology is the scientific concept that Lake Naivasha is approximately 160 km2, is promotes the integration of hydrology and ecolo- 190 km south of the equator at an elevation of gy and the control of these processes to enhance 1890 m a.s.l. and is semi-arid (Agro Climatic Zone the absorption and assimilation capacity of 5). It is bordered by Cyperus papyrus swamp. ecosystems (Zalewski et al. 1997). Methodo- Rainfall is around 680 mm with two rainy seasons. logies include using hydrological processes to Soils are silt loam to clay with a humic topsoil in control biota and as such can provide an effective many places and well drained (although in places, and efficient mechanism for the restoration of poorly drained). Particles from eroded topsoil are degraded environments and associated sustainabil- intercepted the fringing papyrus. Riparian owner- ity of society. Freshwater ecosystems accumulate ship of lake Naivasha is private. the impacts of human activities and consequently Lake Baringo is approximately 140 km2, is the quality of fish habitat depends to a large extent 60 km north of the equator, lies in the Rift Valley upon the density of the human population and its floor at 900 m a.s.l., and has a semi-arid climate. activities within the basin (Zalewski, Welcomme The littoral community is poorly developed 2001). This paper describes the detrimental impact around much of the lake. Being in the Intertropical of degraded environments on two important Convergence Zone (ITCZ) the lake has two rainy Kenyan fisheries and how starting to control the seasons with annual rainfall around 600 mm. Soils key catchment and riparian processes could lead to are clay and clay loams and the risk of soil erosion fishery restoration and sustainability. is high because of the soil properties (clay fills pores or seals the surface giving low infiltration Site descriptions capacity). Forty-six percent of the land in the district is too steep or too dry for agriculture and pas- Lakes Naivasha and Baringo are shallow, toralism is the main source of family income. freshwater lakes in the eastern Rift Valley of Riparian use and ownership is variously private Kenya. Other lakes in the Kenyan Rift Valley are and communal.

LAKE LAKE NAIVASHA BARINGO N

4 m 2.5 m 3.5 m 3 m 2 m 3 m 4.95 m 1.5 m 4 m 2 m 1 m

5 m 16 m

6 m 7 m

Depth contours 5.95 m at 1 m intervals

Depth contours 3 m 6.5 m 2.5 m at 0.5 m intervals (excluding 0.5, 1m)

0123 4 5 km Islands 0123km 2 m 1.5 m Scale: Scale:

Fig.1. Bathymetric maps for Lake Naivasha (August 1991; after Hickley et al. 2002) and Lake Baringo (August 2003). Contour lines are 1 m and 0.5 m (commencing at 1.5 m) intervals respectively. Spot depths for deepest points shown; contour lines <3m around Baringo islands not shown. Habitat degradation and fishery collapse in L. Naivasha and Baringo, Kenya 505

% lake surface area (a) Cumulative % lake area (b) 05010001020304050 0 0-1 10

20 Naivasha : 1-2 Dmean = 3.35 m 30 Dmax = 7.0 m D50 = 3.8 m 2-3 40

50 3-4

% depth (m) 60 Depth zone (m) Baringo : 4-5 70 Dmean = 2.65 m Dmax = 4.1 m 5-6 80 D50 = 3.05 m

90 6-7 Naivasha Baringo Naivasha Baringo 100

Fig. 2. (a) Percentage hypsographic curves for Lakes Naivasha (August 1991, solid line) and Baringo (August 2003, dotted line) with values for mean depth (Dmean), maximum depth (Dmax) and 50 percentile depth (D50); (b) Percentage of lake area underlain by each depth zone for Lakes Naivasha (plain bars) and Baringo (black bars).

Pressures on the lakes and cultivation of sisal gave way in recent years to the present irrigated horticulture. Horticultural The Rift Valley lakes of Kenya are of impor- technology, with its attendant use of pesticides tance to national economy as natural assets, par- has grown dramatically over the past two decades ticularly in the context of tourism. In addition, and this has been accompanied by cultivation those such as Naivasha and Baringo, which are fresh enough to support Happiness Nothing fish populations, provide support to Beauty the local economy by enabling a Power Fishing fishing industry. Local opinion on Climate the importance of Lake Naivasha is Tourism & Foreign represented by Fig. 3. Lake exchange Naivasha became a Ramsar site in April 1995 (Wetlands International Ecology & 2001) and Baringo in January 2002 Science (Ramsar 2001) but the pressures on Washing the lakes' ecosystems and thus on the sustainability of their fisheries are considerable. Both lakes are subject Wildlife to major fluctuations in water level Livestock Drinking and habitats are degraded as a con- watering water sequence of riparian activity. The most significant riparian activity on Lake Naivasha is the Farming large scale production of flowers for Irrigation the European market and at least Employment 50% of the perimeter of the lake is Fig. 3. Principal reasons for Lake Naivasha being considered by the under agriculture of some descrip- local community to be important. Pie chart is a breakdown of 119 tion. Former stock-rearing, ranching answers given by 62 people (see methods). 506 P. Hickley et al.

close to the lake. Also, as the labour intensive stant speed along 32 tracks of total distance 104 flower industry developed, so did the need for km, the positions of which were monitored with a housing, water and latrines (Enniskillen 2002). Garmin 12 hand held GPS receiver. To optimise The local population is approximately 250 000. survey accuracy, a system of tracks crossing the The lake resources are also of critical importance main track was used (Hakansono 1981). For each to geothermal electricity generation, tourism, paper echo-recording chart, the positions along wildlife and conservation (Harper et al. 1990). the track line at which depths at 0.5 m intervals Lake Baringo's most noticeable feature is its occurred were plotted on a plan outline of the lake extreme turbidity which is mainly due to soil ero- and joined to form contour lines. The lake outline sion. This results from low vegetation cover, was derived from a 1:50 000 scale map (Kenya caused by deforestation and overgrazing, exacer- Government 1982) which was based on aerial bated by high intensity, sporadic rainfall on steep photographs taken in August 1977 and January slopes. In recent years two thirds of the total 1978. Depth profiles were transcribed directly catchment (8655 km2) has been grazing and the from the echo-recording charts. Morphometrical rest of the land under agriculture, apart from <1% values were determined according to Hakanson forest. All the grazing area and two thirds of the (1981). agricultural land is environmentally degraded; The area of cover by aquatic macrophytes in almost 90% of the catchment. The Baringo Lake Naivasha was derived from source data in District population is at least 360 000 and grow- Hickley, Harper (2002) and that for papyrus from ing (2.65% p.a.). Associated livestock numbers source data in Boar et al. (1999), Boar, Harper are correspondingly large; approximately 900 000 (2002) and Everard, Harper (2002). goats, 200 000 sheep and 300 000 cattle. At the Water transparency of Lake Baringo in most end of the annual dry season it is usual to receive years was measured monthly using a Secchi disc reports of livestock deaths due to starvation. An in 10 locations chosen at random in each of the additional problem is the installation of dams on southern, central and northern sectors of the lake the inflow rivers and the impounding or diversion and the mean of means used for an annual value. of water otherwise destined for the lake. On 16th August 2003, total light remaining at depths in the water column was measured with a LI-COR LI 190SA Quantum Sensor. Measure- 2. Methods ments in µmol photons m-2 s-1 were converted to the percentage of incident light received at the Commercial fish catch statistics were water surface. Readings were taken mostly at 1 obtained from the Fisheries Department of the cm intervals at the inflow of the Molo River and Kenya Government. In Lake Naivasha, addition- in the open water and at intervals of up to 5 cm at al annual sampling of juvenile fish was carried the lakeward edge of a Typha swamp in the south- out using mono-filament survey gill nets set dur- ern part of the lake. ing the daytime. The survey net catches were con- Public opinion of the importance of Lake verted to a catch per unit effort (CPUE) where the Naivasha was gauged by interviewing sixty-two measure was total catch converted to numbers of residents of the southern shore during August fish per standard gill net per hour. A standard gill 2000. Interviews were semi structured using a net comprised five 30 m long x 1.5 m deep sec- variety of approaches to the question asking. tions, being one each of 11, 15, 20, 24 and 35 mm Each interview lasted about 45 minutes. Category bar mesh. percentages were assigned to a total of 119 The theoretical annual yield (Y) of fish was answers given. estimated using a range of models suitable for use with non-temperate data as described by Hickley et al. (2002). In the equations the following val- 3. Results ues were used for Lakes Naivasha and Baringo respectively: Area (A) = 163, 141 km2; Mean Habitat depth (Z) = 3.35, 2.65 m; Total dissolved solids (TDS) = 231, 280; Air temperature (T) = 21, Recent surface water levels for Lakes 26oC. Naivasha and Baringo are shown in Fig. 4. The The bathymetry of Lake Naivasha was deter- mean depth of Lake Naivasha, when the bathy- mined in August 1991 as described by Hickley et metric survey was carried out in 1991, was 3.35 m al. (2002). Lake Baringo was surveyed during the but this was at a time when the lake level, period 14th - 20th August 2003 using a Lowrance although low, was not exceptionally so. X-15A chart recording echo-sounder with a 20o Notwithstanding the temporary rise in level fol- transducer beam as in the bathymetric survey of lowing El Nino~ rains, the overall trend remains Lake Naivasha (Hickley et al. 2002). The sound- downward and contemporary newspaper reports ings were made from a small boat driven at con- continue to express concern (e.g. Mwakugu Habitat degradation and fishery collapse in L. Naivasha and Baringo, Kenya 507

600 1892 Lake BaringoBaringo 500 Lake NaivashaNaivasha 1890

400 1888

300 1886

200 1884 Water level (cm) Lake level (m.a.s.l.) 100 1882

0 1880 1965 1970 1975 1980 1985 1990 1995 Year

Fig. 4. Recent water surface level fluctuations for Lakes Naivasha (m a.s.l., solid line) and Baringo (cm, dotted line).

2003). The mean depth of Lake Baringo is report- In Lake Naivasha there has been a considered to have been 5.6 m in the 1960s (Ssentongo able reduction in the total area of C. papyrus over 1995), over 8 m in the late 1970s (Meyerhoff the last 40 years as a consequence of both water unpubl.), decreasing to just below 3 m in 1994 but level fluctuation and the reclamation of exposed reaching 4.5 m again in 1998 following El Nino~ stands to increase areas of cultivation. Fig. 5 rains. At the time of the 2003 bathymetric survey shows the area of papyrus cover and, notwith- (Fig. 1) the mean depth was 2.65 m. Satellite standing the short term increase in early 1988 in imagery (Johansson, Svensson 2002) confirms an response to a temporary increase in water level, allegation that the lake is silting up because whilst the trend is an overall and serious decline. Lake the lake outline in August 2003 was comparable Naivasha also has experienced an almost extirpa- with that in 1977/78 when the source data were tion of its submerged macrophytes which, at best, collected for the 1:50 000 map (see methods), the are in a state of flux. An episode of disappearance mean depths are certainly not. is plotted in Fig. 5. Similarly, Lake Baringo has

50 45 40 C. papyrus ) 2 35 Submerged 30 macrophyte 25 20 15 Area of coverage (km 10 5 0 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 Fig. 5. Change in area of Lake Naivasha covered by C. papyrus (data from Boar et al. 1999; Boar, Harper 2002; Everard, Harper 2002) and by submerged macrophytes (data from Hickley, Harper 2002). 508 P. Hickley et al.

Percentage surface light remaining 0.0001 0.01 1 100 0 (1) 0.1 R2 = 0.97 0.2 0.3 (2) 0.4 2 0.5 R = 0.77 0.6 0.7 0.8

Depth (m) 0.9 1 1.1 1.2 1.3 1.4 R2 = 0.94 (3) 1.5 Fig. 6. Total light penetration in Lake Baringo presented as percentage of surface light remaining at depth (August 2003) for three locations (1) River Molo inflow; (2) Main lake; (3) Adjacent to swamp edge.

almost no submerged macrophytes at the present Fisheries time although clear water and macrophytes have been present in the past (Roberts unpubl.). Lake Naivasha originally contained only the Whereas loss of macrophytes in Baringo is endemic Aplocheilichthys antinorii (Vinc.) which assumed to be as a result of low light penetration was last recorded in 1962 (Elder et al. 1971). (see below) this is not the principal cause in Lake Since 1925 various fish introductions have been Naivasha. A certain degree of eutrophication has made, some successful and some not (Muchiri, been recorded for Lake Naivasha (Kitaka et al. Hickley 1991) and, prior to the appearance of carp 2002), induced by less papyrus to interrupt nutri- (Cyprinus carpio L.) in catches during 2002 ent and noxious run off, and the associated (Hickley et al. 2004), the only fish species in the decrease in transparency is likely to have con- lake were Oreochromis leucostictus (Trewavas), tributed to macrophyte disappearance. It has been Tilapia zillii (Gervais), Micropterus salmoides demonstrated, however, by Hickley, Harper Lacépede\ (largemouth bass), Barbus amphigram- (2002) that the submerged plants are decimated ma Blgr. and Poecilia reticulata Peters. Also pres- by crayfish when bass predation is not sufficient ent is the Louisiana red swamp crayfish, to control their density. Procambarus clarkii (Girard). A commercial gill Secchi disk readings for Lake Baringo over net fishery opened in 1959. The mean annual the past twenty years have been in the range 3.5 species composition of the fin-fish catch for the - 13 cm; the clearest period being in 1998 fol- period 1987-2000 was O. leucostictus 71.7%, T. lowing the El Nino rains. Kallqvist (1987) zillii 8.8% and M. salmoides 19.5%. There have recorded 5-7 cm in 1976-77 and 20 cm in 1979. been great fluctuations in the amount of fish land- The 1979 value was the same as that reported by ed from the fishery of Lake Naivasha. Its status Beadle (1932) for 1931. In August 2003 the com- and future was examined by Hickley et al. (2002) parative secchi disk reading was approximately 3 and three phases of development identified: an ini- cm. Detail of this extreme turbidity is indicated tial "boom and bust", a period of stability and, by the plots of total light penetration (Fig. 6). For most recently, a fishery performing poorly. Annual the main lake the extinction coefficient, K (i.e. catches for the last 20 years are presented in Fig. the amount of the available light absorbed), was 7a and selected statistics in Table I. The maximum 3.3 m-1. The transparency of the lake at the recorded total catch of 1150 t yr-1 was attained inflow of the Molo River was even lower at K = during 1970 (Hickley et al. 2002) in contrast with 25 m-1 and only at the very edge of the swamp 21 t yr-1, the lowest ever, in 1997. Survey gill net did the transparency increase (K = 0.71 m-1). catch per unit effort (CPUE) values since 1987 are Secchi disk readings at the same sites were: main shown in Fig. 7b and the trend prior to fishery clo- lake, 3.1 cm; Molo River, 0.6 - 1.5 cm; swamp sure, like that for the commercial fishery for the edge, 32 cm. same period, is downward (p<0.05). Habitat degradation and fishery collapse in L. Naivasha and Baringo, Kenya 509

(a) NAIVASHA NET CATCH 600 FISHERY CLOSED 500 400 300 200 Catch (t) 100 0 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

FISHERY CLOSED 10 (b) NAIVASHA SURVEY CATCH 8 r2 = 0.375 6 p = < 0.05 4 CPUE 2 0 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

600 (c) BARINGO CATCH FISHERY CLOSED FISHERY CLOSED 500 400 300 200 Catch (t) 100 0 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Fig. 7. (a) Total annual fish catch (t) for the Lake Naivasha gill net fishery since 1982; (b) Catch per unit effort (CPUE) for Lake Naivasha survey gill nets since 1986 with trend line fitted for the pre-closure period; (c) Total annual fish catch (t) for the Lake Baringo gill net fishery since 1982. Periods of fishery closure are indicated with arrows.

The fish community of Lake Baringo com- numbers, landings of the lungfish can exceed prises only seven species. The fishers use gill nets 50% of the catch. In 1957, enough fish were being and long lines. Predominant is the endemic caught for a fish processing factory to be opened; Oreochromis niloticus baringoensis and this North-Lewis (1998) states 1500 good specimens a forms the basis of the commercial fishery. This day. For the next twenty years or so fish remained was the only cichlid fish present in 1931 plentiful at about 600 - 1000 t yr-1 but by 1989 (Worthington, Ricardo 1936) and again in 1969 catches had dropped to the point where the fillet- (Ssentongo, Mann 1971), reports of other species ing plant was closed (TVE 2003). Fig. 7c and being unsubstantiated. Aloo (2002) reports O. n. Table I show the comparatively low annual catch baringoensis to comprise 80% of the catch. Other in recent years. species present in the fishery (Aloo 2002), % Fishing pressure has added to the environ- catch in brackets, are: Protopterus aethiopicus mental stresses on the fish populations and com- (7.6%), Clarias gariepinus (8.9%), Barbus gre- mercial catches have been affected detrimentally. gorii (3.1%), Labeo cylindricus (0.1%), Barbus Accordingly, periods of fishery closure are now lineomaculatus and Aplocheilichthys spp. Of the imposed upon both lakes. After lengthy consulta- last three species, L. cylindricus is classed as tions with fishers and other stakeholders, fishing endangered and the other two, rare (Ramsar in Lake Naivasha was banned throughout 2001. 2001). It should be noted, however, that measures On 18th February 2002 the fishery was re-opened of abundance and the consequent conservation with 43 canoes, just over one third of the previous designations can vary according to mesh size and fleet, being allowed to fish. A second period of season. For example, B. lineomaculatus is small closure was imposed from 1 June - 1 October and largely riverine so can give the impression 2003 and this closed period is now expected to be that it is rarer than other species simply through an annual event. The Lake Baringo fishery was not being caught by commercial fishers. De Vos et closed for from May 1993 to April 1994 and at a al. (1998) commented that the first landings of stakeholder workshop in October 2001 it was Protopterus were noticed in 1984 but that nowa- agreed to suspend fishing again for an indefinite days, alongside a decrease in the tilapia size and period from February 2002. 510 P. Hickley et al.

Table I. Estimates of theoretical and actual fish yields calculated for Lakes Naivasha and Baringo:

(a) Theoretical yields based on various predictive models. For formulae and data sources see methods; Theoretical annual Yield t yr-1 Model Parameters used Naivasha Baringo

Crul 1992 Surface area 902 790 Henderson, Welcomme 1974 Morpho-edaphic index (MEI) 716 729 Schlesinger, Regier 1982 Air temperature oC 344 600 Schlesinger, Regier 1982 Air oC & MEI (for depth<25m) 1105 1949 Schlesinger, Regier 1982 Air oC, MEI & effort (low) 358 597 Schlesinger, Regier 1982 Air oC, MEI & effort (high) 727 1190 MEAN of estimates 692 976

(b) Average annual catches compared with mean theoretical estimates. Maximum catch Average catch Theoretical annual LAKE YEAR t yr-1 t yr-1 Yield t yr-1 Naivasha 1962-1973 1150 639 692 Naivasha 1974-1987 576 279 692 Naivasha 1988-2002 439 158 692 Baringo 1982-1998 380 200 976

4. Discussion lake bed virtually devoid of benthic fauna (Ssentogo 1995). In Lake Victoria, turbidity inter- Both lakes are shallow enough to be affected feres with mate choice in cichlid fishes by con- adversely by water level fluctuations. The different straining colour vision (Seehausen et al. 1997). shapes of the hypsographic curves (Fig. 2) suggest The installation of dams on the inflow rivers is that, as a consequence of reduction in water level, likely to have contributed to the severe decline of Lake Naivasha is more prone to exposure of the lit- Labeo cylindricus and Barbus gregori, these fish- toral zone whereas for Lake Baringo the greater es being potamodromous and needing to undertake impact is on lake volume. spawning migrations. Clarias gariepinus and the Muchiri, Hickley (1991) attributed some of lungfish are both much more tolerant of environ- the observed fluctuations in the Lake Naivasha fish mental stresses than the other species. catches to both excess fishing pressure and chang- Although submerged macrophytes are impor- ing lake levels, the latter because the pattern of tant in terms of overall lake quality (Jeppesen et al. overall fish catches appeared to more or less fol- 1994, 1997), their absence in Lake Naivasha low level changes. Similarly, the lake level fluctu- would be unlikely to have any direct effect on the ations, siltation and the peturbed inflow regime availability of fish spawning sites. This is because have influenced the downturn of the Baringo fish- O. leucostictus is a female mouth-brooder ery. Fig. 8 for Lake Naivasha implies that change (Greenwood 1966) and both bass (McClane 1978) in water level has an impact (p<0.01). Fish catch- and T. zillii (Greenwood 1966) excavate nests in es increasing with increase in lake level is likely to the substratum. Macrophyte beds are, however, be because the legal nets can better exploit the likely to be important as refuges and feeding areas. margins in a period of deeper water. A similar link Whilst only suppositions can be made about sub- with water level fluctuation can be demonstrated merged macrophytes, the importance to fish of the for Lake Baringo (Fig. 9) although the relationship papyrus marginal fringe can be demonstrated. is not as strong (p<0.05) whereas water clarity Table II presents CPUE for the survey gill nets (Fig. 10) does relate closely to total fish catch used in three different habitats. For all three (p<0.01). species, the numbers of fish caught were greatest Lake Baringo's extreme turbidity has led to in the nets which were set proximal to the papyrus near extinction of submerged macrophytes and a fringe. In addition, the difference in catches Habitat degradation and fishery collapse in L. Naivasha and Baringo, Kenya 511

600

500

400

300

200 R2 = 0.71 All species catch (t) 100 (p <0.01)

0 1883 1884 1885 1886 1887 1888 Water level (m asl)

Fig. 8. Annual total fish catch (t) from Lake Naivasha plotted against water level (after Hickley, Harper 2002). between nets set near papyrus and those set in open bottom substrata such as sand or gravel, bass are water was significant for both O. leucostictus known to nest and guard their young within vege- (p=<0.05) and T. zillii (p=<0.01). Catches of M. tation as an alternative where, as in Lake Naivasha, salmoides and T. zillii near rocky shores were also firm substrata are not available. In Florida lakes, higher than from open water but the amount of bass are known to migrate to vegetated areas out- rocky shore available as habitat is restricted to only side their normal home range for spawning in pref- a few km (less than 10% of the lake perimeter). erence to remaining in open water (Mesing, Although the difference in survey net catches Wicker 1986). Bruno, Gregory (1990) found bass of bass from papyrus and open water was not sig- nests on rhizomes or on firm detritus in stands of nificant, marginal vegetation is a crucial habitat for Nuphar luteum and emergent grasses Panicum nest sites during the bass spawning season. hemitomon and Paspalidium geminatum. Bass Traditionally considered as building nests on firm nests were observed in the papyrus fringe of Lake

500 450 400 350 300 250 200 R2 = 0.57 Fish catch (t) 150 100 (p <0.01) 50 0 0 2 4 6 8 10 12 14 Transparency (cm)

Fig. 9. Annual total fish catch (t) from Lake Baringo plotted against water level. 512 P. Hickley et al.

500 450 400 350 300 250 200 R2 = 0.57 Fish catch (t) 150 100 (p <0.01) 50 0 0 2 4 6 8 10 12 14 Transparency (cm)

Fig. 10. Annual total fish catch (t) from Lake Baringo plotted against water transparency (Secchi disc, cm).

Naivasha during the late 1980s when samples of (Muchiri et al. 1994) that more detailed work is bass fry were being collected for gut analysis needed, the order of magnitude of the derived val- (Hickley et al. 1994). ues is likely to be correct. For example, Muchiri et For recent years, both the average and maxi- al. (1995) showed that calculations of maximum mum actual catches for the Naivasha and Baringo sustainable yield (MSY) using sophisticated catch- fisheries were considerably lower than the calcu- effort analysis software (MRAG 1992) produced lated theoretical annual yields of 692 t yr-1 and 976 similar estimates to the simple Shaefer model t yr-1 respectively (Table I). This suggests that the (Pauly 1983; Ricker 1975). Also, although the the- two fisheries could be under-performing, as oretical fish yields given in Table I were based addressed for Lake Naivasha by Hickley et al. only on limnological data, this is acceptable (2002). Whilst the equations used were basic, and according to Welcomme (1999) who considered for Lake Naivasha it has been advocated already methods for estimating the potential output from

Table II. Catch per unit effort (CPUE) as No. fish per standard net gang per hour 1987-1999. SPECIES HABITAT CPUE % CATCH Micropterus salmoides Papyrus margin 2.94 37.3 Open water 1.96 28.2 Rocky shore 2.38 34.5 Oreochromis leucostictus Papyrus margin 1.27 62.1 Open water 0.17 13.0 Rocky shore 0.41 24.9 Tilapia zillii Papyrus margin 2.45 41.6 Open water 0.92 18.2 Rocky shore 2.37 40.2

SPECIES P values for t test: Open water Rocky shore M. salmoides Papyrus margin 0.085 0.199 Rocky shore 0.037 O. leucostictus Papyrus margin 0.027 0.083 Rocky shore 0.091 T. zillii Papyrus margin 0.007 0.445 Rocky shore 0.003

Habitat degradation and fishery collapse in L. Naivasha and Baringo, Kenya 513

O.leucostictus T.zillii M.salmoides C.carpio

Fig. 11. Average composition (%) of the catch from the Lake Naivasha gill net fishery (a) overall, 1987 - 2000; (b) three years before closure of the fishery, 1998 - 2000; (c) the year after re-opening of the fishery, 2002. fisheries and concluded that there is no system as Lake Naivasha yet to predict the symptotic level at which Ymax will occur other than the generalised predictors A plan for the management of Lake such as MEI. The higher values for Lake Baringo Naivasha has been written by the Lake Naivasha produced by the Schlesinger, Regier (1982) equa- Riparian Association (LNRA), an association of tions are because, although Lakes Naivasha and about 150 owners of riparian land. The Plan Baringo are of similar physical size, the latter is (LNRA 1999) was adopted by Nakuru District approximately 5oC warmer. Tropical lakes typical- Development Committee in July 1997 and ly support many species of fish, the number being endorsed by the President of Kenya in August broadly correlated with basin area (Welcomme 1997. The Association's Management Implemen- 1999). Lakes Naivasha and Baringo can each be tation Committee has members from Kenya expected to contain 20 or so species according to Wildlife Service, the Kenya Electricity Vanden Bosch, Bernacsek (1990) where, based on Generating Company, The World Conservation 160 tropical lakes, No. species = 5.9 lake area Union, the Municipal Council of Naivasha, (km2) 0.2684. Therefore, some under-performance Government Ministries and the horticultural, could be assigned to low species richness. By far tourism, livestock and fishing industries. Using the most likely detrimental influence, however, international standards, Kenyan law and scientif- remains ecohydrological habitat degradation. ic studies of Lake Naivasha, the LNRA has made Moreover, this view is supported by the evidence recommendations about how the lake's resources presented thus far. should be used, the aim being to prevent the lake Fig. 11 shows species composition of the and its fore shore from being "damaged, polluted Lake Naivasha fish catch overall, immediately or wasted". All categories of users have been before closure of the fishery and after fishing encouraged to write their own Codes of Conduct recommenced. Species composition in the 3 years for their industry according to the Plan's guide- before closure reflects over-fishing of T. zillii with lines and these Codes are included in the Plan. illegal seine nets during a period of low lake lev- Codes of Conduct and LNRA recommendations els, the effect of which was to reduce the size of are followed voluntarily. Regrettably, not all users the legitimate catch. Following reopening of the follow the Plan and not all have joined the LNRA fishery, and notwithstanding the new appearance but they are being encouraged to do so. The Plan of carp (Hickley et al. 2004), the species composi- targets everyone: residents, farmers, growers, tion of the catch seems to have returned to a more hotel and curio shop owners, casual workers, peo- equitable distribution. ple watering their animals, tourists, scientists, fishers and the purchasers of fish. Of particular importance in terms of lake quality and the fish- 5. The way forward ery are the three Codes of Conduct for growers, the power industry and fishers. Key features of Limited remedial action is feasible and, sup- the riparian guidelines contained therein are: ported by Government in the context of both lakes' - Papyrus should not be cleared or burned; Ramsar Wetland status, some local stakeholders - A minimum of 100 m of land must be left are introducing mitigation measures. between farmed land and the water and there 514 P. Hickley et al.

should be at least 50 m of fringing papyrus (RAE) Trust. The RAE is a charitable trust that swamp; has been active in the Baringo District since 1982. - Untreated sewage and chemicals such as fungi- RAE partnerships with the community develop cides, pesticides and fertilizers must not enter practical solutions to dryland restoration that the lake or groundwater; involve fencing degraded areas from livestock, - Water must be used responsibly, efficiently and rain-fed water harvesting and planting of indige- economically; nous grasses and trees. The grasses used to curb - Water use and lake level and weather should be erosion whilst providing a worthwhile product are monitored; the thatching grasses Cenchrus ciliaris and - Soil erosion will be managed and minimised. Eragrostis superba. The trees planted comprise mainly introduced Prosopis spp. Partnerships Specifically for the protection of the fishery, have been successful and some 1820 ha of land when open, the regulations and recommendations across 175 sites have now been reclaimed and adopt a traditional approach (Gulland 1971) 430 000 drought-resistant seedlings have been including gear specification, size limits and distributed both locally and elsewhere in Kenya. closed areas: The aims of the RAE initiative are to produce sus- - Fish must be landed at one of the designated tainable returns from livestock grazing, fuelwood, landing beaches; thatching grass, seed, honey and traditional food, - Numbers and size of fish caught must be report- fruits and medicines. Environmental aims are ed to the Fisheries Department; centred upon reducing soil erosion in the catch- - Each licensed fisher can use not more than ten ment of Lake Baringo and protection of the Lake's 100 m gill nets with a minimum mesh size freshwater resources. (stretched) of 4 inches (100 mm); - Tilapia landed or sold must exceed 18 cm in Pathways to rehabilitation length; - Seine netting and fishing in the dark is prohibit- An ecohydrological approach is appropriate ed; to resolving the environmental quality and fish- - No fishing in bays or lagoons or within 100 m of eries issues outlined and discussed in this paper. A the shoreline or papyrus; schematic diagram showing feasible, albeit opti- - The Department of Fisheries will educate the mistic, pathways to rehabilitation for Lakes public and use community participation to curb Naivasha and Baringo is presented in Fig. 12. illegal fishing. Although highlighting only the more obvious mechanisms, both pathways work on the principle Lake Baringo of addressing riparian activity, aiming towards the ultimate consequential benefit of restored and For Lake Baringo there is a sustainable man- sustainable fishery performance. agement approach being adopted by the stake- For Lake Naivasha one of the most impor- holders including empowering local communities tant contributions will be to facilitate recovery of for natural resource management, diversification the papyrus fringe. C. papyrus seeds germinate of agriculture and agro-forestry systems. Much quickly and, given a return to favourable condi- has been facilitated through UNEP and the Global tions, a papyrus fringe is capable of regenerating Environmental Facility (GEF) through the Lake and reaching maturity within six months (Muthuri Baringo Community Based (LBCB) Land and 1989). As well as providing important habitat for Water Management Project. The management all fish species, this should stimulate recruitment institutions include the local councils, community to the bass population. Increased predation by based organisations, hotels, the Fisheries bass should better control crayfish numbers to Department and the Kenya Forestry and Marine relieve grazing pressure on the aquatic macro- & Fisheries Research Institutes (KFRI and phytes and thus aid their recovery. In parallel, a KMFRI). These institutes need to operate to a restored papyrus margin should induce an overall management plan in a similar way to the riparian reduction in eutrophication and a reduction in interests of Naivasha and accordingly, since the algal blooms which will also aid recovery of recent designation of Lake Baringo as a Ramsar aquatic macrophytes. Management of the fishery site, a management plan is in preparation. on a sustainable basis then becomes more realis- Meanwhile, some steps have already been taken tic. Indeed, according to some e.g. Higgins to address soil erosion such as the creation of over (unpubl.) all conditions are present for Lake 30 km of terraces for slope stabilisation, facilitat- Naivasha to become one of the first basins in ed by the Department of Agriculture. In addition, Africa with a sustainably managed lake according certain Non-Governmental Organisations to World Lake Vision principles (International (NGOs) have been pro-active, a good example Lake Environment Committee Foundation & being the Rehabilitation of Arid Environment United Nations Environment Programme 2003). Habitat degradation and fishery collapse in L. Naivasha and Baringo, Kenya 515

NAIVASHA BARINGO

Protection of Reduction of livestock and planting C. papyrus fringe of grass & trees

C. papyrus fringe recovers Reduction of suspended solids from run-off-

Bass recruitment improves and eutrophication reduces Transparency of water improves

Increased predation on crayfish and algal blooms reduced Increased light penetration

Aquatic macrophytes recover Aquatic macrophytes recover

Fishery restored Fishery restored

Fig. 12. Schematic diagram indicating primary rehabilitation pathways for fishery restoration for Lakes Naivasha and Baringo.

For Lake Baringo the ecohydrological also go to the Fisheries Department of the Kenya approach to fishery restoration is equally chal- Government, the Lake Naivasha Riparian lenging but the pathway is more straightforward Association and the Rehabilitation of Arid and better understood. It is quite clear, therefore, Environments Charitable Trust for provision of that mitigation work in the Lake Baringo catch- information. The views expressed are those of the ment will produce noticeable environmental ben- authors and not necessarily those of their parent efits. Reducing erosion, and thus suspended organisations. solids, should ultimately improve transparency and allow aquatic macrophytes to return to the point where a sustainable fishery can be managed. 6. References It must be remembered, however, that anthro- pogenic activity is not the sole contributor to the Aloo, P.A. 2002. Effects of climate and human activities lake's turbidity; the landscape is such that natural on the ecosystem of Lake Baringo. In: Odada, E.O., erosion risk will remain high. Olago, D.O. [Eds] The East African Great Lakes: Although physical habitat improvement is Limnology, Paleolimnology and Biodiversity. Advan- the key component of the rehabilitation pathway ces in Global Research 12. Kluwer Academic Publis- for both lakes it should be noted, however, that hers. Dordrecht. pp 335-348. fishery restoration cannot be successful without Beadle, L.C. 1932. Scientific results of the Cambridge stakeholder consultation and support. Indeed, the Expedition to East Africa Lakes 1930-1. The waters sustainable development of society is very much of some East Africa lakes in relation to their flora and part of the over-arching ecohydrology concept fauna. Journal of Linnean Society (Zoology) 38, 138- (Zalewski et al. 1997). 211. Boar, R.R., Harper, D.M., Adams, C.S. 1999. Biomass Allocation in Cyperus papyrus in a Tropical Wet- Acknowledgements land, Lake Naivasha, Kenya. Biotropica 31, 411- 421. Some of this paper draws on work carried Boar, R.R., Harper, D.M. 2002. Magnetic susceptibilities out as part of the "Lakes of the Rift Valley" proj- of lake sediment and soils on the shoreline of Lake ect sponsored by the Earthwatch Institute and the Naivasha, Kenya. Hydrobiologia 488 (Develop- authors are grateful to the many Earthwatch vol- ments in Hydrobiology 168), 81-88. unteers for their assistance in the field. Thanks 516 P. Hickley et al.

Bruno, N.A., Gregory, R.W. 1990. Nest Sites Used by Hickley, P., Muchiri, S.M., Britton, J.R., Boar, R.R. Radio-Tagged Largemouth Bass in Orange Lake, 2004. Discovery of carp (Cyprinus carpio) in the Florida. North American Journal of Fisheries already stressed fishery of Lake Naivasha, Kenya. Management 10, 80-84. Fisheries Management and Ecology 11, 139-142. Crul, R.C.M. 1992. Models for estimating potential fish International Lake Environment Committee Foundation yields of African inland waters. FAO Committee for and United Nations Environment Programme 2003. the Inland Fisheries of Africa (CIFA) Occasional World Lake Vision. http://www.ilec.or.jp/ Paper 16, 22pp wwf/eng/wlv_contents/WLV_Final.PDF De Vos, L., Pertet, F., Vanlerberghe, K., Nuguti, S., Jeppesen, E., Bgestrand, J. Skiver, P., Soendergaard, M. Ntiba, M.J. 1998. Present status of the fish fauna and 1994. Impact of macrophyte coverage on fish-zoo- fisheries of Lake Baringo, Kenya. Proceedings of plankton-phytoplankton interactions in a shallow International Conference for the Paradi Association temperate lake. Lake and Reservoir Management 9, and the Fisheries Society of Africa, Grahamstown 84. (South Africa). FISA/PARADI. Grahamstown. pp Jeppesen, E., Jensen, J.P., Soendergaard, M., Lauridsen, 13-18. T., Pedersen, L.J., Jensen, L. 1997. Top-down control Elder, H.Y., Garrod, D.T., Whitehead, P.J.P. 1971. Natu- in freshwater lakes: The role of nutrient state, sub- ral hybrids of the African cichlid fishes Tilapia spiru- merged macrophytes and water depth. Hydrobiologia lus nigra and T. leucosticta, a case of hybrid intro- 342-343, 151-164. gression. Biological Journal of the Linnean Society 3, Johansson, J., Svensson, J. 2002. Land degradation in 103 146. the semi-arid catchment of Lake Baringo, Kenya: a Enniskillen 2002. The Lake Naivasha Management Plan minor field study of physical causes with a socio-eco- - consensus-building to conserve an international nomic aspect. Earth Sciences Centre, Göteborg Uni- gem. Hydrobiologia 488 (Developments in Hydro- versity. Göteborg. 79pp. biology 168), ix-xii. Kallqvist, T. 1987. Primary production and phytoplank- Everard, M., Harper, D.M. 2002. Towards the sustaina- ton in Lake Baringo and Naivasha. Norwegian Insti- bility of the Lake Naivasha Ramsar site and its catch- tute for Water Research (NIVA) Report No. ment. Hydrobiologia 488 (Developments in Hydro- E8041905. 59pp. biology 168), 191-203. Kenya Government 1982. East Africa 1:50 000 (Kenya), Greenwood, P.H. 1966. The fishes of Uganda. The Ugan- Lake Baringo. Survey of Kenya. Nairobi. da Society. Kampala. 131pp. Kitaka, N., Harper, D.M., Mavuti, K.M. 2002. Phospho- Gulland, J.A. 1971. Appraisal of a Fishery. In: Ricker, rus inputs to Lake Naivasha, Kenya, from its catch- W.E. [Ed.] Methods for assessment of fish production ment and the trophic state of the lake. Hydrobiologia in freshwaters. Blackwell Scientific Publications. 488 (Developments in Hydrobiology 168), 73-80. Oxford. pp. 259-269. LNRA 1999. Lake Naivasha Management Plan. Lake Hakanson,o L. 1981. A Manual of Lake Morphometry. Naivasha Riparian Association, P.O. Box 1011 Nai- Springer-Verlag. Berlin, Heidelberg & New York. vasha, Kenya. 77pp. ix+78pp. McClane, A.J. 1978. Field Guide to Freshwater Fishes Harper, D.M., Mavuti, K.M., Muchiri, S.M. 1990. Eco- of North America. Henry Holt. New York. 212pp. logy and management of Lake Naivasha, Kenya, in Mesing, C.L., Wicker, A.M. 1986. Home Range, Spaw- relation to climatic change, alien species introduc- ning Migrations, and Homing of Radio-Tagged Flo- tions and agricultural development. Environmental rida Largemouth Bass in Two Central Florida Lakes. Conservation 17, 328 336. Transactions of the American Fisheries Society 115, Henderson, H.F., Welcomme, R.L. 1974. The relations- 286-295. hip of yield to morpho-edaphic index and number of MRAG 1992. Catch-effort data analysis: Software and fishermen in African inland fisheries. UN/FAO CIFA manual. MRAG Ltd & ODA. London. Occasional Paper 1, 19pp. Muchiri, S.M., Hart, J.B., Harper, D.M. 1995. The persi- Hickley, P., Bailey, R.G., Harper, D.M., Kundu, R., stence of two introduced species in Lake Naivasha, Muchiri, S.M., North, R., Taylor, A. 2002. The status Kenya, in the face of environmental variability and and future of the Lake Naivasha fishery, Kenya. fishing pressure. In: Pitcher, T.J., Hart J.B. [Eds] Hydrobiologia 488 (Developments in Hydrobiology Impact of Species Change in African Lakes. Chap- 168), 181-190. man & Hall. London, pp 299-319. Hickley, P., Harper, D.M. 2002. Fish community and Muchiri, S.M., Hickley, P. 1991. The fishery of Lake habitat changes in the artificially stocked fishery of Naivasha, Kenya. In: Cowx, I.G. [Ed.] Catch Effort Lake Naivasha, Kenya. In: Cowx, I.G. [Ed.] Sampling Strategies: Their application in freshwa- Management and Ecology of Lake and Reservoir ter fisheries management. Fishing News Books, Fisheries. Fishing News Books, Blackwell Scientific Blackwell Scientific Publications. Oxford, pp 382- Publications. Oxford. pp 242-254. 392. Hickley, P., North, E., Muchiri, S.M., Harper, D.M. Muchiri, S.M., Hickley, P., Harper, D.M., North, E. 1994. The diet of largemouth bass, Micropterus sal- 1994. The potential for enhancing the fishery of moides, in Lake Naivasha, Kenya. Journal of Fish Lake Naivasha, Kenya. In: Cowx, I.G. [Ed.] Reha- Biology 44, 607-619. Habitat degradation and fishery collapse in L. Naivasha and Baringo, Kenya 517

bilitation of Freshwater Fisheries. Fishing News Freshwater Fish Research Organisation for 1970, Books, Blackwell Scientific Publications. Oxford. 20-27. pp 348-357. TVE 2003. Earth Report: Healing the Rift. Muthuri, F.M. 1989. Primary productivity of papyrus http://www.tve.org/earthreport/archive/doc.cfm?ai (Cyperus papyrus L.) in a tropical swamp - Lake d=826 Naivasha, Kenya. Biomass 18, 1-14. Vanden Bosch, J.P., Bernacsek, G.M. 1990. Source Mwakugu, N. 2003. Anxiety as Kenyan lake dries up. book for the inland fishery resources of Africa. BBC News, Nairobi. http://newsvote.bbc.co.uk/2/ FAO Committee for the Inland Fisheries of Africa hi/africa/3241049.stm (CIFA) Technical Paper 18. FAO. Rome. Vol. 1, North-Lewis, M. 1998. A guide to Lake Baringo and 411pp; Vol.2, 240pp; Vol.3, 219pp. Lake Bogoria. Horizon Books. v+161pp. Welcomme, R.L. 1999. A review of a model for quan- Pauly, D. 1983. Some simple methods of assessing fish titative evaluation of exploitation levels in multi- stocks. FAO Fisheries Technical Paper 234, 52pp. species fisheries. Fisheries Management and Eco- logy 6, 1-19. Ramsar 2001. Information Sheet on Ramsar Wetlands (RIS) - Kenya, Lake Baringo 2002. Wetlands International 2003. Ramsar Sites Database, http://www.ramsar.org/ris_kenya_baringo1.htm Lake Naivasha 1KE002. http:/www.wetlands.org/ reports/ris/1KE002en.pdf RAE 2003. Rehabilitation of Arid Environments Cha- ritable Trust. http://www.michna.com/rae/back- Worthington, E.B., Ricardo, C.K. 1936. Scientific ground_ home.htm results of the Cambridge Expedition to the East African lakes, 1930-1. No.15. The fish of Lake Ricker, W.E. 1975. Computation and interpretation of Rudolf and Lake Baringo. Journal of the Linnean biological statistics of fish populations. Bulletin of Society of London (Zoology) 39, 353-389. Fisheries Research Board of Canada 191, 382pp. Zalewski, M., Janauer, G.A., Jolankaj, G. [Eds] 1997. Schlesinger, D.A., Regier, H.A. 1982. Climatic and Ecohydrology: A New Paradigm for the Sustaina- morphoedaphic indices of fish yields in natural ble Use of Aquatic Resources. Conceptual Bak- lakes. Transactions of the American Fisheries kground, Working Hypotheses, Rationale and Society 111, 141-150. Scientific Guidelines for the Implementation of the Seehausen O., van Alphen, J.J.M., Witte, F. 1997. IHP-V Projects 2.3/2.4. International Hydrological Cichlid fish threatened by entrophication that curbs Programme UNESCO. Paris Technical Documents sexual selection. Science (Washington) 277, 1808- on Ecohydrology No.7.. 1811. Zalewski, M., Welcomme, R. 2001. Restoration of Ssentongo, G.W. 1995. Report on the present fisheries physically degraded fish habitats - the Model of situation of Lake Baringo. FAO. Rome. Intermediate Restoration. Ecohydrology & Hydro- Ssentongo, G.W., Mann, M.J. 1971. On the fish spe- biology 1, 279-282. cies of Lake Baringo. Report of the East Africa