Impact of land use and tenure changes on the Kitenden wildlife corridor, Amboseli Ecosystem, Kenya
Abstract
enThis study assesses the ecological pressure exerted by changing land use and tenure on the Kitenden wildlife corridor, a critical cross-border link between the Amboseli and Kilimanjaro national parks. The implications on viability of the two high-value protected areas and their respective dispersal areas are both negative and serious. The extent of land use change and its impacts were assessed through household and vegetation surveys, while wildlife abundance was measured using transect walks. Approximately 30% of the study area has shifted from community to private land ownership over the last two decades. Except for baboon and vervet monkey, most wildlife avoided the cultivated area. Vegetation composition on the noncultivated area has been greatly altered by intense wildlife and livestock use, where mean herbaceous vegetation cover differed significantly among range-plant categories (F3, 524 = 29.015, p < 0.05). The frequency of increaser I (21.4%) differs greatly from that of decreaser and forbs, at 8.3% and 7.4%, respectively Tree recruitment was low, with a significant difference in mean density among age classes (F2, 110 = 3.663, p < 0.05). Only through land leasing agreements between landowners and conservation organisations, and a widely supported land use plan, can the spread of cultivation be controlled and complete cessation of wildlife movement be prevented.
Résumé
frCette étude évalue la pression écologique exercée par le changement dans l'utilisation des terres et du régime foncier sur le corridor faunique de Kitenden, un lien transfrontalier essentiel entre les parcs nationaux d'Amboseli et de Kilimanjaro. Les conséquences sur la viabilité des deux aires protégées de grande valeur et de leurs zones de dispersion respectives sont à la fois négatives et graves. L'ampleur du changement dans l'utilisation des terres et ses impacts ont été évalués à l'aide d'enquêtes auprès des ménages et de la flore, tandis que l'abondance de la faune été mesurée à l'aide de marches transect. Environ 30% de la zone d'étude est passée de la propriété foncière communautaire à la propriété privée au cours des deux dernières décennies. À l'exception du babouin et du singe vervet, la plupart des animaux sauvages évitaient les zones cultivées. La composition de la végétation dans la zone non cultivée a été fortement modifiée par une utilisation intense de la faune et du bétail. La couverture végétale herbacée moyenne différait considérablement entre les catégories de plantes (F3, 524 = 29.015, p < 0.05). La fréquence de plantes de catégorie évolutive I (21.4%) était très différente de celle des catégories regressive (8.3%) et herbacées non graminoïde (7.4%). La régénération des arbres était faible, avec une difference significative dans la densité moyenne entre les classes d'âge (F2,110 = 3.663, p < 0.05). Ce n'est que par le biais d'accords de location de terres entre les propriétaires fonciers et des organisations de protection de la nature, et d'un plan d'utilisation des terres largement soutenu par toutes les parties prenantes que l'expansion de l'agriculture peut être contrôlée, et la cessation complète des mouvement des espèces sauvages empêchée.
1 INTRODUCTION
Over the past century, African nations have established an extensive network of protected areas, which play an essential and central role in conserving species and ecosystems (Newmark, 2008). While the expansion of protected areas coverage in Africa over the past 30 years has been encouraging, the capacity of some to maintain viable populations of many wildlife species over the long term is threatened by a combination of human-influenced activities including land privatisation, fencing, cultivation, livestock rearing and road development, within and outside protected areas (Newmark, 2008; Western, 1982).
Progressive insularisation heralds a bleak future for protected areas. Ecological isolation caused by land use and change has numerous consequences on wildlife populations. These include limiting gene flow leading to genetic erosion and reduce dispersal which enhances the risk of species extinction in case of natural disasters or extreme climate events such as disease outbreak, drought, fire, floods or earthquakes. Moreover, ecological isolation may be damaging to species requiring large home ranges and to those having complex needs at different stages in their life cycles or seasons (Kideghesho, Nyahongo, Hassan, Tarimo, & Mbije, 2006).
Terrestrial wildlife corridors can be defined based on functionally as an area used by animals to pass from one “habitat patch” to another and structurally as an area connecting two patches of suitable habitat by passing through a matrix of unsuitable ones (Hilty, Lidicker, & Merelender, 2006). According to Caro, Jones, and Davenport (2009), wildlife corridors are classified into five categories including unconfirmed corridors, uncultivated lands or patches of natural vegetation between protected areas without documented animal movement, continuous or semi-continuous uncultivated land between protected areas with anecdotal information on animal movements, known animal movement routes between two protected areas, as in the case of the Kitenden Wildlife Corridor (KWC) and, potentially connected important habitats. These can be especially important if they contain endangered species.
According to Crooks and Sanjayan (2006), wildlife corridors are of tremendous importance because they help avoid local extinction through immigration of individuals from elsewhere. They can also boost the genetic variability of a small isolated population through inflow of new individuals hence reducing inbreeding and genetic drift and increase the area and habitat diversity available for wildlife beyond the two habitat patches being connected. In the event of a habitat becoming unsuitable due to natural or anthropogenic conditions, plant and animal species from that habitat can move to a suitable habitat through the corridor. Finally, as protected areas do not encompass the total range of ecosystem requirement for certain species, corridors are used by these species to acquire what cannot be found within the area.
In Kenya, wildlife is protected primarily in national parks, national reserves, marine parks/reserves and sanctuaries, but the majority of large mammal populations occur outside such protected areas. This proportion is declining at a high rate, such that the total wildlife population across Kenyan rangelands declined by 39% between 1970 and 1990 (Okello, 2009; Okello & Kioko, 2010; Ottichilo, Grunblatt, Said, & Wargute, 2000) with reductions in the population of individual species ranging from 2% to 72%. Dispersal areas/movement corridors loss, poaching and overexploitation for bushmeat were believed to be the most important drivers of that decline (Newmark, 2008; Ottichilo et al., 2000). In the state protected areas, and on communally owned rangelands, numbers have declined by 60%–70% since the 1970s (2007; Western, Russell, & Cuthill, 2009). These losses have been significantly greater outside protected areas than inside and in trust land compared with elsewhere (Bourn & Blench, 1999), higher still in regions with little tourism (Norton-Griffiths, 1998). Extreme drops in numbers of 14 of the 18 common wildlife species have been registered nationwide since 1977, averaging 68.1% (1.7% per year) (Ogutu et al., 2016).
The Kitenden Wildlife Corridor is critical for dispersal and movement of many wildlife species in the Amboseli ecosystem and out to west Kilimanjaro across the border with Tanzania. However, a proliferation of crop cultivation, settlements and land fencing threatens to block these movements. The progressive change in land use and land subdivision is a major threat to long-term viability of this corridor. Previous studies have showed that habitat fragmentation, degradation and loss of dispersal and migration corridors are important causes of decline in wildlife numbers (Kiringe & Okello, 2007; Ottichilo et al., 2000; Western, 1982). However, the level of threat to Kitenden Wildlife Corridor is poorly documented. If this trend is not controlled urgently, Amboseli National Park (ANP) and Kilimanjaro National Park (KNP) are likely to lose their connectivity, making them less ecologically viable.
The aim of this study was to underpin ecological pressure exerted on KWC by changing land use and tenure, and the implications of these changes on the viability of wildlife movement between the two protected areas and their respective buffer zones.
1.1 Study area
Kitenden Wildlife Corridor is located in the Amboseli ecosystem, which broadly refers to the dry lake basin, permanent wetlands, gently rolling plains and volcanic hills located in the southeastern part of Kajiado County in Kenya. It covers a total area of 167.53 km2 southeast of ANP in the Olgulului–Ololorashi group ranch that surrounds about 90% of the park. It lies across Kenya–Tanzania border between latitudes 2° 44' 49" and 2° 52' 13" S and longitudes 37°21'45" and 37°21'44"E. It is also part of the buffer zone of a Man and the Biosphere (MAB) Reserve declared in 1991 by the United Nations Educational, Scientific and Cultural Organization (UNESCO) with the park as its core. Nyeki (1993) reported a total of 36 mammal species that were found in the Amboseli ecosystem.
Kitenden Wildlife Corridor falls broadly under ecological zone V of the East African rangeland classification (Pratt, Greenway, & Gwynne, 1966) (Figure 1). Climate is warm and dry with 250–300 mm of rainfall per year. Rainfall pattern is bimodal with long rains occurring in March–May, while the short rains are in October–December. Rainfall tends to be patchy, unpredictable and erratic and is highly dependent on altitude (Western & Maitumo, 2004). Daily temperature ranges between 30 and 35°C with the temperature being low during the months of May and July. The rest of the months experience high temperatures that are characteristic of semi-arid conditions (Western & Ssemakula, 1981).
Source:
AWF spatial analysis laboratory, May 2012) [Colour figure can be viewed at wileyonlinelibrary.com]Plant communities in the Amboseli basin are characterised based on species composition and woody cover, typically Acacia-Commiphora bushland varying along the elevation gradient and open grassland dominated by Pennisetum stramineum. The woody cover provides the main source of domestic energy, which comprises predominantly firewood and charcoal.
Land ownership in the Maasai land was initially communal or group ranches (Ntiati, 2002). However, a policy enacted by the Government of Kenya to promote the subdivision of group ranches to individually owned parcels marked the beginning of dramatic change in land ownership (Mbote, 2005).
2 MATERIALS AND METHODS
2.1 Determination of land use and tenure changes
To capture the diverse opinions of the local people who were the main target for the study, KWC was divided into two settlement areas, the lower belt represented by Eldepen, Olmoti, Emurraa, Oldule, Elmarba and Enkong Narok villages, and the upper belt comprising the Irkaswa and Imisingiyio villages within the higher or rain-fed agricultural zone. A total of 169 questionnaires were administered on adult respondents of both gender, in households within or in the immediate vicinity of sampling plots, which represented approximately 30% of all the households in the study area as described above. These were selected using random tables and a numbered list of households provided by the local authority and leadership of the Olgulului–Ololorashi group ranch.
Interviews conducted between October and November 2011 aimed at getting information on (1) mainland uses, (2) most beneficial land use and (3) human–wildlife conflict incidences. Only one adult (male or female) household head first encountered from each homestead was interviewed. All interviews were conducted face to face using a semi-structured questionnaire, with each lasting 10–15 min. Interviewers conversant in the languages spoken locally and familiar with KWC, and locations of homesteads posed each question and guided discussion to ensure precise understanding of the study questionnaire. Interviewers translated the questions to respondents into the Maasai language if necessary. Secondary data on the extent of cultivation and private ownership of land were obtained from the African Wildlife Foundation (AWF) spatial analysis laboratory.
2.2 Woody species characterisation
Five vegetation stands characterised by homogeneity in the general species composition and physiognomic structures were selected for stratified sampling of habitat types. Twenty line transects, each 1km long, were established and sampled along east–west oriented topographic map grids. Transects were systematically placed at 3-km intervals distributed in the whole study area. Transect orientation from the starting point was determined by the nature of the topographic features, with transects cutting across major drainage channels.
The number of transects sampled was established proportionately to the area covered by each habitat type. Eight transects were set up in Licium europeaneum grassland, three in Acacia tortilis bushland, five in Acacia mellifera bushland and two in Commiphora schimperi bushland. For these four habitats, sampling was done systematically at every 100-m interval using Point Centered Quarter (PCQ) technique (Brower & Von, 1990). Woody species density, frequency, dominance, their relative values and the important value index (IVi) (Brower et al. 1990) were estimated.
The same variables were sampled on farmland using belt transect method. Two transects were established in this habitat type as explained above and used as baselines, along which a total of 10 (100x20m) belts were systematically set at 100-m intervals. The data were used to calculate the same parameters as in PCQ technique.
2.3 Herbaceous vegetation composition and cover
Transects used for woody vegetation sampling were also sampled for the herbaceous cover during rainy season of November and December. The herbaceous layer was sampled using 2x2m quadrats established along each 1-km transect. Quadrats were systematically placed at 100-m intervals on each of the 20 transects, resulting in 10 quadrats per transect and a total of 200 quadrats. Individual dicots and grass (monocots) species were sampled for percentage cover in each quadrat. Herbaceous species cover was estimated in a semi-objective manner while individuals of each species were enumerated. Following the classification by Trollope and Trollope (1999), herbaceous species were classified as decreaser, increaser I, increaser II and forbs. This was done to determine the level of grazing pressure.
2.3.1 Wild mammals and livestock distribution and composition
Foot count method was used to sample animal abundance as described by Norton-Griffiths (1978), by walking along transects laid across the terrain. South–north oriented map grid lines were used as baselines, the first of which was chosen at the extreme west of the KWC while the last was located at the eastern end. Five baselines were systematically established at 3-km intervals and sampled once every month between October and December 2011. Along each baseline, a 1-km sampling unit was established from Tanzania/Kenya border. The research team walked along each sampling unit, searching and counting mammals the size of a Kirk's dik-dik (Madoqua kirkii) and above sighted within 100m on either side using a pair of binoculars for identification. The width and length of the sampling units were monitored and maintained using a range finder and GPS, respectively. Between one sampling unit and the next, a buffer zone of 500m was left where no counting was done to minimise the possibility of double counting. Whenever habitat changed, the sampling unit was terminated and a new one begun in the new habitat.
All the wild ungulates, carnivores and primates encountered during the counts were identified through direct observation or distinctive signs. When animals or their signs were spotted, habitat, species and number of individuals were recorded. The data obtained were used to estimate population density following Norton-Griffiths (1978).
2.4 Data analysis
All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS version 18.0). Data obtained from questionnaire interviews were subjected to a contingency table chi-squared test to establish differences in land uses between the upper and lower belts. Descriptive statistics were used to generate means and percentages. Percentage cover for different herbaceous species categories was calculated for each quadrat sampled and data transformed using Arcsine, for one-way ANOVA, following which the F-statistic was calculated to test for differences in density of plants and animals among habitats. All statistical test results were considered significant at p < 0.05 confidence level (Zar, 2010).
3 RESULTS
3.1 Land use, tenure and agricultural expansion
Cultivation and livestock raising dominated land use in KWC, with the numbers involved in those two activities showing significant difference between the upper and lower belt (X23 = 57.7,) (Table 1). Residents of the upper belt were primarily engaged in either mixed farming (livestock and cultivation) or cultivation alone, while those of the lower belt practiced either mixed farming or pastoralism (Table 1). Although inhabitants of the lower belt were also engaged in cultivation, their farms were located at the upper belt where rainfall was more plentiful. A large proportion of the population (85%) considered cultivation more beneficial than pastoralism, 11% thought vice versa while a small portion at 4% did not have a choice. In the upper belt, about 5,046.3 hectares of land, representing 30.1% of the study area were demarcated into 1,246 privately owned individual parcels. By the time of this study, 25.8% of this land was under cultivation (Figure 1).
| Land uses | Location of respondent | Total | |
|---|---|---|---|
| Upper belt | Lower belt | ||
| Livestock only | 2 | 31 | 33 |
| Livestock and cultivation | 55 | 39 | 94 |
| Cultivation only | 37 | 3 | 40 |
| Others strategies | 0 | 2 | 2 |
| Total | 94 | 75 | 169 |
Note
- X2 0.05, 3 = 57.701 comparing across locations
3.2 Human–wildlife conflicts and problem animals
Livestock depredation was mentioned as a source of conflict most frequently (32%) followed by crop raiding (28.3%) and disruption of social activities (23.9%). Elephant (Loxodonta africana) and buffalo (Synceru scaffer) were the most important crop raiders, while lion (Panthera leo) and hyaena (Crocuta crocuta) were the most important livestock predators. Other prevalent sources were destruction of properties, human injury and death, representing 15.8% of all the types of conflicts mentioned.
3.3 Vegetation characterisation
3.3.1 Woody species composition and density
A significant difference was found in woody species mean density among habitats (F4, 52 = 3.57, p < 0.05). Commiphora schimperi bushland, with a mean of 56 ± 13 trees/ha, differed from Licium europeaneum grassland with 15 ± 6 trees/ha, and farmland with 19 ± 9 trees/ha (Figure 2). In the upper belt, farmland had a very low woody density, contrasting with the neighbouring Commiphora schimperi bushland.
There was a significant difference in woody species mean density among the three age classes (F2, 110 = 3.66, p < 0.05). Tukey post hoc test revealed that the mature age class (>3m height) with 18 ± 5 trees/ha differed from the sapling age class (<1m height) 5 ± 1 trees/ha (Figure 3). No difference was found in sapling density among habitats (F4, 22 = 2.75, p > 0.05), meaning that recruitment rate was low overall. Density of mature trees showed a significant difference among habitats (F4, 30 = 6.09, p < 0.05), with C. schimperi bushland accounting for 50 ± 13 trees/ha and differing significantly from L. europeaneum grassland (4 ± 2 trees/ha) and farmland (8 ± 3 trees/ha), further pointing out the modification on farmland.
3.3.2 Difference in percentage cover among herbaceous vegetation categories
A total of 43 herbaceous species were found in the study area and classified in four categories (decreaser, increaser I, increaser II and forb). There was a significant difference in mean percentage cover among these categories (F3, 524 = 29.01, p < 0.05). Tukey post hoc test indicated that increaser I with 21.4% was different from decreaser (8.3%) and forbs (7.4%). The most affected area was farmland with the highest presence of increaser II (24%) and low decreaser (2%). The increaser I category was dominated by Pennisetum stramineum and Pennisetum mezianum grass species. Only one decreaser (Cenchrus ciliaris) was found in the whole study area and at low cover (8.3%).
3.4 Wildlife and livestock composition and distribution
Twenty-five wildlife species (Table 2) and four livestock species were counted in the study area. Twenty-two out of the total were sighted, and seven were identified through various signs (footprint, dung, vegetation disturbance). Of the species sighted, sixteen were wild herbivores, two carnivores and four livestock. Density of wild herbivores among habitats showed no significant difference (F3, 37 = 2.23, p > 0.05). The most dominant wild herbivore was Grant's gazelle (Nanger granti) (38.88 animals/km2) and was found in L. europeaneum grassland. In farmland, only yellow baboon (Papio cynocephalus) and vervet monkey (Cercopithecus aethiops) were found; thus, this habitat was not included in analysis. Wildlife avoided areas with high human density.
| Species name | Scientific names |
|---|---|
| Africa crested porcupinea | Hystrix cristata |
| Aardvarka | Oryctoropus afer |
| African elephant | Loxodonta africana |
| Bat-eared fox | Otocyon megalotis |
| Black-backed jackala | Canis mesomelas |
| Plains zebra | Equus quagga |
| Buffalo | Syncerus caffer |
| Cheetaha | Acinonyx jubatus |
| Eland | Tragelaphus oryx |
| Warthog | Phacochoerus aethiopicus |
| Wildebeest | Connochaetes taurinus |
| Common duikera | Sylvicapra grimmia |
| Gerenuk | Litocranius walleri |
| Grant's gazelle | Nanger granti |
| Impala | Aepyceros melampus |
| Kirk's dik-dik | Madequa kirkii |
| Leoparda | Panthera pardus |
| Lesser kudu | Tragelaphus imberbis |
| Liona | Panthera leo |
| Maasai giraffe | Giraffa camelopardalis |
| Yellow baboon | Papio cynocephalus |
| Ostrich | Struthio camelus |
| Spotted hyaena | Crocuta crocuta |
| Thomson's gazelle | Eudorcas thomsonii |
| Vervet monkey | Cercopithecus aethiops |
- a Identified from signs
A significant difference was found in mean livestock (sheep (Ovis aries), goat (Capra hircus), cattle (Bos indicus) and donkey (Equus asinus) combined) densities among the four habitats considered (F3, 57 = 3.973, p < 0.05). The density was highest in A. tortilis bushland with 301 ± 69 animals/km2 and lowest in farmland with (91 ± 18 animals/km2), the two habitats that contributed to the significance in differences according to Tukey post hoc test. Most pastoralists were found in the lower belt, probably explaining the high livestock density in grassland, A. tortilis and A. mellifera bushlands.
4 DISCUSSION
4.1 Land use and tenure
This study shows that the Maasai community in the area is slowly getting engaged in cultivation of horticulture products (beans, tomatoes, watermelon, potatoes and sunflower) and maize, with pastoralism becoming a secondary livelihood strategy. Cultivation is considered a more beneficial form of land use according to local community perceptions, as it is a source of food and income for households. More people are likely to engage in this activity, causing its further expansion and resulting in wildlife displacement from their former habitats.
The desire for direct household benefits and alternatives to the less predictable and declining pastoral lifestyle, and availability of market for horticultural products, encourage expansion of cultivation in KWC. Natural habitat is being converted into farms, reducing food availability for wildlife, causing constriction of the wildlife dispersal area and grazing range. This has consequently resulted in increased human–wildlife conflicts and retaliatory killing of endangered wildlife.
Community members are generally poor and often subjected to more suffering when their livestock are killed or their only annual crop destroyed by wildlife without commensurate compensation from either government or other conservation partners. As a result, there is a high resentment for KWS, promoting negative attitudes towards wildlife conservation by local community. According to Mulholland and Eagles (2002), where wildlife compromises the people's livelihoods and adequate solutions to conflicts are not found, local support for conservation efforts is eroded. Furthermore, previous research in Amboseli and adjacent ecosystems by Campbell, Gichohi, Mwangi, and Chege (2000), Ottichilo et al. (2000), Okello and Kioko (2010) and Makindi, Mutinda, Olekaikai, Olelebo, and Aboud (2014) also reported agricultural expansion as a threat to wildlife conservation.
Another important threat to wildlife conservation is the land tenure change from communal to private ownership leading to land subdivision into smaller unsustainable portions and fencing. With about 30% of the study area already under private ownership, further subdivision will lead to these private lands being sold soon and rapidly turned into cultivation. If settlement and cultivation expand to the whole area as will happen if nothing is done urgently, this will curtail wildlife dispersal and movement. Consequently, the corridor will become nonfunctional in facilitating wildlife movement, threatening the ecological integrity of the two globally important parks that depend on it. Western (1997) reported a similar trend in the area around Nairobi National Park, whose prediction has since come to pass. Although no information was available on bushmeat utilisation in the study area, there is anecdotal evidence that subsistence poaching occurs considerably and is attributed mainly to persons coming from outside the local community.
4.2 Vegetation characterisation
That woody vegetation density varied significantly from one habitat to another is due to human activities (cultivation, pastoralism, cutting trees for building material) and wildlife browsing. Despite a very low tree density, farmland had species composition very similar to C. schimperi bushland. If the current trend of clear felling of trees for building materials, fuel wood and to give way to cultivation continues, the C. schimperi bushland will be lost, further causing displacement of wildlife and loss of biodiversity.
The study found that recruitment of saplings to upper age classes was not adequate to ensure maintenance of trees density. Woody vegetation in the corridor is therefore highly threatened. The high density of livestock, elephants and other animals is leading to over-browsing, trampling and uprooting of seedlings/saplings, causing low recruitment. Elephants are particularly known to affect woody species recruitment (Western & Maitumo, 2004). The woody habitat in KWC will be converted into grassland, resulting in disappearance or local extinction of bushland or woodland-dependent animal species, as happened to gerenuk (Litocranius walleri) in Amboseli National Park (Western & Maitumo, 2004).
KWC was dominated by low palatable increaser species, indicating a deteriorating range. Highly palatable grasses are being replaced by less palatable plant species—increasers and forbs which could be seen as invaders taking advantage of disturbance to colonise the area. This trend is explained by the high population of livestock overgrazing the area. Only one decreaser species (Cenchrus ciliaris) was found in the whole area and at a low cover. If overstocking is not urgently controlled, wildlife species utilising this corridor will soon have to move elsewhere due to lack of good grasses. Hobbs and Huenneke (1992) argued that disturbance is known to increase the invasibility of a community and can affect the community structure and functions. A similar trend was also reported by Njenga (2007) who found that highly palatable decreaser species were being replaced by increaser at Il Ngwesi community conservancy in Laikipia District.
4.3 Wildlife distribution and composition
The twenty-two mammal species identified in the study area can be generally considered a high number and compares favourably to the twenty-six species reported by Okello (2005) in Maasai Kuku Group Ranch. Such species richness was the result of a mosaic of habitats but due to high concentration of humans and their associated activities, habitat heterogeneity is affected resulting in decline of ungulates species. The study has shown that wildlife was displaced by human encroachment. Okello (2009) reported similar factors as causing range contraction and wildlife displacement in Kimana Group Ranch.
Overall, it is clear that the viability of Kitenden Wildlife Corridor is highly threatened by human-related activities including expansion of settlement and cultivation, overstocking/overgrazing and land tenure change. These threats must be addressed urgently if the corridor is to be saved to ensure the integrity of the two parks it connects is preserved. It is, therefore, recommended that appropriate action be taken to minimise negative impacts, which should include re-seeding with decreaser species. A broadly negotiated land use plan and land lease programme between landowners and conservation partners are urgent priorities, with the local community being empowered to participate in the conservation. Enhanced conservation awareness and enacting of a transparent benefits sharing system would enable them view wildlife as a viable land use option.
ACKNOWLEDGEMENTS
We are grateful to the African Wildlife Foundation (AWF) for financial and logistic support that permitted the realisation of this work. Thanks to Daniel Moonka and Julius Kayiai who assisted with the fieldwork, all staff and graduate students of the University of Nairobi for providing an enabling and academic environment.
