Distribution and habitat requirements of Aders duiker (Cephalophus adersi) in the coastal forests of Kenya

Transcription

1 Distribution and habitat requirements of Aders duiker (Cephalophus adersi) in the coastal forests of Kenya Principal investigators: Wayne Edwards 1, 2 and Simon Tollington 2 1 School of Biological Sciences University of Plymouth Plymouth PL4 9AA UK Report compiled by: Andrew Bowkett 2 and Amy Plowman 2 2 Whitley Wildlife Conservation Trust, Paignton Zoo Environmental Park, Totnes Road, Paignton, Devon TQ4 7EU, UK Correspondence: andrew.bowkett@paigntonzoo.org.uk Affiliating institution: Kenya Wildlife Service Local contact: Erustus Kanga, KWS

2 Project background The Whitley Wildlife Conservation Trust (Paignton Zoo Environmental Park) has funded and undertaken research and surveys for Aders duiker since 2002 in collaboration with Erustus Kanga, Kenya Wildlife Service. Schrodt (2005) quantified potential duiker habitat in the Arabuko-Sokoke Forest in terms of plant species composition, vegetation structure and disturbance factors. She found that habitat plots located close to recent Aders duiker sightings (post-1999) were mostly within Cynometra vegetation and associated with high food plant abundance, low visibility and low elephant disturbance. Indirect field sign (dung pellets and paths) showed similar relationships with these habitat variables. However, these conclusions were limited by small sample size and an inability to distinguish Aders duiker field sign from that of other small antelope species. These are common problems in studying threatened duiker species, which are normally shy and occupy dense habitats (Bowkett et al. 2006). In 2006, Neelakantan & Jackson (2007) investigated whether camera trapping would be a suitable technique for surveying Aders duiker. Ten camera traps were set up within the Arabuko Sokoke Forest for a mean of 62.6 days (total trapping effort was 626 days). Cameras were set up close to areas where Aders duiker had been sighted. Seventeen animal species were recorded in total including Aders duiker. These are possibly the first photographs of Aders duiker in the wild (Anon 2007). This report documents our efforts to establish the distribution of this species in Kenya outside of of its known range. 2

3 Introduction The critically endangered Aders' duiker Cephalophus adersi is restricted to a very few fragmented forests in Zanzibar and on the coast of mainland Kenya (Finnie 2004) and has been referred to as Africa's most endangered antelope (Baldus 2004). This species is endemic to the coastal forests of eastern Africa, an internationally recognised hotspot for exceptional biodiversity and species endemism (Mittermeier et al. 2005). On Zanzibar, Aders duiker has undergone a dramatic decline in response to hunting and habitat loss. The population size was estimated at 614 ± 46 at the last survey in 1999 (Kanga & Mwinyi 1999). The species is now probably restricted to three forests on Unguja Island (Finnie 2004) although small numbers have been translocated to offshore islets as a conservation measure (Mwinyi & Wiesner 2003). As recently as 1982, Aders duiker was described as widespread in forest and woodland along the Kenyan coast north of Mombassa almost up to the Somali border (Kingdon 1982). By the mid-1990s the species was thought to be restricted to one site, the Arabuko-Sokoke Forest (ASF), and to be extremely rare. Sightings reported during research by Erustus Kanga from 1999 confirmed the continued existence of Aders duiker in Kenya (Kanga 2002). Since that time infrequent sightings continue to be recorded by researchers and forest guides and the first photographic evidence has been obtained by Neelakantan & Jackson (2007; Anon. 2007). A reliable sighting from Dodori Forest (c. 250 km north of ASF) was reported in 2004 raising the possibility that viable Aders duiker populations may persist on the mainland outside of the Arabuko-Sokoke Forest (Andanje & Wacher 2004). Therefore, this study aimed to survey the remaining coastal forests of Kenya for Aders duiker using heat and motion triggered camera-traps. In addition, forest habitat was described in terms of vegetation structure, plant diversity and disturbance in an attempt to explain the observed distributions of photographed species. 3

4 Methods Study site The coastal forests of eastern Africa represent a distinct vegetation type with high biodiversity value in terms of species richness and endemism (Burgess & Clarke 2004, Mittermeier et al. 2005). The coastal forests of Tanzania and Kenya form part of the Zanzibar-Inhambane Regional Mosaic (White 1983) and the Northern Zanzibar-Inhambane Coastal Forest Mosaic eco-region (WWF-US 2003). The major historical and ongoing threats to this ecosystem are deforestation and habitat fragmentation due to clearance for agriculture and settlement and unsustainable use of forest resources such as charcoal (CEPF 2003). The internationally important biological value and ongoing threats to these forests have lead to their designation as part of the Coastal Forests of East Africa biodiversity hotspot by Conservation International (Mittermeier et al. 2005). Ongoing security issues prevented us surveying the adjacent Boni and Dodori National Reserves, Lamu District. Andanje & Wacher (2004) documented the presence of Aders duiker in Dodori forest but further surveys of this area are a conservation priority for this species. Although Aders duiker is already known to occur in the Arabuko-Sokoke Forest we deployed camera-traps in the southern section not previously surveyed by Neelakantan & Jackson (2007). Camera-trap surveys were conducted in six forest areas: 1. Kinondo: Kaya Kinondo (4 23` S, 39 32` E) is a small sacred forest (approx. 0.3 km 2 ) south of Mombassa in Kwale District. 2. Chale: Kaya Chale (4 26` S, 39 31` E) is approximately 7 km south of Kinondo, Kwale District. Also a small sacred forest (approx. 0.5 km 2 ). 3. ASF: The Arabuko-Sokoke Forest is situated in Kilifi and Malindi Districts (3 20` S, 39 55` E). At c. 416 km 2 this forest is the largest remaining block of tropical coastal forest in east Africa. The forest vegetation has been classified into three main habitat types dictated by local soil conditions: Cynometra 4

5 forest and thicket, Brachystegia woodland and mixed forest (KIFCON 1995). Other small antelope species found in the ASF include Harvey s duiker C. harveyi, blue duiker C. monticola, grey or bush duiker Sylvicapra grimmia and suni Neotragus moschatus. (Fitzgibbon et al. 1995, Neelakantan & Jackson 2007). The Kipini Wildlife and Botanical Conservancy Trust own a large area of natural coastal habitat mosaic (Nairobi Ranch; 220 km 2 ) situated to the north-east of the Tana River Delta in the area between the towns of Kipini, Witu and Mpeketoni, Lamu District. This area includes numerous forest patches and coastal thicket that are potential habitat for Aders duiker (Wacher & Andanje 2004). 4. Lumshi: Lumshi Forest (10-15 km 2 ) is found in the north-west of Nairobi Ranch (2 24` S, 40 34` E) and is continuous with Witu Forest. 5. Bahareni: Within a continuous area of coastal thicket (approx. 49 km 2 ; 2 31` S, 40 37` E). 6. Kangawati: Forest patch near the centre of Nairobi Ranch (2 27` S, 40 35` E). Camera-trapping Heat and motion, infra-red triggered camera-traps (Stealthcam, USA) were deployed during July to September 2007 (see Table 1 for camera-trap sampling details). Cameras were set to take pictures 24 hours a day on 400 ASA colour print film, with a 1 minute delay between exposures. Cameras were kept in the field for a mean of 10.3 nights per forest (we use trap nights as our unit of time rather than actual 24 periods as accurate records of camera removal times were unavailable). 5

6 Table 1. Camera-trap sampling details for six forest areas on the coast of Kenya. Forest Sampling period No. of cameras No. of trap-nights Kinondo 02-09/07/ Chale 09-21/07/ ASF 25/07-13/08/ Lumshi 30/08 10/09/ Bahareni 17 22/08/ Kangawati 22 30/08/ Camera-traps sites were located on vehicle tracks (ASF and Bahareni only) or animal trails through the forest. These locations were selected to maximise the probability of detecting medium to large-sized forest mammals using trails to traverse dense vegetation. Cameras were set at a height of 30 to 60 cm and positioned at distance from, and angle to, the trail with the aim of obtaining full-body lateral images of small antelope. For any series of photographs of the same species the first picture was considered an independent capture event but subsequent pictures were excluded up to a maximum period of 1 hour. This was a compromise between scoring the same individual multiple times and missing individuals, following Bowkett et al. (2008). Camera-trap capture rate was calculated across all locations as the number of independent photographs of a species per trap-night (expressed as a percentage). The camera model used in this study only prints the date or time (24 hour format) on each exposure. We chose to record the time to allow the identification of independent events (see above). Careful inspection of the photographs combined with regular checks of cameras whilst in the field enabled us to calculate accurate trap-rates despite the date not being printed on the photographs. 6

7 Habitat description Habitat variables were quantified in 10 m 2 quadrat plots located every 50 m along linear transects. Transects were positioned to evenly sample the available area of closed canopy forest at each study site. Distances and bearings were measured using a 250 m tape measure and compass and verified with a Garmin e-trex GPS unit. Woody plants species (trees and shrubs) were identified by a locally-based botanist at each study site (see Acknowledgements). Woody plant abundance refers to the number of individuals within a plot whereas species richness refers to the number of species. Shannon species diversity and evenness were calculated for each plot using the Diversity Index Calculator, an MS-Excel add-in, provided by the University of Reading (available at Average canopy height was estimated visually. Substrate depth was measured at the centre of each plot. A visibility index was calculated for each plot by counting the number of visible squares (out of eighteen) on a 1 by 1 m checked board at a distance of 7 m from the centre of the plot. This was repeated four times at different bearings to give a mean value for each plot. This method was taken to incorporate the extent of horizontal vegetation growth not measured by stem density. The number of cut woody plant stems per plot was recorded as an index of human habitat disturbance. Elephant disturbance was quantified as the number of elephant trails and/or dung piles per plot. Plant species assemblages and species dominance were analysed by Wayne Edwards for Kinondo, Chale and Lumshi and are described in Edwards (2008; submitted as part of this report). Several variables were not normally distributed. Therefore, non-parametric statistics were used to test for differences in the mean value of habitat variables between forest areas (Kruskal-Wallis test) and for relationships between habitat variables and number of species photographed (Spearman rank correlation). Statistical analysis was undertaken using the software SPSS 16.0 (SPSS Inc., Chicago, IL, USA). 7

8 Results Camera-trapping A total of 70 independent photo-capture events of eleven animal species were recorded across all forest areas (Table 2). No Aders duiker records were obtained from any of the selected study sites. The most frequently recorded species was the yellow baboon Papio cynocephalus (27 independent events) followed by small antelope species, Harvey s duiker and suni (six each). Hippopotamus Hippopotamus amphibius and Gambian pouched rats Cricetomys gambianus also had relatively high camera-trap rates but these species were only recorded from one forest area each. Bahareni, the costal bush area in Nairobi Ranch, had the highest number of species recorded (10) with only two species recorded at Chale and Kangawati. Number of species recorded per forest was not correlated with any of the habitat variables (P >0.05). Habitat description There were highly significant differences between forest areas for all recorded variables (P <0.001; Table 3). This illustrates the diversity of conditions and habitat types found within east African coastal forests. See Edwards (2008) for a more detailed comparison of Kinondo, Chale and Lumshi. Generally, too few photo-capture events (see Discussion) were obtained to test for relationships between camera-trap rates and habitat variables within or between forest areas. For example, there were no significant correlations between the capture rate for yellow baboon, the most commonly encountered species, and any of the recorded habitat variables, probably due to the small sample size (data not shown). 8

9 Table 2. Camera-trap capture rates (independent capture events per trap-day x 100) for mammal and bird species from six forest areas on the coast of Kenya. Number of independent photo-capture events in parentheses. See Appendix 1 for scientific names. Kinondo Chale ASF Lumshi Bahareni Kangawati Harvey s duiker 5.71 (4) 0.83 (1) (1) 0 0 Suni 1.43 (1) (1) 8 (4) 0 African buffalo (1) (1) Hippopotamus (5) 0 Bushpig 1.43 (1) (2) 0 0 Yellow baboon 7.14 (5) 7.50 (9) 3.86 (7) 0.91 (1) 8 (4) 1.25 (1) Leopard (2) 0 Spotted hyena (1) 0 White-tailed mongoose (1) Mongoose spp (1) 0 African civet 2.86 (2) (1) 0 Crested porcupine 1.43 (1) (1) 2 (1) 0 Gambian pouched rat Crested guineafowl 7.14 (5) (1) 1.82 (2) 2 (1) 0 Hadada ibis (1) 0

10 Table 3. Habitat and disturbance variables for six forest areas on the coast of Kenya. Values are means with standard error (in italics). Differences in the mean between forest areas were tested for using the Kruskal-Wallis test (χ). Canopy (m) Visibility (%) Substrate (mm) Woody plant abundance Species richness Shannon diversity (H) Shannon evenness Human disturbance Elephant disturbance Kinondo (n = 30) Chale (n = 59) ASF (n = 81) Lumshi (n = 60) Bahareni (n = 30) Kangawati (n = 25) < < < < < < < < out of 59 plots* <0.001 * Elephant disturbance recorded as presence / absence only. χ P 10

11 Discussion The primary aim of this study was to establish the distribution of the critically endangered Aders duiker outside of its documented range in Kenya. Given that no new records of Aders duiker were obtained it is worth considering whether our results reflect a genuine absence of the species in the surveyed forests. While a lack of camera-trap pictures may suggest the absence of a species it cannot confirm it. The main limitation of our survey results is the short sampling period at each forest (5-19 days). While survey effort is always a compromise between the length of time spent at each location and the number of locations surveyed it is clear in this case that cameras were not left in place for long enough to detect rare species. This is reflected in the low number of independent photo-capture events and the low overall number of species recorded at any one site. For example, while this study only recorded three species from the ASF a longer camera-trap survey (mean trapping effort of 62.2 days per camera) recorded seventeen species including Aders duiker and other rare mammals such as the golden-rumped elephant shrew Rhynchocyon chrysopygus (Neelakantan & Jackson 2007). The small number of records also prevented an analysis of relative abundance or habitat use as undertaken in other camera-trap studies (O Brien et al. 2003, Bowkett et al. 2008, Rowcliffe et al. 2008). Therefore, we were also unable to fulfil our secondary aim of relating camera-trap records to habitat variables. Camera-trap rates for Harvey s duiker in Tanzania only reach an asymptote after approximately 30 days (Rovero & Marshall submitted) and this may represent a useful guideline for future surveys. Despite the limitations, our results highlight the importance of Kenya s remaining coastal forests for conservation. Results from the coastal habitat mosaic of Nairobi Ranch illustrate how forest patches act as important refuges for savannah woodland species including elephants and large predators. In Bahareni our survey adds five mammal species to the seven recorded in the area s coastal thicket by Wacher &

12 Andanje (2004). One of these is a currently unidentified mongoose which is not the previously recorded white-tailed mongoose (Ichneumia albicauda). The botanical importance of kaya forests such as Kinondo and Chale has been recognised elsewhere (Robertson & Luke 1993) and is reflected here in high species diversity scores (Table 2) and the presence of globally threatened species endemic to east Africa s coasts (Edwards 2008). However, our results also show how these forests also provide important habitat for medium to large-sized mammals despite their small size. The protection afforded these forests by local communities may allow cover-dependent and/or hunted species such as Harvey s duiker to persist within the wider landscape despite habitat fragmentation and high human population density. Our results do not allow us to rule out the presence of Aders duiker in the surveyed forests. However, based on personal observation and informal discussion with local communities and wildlife workers we feel the species is unlikely to persist south of Mombassa, is present but rare in the ASF and possibly present in the Nairobi Ranch area subject to further survey efforts. Given that a single sighting in Dodori National Reserve remains the only recent record of Aders duiker outside of the ASF (Andanje & Wacher 2004) the clear priority for future survey work is the Boni-Dodori area near the border with Somalia. It is hoped that this area will be subject to increased conservation effort from all stakeholders in the near future. Acknowledgements This project would not have been possible without the identification skills and field assistance of botanical experts at each study site. We are therefore very grateful to Saidi Ali Chidzinga (NMK Office, Ukunda) in Kaya Kinondo and Chale, Willington Kombe (ASF bird guide and research assistant) in Arabuko-Sokoke and Japhet (Bakari) Kakawa (KWBC) in Nairobi Ranch. We would also like to thank the following organisations and individuals: Kenya Wildlife Service and the Ministry for Education for permission to conduct research; Dr. Farouk Sherman for permission and assistance with surveying within the Kipini Wildlife and Botanical Conservancy 12

13 (KWBC); all other KWBC staff especially Ben, Aden, Big John, Masai John, Kidonga and Chef; the staff of the ASF and the Mwamba Arocha Field Centre; Ann Robertson; Bernard (Risky) Agwanda at National Museums Kenya; the elders and communities of Kaya Kinondo and Chale for permission to work in those forests. Special mention should go to Erustus Kanga for his help and great enthusiasm for Aders duiker research and Natasha de Vere. This project was funded by the Whitley Wildlife Conservation Trust. References Andanje, S. & Wacher, T Aders duiker in Dodori Forest Reserve, North Kenya Coast. Gnusletter (Antelope Specialist Group) 23: 4-5. Anon First sight. Current Biology 17: Baldus, R Africa s most endangered antelope lives in Zanzibar. Gnusletter (Antelope Specialist Group) 23: 5-6. Bowkett, A.E., Lunt, N., Rovero, F. & Plowman, A.B How do you monitor rare and elusive mammals? Counting duikers in Kenya, Tanzania and Zimbabwe. In Animals, Zoos and Conservation. E. Zgrabczyñska, E., Ćwiertnia P. & Ziomek, J., editors. Zoological Garden in Poznan, Poland Bowkett, A.E., Rovero, F. & Marshall, A.R The use of camera-trap data to model habitat use by antelope in the Udzungwa Mountain Forests, Tanzania. African Journal of Ecology 46: Burgess, N.D. & Clarke, G.P Coastal Forests of Eastern Africa. IUCN, Gland, Switzerland and Cambridge, UK. 13

14 Edwards, W.M A biodiversity evaluation of Kenya s coastal forest systems. BSc Thesis, University of Plymouth. Finnie, D Cephalophus adersi. In IUCN IUCN Red List of Threatened Species. < Downloaded on 21 September Fitzgibbon, C.D., Mogaka, H., Fanshawe, J.H., Subsistence hunting in Arabuko- Sokoke Forest Kenya, and its effects on mammal populations. Conservation Biology 9: Kanga, E.M., Some ecological aspects and conservation of duikers in Arabuko- Sokoke Forest, Kenya. MSc Thesis, University of Nairobi. Kanga, E. M. & Mwinyi, A. A Population survey of Zanzibar mini antelope. Unpublished report, Commission for Natural Resources, Zanzibar. KIFCON (Kenya Indigenous Forest Conservation Programme) Arabuko-Sokoke Forest and Mida Creek: The Official Guide. Majestic Printing Works Ltd., Nairobi, Kenya. Kingdon, J East African Mammals: An Atlas of Evolution in Africa. Vol. IIIC. Academic Press, London, UK. Mittermeier, R. A., Gil, P. R., Hoffman, M., Pilgrim, J., Brooks, T., Mittermeier, C. G., Lamoreux, J. & da Fonseca, G. A. B Hotspots revisited: earth's biologically richest and most threatened terrestrial ecoregions. Conservation International, Washington, D.C., USA. Mwinyi, A. A. & Wiesner, R Aders duiker (Cephalophus adersi): hunt, capture and translocation in Zanzibar. In Ecology and Conservation of Small Antelope. Plowman, A. editor, Filander Verlag, Fürth, Germany

15 Neelakantan, A. & Jackson, C Abundance, distribution and habitat requirements of duiker (Cephalophus spp.) in the Arabuko-Sokoke Forest, Kenya: A pilot study using camera-traps to record the critically endangered Aders duiker Cephalophus adersi. Unpublished report to Kenya Wildlife Service. Whitley Widlife Conservation Trust, UK. O'Brien, T.G., Kinnaird, M.F. & Wibisono, H.T Crouching tigers, hidden prey: Sumatran tiger and prey populations in a tropical forest landscape. Animal Conservation 6: Robertson, S.A. & Luke, W.R.Q Kenya Coastal Forests: the report of the NMK/WWF Coast Forest Survey. National Museums of Kenya, Nairobi. Rovero, F. & Marshall, A.R. Submitted. Camera trapping photographic rate as an index of density in forest ungulates. Journal of Applied Ecology Rowcliffe, J.M., Field, J., Turvey, S.T. & Carbone, C Estimating animal density using camera traps without the need for individual recognition. Applied Journal of Ecology 45: Schrodt, F Estimating distribution and habitat associations of the endangered Aders duiker (Cephalophus adersi): Implications for research and conservation. MSc Thesis, University of Plymouth. Wacher, T. & Andanje, S The potential role of Kipini Wildlife and Botanical Conservancy in the conservation and development of wildlife and natural resources; Lamu, Ijara and Tana Districts, Kenya. Kipini Wildlife and Botanical Conservancy Trust. White, F The vegetation of Africa. A descriptive memoir to accompany the UNESCO/AETFAT/UNSO Vegetation Map of Africa. Paris, UNESCO 15

16 WWF-US Ecoregional reports: Northern Zanzibar-Inhambane coastal Forest Mosaic. Eastern and Southern Africa Bioregions. J. Schipper & N. Burgess, authors. 16

17 Appendix 1. Class, order, English and scientific names of mammal and bird species photographed using camera-traps in six forested areas in coastal Kenya, Taxonomy follows IUCN Red List Accessed 23 rd November Class Mammalia Aves Order (English name) Cetartiodactyla Genus Harvey s duiker Cephalophus harveyi Species Suni Neotragus moschatus African buffalo Syncerus caffer Hippopotamus Hippopotamus amphibius Bushpig Potamochoerus larvatus Primates Yellow baboon Papio cynocephalus Carnivora Leopard Panthera pardus Spotted hyaena Crocuta crocuta White-tailed mongoose Mongoose spp. Ichneumia albicauda African civet Civettictis civetta Rodentia Crested porcupine Hystrix cristata Gambian pouched rat Ciconiiformes Cricetomys gambianus Hadada ibis Bostrychia hagedash Galliformes Kenya guineafowl crested Guttera pucherani 17

18 Appendix 2. Camera-trap images obtained by the authors during coastal forest surveys in Kenya. Leopard, Bahareni, Nairobi Ranch (Kipini Wildlife and Botanical Conservancy). 21 st or 22 nd August, Hippopotamus, Bahareni, Nairobi Ranch (Kipini Wildlife and Botanical Conservancy). 20 th August,

19 Harvey s duiker, Lumshi, Nairobi Ranch (Kipini Wildlife and Botanical Conservancy). 30 th August 7 th September, Suni, Bahareni, Nairobi Ranch (Kipini Wildlife and Botanical Conservancy). 17 th 21 st August,