Breaking News
Home » Research » Predators » Results Of A Motion Camera Survey In Jwana Game Park, Jwaneng, Botswana

Results Of A Motion Camera Survey In Jwana Game Park, Jwaneng, Botswana

Jwana Game Park is a small game reserve (180km2) located in the Southern district of Botswana. The reserve surrounds the Jwaneng Diamond mine and is owned and managed by the Debswana diamond company. It is surrounded by a game fence and
stocked with small and large game. Un-stocked mammal species including predators freely move between the reserve and the surrounding communal farming areas, by utilising holes under the game fence. The reserve has always been an important site for
the conservation of large predators such as cheetah and leopard and was believed to form a refuge for female cheetahs to raise their cubs (Houser pers comm.). Previous spoor surveys in 2006 had estimated cheetah spoor densities as 2.32 cheetah spoor per 100km2 (Houser, Somers & Boast 2009b). When calibrated (Funston et al. 2001) this was equivalent to 0.75 cheetahs per 100km2, higher than the estimate of 0.57 per 100km2 found in the Kgalagadi Transfrontier Park (KTP) (Funston et al. 2001). However, a reduction in cheetah sightings since completion of the study had raised concerns that the cheetah population may have decreased within the area.

A motion camera survey was conducted in October – December 2010 to estimate current predator populations, with the aim of conducting future follow up studies to determine trends in large predator numbers. Motion camera surveys have been widely
used worldwide to estimate large felid populations. The cameras are triggered remotely by movement sensors as an animal crosses its field of view. Species such as the cheetah, leopard, African wild dog Lycaon pictus and brown hyena have individually recognisable coat patterns enabling the identification of individuals. Using a capture-mark and recapture framework it is possible to estimate species abundance and density.

Survey Design
Cuddeback capture motion cameras were placed in pairs at nine locations throughout the reserve for 60 days (Figure 1).  Locations were chosen based upon cheetah signs, such as tracks or scat recorded in 2006 – 2007. Cameras were placed on either side of roadways or water points to photograph both flanks of the animal for accurate identification. Cameras were offset by 1 – 2 m to prevent interference from the opposing camera flash. No bait or lures were used and obstructions such as grass or branches were removed from the site. Cameras were mounted on wooden poles, fence posts or trees at approximately 50 cm height. Branches from various acacia species were placed around the camera in an attempt to prevent animal interference. The mean distance to the next nearest camera station was 3.84 ± 1.46 km. No gaps within the sample area were large enough to contain a cheetah’s home range, based upon data from the smallest home range recorded in Botswana (241km2, radius = 8.8km, (Houser, Somers & Boast 2009a), thereby ensuring every individual had an opportunity for capture. Cameras were active 24 hrs a day, with a 30 second delay between photos. They were checked every 5 – 12 days.

Photographs of all species captured were stored and identified on Camerabase 1.3 (Tobler 2007) for storage and analysis.  Estimates of species abundance were calculated based upon the number of captures per 100 camera trap nights. Captures were considered independent if recorded more than one hour since the previous capture of that species. Any photographs taken of the same species within the hour (with the exception of individually recognisable species) were assumed to be the same individual and were discarded. Individually recognisable species such as cheetah, leopard and brown hyena were identified for capture-mark-recapture analysis. The sex of the individual was determined by the presence or absence of testicles. Abundance was calculated using both the Bayesian spatially explicit capture – recapture (SECR) model and the more commonly used capture-mark-recapture method with the program CAPTURE. Although the SECR model is thought to be more applicable and accurate for small sample numbers (Royle et al. 2009a, Royle et al. 2009b), both techniques were conducted for comparison purposes. Both adults and cubs were included in the analysis.

Read more: Download PDF Document

About Peter Katt

Leave a Reply

Your email address will not be published. Required fields are marked *