From big spots to little spots: Influence of camera trap deployment on spatial capture-recapture estimates of servals (Leptailurus serval) in Ithala Game Reserve

Servals (Leptailurus serval) face a range of threats which can impact their populations, but we have little information on their conservation status across much of their range. Repeated population density estimates are the most useful parameter for assessing population trends and the impacts of anthropogenic changes (e.g. habitat loss and poaching) on serval densities. These could further be used to establish a correlation between any changes in this population and relevant highlighted anthropogenic influences that may exist with relevance to their conservation vulnerability. However, such surveys for small cryptic carnivores are rare, largely because funding and hence research is heavily biased towards large, charismatic and threatened species. Fortunately, servals and other mesocarnivores are frequently recorded as by-catch in camera trap surveys designed for larger carnivores which offers a unique opportunity to explore the viability of using these 'bi-catch' data for the determination of population estimates of cryptic carnivores. Spatial capture-recapture models are the most robust means of estimating the densities of individually identifiable species like servals. In this study, I investigate whether the ongoing leopard (Panthera pardus) surveys in Ithala Game Reserve can be used to accurately estimate serval density and thus provide the first long term assessment of serval population trend within a protected area in South Africa. To achieve this, I designed a camera trap array to estimate serval density specifically (i.e. smaller intertrap distances and the inclusion of wetland habitat) and ran it simultaneously with a less intensive survey designed to estimate leopard population density in Ithala. The leopard array produced an estimate of 1.73 ± 0.80 (0.76-3.97) servals/100 km2 compared to an estimate of 2.49 ± 0.81 (1.24-4.63) servals/100 km2 from the serval array. In line with standard analysis of the results, the approximately 75% overlap in the 95% confidence intervals suggests the two density estimates are comparable. The inclusion of vehicle traffic (as a measure of anthropogenic disturbance) and vegetation (as a proxy for habitat suitability) as covariates did not improve the serval specific density estimate. Based on these findings I proceeded to use the long-term leopard survey data to produce annual density estimates for serval over a seven-year period (2013-2019). Serval density has decreased from the high of 9.66 (± 2.1) servals/100 km2 recorded in 2014 to a low of 1.42 (± 0.6) in 2018. A similar decline was evident in the leopard density estimates, suggesting that both these two carnivore species are facing some form influence that is threatening their population numbers in Ithala. Recent social surveys in nearby neighbouring communities reveal that snaring and hunting with dogs are both common methods of illegal hunting and such activities may be greatly facilitated in the northern section of Ithala owing the absence of a boundary fence. This study suggests that serval density can be reliably estimated using data collected as part of ongoing leopard surveys in protected areas throughout South Africa. Given the paucity of such data the approach used in this study should be expanded to provide a more comprehensive assessment of serval population status and the generality of the finding that serval density is declining within a protected area previously considered to be a stronghold for this species.

Grantor

Department of Biological Sciences

Year dc:date.issued

2020

Author dc:creator

• Taylor, Johanna

Advisors dc:contributor.advisor

• O'riain, Mannus Justin
• Mann, Gareth

• Conservation Biology

Handle dc:identifier.uri

http://hdl.handle.net/11427/33049

OAI identifier oai:identifier

oai:open.uct.ac.za:11427/33049