Botswana Leopard Scent and Spatial Study

The overall aim of this project is to aid the development of scent- and acoustic-based leopard deterrents that could be used to reduce human-leopard conflict. This project has been developed in collaboration with the Botswana Predator Conservation Trust (BPCT; www.bpctrust.org), an NGO tasked by the Government of Botswana in leading large carnivore research initiatives in Northern Botswana.

Country: Botswana

Leopards are economically and ecologically important but they have disappeared from an estimated 48 – 67% of their range in Africa and their IUCN conservation status was recently downgraded to Vulnerable.

Human persecution is a key driver of carnivore declines with most mortality occurring when carnivores come into contact with humans outside of reserve boundaries. Unfortunately current strategies to mitigate human-carnivore overlap, such as fences and translocations, have low efficacy for leopards; there is little innovation in mitigating human-leopard conflict; and there is a deficiency in leopard research with conservation applications.

This project addresses all of these shortcomings and its output will directly aid leopard conservation. Supporting the development of tools that reduce the potential for conflict by restricting leopards from areas where they are likely to be persecuted will help stem the species’ decline.

1. To quantify whether there is spatial avoidance between leopards that could be exploited for conservation purposes. Six territorial leopards to be radio-collared with GPS-loggers.

2. To determine the role of scent marks and vocalisations in mediating interactions between leopards:
• Set camera trap stations at leopard scent marks
• Record leopard vocalisations from collar-mounted audio recorders

3. To field-test a scent signal (tomcat-thiol) that could be used to repel leopards from entering undesirable areas. Camera traps to be set at tomcat-thiol scent stations in leopard territories.

The Trust funded radio collars, veterinary fees and drop-off units for this project.

Study site
This project focuses on a protected, undisturbed population of leopards in Botswana’s Okavango Delta. This is an ideal study location because detailed life history data is available for the area’s leopards, and human-leopard conflict takes place in areas surrounding the study location. Data from a protected population is being collected, as a sound understanding of the processes aiding spatial separation in an undisturbed group is needed before considering how these change from the norm in disturbed populations.

Leopard radio collars
VHF radio collars are used, which are attached with commercially available GPS and audio loggers that can record up to eight months and 30 days of data, respectively. All radio collars have timed mechanical drop-off units (set for July 2017). GPS-operating instructions can be uploaded and data downloaded remotely by handheld base stations when within sight of radio collars.

Leopard capture
Six territorial leopards (three males; three females) with overlapping and adjacent territories have been identified; these will be fitted with the radio collars (Figure 1) .

Spatial avoidance between territorial leopards
GPS-loggers will be initially configured to take fixes at five-minute intervals. This data will be used to define home ranges, quantify home range overlap, and quantify fine-scale leopard space use. These results will then be used to identify whether there is spatial separation between leopards (i.e. as discrete home ranges or as the use of shared areas at different times). After four months of data collection, the fix rate for each logger will be reduced to one fix every 60 minutes to increase battery life.

Vocalisations and competitor avoidance
Collar microphones will begin recording 24 hours after leopard capture and record continuously for 30 days. Data will be stored on the device and collected after collar drop-off.

Leopard vocalisations will be processed using a programme that can be trained to detect and classify leopard calls. Audio data will be combined with GPS data to identify where and when leopards are vocalising, which will then be compared with potential indicator variables, such as location within home range and proximity to radio-collared leopards, to infer reasons for vocalisations.

If the results suggest that acoustics deter territorial leopards from entering neighbouring territories, the deterring calls will be analysed to identify features of the calls that leopards avoid. Vocalisations will then be field trialled in future studies through playback experiments.

Scent marks and competitor avoidance
Scent-marking data collected from focal follows (leopards in the study area are habituated to vehicles) will be used to quantify the spatial distribution of scent marks. Camera traps will be placed at fresh scents and scent samples collected. Videos from traps will be used to investigate whether there is avoidance or attraction to certain scents. Avoided scents will be analysed by the BPCT laboratory and chemical components common between avoided scents will be field trialled in future studies as possible leopard deterrents.

Tomcat-thiol as a leopard deterrent
Preliminary work by the BPCT bio-boundary laboratory has identified tomcat-thiol as a potential leopard deterrent. Camera traps will be deployed at scent stations for one month at 20 different trails frequented by leopards. In contrast to previous work, camera positioning will allow the detection of leopards that don’t reach the dispenser, and stations will be placed across the ranges of a larger sample of leopards.

Scent dispensers will be empty for the first month as controls and then filled with tomcat-thiol. Encounter rates for control and tomcat stations will be compared to identify whether leopards avoid tomcat-thiol. Data from GPS-loggers will be used to investigate leopard space use after encountering stations.

Annual Report 2017
Over the past few months, four of the targeted six leopards have been radio-collared with high-resolution GPS-tracking tags. Two of these radio collars also have collar-mounted audio recorders that will provide insights into where and when leopards vocalise and the role played by vocalisations in mediating interactions. The data will be retrieved and downloaded once the collars drop off the animals in mid-June.

In order to investigate the responses of large predators to leopard scents, camera traps are being used to record the leopards’ scent markings. The majority of this data will be collected during the dry season when the weather and vegetation are better suited for following the cats.

In January the control phase of the project began, testing the suitability of tomcat-thiol as a leopard deterrent. Five camera-trap scent stations have been placed on roads within the territories of the radio-collared leopards. Two scent dispensers have been placed within the centre of each station on either side of the road – currently empty at this point to allow the project to quantify baseline camera capture rates of predators and baseline responses to empty scent stations. Once two months’ worth of camera trap videos have been collected at each site, vials of tomcat-thiol will be placed within the dispensers to see if whether leopards and other predators show avoidance behaviours to the scent, or if capture rates decrease.

December 2016
Camera traps placed at pre-baited cage traps were used to decide when cages should be armed, i.e. when one of the six targets were regularly visiting. Armed traps were checked three times per day (noon, dusk and dawn) and a Botswana-registered veterinarian was on standby in camp, ready to immobilise and radio collar target leopards.