An evaluation of two methods of assessing feral cat and dingo abundance in central Australia

1. What they did

Small and medium-sized carnivores are hard to count: they are shy, secretive, often solitary and usually live at low densities. In Australia there is interest in monitoring both feral cats, which are believed to harm native wildlife, and dingoes, because of possible impacts on native wildlife and on the pastoral industry. The authors compared two ways of measuring relative abundance (how numbers change over time or between places) at a semi-arid rangeland site north-west of Alice Springs, from July 1995 to October 1997.

• Spotlight surveys. About every three months, observers drove station roads at night at 10 to 15 km per hour, scanning with a 100-watt spotlight from the back or roof of a four-wheel-drive vehicle. Each survey had 5 to 7 counts over 14 days, covering 107 to 178 km in total. They took about 42 person-hours per survey.
• Track surveys. Four permanent 10 km transects were set on sandy dirt roads, at least 5 km apart. A day before counting, each transect was cleared by dragging a heavy steel bar behind a vehicle. Next morning, two observers rode All-Terrain Vehicles (ATVs) at 5 to 8 km per hour, stopping at each fresh set of carnivore tracks to inspect and identify them. They recorded where the tracks were, how far they stayed on the road, whether the animal walked along or across the road, whether the prints were small or large, whether one or more animals made them, and the habitat. The road was dragged again for the next day’s count. Counts ran for 2 to 3 days and took about 18 person-hours per survey.
• Two track indices. Track Index 1 counted every set of tracks, giving the maximum number of animals that could have made them. Track Index 2 estimated the minimum number of individuals: cat tracks less than 500 m apart were treated as one cat, and dingoes were counted from the most sets of tracks seen travelling side by side (less than 1.5 m apart).

2. What they found

Feral cats

• Few cats were seen by spotlight, and the counts varied hugely (coefficients of variation 68 to 224%).
• Track surveys recorded 5 to 55 times more cats per kilometre (minimum index) and 8 to 114 times more (maximum index) than spotlight surveys.
• The two track indices were related to each other (r = 0.75), but neither was related to the spotlight counts.
• Both track indices were much more precise than spotlight counts in seven of the nine surveys.

Dingoes

• Dingoes were seen very rarely by spotlight; in three surveys none were seen at all, although tracks showed they were present. Spotlight counts varied even more (coefficients of variation 137 to 265%).
• Track surveys recorded 8 to 50 times more dingoes per kilometre (minimum index) and 10 to 125 times more (maximum index) than spotlight surveys.
• The two track indices were closely related (r = 0.90), but again unrelated to spotlight counts.
• Wherever a comparison was possible, both track indices were much more precise than spotlight counts.

3. What they concluded

• Spotlight counts appear to be generally unsuitable for monitoring mammalian carnivores. Some animals do not look at the light, the counts miss much of the time the animals are active, and the results are very imprecise.
• Where the ground suits it, passive track surveys are simple, quicker, and sample the whole active period of the animals, giving larger samples and higher precision. They can also pick up other species.
• Track counts need care: repeated counts on the same road are not independent, and animals that follow roads (such as dingoes) can inflate the counts. The authors summed counts over days and spaced transects at least 5 km apart; this made them independent for cats, but perhaps not fully for dingoes, which range more widely.
• The minimum index for cats should level off and then drop at high cat densities, as tracks run together, but at the low densities typical of arid Australia it should rise and fall with true numbers.
• Track counts reflect how active animals are as well as how many there are. Activity changes with breeding season, food, shelter and other species, and in territorial animals such as dingoes, fewer animals may move about more, so a track index might stay the same or even rise after control.
• The accuracy of the methods was not tested. Scat counts and live capture each have their own drawbacks; for example, some cats bury their scats.

4. How tracking was or was not used

Tracking was the core of one of the two methods. Observers read fresh tracks on roads cleared the day before, identified them to species, judged small or large prints, and decided whether one or several animals had made them. They also interpreted the trail on the road surface: whether the animal crossed or walked along the road, how far, and whether dingoes travelled side by side. These readings were used to separate individual cats (tracks at least 500 m apart) and to count dingoes travelling together. No trailing beyond the road was reported.

5. Tracker or observer certification

Not reported. The paper does not say who identified the tracks or what experience they had; it thanks many volunteers, including students of Centralian College, for help with fieldwork. This matters because both track indices depend on the observers telling cat from dingo tracks and judging how many animals made each set.

6. Other technologies, skills or innovations

• Road transects cleared with a steel drag towed behind a vehicle, then counted from All-Terrain Vehicles (ATVs).
• A 100-watt spotlight, binoculars, a range-finder and a microcassette recorder for the spotlight surveys.
• Simple rules for turning track counts into a maximum and a minimum number of animals.

7. Implications for conservation (from the paper)

• Passive track surveys offer a more efficient and precise way than spotlighting to follow changes in feral cat and dingo numbers in arid Australia.
• Because animals may move about more when their numbers drop, track indices could give misleading results when used to judge the success of control programmes. Careful design (for example surveying outside the breeding season) or suitable statistics can help.