Monitoring bait removal in vertebrate pest control: a comparison using track identification and remote photography

1. What they did

In Australia, red foxes and wild dogs (dingoes, feral dogs and their hybrids) are widely controlled with poison baits. Bait removal is often checked by identifying tracks in sand plots around the bait, but sand plots can fail in bad weather, and there is concern that baits taken by non-target animals, especially the spotted-tailed quoll, may be underestimated. Earlier camera methods worked but were too expensive to use at many bait stations.

The authors ran simulated trail-baiting campaigns at two sites in New South Wales between November 2000 and February 2001. They built 117 bait stations along roads and trails, each with a sand plot of raked sand 1 m across, and used non-toxic fox baits buried from 0 to 12 cm deep. Each station was checked and raked daily, and any removed bait was replaced.

At the same time, every station had a cheap, mechanically triggered 35 mm camera, costing about 30 Australian dollars per unit. A fishing line ran from the shutter release to the bait, so that pulling the bait took a flash photograph. They then compared what the tracks showed with what the photographs showed.

2. What they found

Bait removal

• Over 1126 bait-nights, baits were removed 106 times (9.4%).
• Spotted-tailed quolls took 49 baits (46.2%), Australian brush-turkeys 12, wild dogs 9, red foxes 7 and superb lyrebirds 2. Small mammals such as rats took 14, and 13 were taken by unknown animals.
• Only 15% of baits were taken by the target animals (foxes and dogs).
• The 13 unknown removals happened when heavy rain wiped out tracks and the camera flash also failed.

Cameras

• Cameras photographed every species that took baits except small mammals, which were presumably not strong enough to trigger them.
• Early on, many failures came from batteries leaking and the flash failing. Wrapping cameras tightly in plastic bags and placing them in shade fixed this.
• Leaving aside battery failures and small mammals, cameras missed 5 animals (4.7%): the trigger jammed 3 times, and twice an animal walked into the line and set the camera off, so the actual removal was not photographed.

Tracks in sand plots

• Track identification failed for 27 animals (25.5%).
• On 19 occasions (17.9%) heavy rain obscured the tracks.
• On 7 occasions (6.6%) tracks were misidentified. Six of these were in the first week, from inexperience with brush-turkey and lyrebird tracks. Once the observer had seen the photographs, these birds were identified correctly every time afterwards.
• One misidentification (0.9%) happened when very dry sand left prints without definition: tracks read as a quoll’s were shown by the photograph to be a fox’s.
• Once (0.9%), tracks of both a dog and a quoll were on the plot. The photograph showed the quoll took the bait. The dog’s prints were easy to see because they were bigger, while the quoll’s were seen only on close inspection.
• Quolls often dug holes 10 to 15 cm across to get at buried baits, and these looked no different from holes dug by foxes.
• Photographs showed that the same quoll visited more than one bait station in one night.

3. What they concluded

• Many baits can be taken by non-target animals, so accurate monitoring matters in pest control.
• Photographs are less open to misinterpretation than sand plots, need no expert to interpret, do not need checking every day, and can identify individual animals.
• Remote photography costs more and takes more work than sand plots, so it is unlikely to suit routine monitoring, but it may be useful for research or as an audit tool to check how accurate sand plots are.
• An experienced observer can identify tracks in sand plots with high accuracy in fine weather, but sand plots are highly unreliable in rain. Sand that is moist but not saturated shows tracks best; sand that is too coarse, soft or firm (or very dry) can mislead.
• Observers checking sand plots should be skilled in the tracks of the target animals and also know, and be experienced with, the tracks of all non-target species of concern in the area.
• Identifying animals by the size and look of their diggings may also be unreliable.
• Because the cameras are cheap, large numbers of bait stations can be watched at once. With a larger battery pack, cameras could be left set for longer without daily checks, and might also be used in wildlife surveys as an alternative to trapping or hair tubing.

4. How tracking was or was not used

Tracks and signs identification was one of the two methods being compared. Each day, the observer read the tracks in the raked sand plot at each bait station to identify which animal had removed the bait. The paper also discusses diggings at bait stations as signs. No trailing was used. Camera photographs served as the check on the track identifications.

5. Tracker or observer certification

The sand plots were read by the first author (Alistair S. Glen). No certification or prior tracking experience is reported. The observer’s accuracy was tested in the study itself by comparison with the photographs: six of seven misidentifications came from inexperience with bird tracks in the first week. Triggs (1996), a field guide to Australian mammal tracks and signs, is cited on sand texture.

6. Other technologies, skills or innovations

• A low-cost, mechanically triggered camera (a modified Hanimex Handy Panorama 35 mm camera, based on a design by Major 1991), about 30 Australian dollars per unit, mounted on a steel fence post 2 to 3 m from the bait.
• Heavy fishing line (13.6 kg breaking strain) ran from the shutter release, through a tent peg in front of the camera and a second buried peg under the bait that acted as a pivot, so any pull on the bait fired the camera.
• The flash-ready light was disabled to save batteries; tight plastic bags kept out moisture; packing tape held the camera in its box; black and white 400 ASA film was used.
• Up to 60 bait stations were monitored at the same time.

7. Implications for conservation (from the paper)

• Large numbers of baits may be taken by non-target animals such as spotted-tailed quolls, so baiting programmes need accurate monitoring.
• Knowing that one quoll can visit several bait stations in a night is crucial for planning pest control programmes.
• A careful free-feed period before poison baits are laid may make control more effective and reduce harm to non-target animals.