Track-based monitoring for the deserts and rangelands of Australia

1. What they did

This is a scoping study written for the Threatened Species Network at WWF-Australia. It sets out how a national, coordinated program could use animal tracks and signs to monitor threatened and invasive species across the sandy deserts and rangelands of inland Australia, and how such a program could give meaningful work to Indigenous people, many of whom have strong tracking skills.

The authors drew on:

• Their own experience. They had used track-based monitoring since the mid 1980s, including plot surveys in the Tanami Desert, the Lake Eyre region and the Maralinga Tjarutja lands.
• A workshop and interviews. A workshop at Alice Springs Desert Park on 2 April 2008 brought together traditional owners, land management staff, scientists, government agencies and conservation organisations. They also interviewed people who could not attend, in person, by phone and at conferences.
• A review of monitoring options, track-based methods, community monitoring programs, training, accreditation and data management.

They then proposed a standard field method, a draft monitoring protocol and datasheet, a training plan for Indigenous communities, and a structure for a national organisation, tentatively called “Tracking Australia”.

2. What they found

Why tracks

• Over 33% of native mammal species in arid Australia are now extinct, and 90% of medium-sized mammals have become extinct, endangered or vulnerable. Very little biodiversity monitoring takes place in the deserts and rangelands.
• In their comparison of monitoring methods, no single method suits all animal groups, but track transects and track plots were feasible for medium and large mammals and some other groups, although expertise is required.
• Tracks are often the only practical way to monitor elusive or rare animals. Track surveys can be simple, cheap, cover many species at once (both introduced predators and the threatened species they affect), and do not disturb or handle the animals.
• Tracks show where a species occurs; they rarely give true density. High track numbers do not always mean many animals.

Two kinds of track survey

• Prepared surfaces: sand is raked or brushed into small plots (often 1 x 2.5 m), or a chain is dragged behind a vehicle along 10 to 20 km road transects. These take a lot of preparation, and the small area means trackers rely more on single footprints than on gait.
• Unprepared surfaces: observers read tracks on natural sand in large plots. No preparation is needed, and observers can follow a trail until the prints become clearer or the sun angle is better, using both gait and footprints. Skill is needed to age tracks.
• Indigenous groups already used unprepared surfaces, but search time and area varied, people tended to go where they expected an animal, and absences were seldom recorded. The use was more for “keeping an eye” on species than for measuring change.

Errors

• A false negative is missing a species that left sign; a false positive is putting tracks down to the wrong species. These come from the observer (identification) and from conditions (rain, wind, other tracks, hard or vegetated ground, poor light).
• Ways to reduce observer error: photos checked by experts, testing whether results differ by observer, training and accreditation, and several observers sampling the same plot independently.
• Ways to reduce detection error: checking that tracking conditions are good (insect tracks show the surface can hold small prints), recording conditions, and repeat visits.

Examples from earlier surveys

• Tanami Desert, late 1990s: over 215 plots in a 230,000 km² region, with Indigenous people from four language groups, produced a map of where bilbies were likely to occur. Bilbies were linked to substrate and rainfall, positively to dingoes and negatively to foxes.
• Fresh dingo and fox sign often occurred on the same plots in the two southern study areas, where rabbits are common, but seldom in the Tanami, where rabbits were less common.
• In the Maralinga Tjarutja region in 2007, four Aboriginal women and eight men were employed for three to five days to help with tracking work on 300 plots.
• A trial in the Simpson Desert in July 2007 found that volunteers aged 50 to 70 could be trained to identify basic mammal species in one to two days.

Current situation

• Different groups used different plot sizes (2 ha by most consultants, 6 ha in the Tanami ranger programs) and plot spacing. Data were mostly kept by those who collected them, not shared, and unknown to state and federal agencies.
• Concern over validation of the data was limiting the uptake of track-based monitoring.

3. What they concluded

The authors propose a standard “big plot” method:

• Plot: 2 ha (200 x 100 m) of natural sand, searched for 30 minutes of observer time (one observer for 30 minutes or two for 15 minutes each), zig-zagging to cover the area.
• What to record: tracks, burrows, diggings, sightings and scats of all terrestrial vertebrates over 35 g, to species or group; smaller mammals as mice or small dasyurids. Record how abundant the sign is and whether juveniles are present. Use a list of species expected in the region so that absences are recorded with equal effort.
• Age of the most recent sign: 1 to 2 nights, 3 to 7 nights, or over a week. Ageing lets small animals, whose tracks fade quickly, be compared fairly with large animals.
• Conditions: record shadow length, how good the tracking surface is, how much of the plot can hold tracks, and the habitat, fire history and distance to water.
• Layout: plots more than 5 km apart for general fauna surveys, chosen by stratified random sampling where possible, and revisited (within a season to estimate detection error, and every 3 to 5 years for long-term trends).
• Field tips: walk into the sun with your shadow behind you; do not rely on a single print; follow the animal’s trail and look at several gaits until certain; when in doubt, photograph (angled, no flash), draw and measure the prints and gait.

They also recommend:

• Training suited to each group: small field teams for Indigenous communities, with an experienced Indigenous tracker and a coordinator (a “Green Corps”-type model), and study camps with remote, photo-based support for volunteers and consultants.
• One accreditation scheme for everyone. At a baseline level, trackers should reliably tell apart feral cat, red fox, dingo/dog, rabbit, camel, birds, lizards (dragons and goannas), hopping mice and small mice or dasyurids.
• A central data management body, clear agreements on data ownership and intellectual property, and results returned to communities. The method collects track data, not Indigenous ecological knowledge, which needs separate approval.
• Partnership with a major conservation organisation, research centre or university, since without high-level federal support the method would stay ad hoc.

A two-week visit to a community could train and employ 8 Indigenous people for 8 days each and sample up to 112 plots twice.

4. How tracking was or was not used

Tracking is the subject of the report. It reviews track-based survey methods, and the proposed protocol relies on tracks and signs identification on natural sand (footprints, gaits, scats, burrows and diggings), ageing the most recent sign, measuring and photographing unclear prints, and following an animal’s trail until its identity is certain. The worked examples come from plot surveys in which Indigenous people took part in the tracking.

5. Tracker or observer certification

Reported but not validated. The report speaks of “experienced trackers” and says many Indigenous people have exceptional tracking skills, but no certification or test of the observers in the example surveys is reported. The report proposes training and an accreditation scheme, and describes the South African CyberTracker tracker evaluation (five tracker levels and four evaluator levels) as one model. It also suggests a quantitative accreditation in which several trackers sample the same plot independently and their false positives and false negatives are compared with the group tally or an expert tracker.

6. Other technologies, skills or innovations

• A draft field datasheet for 2 ha plots, including a simple test of tracking surface quality (which animals’ tracks could be told apart on it) and of light, using the length of the observer’s own shadow.
• Global Positioning System (GPS) units, cameras and CyberTracker in the proposed training for Indigenous trackers.
• Predictive occupancy maps built from presence-absence plot data and spatial data layers.
• An outline structure for “Tracking Australia”, covering coordination, data and science, training, accreditation and employment.

7. Implications for conservation (from the paper)

• A standard national method would let land managers and governments see whether threatened species are declining or recovering, and whether management, such as control of foxes, cats and rabbits, is working.
• Track plots monitor threats (invasive species) and threatened species at the same time and can detect species at low density.
• The program could give Indigenous communities work, especially for women and older people with low literacy, help keep Indigenous ecological knowledge strong, and let Indigenous people contribute to national understanding of biodiversity.
• Independent assessment of the method’s accuracy and precision will probably be needed before it is widely accepted.