Determining the numbers of a landscape architect species (Tapirus terrestris), using footprints

1. What they did

The lowland tapir is a large, mostly nocturnal browser and an important seed disperser, sometimes called a “landscape architect” of tropical forests. In Brazil’s Atlantic Forest they are considered Endangered, so knowing how many there are matters. Adult tapirs have no natural markings to tell them apart on camera, and collaring means capturing and immobilizing them, which is costly and risky. The authors tested whether the footprint identification technique (FIT), which identifies individual animals from digital photos of their footprints, could count tapirs in one of the last areas in the south-eastern Atlantic Forest with a viable population.

• Finding the footprints. On 10 field trips between March 2014 and June 2015, the team searched 17 dirt roads (about 35 km in total) in and around the Recanto das Antas reserve in Espírito Santo, early in the morning and late in the afternoon.
• Photographing trails. They looked for trails (an unbroken series of footprints made by one animal) and photographed the left hind footprints from directly overhead, with rulers and a data slip beside each print, following the WildTrack FIT protocol. Only clear, undistorted prints were used, and each print was wiped out after photographing so it could not be collected twice.
• Measuring the footprints. In the FIT software, 15 landmark points were placed by hand on each image, and the software calculated 121 measurements. A tapir-specific algorithm, built earlier from 426 footprint images of 36 known captive tapirs, compared trails to decide which belonged to the same animal.
• Counting at different scales. They split the study area into six locations (A to F) and compared estimates for the whole area pooled, for each location separately and summed, and for each pair of locations. A pooled count lower than a summed count means some tapirs were found in more than one location.

2. What they found

The footprints

• They photographed 547 footprints from 48 trails. After removing poor-quality images, 440 footprints from 46 trails were analysed (4 to 23 usable prints per trail, about 9.6 on average).

How many tapirs

• FIT identified at least 29 different tapirs in the study area when all the data were pooled.
• Counting each location separately and adding them up gave 35. Location A had the most tapirs (12), location E had 8, F had 7, B had 6, and C and D had only one each.
• Locations A and E together held more than 65% of the tapirs identified.

Tapirs moving between locations

• The difference between 29 and 35 suggests that six tapirs used more than one location. The paper reports three tapirs at both A and B (1,110 m apart), one at A and E (1,567 m), three at B and F (3,868 m) and two at A and F (4,494 m). These numbers add to more than six, and the table shows one, not two, for A and F.
• So the same individuals were found at sites from 1,110 to 4,494 m apart, crossing a mix of forest, eucalyptus, farmland and pasture.

Trails as an index

• Across the six locations, the number of trails predicted the number of tapirs identified very closely (R² = 0.97).

3. What they concluded

• FIT is an effective, non-invasive way to identify individuals of a threatened species that has no visible natural markings, and it can give managers of protected areas rapid feedback.
• Tapirs in the area have overlapping ranges, and some move between parts of the study area. This suggests the farm and plantation landscape still lets tapirs move through it, and so carry seeds between forest patches.
• FIT has several benefits for tapirs: their footprints are easy to find and abundant on dirt roads; it avoids the risks of capture and immobilization, and the possibility that collared animals behave differently; a survey can be done fairly quickly; it can employ local trackers’ expertise and support their traditional knowledge; and it can be used alongside camera traps or line transects.
• Simply counting trails might one day estimate tapir numbers, but this needs testing at other sites, with detection probability taken into account, before it is used.
• At least 29 tapirs confirm the area’s value as a stronghold for lowland tapirs. This was the first census using footprints in the Tabuleiros Forest, and the authors plan to repeat it over time.

4. How tracking was or was not used

Tracking was the core of the study. The team searched dirt roads for tapir footprints, identified trails made by single animals, and chose clear left hind footprints to photograph and measure. The paper notes that tapirs usually place the hind foot in the front foot’s print when walking, that the small fifth toe of the front foot shows only on soft ground, and that clear prints usually come from animals walking at a relaxed pace. Dry weather made prints less clear and rain wiped them out. Prints were much easier to find on the sand and clay roads than on game trails in the forest, where leaf litter hides them. Individual identification was done by the FIT software from measurements, not by eye.

5. Tracker or observer certification

Not reported. The paper does not say who found and selected the trails and footprints, or describe any tracking experience or evaluation, apart from author contributions saying all six authors “performed the experiments”. Which prints were chosen as part of one trail and as “undistorted” affects the FIT results.

6. Other technologies, skills or innovations

• Footprint identification technique (FIT), an add-in for JMP statistical software (SAS). The FIT software is free from WildTrack; JMP is commercial, but the authors help non-profits apply for a free licence, and they offer FIT training workshops.
• High-resolution digital photographs (2,248 × 4,000 pixels or more, although FIT can work at 1,600 × 1,200), with metric scales and a data slip in each photo.
• GIMP image software to adjust contrast and crop; ArcMap 10.1 for the map.
• Footprints were wiped out after photographing to avoid recording them twice.

7. Implications for conservation (from the paper)

• Tapirs are decreasing because of habitat loss and fragmentation, fires, hunting and road kill, and their loss can threaten forest regeneration. Reliable numbers are needed to plan their conservation.
• A non-invasive, low-cost method like FIT can rapidly provide data on the numbers and distribution of a key seed disperser from footprints alone, and give those data to protected-area managers.
• The farm and plantation matrix around the reserve appears to allow tapirs to move between forest patches, which matters for the plants whose seeds they carry.
• The area, with at least 29 tapirs, is a stronghold for the species.