Spoor density as a measure of true density of a known population of free-ranging wild cheetah in Botswana

1. What they did

Counting tracks along roads is a cheap and practical way to monitor large carnivores, but spoor counts only give a relative index. To turn them into a real density, the index has to be calibrated against a population whose true size is known. Earlier work had done this for lion, leopard and wild dog, but not for cheetah, whose home ranges, daily movements and road use differ.

The authors did this in Jwana Game Reserve, southern Botswana, where the cheetah population had been monitored since 2003:

• Knowing the true population. Cheetahs were caught in box traps, ear-tagged, microchipped and photographed so they could be recognised by their spots. One female (F5) wore a cell collar that sent Global Positioning System (GPS) locations one to four times a day. A second family (F6) was seen about every 12 days. On average 89% of cheetahs present each month were known individuals. Researchers, game park staff and tourist photographs accounted for the rest. Spoor that seemed to belong to a new animal was followed to try to see and trap it.
• The spoor survey. The reserve was split into four sections with three 9 km transects each (12 transects, 108 km). One transect per section was driven each day, so each was left at least 48 hours between surveys. One or two four-wheel-drive vehicles drove at 8 to 12 km per hour from one hour after sunrise until 11:30. The roads were not swept beforehand. All large predator spoor was recorded, with a minimum of two people per vehicle. Each cheetah counted as a separate spoor, even when walking in a family group, and each animal was counted only once a day.
• Calibrating and testing. They compared spoor density with true density for the first wet season (October 2005 to March 2006) and the dry season (April to September 2006). They did this for one family, for two families and for all cheetahs. They then used a second wet season (October to December 2006) to test the wet season correction factor.
• Prey counts. From March to December 2006 an extra observer counted prey animals seen from the transects.

2. What they found

Survey effort

• 8226 km were driven, and 191 individual cheetah spoor were found: one spoor every 43.07 km, or 2.32 spoor per 100 km.
• There was no significant difference in spoor frequency between transects, although some were driven about twice as often as others.
• Results became reliable once there was at least 1 km of road surveyed for every 3.34 km² of study area.
• In the wet season about 38 spoor (1296 km of driving) gave a precise and accurate result. The dry season needed about 90 spoor (3636 km). The second wet season needed 50 spoor (864 km).

True density versus spoor density

• True density averaged 5.23 cheetahs per 100 km². It ranged from 3.33 to 7.78 and rose from 3.80 in the wet season to 6.67 in the dry season, when one female’s five cubs joined the counts.
• Spoor density also rose, from 1.45 spoor per 100 km in the wet season to 3.10 in the dry season.
• Spoor density was lower than true density in both seasons, unlike earlier lion and leopard studies, where spoor overestimated density.
• Trend lines linked spoor and true density in both seasons, but neither was statistically significant on its own.

The test

• In the second wet season spoor density was 4.44 per 100 km. The first wet season correction factor predicted 10.84 cheetahs per 100 km², but the true density was 7.22. The correction factor overestimated the population.
• Spoor counts rose by 306% while true density rose by only 190%. The first wet season had 364 mm more rain than the second.
• Combining both wet seasons gave a significant relationship between spoor and true density (r = 0.968, P = 0.002).

Prey and movement

• Prey density was 52 animals per km² in the wet season and 21 per km² in the dry season.
• Collar data showed female F5 used a larger area and spent more time outside the reserve in the dry season.
• The two main families followed a regular cycle: each month they overlapped in one area for up to a week, then moved to different parts of the park.
• Cheetahs in families and coalitions did not walk in a line along roads. They often spread out up to 50 to 100 m apart.

3. What they concluded

• A measurable relationship between spoor density and true density exists for cheetahs, separately for wet and dry seasons. Spoor counts underestimated the population.
• The authors suggest the underestimate may be because family members spread out, so all of them are rarely on the road at the same time. The differences from lion and leopard studies may be species related.
• The dry season needs more sampling. In the wet season, cheetahs may stay in smaller areas, with taller grass for stalking and thick bush for resting. In the dry season, prey moves to find water, so cheetahs may need to travel further and use a larger area.
• The wet season correction factor had to be revised with the second wet season’s data. Rainfall changed the habitat and water points, which may have changed prey movements and how cheetahs behaved and used roads.
• Limitations: a small study area, few animals, and an open population with breeding females. Cubs arriving and sub-adults and females leaving may have changed spoor density, so studies need to run long enough not to underestimate the population. The dry season line could not be tested with a second dry season.
• Spoor surveys with correction factors can work for long-term monitoring, but should be used with caution. Calibration studies need repeating many times so the correction factor captures changes in season, habitat, prey and behaviour.

4. How tracking was or was not used

Tracking was the core of the study. Teams drove sandy roads and identified the spoor of cheetah, leopard, brown hyena and domestic dog. Where possible they identified individual cheetahs from their spoor, using group size, cub age, known GPS locations and, for one family (F6), the females’ distinctive round hind feet. Spoor thought to come from a new cheetah was followed to try to see and trap the animal. The paper does not describe how cheetah spoor was told apart from leopard or other spoor.

5. Tracker or observer certification

Reported but not validated. Spoor was identified by “trained researchers”. The lead researcher had three years of experience in spoor tracking, and trained and tested the other researchers on spoor identification. No certification, test method or test results are reported. Staff members thanked in the acknowledgements guided international volunteers on the daily spoor surveys. The authors accept there may have been some errors in spoor detection and identification.

6. Other technologies, skills or innovations

• GPS cell collar on one female, radio telemetry, ear tags and microchip transponders, and spot-pattern identification from photographs, including tourists’ photographs.
• Double-ended box traps with live bait in a separate holding cage.
• Monte Carlo resampling in PopTools (an Excel add-in) to decide how much road to survey for a reliable spoor frequency.

7. Implications for conservation (from the paper)

• Cheetahs were listed as vulnerable, and like most large carnivores their numbers are declining, so accurate, repeatable and affordable ways to measure population trends and density are needed.
• Spoor surveys are less invasive and cheaper than direct methods and suit low-density species such as cheetah, but they need a correction factor to give true density.
• Correction factors must be tested and refined over time, because animals change their behaviour with season, habitat and prey.
• Cheetah density here was higher than in a Kgalagadi Transfrontier Park spoor study, perhaps partly because there were no lions or spotted hyenas. Cheetahs are known to survive better outside protected areas, where there is less competition from these large predators.