A Track Count for Estimating Mountain Lion Felis concolor californica Population Trend

1. What they did

Large carnivores are hard and costly to count over big areas, so their population trends are often unknown. In California, mountain lion numbers had been estimated in different ways by different investigators. The authors set out to build a practical, low-cost index of the mountain lion population trend based on counting sets of tracks on dirt roads.

• Survey layout. California was divided into quadrats along 0.5 degree lines of latitude and longitude, and quadrats were picked at random. In each, local wildlife biologists chose up to three dirt-road transects of 11.3 km, at least 16 km apart, using the authors’ written guidelines. East-west roads were requested, because low sun makes tracks easier to see.
• Two ways of surveying (1986). Local teams (a driver and a professional wildlife biologist) drove the transects at 5 to 8 km per hour and recorded only whether mountain lion tracks were present. They sketched one track from each set and measured heel pad and toe width, so the authors could check the identification. Separately, Smallwood rode the same transects on a motorcycle at 10 to 16 km per hour and counted every track set (a continuous series of footprints, presumably from one animal). He traced tracks onto acetate and noted direction of travel and which side of the road the animal used.
• Other carnivores and habitat (1992). Smallwood repeated the survey in 1992 and also recorded tracks of black bear, coyote, dog, bobcat and fox. He mapped habitat and topography every 0.8 km, and the authors tested where tracks were found more or less often than expected by chance.
• Costs and sample size. They compared the costs, ease and reliability of the surveys, and estimated how many quadrats a future survey needs.

2. What they found

The two surveys

• Local biologists found mountain lion tracks on 20 of 36 completely surveyed quadrats (55.6%); Smallwood found them on 21 of 38 (55.3%).
• On the 29 quadrats surveyed by both, tracks were found on 72% (84% of the 25 non-desert quadrats).
• Local biologists misidentified 9.5% (three) of the tracks they sketched, and five teams (14.3%) did not send sketches.
• The cheapest option was one person surveying all quadrats. All three options costed were inexpensive: about 4,600 to 11,300 US dollars per statewide survey, based on 1986 costs.

Sample size

• Track counts were clumped (they fitted a negative binomial distribution), with a mean of 1.26 track sets per quadrat.
• About 44 quadrats would give a mean within 40% of the true mean at 95% confidence; 52 quadrats would give 30% at 90% confidence.
• This estimate was conservative: the 15 most productive quadrats were not fully surveyed, and they held about six times the track density of the rest. With them completed, 37 quadrats might have been enough.

Where the tracks were

• Carnivore tracks were absent on most road segments and turned up together mostly where adult mountain lion tracks were.
• Mountain lion tracks were found 2.3 times more often than expected along streams and 1.9 times along mountain slopes, and less often on ridge tops. They were found more often in oak woodland and montane hardwood-conifer forest.
• Tracks of mountain lion families and black bear families occurred together 26 times more often than expected by chance, but they avoided segments with tracks of most other carnivores.
• Coyote and dog tracks were found more often on basins and plateaus, fox tracks on knolls and peaks.
• Species used different sides of the road. Adult mountain lions and coyotes preferred the uphill side along streams; adult black bears the middle; mountain lion and black bear families the downhill side.
• Direction of travel differed by region and species. For example, adult mountain lion tracks showed east-west travel on 67% of segments in the Coast Range, southward on 70% in the Sierra Nevada, and eastward on 76% in the Southern Mountains.

Age and sex from tracks

Using rear heel pad width (over 52 mm adult male, 43 to 52 mm mostly adult female, under 43 mm juvenile), the ratio of adult males to adult females to juveniles was 1.0 : 1.0 : 0.1 (45 track sets). Far fewer juveniles showed up than in a radio-telemetry study in Idaho. The authors suggest juveniles do not travel on roads in proportion to their numbers.

3. What they concluded

• The track survey was inexpensive, easy to run and able to cope with changes in staff and road conditions. Dividing each quadrat into three transects spread the effort and let one person or team finish a quadrat in a day. A statewide survey could be done in 30 days.
• Using the same survey, Smallwood (1994) had found a population decline and a change in age and sex make-up between 1986 and 1992, both linked to timber loss.
• Many local teams placed transects on ridge tops hoping to find more tracks, but mountain lions avoided ridge tops. Choosing transects along streams, on mountain slopes and in the habitats lions favour would cut the sample size, cost and effort needed.
• Track counts would be more useful if calibrated against direct counts of animals of known age and sex in smaller intensive study areas, because transient animals and different age groups are detected at different rates.
• Recommendations: keep methods consistent between years rather than trying to survey every quadrat on the same day; use about 44 quadrats; exclude desert quadrats, where coarse dry sand hides track detail; use two trained people on motorcycles or vehicles with the best view; fewer trackers using standard methods would make results more consistent between years. Non-parametric tests that account for differences between quadrats are suggested for detecting trends.

4. How tracking was or was not used

Tracking was the core of the method. Surveyors identified mountain lion footprints on dusty dirt roads and counted track sets. Local teams sketched and measured tracks, and these sketches were used to check identifications: a sketch counted as an error when it showed claw marks, fewer than three heel lobes or the wrong size. Smallwood traced tracks onto acetate and read direction of travel and which side of the road each animal used. In 1992 he also identified tracks of black bear, bobcat, coyote, fox and dog. Rear heel pad width was used to sort mountain lion tracks into age and sex classes. No trailing was reported. The authors mention their earlier method for identifying individual mountain lions from tracks (Smallwood and Fitzhugh 1993) but do not apply it here.

5. Tracker or observer certification

Reported but not validated. The local teams were professional wildlife biologists given written instructions and tracings of tracks of mountain lion, bobcat, dog, coyote and black bear; their identifications were checked from sketches (9.5% wrong). Smallwood, who did the counts, is referred to as an “expert tracker” in the costing table, but his skill was not tested or certified in this study.

6. Other technologies, skills or innovations

• Motorcycle surveys, for a wider field of view and faster travel.
• Track sketches on pre-printed grids and tracings on acetate.
• Dust ratings at the start, middle and end of each transect and at tracks (no link was found between dust ratings and success in finding tracks, Smallwood 1994).
• An observed/expected association measure for linking tracks with topography, habitat and other species on 0.16 km road segments.

7. Implications for conservation (from the paper)

• Knowing the status and trend of large carnivores is critical for their conservation, and standard monitoring exists for only a few carnivore populations worldwide. A periodic road track survey can detect a changing mountain lion population cheaply over a whole state.
• The same survey can collect information on other carnivores, their habitats and their associations, which helps explain trends.
• The method can be adapted for some other large carnivores.