# An Improved Estimate of Trail Detectability for Snow-Trail Surveys

> Is a meadow with few tracks really empty, or have the tracks faded? Using wooden blocks to make artificial tracks in Wyoming’s Big Horn Mountains, researchers showed that small tracks in exposed, windy sites disappear much faster than large tracks in sheltered forest. They offer a simple correction so that snow-track counts can be compared fairly across species and habitats.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/1999-beauvais-buskirk-improved-estimate-trail-detectability/

## Citation and access

- **Authors:** Gary P. Beauvais and Steven W. Buskirk
- **Year published:** 1999
- **Journal or source:** Wildlife Society Bulletin
- **Full citation:** Beauvais, G. P. and Buskirk, S. W. 1999. An improved estimate of trail detectability for snow-trail surveys. Wildlife Society Bulletin 27(1): 32–38.
- **Access status:** Publisher link only
- **Where to find it:** https://www.jstor.org/action/doBasicSearch?Query=%22An+improved+estimate+of+trail+detectability+for+snow-trail+surveys%22

## Tags

- **Region:** North America > United States
- **Tracking:** Snow tracks; Track counts; Tracking; Tracks
- **Methods and technology:** Analysis of variance (ANOVA); Detectability; Linear regression
- **Data type:** Modelling or theoretical; New field data
- **Observer qualification:** Not reported
- **Discipline:** Wildlife tracking
- **Year published:** 1999

## Detailed keywords

- **Species:** No live animals were studied: the trails were artificial, made with wooden blocks sized to match three track-size groups. The track-size groups were based on 16 winter mammals over 75 g in the Big Horn Mountains, named in the paper by scientific name: Mustela erminea and M. frenata (small group, averaged together), Tamiasciurus hudsonicus, Mustela vison, Sylvilagus nuttallii and S. auduboni (averaged together), Martes americana, Lepus americanus (snowshoe hare); medium group: Lepus townsendii, Vulpes vulpes, Procyon lotor, Erethizon dorsatum (North American porcupine), Lynx rufus (bobcat), Canis latrans (coyote); large group: Puma concolor, Odocoileus hemionus, Cervus elaphus, Alces alces. A photograph shows snowshoe hare tracks crossed by red squirrel tracks. Habitat tree: lodgepole pine (Pinus contorta).
- **Location and habitat:** Six study locations in the Big Horn Mountains, about 20 km southwest of Buffalo, north-central Wyoming, United States, at 2386 to 2549 m. Each location was about 4 ha on a fairly flat slope (0 to 10 degrees), with a meadow of about 1 ha surrounded by closed lodgepole pine forest (trunks 15 to 25 cm across at breast height). Four exposure classes at each location: open meadow; meadow and forest edge; closed forest near the meadow; closed forest at least 10 m from the meadow.
- **Topics:** snow-trail surveys; trail detectability; trail visibility and persistence; snow settling and hardening; trail weathering (wind, sublimation, melting); track size; weather exposure; kilometer-day; modified kilometer-day; indices of relative occurrence; sampling effort; trail visibility versus trail identifiability; movement rates and comparisons between species; road bias in transect placement. Possible inconsistency: the Methods describe exposure class 3 as closed forest “10 m” from the nearest clearing, while the Figure 2 caption and Table 3 give “less than 10 m”.
- **Statistics or models:** Trail-visibility index (average visible proportion of all trails made on and before a given day); linear least-squares regression of the index on days since snowfall; two-way analysis of variance (ANOVA) in SPSS/PC+; 95% confidence intervals; six trail-detectability equations (modified kilometer-day = km x days x [intercept - slope x days]).
- **Data type:** New field data, late November 1994 to early February 1995, once at each of 6 locations after a new snowfall of at least 5 cm; data collection ran 4 to 8 days per location. The detectability models are built from these data.

## Synopsis

### 1. What they did

Counting animal trails in snow along transects is a long-standing way to learn where mammals live and how common they are. Counts are usually standardized as trails per “kilometer-day” (kilometers surveyed multiplied by days since snowfall). The authors point out that this leaves out a third factor: how visible a trail stays. Snow settles and hardens, so trails made later after a snowfall leave fainter marks, and wind, sublimation and melting wear trails down once they are made. These effects probably differ between species and habitats.

To measure this, they:

- Grouped tracks by size. They plotted the track area against body weight for 16 winter mammals in the Big Horn Mountains, Wyoming, and found three clusters: small (about 20 cm²), medium (about 50 cm²) and large (about 100 cm²).

- Made artificial tracks. They dropped wooden blocks onto the snow: a 130 g block from 2 cm for small tracks, a 675 g block from 5 cm for medium, and a 1620 g block from 35 cm for large.

- Laid out trails after each snowfall. At 6 locations, each with 4 sites of differing exposure to weather (open meadow, meadow edge, forest near the meadow, and forest at least 10 m from the meadow), they made a trail of 10 tracks of each size 12 to 24 hours after a snowfall of at least 5 cm. Each day they added a new trail beside the last and rescored all older trails as “visible” or “not visible” until the next snowfall. Observers kept at least 1 m from trails and used the same paths each day.

- Built a new measure. They modelled how the visibility of trails fell with days since snowfall for each track size and exposure, and used this to adjust the kilometer-day into a “modified kilometer-day”.

### 2. What they found

- Data collection lasted 4 to 8 days after a snowfall at each location.

- Trail visibility fell faster for smaller tracks and for more exposed sites (track size: F = 3.4, P = 0.040; exposure: F = 3.8, P = 0.014). The two factors did not interact.

- The three more exposed classes (meadow, edge and forest near the meadow) lost visibility at similar rates, so they were combined as “exposed”, with the deep forest as “protected”.

- Large tracks in protected sites were the only trails whose visibility did not decline from the first to the eighth day after snowfall. Medium and small tracks in exposed sites were fairly hard to see even on the first day, which points to snow hardening quickly after a storm.

- They give six equations, one for each track size and exposure, for the modified kilometer-day. For example, for small tracks in exposed sites: km x days x [1.11 - 0.12 x days].

- Worked example: on day 5 after snowfall, one trail of a medium-sized species on a 1 km transect in a meadow and one on a 1 km transect in dense forest both give 0.20 trails per kilometer-day, suggesting equal occurrence. Corrected for visibility, the species occurs 45% more often in the exposed meadow (0.32 versus 0.22 trails per modified kilometer-day).

- Planning surveys: a 2 km survey in protected forest on day 5 equals 9.2 modified kilometer-days for medium tracks; to match it, a survey in exposed habitat on the same day would need to be 2.97 km long.

### 3. What they concluded

- The modified kilometer-day adjusts for differences in trail visibility, so indices of relative occurrence become more comparable across species and habitats.

- Exposed sites are probably affected by windblown snow: forest edges change wind flow and redistribute snow, which may reach into nearby forest. In open meadows, tracks often froze into the snow after daytime thawing and night refreezing; shallow frozen tracks then stayed fairly visible even in wind.

- Detailed weather measurements (wind speed, sunlight) might predict visibility better, but measuring them along real transects would be impractical.

- Limits: the models apply only to surveys that record trails made since the last snowfall, and probably not beyond 8 days after a snowfall, when the relationship is expected to level out.

- The study measured whether trails could be seen, not whether they could be identified. Degraded trails in exposed areas often cannot be identified to species, especially where species have similar tracks, such as small and medium-sized canids.

- Other issues remain: species that move more (such as coyotes) leave more trails than sedentary ones (such as porcupines), which confounds comparisons between species, and surveys along roads can be biased because some species prefer roads (coyote) and others avoid them (bobcat).

- With slight adaptations, such as careful control of environmental conditions, traditional field naturalist techniques can meet the demands of modern wildlife science.

### 4. How tracking was or was not used

The study is about snow tracking itself: it measured how long trails in snow stay visible, so that snow-trail surveys (counts of trails along transects) can be corrected. The trails were artificial tracks made with wooden blocks, scored each day as visible or not visible; tracks were “not visible” when they left no marks or only marks that could not be told from natural irregularities in the snow. The authors note that, in their experience, a tracker can usually follow trails that cannot be identified into sheltered spots, such as the lee side of fallen trees, where identification is possible, although this may be difficult in very exposed areas.

### 5. Tracker or observer certification

Not reported. The paper does not say who scored the tracks or what tracking experience they had; R. Hay and G. Vos are thanked for help with data collection. Because visibility was scored by eye, the results depend on the observers’ judgement.

### 6. Other technologies, skills or innovations

- Artificial tracks made by dropping weighted wooden blocks of set size and height, matched to the area and depth of real tracks in three size groups.

- A trail-visibility index that captures both snow hardening (new trails each day) and trail weathering (rescoring older trails).

- The modified kilometer-day, a correction that can also be used to plan how much longer transects in exposed habitat need to be.

### 7. Implications for conservation (from the paper)

- Snow-trail surveys are especially useful for finding where rare and wide-ranging species occur, and they do not change animals’ natural behaviour or habitat use.

- Correcting for trail visibility allows snow-trail survey data to be used in rigorous tests of hypotheses about the distribution and abundance of mammals.

### 8. Implications for tracking (Original Wisdom's observations)

_These are Original Wisdom's observations, not the authors'._

- The study shows in numbers what experienced trackers know: small, light tracks in open, windy places fade fastest, so a survey that finds fewer trails there may simply be seeing less, not finding fewer animals.

- It separates seeing a trail from identifying it. Visibility was tested here, but identification skill was not, so the real detection of a given species will also depend on the observer.

## Cite this summary

Lawrence, K. 2026. Summary of: An Improved Estimate of Trail Detectability for Snow-Trail Surveys (Beauvais and Buskirk, 1999). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/1999-beauvais-buskirk-improved-estimate-trail-detectability/