# Estimating the costs of locomotion in snow for coyotes

> How do coyotes, animals of open, snow-free country, cope with Quebec’s deep winters? By measuring the heart rates of coyote-sized dogs walking through snow and back-tracking wild coyotes on snowshoes, researchers found that snow makes travel more costly and that coyotes cut the cost by choosing shallower, harder snow and following snowmobile trails and other animals’ tracks, especially in a harsh winter.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2003-crete-lariviere-estimating-costs-locomotion-snow/

## Citation and access

- **Authors:** M. Crête and S. Larivière
- **Year published:** 2003
- **Journal or source:** Canadian Journal of Zoology
- **Full citation:** Crête, M. and Larivière, S. 2003. Estimating the costs of locomotion in snow for coyotes. Canadian Journal of Zoology 81(11): 1808–1814.
- **DOI:** 10.1139/z03-182
- **Access status:** Publisher link only
- **Where to find it:** https://doi.org/10.1139/z03-182

## Tags

- **Species:** Mammals > Coyote; Mammals > Domestic dog; Mammals > Snowshoe hare
- **Region:** North America > Canada
- **Tracking:** Snow tracks; Spoor age estimation; Track measurements; Tracking; Tracks; Tracks and signs identification; Trailing
- **Behaviour and ecology:** Behaviour; Energy budget; Movement
- **Methods and technology:** Analysis of variance (ANOVA); Linear regression; Telemetry
- **Data type:** New field data
- **Observer qualification:** Not reported
- **Discipline:** Behaviour and ecology; Wildlife tracking
- **Year published:** 2003

## Detailed keywords

- **Species:** Studied: coyote (Canis latrans), wild and free-ranging, followed by their tracks in snow; three coyote-sized domestic dogs (Canis familiaris, mixed breed, 15.7 kg on average) used as a model for coyotes in heart-rate trials. Trails used by coyotes included those of snowshoe hare (Lepus americanus), other coyotes and other species. In the study area but not studied: moose (Alces alces), white-tailed deer (Odocoileus virginianus), black bear (Ursus americanus), red fox (Vulpes vulpes), fisher (Martes pennanti), American marten (Martes americana) and Canada lynx (Lynx canadensis). Compared from other studies: mule deer (Odocoileus hemionus), elk (Cervus elaphus), caribou (Rangifer tarandus), polar bear (Ursus maritimus), swift fox (Vulpes velox) and kit fox (Vulpes macrotis).
- **Location and habitat:** About 1200 km² of rolling hills on the south shore of the St. Lawrence River, southeastern Quebec, Canada. A rural landscape (473 km²) of cultivated fields, old fields, regenerating forest, woodlots and open water, and contiguous forest blocks (721 km²) of sugar maple, yellow birch, balsam fir, white birch, trembling aspen and black spruce. Snow usually exceeds 1 m in midwinter and lasts from early December to late April. Snow was monitored at a permanent station in the Pohénégamook white-tailed deer wintering area.
- **Topics:** cost of moving through snow; energy expenditure; heart rate as a measure of energy use; sinking depth of tracks; snow depth and snow hardness; travel speed; use of packed trails (snowmobile trails, roads, animal trails, frozen water bodies); fine-scale selection of snow conditions; mild versus harsh winters; forest versus rural coyotes; penetrometer to mimic a coyote foot; field metabolic rate; slow travel in cursorial carnivores; predator pass-along hypothesis; range expansion of coyotes into northeastern North America; coyote predation on white-tailed deer in deep snow.
- **Statistics or models:** Linear regression (dog as a fixed factor) with a quadratic term for sinking depth; random coefficient (mixed) model with dog as a random factor; log-transformed heart rates; multivariate analysis of variance (MANOVA); analysis of variance (ANOVA) on ranks; repeated-measures ANOVA; regression with landscape as a dummy variable; paired t tests. Inconsistencies: the final model is printed as ln(heart rate) = 4.870 + 0.047(travel speed) − 0.003(sinking depth) + 0.007(sinking depth)²; with 0.007 the model cannot give the heart rates reported (for example 143 beats per minute at about 11 cm sinking depth, or a doubling of heart rate at 33 cm), but a squared-term coefficient of about 0.0007 does, so 0.007 looks like a printing error. In Table 1, the “Other species” row repeats the “Hare tracks” row exactly. The 1979 study used for daily travel distance is cited as “Andelt and Gibson” in the text and “Andelt and Gipson” in the references. The repeated-measures comparison of snow in forest and rural landscapes in winter 1996–1997 is described in the methods, but no results are given.
- **Data type:** New field data. Snow-tracking of wild coyotes in December 1995 to March 1996 (a mild winter) and January to March 1998 (a harsh winter, with about 150% of the previous snowfall). Heart-rate trials with dogs in late January to February 1997 (implants fitted in November 1996). Snow stations measured in winters 1995–1996, 1996–1997 and 1997–1998.

## Synopsis

### 1. What they did

Coyotes evolved in the open, mostly snow-free landscapes of central and southwestern North America, but over the last 50 to 100 years they have spread north and east, into places like southeastern Quebec where snow is often more than a metre deep. The authors asked how much extra energy coyotes spend moving through snow, and whether they behave in ways that cut that cost.

They worked in two parts:

- Measuring the cost with dogs. Three coyote-sized domestic dogs were fitted with heart-rate transmitters, since heart rate closely tracks energy use. After training to walk on a slack leash, each dog walked about 300 m per trial: first on a plowed road at different speeds, then in many different snow conditions. The researchers measured the dogs’ sinking depth, travel speed and heart rate, and built models predicting heart rate from sinking depth and speed.

- Following wild coyotes. In two winters, they searched the study area by truck or snowmobile in the early morning for fresh coyote tracks (less than 12 hours old), then back-tracked them on snowshoes, measuring distance with a hip-chain. Every 50 m they measured the coyote’s track sinking depth, and snow depth and hardness both on the trail and 1 m off it. Hardness was measured with a weighted copper pipe (a penetrometer) dropped from 10 cm, designed to sink like a coyote’s foot. They also recorded how far coyotes travelled on snow packed by snowmobiles, other human activity, animals, or frozen waterways.

Finally, they used the dog models to estimate how much energy coyotes spent, and saved, in the snow conditions they actually used.

### 2. What they found

Dogs: speed and snow

- Resting heart rate was about 72 beats per minute and rose slightly in colder weather (about 0.6% for each degree drop).

- On a hard surface, heart rate rose in a straight line with speed: about 6 beats per minute (about 8% of resting heart rate) for every 1 km per hour faster.

- In snow, sinking depth (1 to 35 cm) affected heart rate much more than speed did, and the cost climbed ever more steeply as the dogs sank deeper.

- At 1 km per hour, heart rate was 136 on a hard surface, 143 on a coyote’s track in the snow, and 149 off the track. At 5 km per hour it was 165, 173 and 180.

- Going from 1 to 5 km per hour raised heart rate by about 20%, the same as sinking 19 cm in snow at 1 km per hour. The model predicted heart rate would double at a sinking depth of 33 cm.

Wild coyotes

- In the harsh winter, coyotes travelled much more on packed snow than in the mild winter: 48% versus 28% of their distance in forest, and 63% versus 17% in rural areas.

- Use of snowmobile trails and other coyotes’ tracks rose in the harsh winter. Rural coyotes used hare tracks more often than forest coyotes, though not often.

- Coyote tracks sank about 10 to 11 cm on average, in snow about 56 to 60 cm deep.

- Even off packed trails, coyotes chose their route closely: snow was about 3 cm shallower on their trail than 1 m away, and harder, with the penetrometer sinking 2.3 cm deeper off the track.

- The penetrometer matched coyote sinking depth well in the rural landscape and less well in forest.

Energy

- Moving through snow raised the predicted heart rate by about 4% to 6% compared with a hard surface, and the coyotes’ choice of harder snow saved a similar amount.

- Assuming 10.9 km travelled a day at 1 km per hour, moving about would make up 27% of a coyote’s daily energy use on hard surfaces and 29% in average snow.

### 3. What they concluded

- Snow raises the cost of moving for coyotes, and sinking deeper raises it ever more steeply, as in deer, elk and caribou. The same body mechanics may apply to dogs and deer.

- Coyotes in deep-snow country reduce this cost by their behaviour: using trails packed by people and other animals, especially in harsh winters, and choosing shallower, harder snow.

- Speeding up is costly, which may explain why coyotes in many studies travel at about 1 km per hour or less. The authors propose that slow travel suits predators that must cover large areas to find prey.

- Fast chases through soft snow are very costly for coyotes, which would explain why they ambush hares in snow and make shorter chases than the large-footed Canada lynx.

- The authors hypothesize that snow packed by people, especially snowmobile trails, helps coyotes move in deep snow, may let them live in marginal forest habitat, and may have helped their spread into northeastern North America.

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

Tracking was central to the field study. The researchers searched for fresh coyote tracks in snow, judged them to be less than 12 hours old, and back-tracked them on snowshoes. Along the trails they measured the coyotes’ track sinking depth every 50 m, compared snow on and off the trail, and recorded how far each coyote travelled on snowmobile trails, roads, frozen water, and the tracks of hares, other coyotes and other species.

### 5. Tracker or observer certification

Not reported. The paper does not say who did the snow-tracking or what experience they had. N. Bergeron and D. Murray are thanked for their work in the field.

### 6. Other technologies, skills or innovations

- Heart-rate transmitters implanted under the skin of the dogs (Telonics HR-300), with heart beats counted from a tape recorder connected to a telemetry receiver.

- A home-made penetrometer (a copper pipe partly filled with lead, 60.3 cm long, 2811 g) dropped from 10 cm to mimic the pressure of a coyote’s foot in snow.

- Hip-chain to measure distances along trails.

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

- Snow packed by human activity, especially snowmobile trails and plowed roads, may make it easier for coyotes to live in forested areas and may have helped their range expansion in northeastern North America.

- Differences in how well canids and deer cope with deep snow may help explain why white-tailed deer are more vulnerable to coyote predation in deep-snow winters.

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

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

- The study shows how much trackers can learn simply by following a trail and measuring along it: how deep the animal sank, what the snow was like just beside the trail, and which trails made by people and other animals it chose to use.

- Coyotes picked snow only about 3 cm shallower and a little harder than the snow 1 m away, so small choices visible in a trail can reflect real energy savings.

## Cite this summary

Lawrence, K. 2026. Summary of: Estimating the costs of locomotion in snow for coyotes (Crête and Larivière, 2003). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2003-crete-lariviere-estimating-costs-locomotion-snow/