# Signs at the Top: Habitat Features Influencing Snow Leopard Uncia uncia Activity in Sagarmatha National Park, Nepal

> Snow leopards have come back to the slopes of Mount Everest after vanishing in the 1960s. By mapping 223 scrapes, scats, footprints, scent marks and rubbing sites, researchers found the cats were most active lower down, near rivers and away from busy trekking trails, suggesting a trade-off between gentler conditions and avoiding people. Their prey, Himalayan tahr, mattered less than expected.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2009-wolf-ale-signs-top-habitat-features/

## Citation and access

- **Authors:** Mosheh Wolf and Som Ale
- **Year published:** 2009
- **Journal or source:** Journal of Mammalogy
- **Full citation:** Wolf, M. and Ale, S. 2009. Signs at the top: habitat features influencing snow leopard Uncia uncia activity in Sagarmatha National Park, Nepal. Journal of Mammalogy 90(3): 604-611.
- **DOI:** 10.1644/08-MAMM-A-002R1.1
- **Access status:** Hosted (freely available elsewhere)
- **PDF:** https://www.originalwisdom.com/wp-content/uploads/2026/11/2009_Wolf_and_Ale_signs-at-the-top-habitat-features-influencing-snow-leopard-activity.pdf

## Tags

- **Species:** Mammals > Snow leopard
- **Region:** Asia > Nepal
- **Tracking:** Scat; Sign surveys; Tracking; Tracks; Tracks and signs identification
- **Behaviour and ecology:** Behaviour; Habitat selection; Human disturbance
- **Methods and technology:** Akaike Information Criterion (AIC); Direct observation; Geographic Information System (GIS); Global Positioning System (GPS); Logistic regression; Resource selection function; Utilization distribution
- **Data type:** New field data
- **Observer qualification:** Not reported
- **Discipline:** Behaviour and ecology; Wildlife tracking
- **Year published:** 2009

## Detailed keywords

- **Species:** Studied: snow leopard (Uncia uncia), a population that has recently returned to the park after disappearing in the 1960s. Main prey, counted as a habitat variable: Himalayan tahr (Hemitragus jemlahicus). In the park but not studied: Impeyan pheasant (Lophophorus impejanus), Tibetan snowcock (Tetraogallus tibetanus), blood pheasant (Ithaginis cruentus), Himalayan pika (Ochotona himalayana), Himalayan weasel (Mustela sibirica), hill fox (Vulpes vulpes), golden jackal (Canis aureus), Himalayan serow (Capricornis sumatraensis), musk deer (Moschus chrysogaster), domestic yak (Bos grunniens) and yak and zebu cattle (Bos taurus) hybrids. Snow leopards were the only large predators known in the study area.
- **Location and habitat:** Sagarmatha (Mount Everest) National Park (1,148 km²), Solu-Khumbu District, northeastern Nepal. Four valleys sampled: Namche, Phortse, Gokyo and Thame; main study area between the villages of Namche and Phortse and Gokyo Lake, in the upper catchment of the Dudh Kosi River (Bothe Kosi and Imja Khola streams) and the lower Cho-Oyu Valley. Open forest of Pinus wallichiana, Abies and Betula utilis at 3,000 to 4,000 m, grading into subalpine grassland and scrubland with thick mats of Juniperus and Rhododendron at 4,000 to 5,000 m. The park coordinates as printed (27°46’19” to 27°6’45”N, 86°30’53” to 86°99’08”E) cannot be right, since a longitude cannot have 99 minutes.
- **Topics:** snow leopard sign (scrapes, feces, footprints or pugmarks, spray or scent marks, rubbing sites); sign-based habitat use; predator trade-off between resources and risk; humans as a risk factor for large carnivores; trails, settlements and tourist traffic; prey (tahr) distribution; elevation, snow depth, slope and distance from rivers; recolonization; territorial marking; limits of sign data compared with sightings and radio telemetry.
- **Statistics or models:** Use-versus-available binary logistic regression, similar to a resource selection function (Manly et al. 2002), with five random “available” sites for every “use” site; model selection by Akaike Information Criterion (AIC); k-fold validation (k = 4) with Spearman rank correlations; McFadden’s rho², Nagelkerke pseudo-r² and area under the receiver operating characteristic curve (AUC = 0.934); Wald tests; adaptive kernel utilization distribution for tahr groups; Hawth’s Tools for ArcGIS 9.x and ArcGIS 9.1 at 50 x 50 m cells; SPSS 13.0. Inconsistencies in the paper: (1) the Results call the k-fold coefficients “Pearson correlation coefficients” but the Methods and Table 2 give Spearman rank correlations; (2) the Methods list annual precipitation and aspect as habitat variables, but the Results and tables use winter snow depth and never report aspect; (3) the Results report a weak correlation between snowfall and elevation (r = -0.34), while the Discussion says winter snow depth was “not significantly related” to elevation; (4) the Results compare slope with “the 3 physical habitat variables already mentioned” when only two (elevation and snow depth) had been mentioned; (5) the distance-from-river coefficient is 0.001 in Table 3 but 0.009 to 0.011 in every Table 2 subset; (6) slope in Table 3 has a Wald value of 3.12 marked P < 0.005, which seems too small for that level; (7) the Results report a significant positive relationship with distance from rivers, yet read it as a preference for being close to rivers or valleys; (8) the abstract calls the tahr effect “only moderate”, while Table 3 shows it as highly significant (Wald 16.36, P < 0.0005), though smaller than elevation, snow depth, river distance and trail distance.
- **Data type:** New field data: sign searches in autumn 2004 (October to November, 58 days), autumn 2005 (August to November, 111 days) and summer 2006 (May and June, 32 days), two people working 06:00 to 16:00 each day, 4,020 person-hours in all. Tahr counts in 2004 and 2006.

## Synopsis

### 1. What they did

Snow leopards disappeared from the Mount Everest region of Nepal in the 1960s. They have recently come back to Sagarmatha (Mount Everest) National Park, apparently following the recovery of Himalayan tahr, their main prey. The authors asked what shapes where these returning cats are active: the land itself, their prey, or people. They predicted that snow leopards would be most active where tahr were plentiful and less active near busy human areas.

- Finding signs. Over three field seasons (2004, 2005 and 2006), two people trekked the four main valleys of the park, visiting every place where they judged snow leopards and tahr to occur. They followed the standard methods of the Snow Leopard Information Management System. They recorded scrapes, feces, footprints, spray or scent marks and places where the cats had rubbed against rock, with a Global Positioning System (GPS) location for each. Signs within 50 m of each other in the same year were counted as one site.

- Describing the habitat. For each sign site, and for five times as many randomly chosen sites in surveyed areas with no sign, they measured elevation, slope, winter snow depth, distance from rivers, distance from trails and from settlements, using a Geographic Information System (GIS).

- Mapping prey. They counted tahr herds from vantage points with binoculars and spotting scopes, and turned the herd locations into a map of where tahr spend their time.

- Modelling. They compared sign sites with the available sites using logistic regression, chose the best model with the Akaike Information Criterion (AIC), and checked it by rebuilding it four times, each time with a quarter of the data held back.

### 2. What they found

The signs

- 223 sign locations were found: 131 scrapes, 55 feces, 23 pugmarks (footprints, 15 of them in 2005), 7 spray or scent marks and 7 rubbing sites.

- The models did not differ between years or sign types, so all signs were pooled. After joining signs within 50 m of each other, 177 sites were left for analysis.

What predicted where signs were found

The best model included every variable and explained the data well (area under the curve 0.934).

- Elevation: signs were more common at lower elevations, which surprised the authors, since snow leopards are generally thought to prefer high ground.

- Snow depth: signs were more common where winter snow was shallower.

- Slope: signs were somewhat less common on steep ground, a weaker effect. The authors note this may come from averaging slope over each 50 x 50 m cell.

- Rivers: the paper reports a significant positive relationship with distance from rivers but reads it as a preference for being near rivers or valleys.

- People: signs were more common farther from trails, a strong effect. Distance from settlements stayed in the best model but was not significant.

- Prey: signs were more common where tahr herds were denser, but this effect was weaker than elevation, snow depth and distance from rivers.

### 3. What they concluded

- Terrain and human activity were the most important factors in where snow leopards were active; tahr had a more moderate effect.

- The cats seem to trade off between milder conditions at lower elevations and avoiding people. Being more active low down puts them in the areas most likely to bring them close to humans.

- Most trails carry tourists, guides, porters and yak caravans, which may disturb these secretive cats, even though hunting is not common in the area. Humans may be a substantial factor in where snow leopards are active.

- The effect of prey was probably underestimated. One possible reason the authors give is that snow leopards might leave fewer scrapes and feces where prey could detect them.

- The study shows where snow leopards are likely to leave sign, not habitat selection itself, but areas the cats find attractive should collect more sign. Because scrapes and spray sites are territorial markings placed near those of other snow leopards, not every site was fully independent.

- Footprint analysis from the same study (Ale 2007) suggests at least 10 different individuals were represented.

- Sign surveys are less accurate than sightings or radio telemetry, but they are cheaper and easier, and remain a valuable way to study rare carnivores when other methods are not feasible.

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

The whole study rests on tracks and signs. Two people searched for and identified snow leopard sign across the park: scrapes, feces, pugmarks, spray or scent marks and rubbing sites on rock. Each sign site became a “use” site in the habitat model. No trailing was reported. The paper also mentions that the footprints found were analysed in another study (Ale 2007) to estimate that at least 10 individuals were present.

### 5. Tracker or observer certification

Not reported. The paper says two people searched for signs following the Snow Leopard Information Management System methods, but does not say who they were or how their sign identification was checked. Because the only large predator known in the area was the snow leopard, the risk of confusing their sign with another large predator’s may have been small, but the paper does not discuss this.

### 6. Other technologies, skills or innovations

- Garmin eTrex Venture GPS receivers (20 m accuracy) for sign and tahr locations.

- ArcGIS 9.1 and Hawth’s Tools for random site selection, distance layers and habitat values at 50 x 50 m cells, from a digital elevation model.

- Tahr counts by scanning opposite slopes with 8 x 21 binoculars and Nikon 15-60x spotting scopes, recording calves, yearlings, adult females and adult males.

- A use-versus-available logistic regression, a method usually applied to radio telemetry locations, adapted here to sign locations.

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

- Most large predators are at risk or endangered, so knowing how physical habitat, human activity and food together shape where they live is critical to conserving them.

- Most habitat studies of large predators look at either prey or people. This study suggests carnivores face trade-offs between physical habitat, human activity and food.

- Understanding how this newly re-established population uses its habitat, in relation to the land, people and prey, is an important step toward its long-term survival.

- Sign surveys are a valuable, low-cost way to assess carnivore presence when more costly monitoring is not feasible.

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

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

- The study shows how tracks and signs alone, combined with mapping, can reveal how a rare and secretive cat responds to people and terrain, without ever seeing the animals.

- The authors raise a point trackers know well: where signs are found is not the same as where animals spend time. Marking signs such as scrapes and spray sites cluster where animals choose to communicate, and may be left less where prey could notice them.

- Who found and identified the signs, and how reliably, is not described. Reporting observer skill would make sign-based habitat studies like this easier to judge.

## Cite this summary

Lawrence, K. 2026. Summary of: Signs at the Top: Habitat Features Influencing Snow Leopard Uncia uncia Activity in Sagarmatha National Park, Nepal (Wolf and Ale, 2009). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2009-wolf-ale-signs-top-habitat-features/