# River Otter Latrine Site Selection in Arid Habitats of Western Colorado, USA

> Where would a river otter leave a message in the desert? Surveying 247 km of canyon rivers in western Colorado, researchers found that otters choose large prominent rocks beside deep water, shaded by cliffs and close to beaver sign, for their latrines. Their simple model can help anyone searching for otter signs in dry country.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2010-depue-ben-david-river-otter-latrine-site/

## Citation and access

- **Authors:** John E. DePue and Merav Ben-David
- **Year published:** 2010
- **Journal or source:** The Journal of Wildlife Management
- **Full citation:** DePue, J. E. and Ben-David, M. 2010. River otter latrine site selection in arid habitats of western Colorado, USA. Journal of Wildlife Management 74(8): 1763–1767.
- **DOI:** 10.2193/2008-065
- **Access status:** Publisher link only
- **Where to find it:** https://doi.org/10.2193/2008-065

## Tags

- **Species:** Mammals > American beaver; Mammals > North American river otter
- **Region:** North America > United States
- **Tracking:** Scat; Sign surveys; Tracking; Tracks and signs identification
- **Behaviour and ecology:** Behaviour; Communication; Habitat selection
- **Methods and technology:** Akaike Information Criterion (AIC); Geographic Information System (GIS); Global Positioning System (GPS); Logistic regression; Pearson correlation
- **Data type:** New field data
- **Observer qualification:** Not reported
- **Discipline:** Behaviour and ecology; Wildlife tracking
- **Year published:** 2010

## Detailed keywords

- **Species:** Studied: North American river otter (Lontra canadensis), at least 21 individuals using the latrines (from DNA fingerprinting of feces and hair in a companion study). Beaver (Castor canadensis): fresh beaver sign recorded as a habitat feature, not studied. Eurasian otter (Lutra lutra) discussed from a study in Spain. Plants recorded as habitat: eastern cottonwood (Populus deltoides), tamarisk (Tamarix ramosissima), willows (Salix spp.), Utah juniper (Juniperus osteosperma) and box elder (Acer negundo).
- **Location and habitat:** Six reaches of five rivers in western Colorado, USA: the Green River (Browns Park National Wildlife Refuge and Dinosaur National Monument), the Colorado River (Grand Junction to the Colorado-Utah state line), and the Dolores, Piedra and Gunnison rivers. Semiarid to arid climate, rivers mostly in large desert canyons with sandstone and granite walls, from calm flat water to class IV whitewater. River otters were reintroduced in the region from the 1970s to 1990s. The paper gives the average temperature as ranging “from 18° C in January to 38° C in July”, which looks unusual for January in western Colorado and may be a printing error.
- **Topics:** latrine sites, scent marking and communication (urine, feces and anal gland secretions), latrine site selection, habitat selection, otter and beaver association, prominent rocks, cliff and rock overstory, shade, river depth and deep pools, drought, reintroduction monitoring, predictive habitat model, sign surveys, limits of remote sensing.
- **Statistics or models:** Logistic regression comparing latrines with random sites; Pearson correlation matrix to remove correlated variables; eight a priori models compared with Akaike’s Information Criterion corrected for small sample size (AICc) and Akaike weights; Hosmer-Lemeshow goodness-of-fit test; receiver operating characteristic curve and area under the curve (AUC); 80/20 training and test split repeated 5 times; odds ratios; frequency of occurrence with 95% confidence intervals; SAS 9.1. Inconsistencies: Table 2’s column is labelled “CV” and “coefficient of variation”, but the values are the regression coefficients (for example 3.935 for beaver matches the odds ratio of 51.166). The text says the variables in the chosen model “occurred at significantly higher frequencies” at latrines, but cliff had P = 0.051 in Table 2, and the text also lists cliff overstory and shade as statistically insignificant in the top model (model 1) even though both are kept in the chosen model (model 2).
- **Data type:** New field data: summer surveys in 2004 (13 July to 5 September) and 2005 (30 May to 14 August); figure captions give May to September 2004 and 2005.

## Synopsis

### 1. What they did

River otters were wiped out from much of southwestern North America by the early 1900s and have since been reintroduced in Colorado, Arizona and Utah. River otters leave urine, feces and anal gland secretions at latrine sites on land, which they use to communicate with each other, and the presence or absence of latrines is used to track where otters live and whether reintroduced populations persist. Latrine site selection had been described in temperate regions but not in arid and semiarid ones. The authors set out to build a model that predicts where otters place latrines in western Colorado, to make sign surveys more efficient and accurate.

- Finding latrines. In the summers of 2004 and 2005 they surveyed 247 km along six reaches of five rivers by raft or kayak and on foot, searching both banks up to 30 m from the water’s edge. A site counted as a latrine if it had at least 3 feces.

- Choosing random sites. For comparison, they generated 30 random points along the shore of each reach (180 in total) on a map and found them in the field with a handheld Global Positioning System (GPS) unit.

- Measuring habitat. At 42 latrines and 119 random sites they described a 10 m circle: overstory and understory cover (trees, shrubs, rock cliff), shade, substrate (sand, grass, mud, rocks or a large prominent rock), water depth near the bank, and whether fresh beaver sign was within 40 m.

- Modelling. They compared eight models with logistic regression and tested how well the best ones predicted latrines, building each model on 80% of the sites and testing it on the other 20%, five times over.

### 2. What they found

- Latrine density ranged from 0.03 to 0.42 latrines per km. Otters used rivers from very gentle to steep gradients, including stretches of class IV whitewater.

- DNA from feces and hair at the latrines (from a companion study) showed that at least 21 different otters used them.

- The authors chose the second-ranked model because it was simpler and its variables appeared again and again in the five top models. Latrines were more likely where there was:

- beaver sign within 40 m: the odds of a latrine were about 51 times higher;

- a large prominent rock: about 24 times higher (found at 41% of latrines and only 5% of random sites);

- deeper water: about 3 times higher for every 20 cm of depth (water 60 cm or deeper at 61% of latrines);

- cliff or rock overhang overhead: about 2 times higher for every 20% more cover;

- shade: a small increase for every extra percent of shade.

- The model predicted well: the area under the curve was 0.95, it classified 87% of test sites correctly on average, and it picked out 72% of the latrines.

- Trees were sparse along these rivers and did not predict latrines; otters chose sites under cliffs and rocks instead.

### 3. What they concluded

- Otters in this dry country choose latrine features much like those chosen in temperate areas, but with cliffs and rocks taking the place of trees as cover. The authors suggest overhead cover gives protection from aerial predators and keeps scent marks from drying out, and that high rocks give a good view and spread scent further.

- Deep pools probably help otters escape predators and matter during low flows and drought, when fish gather in pools. They could not study latrine use at low water because the canyons could not be floated then.

- Beaver activity was the strongest predictor. Beavers probably help otters in dry regions by keeping pools that hold fish during drought and by creating den sites. The authors call for studies of otter and beaver interactions in these rivers.

- The model gives researchers and managers a robust tool for finding latrines and monitoring otters in arid environments.

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

The study is built on a sign survey: researchers searched riverbanks for otter latrines, identified them from otter feces (a site needed at least 3 feces), and found the point with the most otter activity at each one. They also recorded fresh beaver sign near each site, though the paper does not say what kinds of beaver sign were used. Feces and hair from the latrines were collected for DNA in a companion study. No trailing was reported.

### 5. Tracker or observer certification

Not reported. The paper does not say who identified the latrines and beaver sign or what experience they had. Four people are thanked for assistance in the field.

### 6. Other technologies, skills or innovations

- Random sites generated in a Geographic Information System (GIS), ArcView 3.2, from national hydrography data, and located in the field with a handheld GPS unit.

- A spherical densiometer to measure overstory density, checked against visual estimates of shade.

- Water depth ranked by body height (below knee, below waist, above waist) in the first year and measured with a graduated rod in the second. The two methods were closely correlated.

- DNA fingerprinting of feces and hair collected at latrines (reported from a companion study).

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

- The model variables are easy to spot in the field with little training and could improve monitoring of river otter reintroductions in southwestern North America.

- Surveys for latrines in arid areas should concentrate where beavers are present.

- Remote mapping cannot replace field surveys: prominent rocks at the bottom of deep canyons are hidden from above, and beaver activity changes over time. Planning reintroductions or checking for otters will still need surveys for latrines on the ground.

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

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

- The study shows how knowing an animal’s habits can guide where trackers look: in dry canyon country, a search for otter latrines can start at large prominent rocks near deep water, under cliffs, and close to beaver sign.

## Cite this summary

Lawrence, K. 2026. Summary of: River Otter Latrine Site Selection in Arid Habitats of Western Colorado, USA (DePue and Ben-David, 2010). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2010-depue-ben-david-river-otter-latrine-site/