What is the Random Encounter Model (REM)?

REM is a method developed by Rowcliffe et al. (2008) to estimate the absolute density of animal populations from camera trap photos without requiring individual identification. It treats animals as randomly-moving particles and uses their day range plus the camera's detection profile to infer density from raw detection rate.

What is the REM formula?

The Rowcliffe et al. (2008) REM formula is D = (y/t) × pi / (v × r × (2 + theta)) × Eg, where y is the number of detections, t is total camera-time, v is mean daily animal speed (day range), r is effective detection radius, theta is the camera viewshed angle in radians, and Eg is mean group size. For solitary species set Eg = 1.

How is REM different from CTDS?

REM (Rowcliffe et al. 2008) requires a day-range estimate but does not require measuring distance to each detection. CTDS (Howe et al. 2017) requires measuring distance to every detected animal but does not require movement speed. Use whichever data type you have available; both produce absolute density without individual ID.

What detection radius (r) should I use for REM?

r is the average distance at which the camera reliably detects the target species. Typical values: deer-sized 8 to 15 m, large carnivores 10 to 20 m, small mammals 3 to 8 m. Measure empirically using a known-position object or use published species-specific values from reaction-distance trials.

How do I estimate day range (v) for REM?

Three options: (1) Use published values for the species and biome — most reliable. (2) Calculate from GPS-collared individuals (mean daily displacement). (3) Use the REM-v method (Rowcliffe et al., 2014), which estimates v from camera trap data itself when sequential photographs of moving individuals are available.

How many detections do I need for REM?

At least 50 independent detections per species are recommended for reliable estimates. With fewer detections, the coefficient of variation (CV) of density becomes inflated. The Poisson contribution to CV is approximately 1 over the square root of y, so 100 detections gives roughly 10 percent CV from sampling alone.

What is an independent detection in REM?

An independent detection is a single animal-camera contact event. Photos of the same individual within a short time window (typically 30 to 60 minutes) are pooled into one detection. The independence threshold should be reported and justified for the species in any publication.

Does camera placement matter for REM?

Yes — REM assumes random or stratified-random camera placement. Trail-based placement inflates detection rates and violates the random-movement assumption. If trails are unavoidable, the bias should be quantified and reported as a limitation.

Can REM handle group-living species?

Yes — multiply density by mean group size Eg to convert from group density to individual density. This tool includes the Eg factor in the formula. For solitary species, set Eg = 1.

What is the difference between REM and REST?

REM (Rowcliffe et al. 2008) uses day range; REST (Random Encounter and Staying Time; Nakashima et al. 2018) uses animal staying time within the detection zone instead of day range. REST is useful when day range is poorly known but staying time can be measured from camera trap video.

Random Encounter Model Calculator (REM)

📐 Random Encounter Model — Geometry & Measurements

The diagram below shows the REM detection geometry: the camera's detection zone, the effective detection radius r, the viewshed angle θ, and an animal moving through the zone at mean speed v. Updates live as you change config inputs.

The REM treats animals as randomly-moving "gas particles" and uses their movement speed (day range) plus the camera's detection profile (r, θ) to infer density from raw encounter rate y/t.

📥 1. Data Input

📊 Sample Dataset:

🦌 Analysis Mode:

Group / Species Name (editable):

Detection Distances (metres) — one value per animal sighting:

Enter at least 3 detection distances (in metres), one value per animal sighting. Each value = one independent detection. Single-value entries are not valid REM input.

Upload CSV or Excel file:

Supports .csv, .txt, .xlsx, .xls — expects two columns: species name + detection count (per camera or per species).

🦌 Analysis Mode for Upload:

Manual Entry — One Row Per Detection:

#	Detection Label / ID	Distance (m)

⚙️ 2. Analysis Configuration

🏞️ Study Area / Site Name (optional):

🎥 Number of Cameras:

📅 Camera Days (per camera, mean):

🚶 Day Range / Mean Animal Speed v (km/day):

Mean daily distance travelled by the species (its "day range"). Typical values: deer 4–8 km/day, wolves 10–25, lions 6–10, elephants 8–15, small mammals 0.3–1.5, large herbivores (wildebeest, zebra) 10–20.

📏 Effective Detection Radius r (metres):

Average distance at which the camera reliably detects the target species. Deer-sized: 8–15 m; large carnivores: 10–20 m; small mammals: 3–8 m. Measure in the field or use published species-specific values.

📐 Camera Viewshed Angle θ (degrees):

🕐 Active Hours per Day:

Hours per day the cameras are active (24 for always-on; lower if scheduled).

🐾 Mean Group Size Ē(g):

Average number of animals per detection event. Use 1 for solitary species. Use observed mean (e.g., 5.2 for wildebeest, 3.8 for elephant herds) for group-living species.

📐 Output Area Unit (for density):

📊 Results

REM Density Equation (Rowcliffe et al., 2008)

\[ \hat{D} = \frac{y}{t} \cdot \frac{\pi}{v \cdot r \cdot (2 + \theta)} \cdot \bar{E}(g) \]

D̂: Density (D̂) - Animals per output unit
y: Total number of independent detections
t: Total survey time (camera-days)
v: Mean daily animal speed / day range (km/day)
r: Effective detection radius (km)
θ: Camera viewshed angle (radians)
Ē(g): Mean group size

📊 Visualizations

1. Detection Distribution by Camera

2. Density Estimate with 95% Confidence Interval

3. Encounter Rate (y/t) per Species

4. Density Comparison Chart

📋 Detailed Results Table

🔬 3. Methodology

🧠 4. Plain-Language Interpretation

▸ Interpretation Results

▸ How to Write Your Results in Research

▸ Research Poster Panel

🔬 Technical Notes — Formula Derivation, Assumptions & Limitations

📌 When to Use REM — Decision Checklist & Examples

🔍 Conclusion

What Was Found

Ecological Significance

Conservation Implications

Next Steps

❓ Frequently Asked Questions

📖 How to Use This REM Calculator — Step by Step

📚 References

Key peer-reviewed sources on the Random Encounter Model, camera trap density estimation, and animal movement ecology.

Random Encounter Model (REM) Calculator – Free Online Tool