That's not the Mona Lisa! How to interpret spatial capture-recapture density surface estimates

Area of Science:

• Ecology
• Spatial Statistics
• Wildlife Biology

Background:

• Spatial capture-recapture (SCR) models are widely used to estimate animal density and distribution.
• A common misinterpretation arises from confusing spatial variation in density with spatial variation in uncertainty about density.
• This confusion can lead to flawed ecological inferences.

Purpose of the Study:

• To clarify the distinction between estimating the intensity of an activity center process and predicting activity center locations in SCR models.
• To visually demonstrate correct and incorrect inference using a novel simulation approach.
• To provide guidance for accurate interpretation of SCR-derived density surfaces.

Main Methods:

• Simulated SCR data from a grayscale image of the Mona Lisa, treated as an activity center intensity surface.
• Analyzed inferences about point process intensity and predicted activity center locations.
• Compared predictions based on detector placement and survey effort.

Main Results:

• Treating probabilistic predictions of activity center locations as intensity estimates leads to invalid ecological inferences.
• Predictions are highly sensitive to survey design, including detector placement and effort.
• Correct inference requires estimating the intensity of the underlying point process.

Conclusions:

• Researchers must explicitly state whether they are estimating process intensity or predicting locations.
• Predictions of activity center locations should not be conflated with estimates of density or intensity.
• Accurate application of SCR methods is crucial for reliable ecological studies.