A simple framework for maximizing camera trap detections using experimental trials.

Area of Science:

• Ecology
• Wildlife Biology
• Conservation Technology

Background:

• Camera trap data can be biased by missed animal detections, particularly with passive infrared sensors.
• Detection accuracy depends on the interplay between sensor capabilities and animal characteristics.
• Optimizing camera trap deployment requires understanding these detection limitations.

Purpose of the Study:

• To develop and present a general experimental framework for evaluating camera trap detection errors.
• To quantify the effects of distance, camera model, lens height, and angle on detection probability for different mammal sizes.
• To provide a method for improving the design and analysis of camera trap studies.

Main Methods:

• Adapted distance sampling models to estimate detection probabilities.
• Conducted experiments varying distance, camera model, lens height, and vertical angle.
• Tested detection on proxies representing small, medium, and large mammals across different biomes.
• Utilized a half-normal-logistic mixture model to explain detection probabilities.

Main Results:

• Detection probability significantly declined beyond 6 meters from the camera.
• Animal body size and camera model were key factors mediating the effect of distance on detection.
• All experimental covariates (distance, camera model, lens height, vertical angle) influenced detection probabilities.
• Solar position can introduce unmodeled heterogeneity and bias inferences.

Conclusions:

• A robust experimental and analytical framework exists for assessing camera trap detection probabilities.
• Understanding detection heterogeneity is crucial for accurate wildlife population estimates.
• This framework aids in optimizing camera trap protocols and resource allocation in ecological studies.