Simulation-based validation of spatial capture-recapture models: A case study using mountain lions

Area of Science:

• Ecology
• Wildlife Biology
• Statistical Modeling

Background:

• Spatial capture-recapture (SCR) models are vital for estimating densities of rare wildlife.
• Validation of SCR models is infrequent despite diverse formulations and data integration.
• Understanding the link between encounter probabilities, additional data, and density estimate reliability is critical.

Purpose of the Study:

• To validate SCR models using a simulation-based approach with spatially unstructured sampling.
• To assess the accuracy and precision of density estimates under varying data sparsity and potential bias.
• To evaluate the impact of incorporating harvested and telemetry data on SCR model performance.

Main Methods:

• Simulated data under six scenarios varying search effort and its correlation with density.
• Applied four SCR models with increasing amounts of harvested and telemetry data.
• Assessed density estimate accuracy and precision across scenarios and model complexities.

Main Results:

• Density estimates were sensitive to information quantity; low effort yielded biased, imprecise results.
• High search effort produced unbiased and precise density estimates.
• Correlation between effort and density introduced bias, reduced by more informative datasets.
• Harvested and telemetry data improved estimates for low/moderate effort, with minimal impact at high effort.

Conclusions:

• SCR models with spatially unstructured sampling provide reliable density estimates when sufficient information is available.
• Empirical-based simulations are essential for designing studies with appropriate effort and data sources for accurate density estimation.
• Validated SCR models enhance the reliability of wildlife population assessments.