Performance of spatial capture-recapture models with repurposed data: Assessing estimator robustness for...

Area of Science:

• Ecology
• Statistical Modeling
• Wildlife Biology

Background:

• Spatial capture-recapture (SCR) models enhance traditional methods by integrating spatial capture data for direct animal density estimation.
• Repurposing existing capture-recapture data for SCR offers cost-effective ecological insights and supports decision-making.
• However, the suitability of data from studies not originally designed for SCR is often unevaluated.

Purpose of the Study:

• To evaluate the robustness of SCR models for retrospectively estimating large mammal densities using repurposed, non-spatially designed capture-recapture data.
• To assess how varying simulation scenarios, including asymmetrical sampling grids and heterogeneous landscapes, impact SCR model performance.
• To identify limitations and provide guidance on the retrospective application of SCR models.

Main Methods:

• Utilized simulation to test SCR model performance on repurposed capture-recapture data under diverse ecological and sampling conditions.
• Investigated scenarios with asymmetrical sampling grids, broad spatial extents, and heterogeneous landscapes.
• Analyzed bias and precision of density estimates derived from SCR models fitted to simulated, repurposed data.

Main Results:

• SCR model performance using repurposed data was not robust, exhibiting considerable variation in bias and precision across simulation scenarios.
• Relative bias in density estimates ranged across 14 orders of magnitude, with the smallest bias at 3%.
• Detection parameters were identified as the primary factor influencing the variability of density estimates.

Conclusions:

• Casual repurposing of non-spatial capture-recapture data for SCR analysis is cautioned against due to potential unreliability.
• Simulation is crucial for assessing the performance and limitations of SCR models in retrospective applications.
• Future studies should carefully consider data design when planning for SCR analysis or retrospective data use.