Evaluating Multi-Level Models to Test Occupancy State Responses of Plethodontid Salamanders

Area of Science:

• Ecology
• Conservation Biology
• Statistical Modeling

Background:

• Plethodontid salamanders are ecologically important but challenging to study due to habitat complexity and sampling limitations.
• Effective sampling designs and estimators are vital for robust inference on salamander ecology and conservation.
• Hierarchical and multi-scale occupancy models offer potential for analyzing complex ecological data.

Purpose of the Study:

• To evaluate the effectiveness of multi-scale and hierarchical single-scale occupancy models for Plethodontid salamanders.
• To assess the impact of sampling design on estimator precision and bias within a Before-After Control-Impact (BACI) framework.
• To apply these models to empirical data for Oregon slender and Ensatina salamanders.

Main Methods:

• Conducted a simulation study comparing multi-scale and hierarchical single-scale occupancy models.
• Utilized a Bayesian framework for model fitting.
• Applied models to empirical data from 66 forest stands in the Cascade Range, Oregon, over two years.

Main Results:

• Estimator precision improved with increased primary and secondary sampling units for both models.
• Both models exhibited estimator bias at low detection probabilities and sample sizes, more so for the multi-scale model.
• Sufficient sample sizes at all levels can mitigate bias.
• Oregon slender salamander occupancy strongly correlated with coarse woody debris; Ensatina occupancy did not.

Conclusions:

• Both multi-scale and hierarchical single-scale occupancy models are suitable for Plethodontid salamander studies with adequate sample sizes.
• Model choice depends on specific research questions and objectives, as they estimate different parameters.
• Careful consideration of study design is essential for reliable ecological inference and conservation management.