Inferring invasive species abundance using removal data from management actions

Area of Science:

• Wildlife ecology
• Invasive species management
• Statistical modeling

Background:

• Evaluating invasive species management is vital for resource allocation and demonstrating impact.
• Estimating species abundance before and after management is a key metric.
• Current abundance estimation methods are often too costly and labor-intensive for operational programs.

Purpose of the Study:

• To develop a Bayesian hierarchical model for estimating abundance from removal data, accounting for varying effort.
• To assess the conditions for reliable population estimates using simulations.
• To apply the model to estimate feral swine (Sus scrofa) abundance in Oklahoma and Texas.

Main Methods:

• Developed a Bayesian hierarchical model utilizing removal data.
• Conducted simulations to determine conditions for accurate abundance estimates.
• Applied the model to aerial gunning data from 59 site/time combinations in Oklahoma and Texas.

Main Results:

• Abundance estimates were accurate when effective removal rates exceeded 0.40.
• Higher effective removal rates (0.70) were needed for small populations (<50).
• 78% of site/time estimates were accurate; accuracy improved with increased sampling effort and removal rate.

Conclusions:

• The Bayesian model provides reliable abundance estimates from removal data, crucial for invasive species management.
• Effective removal rates, sufficient sampling effort, and closed population assumptions (≤3 months) are key for model accuracy.
• Incorporating auxiliary data like habitat or pilot differences can further enhance model precision.