A new parameterization for estimating co-occurrence of interacting species

Area of Science:

• Ecology
• Conservation Biology
• Statistical Modeling

Background:

• Estimating species co-occurrence with detection errors is challenging.
• Current models struggle with incorporating covariate effects on occurrence probabilities.

Purpose of the Study:

• Develop a new parameterization for species co-occurrence models.
• Improve the estimation of covariate effects on species occurrence.
• Model asymmetric interactions, such as predator-prey dynamics.

Main Methods:

• Developed a novel parameterization for co-occurrence models.
• Simulated data to test model performance under varying detection probabilities and sample sizes.
• Analyzed treefrog survey data from southern Florida to assess invasive species impacts.

Main Results:

• The new model allows natural inclusion of covariate effects without estimation issues.
• Unbiased occurrence probability estimates were achieved with 50-100 locations and high detection rates.
• Invasive Cuban treefrogs significantly reduced the likelihood of native treefrog presence, with no impact on native detection probabilities.

Conclusions:

• The new parameterization offers a robust approach for modeling species co-occurrence and ecological interactions.
• Effective for analyzing impacts of invasive species, like the Cuban treefrog, on native communities.
• Highlights the need for sufficient sample sizes, particularly when species detection is difficult.