Models for analysing species' presence/absence data at two time points.

Area of Science:

• Ecology
• Statistical Ecology
• Environmental Science

Background:

• Species presence/absence data at two time points are fundamental in ecological studies, including succession, monitoring, and climate change research.
• Traditional statistical regression methods have limitations for analyzing this common longitudinal ecological data type.

Purpose of the Study:

• To propose and evaluate the bivariate odds-ratio model for analyzing species presence/absence data.
• To integrate this model within a constrained ordination framework for ecological niche theory dynamics.

Main Methods:

• Application of the bivariate odds-ratio model, seldomly used in ecology, within a constrained ordination framework.
• Exploration of extensions, including complementary log-log links for marginal probabilities with a Poisson abundance model.
• Development of the model based on the zero-inflated Poisson distribution to account for excess absences.

Main Results:

• The bivariate odds-ratio model, within a constrained ordination framework, offers a valuable tool for ecological analysis.
• The proposed framework can describe local extinction and colonization probabilities, contributing to niche theory dynamics.
• Demonstration of the model's utility using two vegetation datasets.

Conclusions:

• The constrained ordination-odds ratio framework provides a powerful approach for understanding ecological processes.
• The zero-inflated Poisson distribution-based model is particularly suitable for ecological data exhibiting excess absences.
• This statistical methodology advances the analysis of longitudinal species occurrence data.