A flexible framework for N-mixture occupancy models: applications to breeding bird surveys

Area of Science:

• Ecology
• Conservation Biology
• Statistical Modeling

Background:

• Estimating species abundance is crucial for biodiversity conservation but challenging due to imperfect detection.
• N-mixture models are widely used but rely on a strict closure assumption, leading to biased results when violated.

Purpose of the Study:

• To develop an extended ecological model that addresses the limitations of N-mixture models under violated closure assumptions.
• To provide a flexible framework that generalizes existing models and improves abundance estimation accuracy.

Main Methods:

• Proposed an extended mixed Gamma-Poisson model incorporating a community parameter.
• This parameter estimates the proportion of individuals consistently present during surveys.
• The framework generalizes zero-inflated occupancy and standard N-mixture models.

Main Results:

• Simulations and real-world data applications demonstrated improved accuracy.
• Tested on 5 species from the North American Breeding Bird Survey and 46 from the Swiss Breeding Bird Survey.
• The model shows adaptability in non-closed population settings.

Conclusions:

• The developed model offers a more robust approach to estimating species abundance with imperfect detection.
• It provides greater accuracy and flexibility than standard N-mixture models, especially when population closure is uncertain.
• This advancement aids biodiversity conservation efforts by providing reliable abundance estimates.