Estimating species occupancy across multiple sampling seasons with autologistic occupancy models via the autoOcc...

Area of Science:

• Ecology
• Statistical Modeling
• Computational Biology

Background:

• Dynamic occupancy models require substantial data for estimating colonization and extinction rates.
• Many ecological studies face limitations in sample size, hindering the application of dynamic occupancy models.
• Autologistic occupancy models present a parsimonious approach by accounting for temporal autocorrelation in species occupancy.

Purpose of the Study:

• Introduce the autoOcc R package for fitting autologistic occupancy models within a frequentist framework.
• Provide a flexible and accessible tool for analyzing detection/non-detection data collected over time.
• Demonstrate the utility of autologistic occupancy models as an alternative to dynamic models, particularly for datasets with limited sample sizes.

Main Methods:

• Developed the autoOcc R package to implement autologistic occupancy models.
• Utilized simulations to compare the performance of autologistic and dynamic occupancy models across various scenarios.
• Included functionalities for model comparison (AIC) and prediction from fitted models.
• Applied the autoOcc package to real-world case studies involving habitat associations and spatiotemporal distribution patterns.

Main Results:

• Autologistic occupancy models demonstrated reduced bias and increased precision in parameter estimation compared to dynamic occupancy models, even with limited data.
• The autoOcc package provides a robust platform for fitting and analyzing autologistic occupancy models.
• Simulations confirmed the effectiveness of autologistic models across diverse sample sizes and ecological scenarios.
• Case studies illustrated the practical application and ecological interpretability of autologistic occupancy models.

Conclusions:

• Autologistic occupancy models are a valuable and efficient alternative to dynamic occupancy models when dealing with limited ecological data.
• The autoOcc R package lowers the barrier for researchers to utilize autologistic occupancy modeling for studying species occupancy dynamics.
• This work expands the toolkit available for ecological research on species distribution and habitat use over time.