Generalized site occupancy models allowing for false positive and false negative errors

Area of Science:

• Ecology
• Wildlife Biology
• Conservation Science

Background:

• Site occupancy models are crucial for wildlife surveys, but traditionally assume no false positive detection errors.
• Existing models do not account for false positives, which can significantly bias occupancy estimates in real-world surveys.
• False positive errors, where a species is detected when absent, can occur due to various sampling issues.

Purpose of the Study:

• To develop and present a novel site occupancy model that incorporates both false negative and false positive error rates.
• To address the limitations of current models that assume perfect detection accuracy (no false positives).
• To provide a robust statistical framework for more accurate species distribution and abundance estimations.

Main Methods:

• Developed a two-component finite mixture model to simultaneously estimate false negative and false positive error rates.
• Utilized maximum likelihood estimation for fitting the proposed site occupancy model.
• Employed a simulation study to evaluate model performance against naive estimators under false positive conditions.

Main Results:

• The proposed model effectively accounts for both false negative and false positive detection errors.
• Simulation results indicate that naive estimators are highly biased in the presence of false positive errors.
• The new model provides more accurate site occupancy estimates when false positives are present.

Conclusions:

• Accounting for false positive errors is essential for reliable site occupancy estimation in ecological surveys.
• The developed finite mixture model offers a statistically sound and accessible method for improving survey data analysis.
• This work has significant implications for conservation planning and wildlife management by enhancing the accuracy of species distribution data.