Accounting for location uncertainty in azimuthal telemetry data improves ecological inference

Area of Science:

• Ecology
• Conservation Biology
• Spatial Ecology

Background:

• Animal space use is key to understanding ecological relationships.
• Radio-telemetry provides valuable spatial data but lacks advanced statistical models for azimuthal data.
• Existing models do not fully integrate location uncertainty into ecological analyses.

Purpose of the Study:

• To develop a hierarchical modeling framework for robust animal location estimates from azimuthal telemetry data.
• To create a method that accounts for azimuthal uncertainty and propagates location uncertainty into ecological models.
• To evaluate the performance of the new model against existing estimators and demonstrate its impact on ecological inference.

Main Methods:

• Developed a hierarchical modeling framework: the azimuthal telemetry model (ATM).
• Used simulation studies to evaluate ATM performance against Lenth (1981) maximum likelihood and M-Estimators.
• Applied the ATM to empirical data for home range and resource selection analyses, comparing results with and without accounting for location uncertainty.

Main Results:

• The ATM accurately estimates animal locations and provides appropriate coverage measures.
• Ignoring location uncertainty leads to overconfident and conservative home range estimates.
• Failure to account for location uncertainty results in incorrect inference and overconfidence in resource selection coefficients.
• Incorporating location uncertainty reduces bias in resource selection coefficients, even with high covariate spatial autocorrelation.

Conclusions:

• The ATM effectively estimates animal locations using one or more azimuths, utilizing all collected data.
• This model has significant implications for interpreting past radio-telemetry studies and designing future research.
• Accounting for location uncertainty is crucial for accurate ecological inference in animal movement studies.