A fast re-sampling method for using reliability ratings of sightings with extinction-date estimators

Area of Science:

• Ecology
• Conservation Biology
• Computational Biology

Background:

• Estimating species extinction dates often relies on sighting patterns, but the reliability of these observations is typically uncertain.
• Traditional methods may not adequately account for the probabilistic nature of sighting data, potentially leading to inaccurate extinction date estimations.
• Physical captures or specimens are rare, making observational data crucial yet ambiguous for extinction inference.

Purpose of the Study:

• To develop and present a computational approach for integrating observational reliability into extinction date estimators (EDE).
• To provide a generalized and robust method for inferring extinction dates that accommodates varying levels of sighting data uncertainty.
• To offer a flexible tool applicable to diverse statistical EDEs without being tied to a specific model.

Main Methods:

• A novel computational approach that probabilistically combines within-year sightings.
• Incorporates observational reliability as an inclusion probability for sampling observations.
• Infers probability distributions and summary statistics for extinction dates using any EDE, and computes extinction date frequency distributions.

Main Results:

• The method was applied to eight diverse sighting records, demonstrating its applicability across various data complexities.
• Comparison with a threshold-based sighting selection approach showed improved performance.
• Validated through real-world examples of rediscovered and extinct species, and simulated data, confirming robust coverage of true extinction dates.

Conclusions:

• The proposed method offers a powerful generalization for extinction date estimation by incorporating observational reliability.
• Its simplicity and compatibility with various EDEs make it a valuable tool for conservation science.
• The R script provided facilitates easy implementation for researchers studying species extinction.