Inferring extinction risks from sighting records.

Area of Science:

• Ecology
• Conservation Biology
• Statistical Modeling

Background:

• Accurate extinction probability estimation is vital for effective conservation resource allocation.
• Existing frameworks often struggle with the variable quality and uncertainty inherent in historical sighting records.

Purpose of the Study:

• To develop a more general and robust statistical framework for estimating species extinction probability.
• To accommodate both definite and uncertain observational data, reflecting real-world sighting record complexities.

Main Methods:

• Formulation of a generalized Bayesian framework incorporating observation probabilities from a probability density function.
• Implementation of two distinct methods ('quenched' and 'annealed') to incorporate randomness in observation probabilities.
• Development of a deterministic calculation for cases with only definite sightings.

Main Results:

• The proposed framework is more general than existing methods, handling both definite and uncertain observations.
• The 'quenched' and 'annealed' approaches can yield significantly different extinction probability estimates.
• The method was successfully applied to the Caribbean monk seal (Monachus tropicalis) and synthetic datasets.

Conclusions:

• The developed framework provides a more realistic approach to extinction probability estimation by accounting for observation uncertainty.
• This generalized method can improve the accuracy of conservation policy decisions by refining extinction risk assessments.
• The flexibility in handling diverse observational data types enhances its applicability across various species and datasets.