Factors influencing soay sheep survival: a Bayesian analysis.

Area of Science:

• Ecology
• Statistical Modeling
• Bayesian Inference

Background:

• Soay sheep populations are crucial for ecological studies.
• Understanding survival rates is vital for population management.
• Previous analyses using classical methods may have limitations.

Purpose of the Study:

• To apply Bayesian analysis to mark-recapture-recovery data for Soay sheep.
• To model survival probabilities using logistic regression with various covariates.
• To identify previously undetected ecological factors influencing survival.

Main Methods:

• Bayesian analysis of mark-recapture-recovery data.
• Reversible jump Markov chain Monte Carlo (MCMC) for age class determination.
• Logistic regression incorporating environmental, individual, and random effects.
• Imputation of missing covariate data using auxiliary variables.

Main Results:

• The Bayesian approach yielded different models compared to classical methods.
• Model averaging identified ecologically significant features not previously detected.
• Survival probabilities were modeled considering complex covariate interactions.

Conclusions:

• Bayesian methods offer a more comprehensive analysis of Soay sheep survival data.
• New ecological insights into sheep population dynamics were uncovered.
• This approach enhances the understanding of factors affecting wildlife survival.