Bayesian hierarchical models for estimating nest survival

Abstract

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Nest survival rate is a critical value in avian study to evaluate the landbirds populations. The widely used likelihood-based logistic regression model was evaluated in the first part of the dissertation. In this part, we investigated the importance of nest age in estimating survival rates and measured the model selection accuracy based on AIC results. Next we extended Bayesian Hierarchical Model to include different nest period lengths which estimated the overall survival rates and survival curves with combined nest period lengths. For unknown nest fate, the nest fate effect and the nest-specific covariates were included in the missing probability estimation. We also compared the results from incomplete data with the results from complete data analysis. The estimated overall survival rates and survival curves all supported the model performance. Finally, we included the spatial effect into the age-specific outcome rates estimation. The point-level nest observations explained the nest-specific spatial effect within each unit; while the grid-level data explained the spatial effect between different units. In each part, a simulation study was conducted to evaluate the performance of the model and an application was also provided. All the programs were written in FORTRAN and a R package (function) was created to make it more user friendly.