Hierarchical bayesian modeling of recruitment: A semiparametric approach

Models of recruitment are important for determining management reference points. Hierarchical modeling of recruitment allows us to 'borrow strength' across data sets to improve inference in data limited situations.  Unfortunately reference points are often highly sensitive to the specific model used to estimate the stock-recruitment relationship.  Semiparametric Bayesian approaches to modeling recruitment are robust to structural uncertainty and allow us to infer the shape of the stock-recruitment model whenever there is sufficient shape information in the data. Here I extend the this method to hierarchical and spatially dependent cases and show that significant improvements in estimating the relationship between stock size and recruitment.