Statistics, Software, and Fish Stock Assessment
J.T. Schnute, L.J. Richards, and N. Olsen
- Price: $0.75 Sale: $0.00
Classical regression, analysis of variance, and experimental design fail to provide adequate tools for building modern fishery models. In the last decade, however, both statistical theory and software have progressed in directions that make realistic stock assessment models possible. We review the modern theory and show how to apply it to complex fishery analyses, in which dual simulation and estimation models relate parameters to available data. In particular, we demonstrate that state space models in a Bayesian framework provide useful design prototypes. More generally, the theoretical framework leads to probability distributions that must be translated into computer code. We discuss software requirements for implementing this framework. We also present benchmark results from a catch-at-age model rendered in four current software environments.
- Item number: AK-SG-98-01i
- Year: 1998
- DOI: https://doi.org/10.4027/fsam.1998.09