Bayesian Analysis of the Ricker Stock-Recruitment Model

Wood, E.F. (1975). Bayesian Analysis of the Ricker Stock-Recruitment Model. IIASA Research Memorandum. IIASA, Laxenburg, Austria: RM-75-065

Preview

Text RM-75-065.pdf Download (460kB) | Preview

Abstract

Bayesian statistics were applied to analyze the Ricker stock-recruitment model. This model is used for salmon fishery management and predicts the resulting recruits for a specified level of spawners. The Ricker model is transformed into a linear regression form, and the uncertainty in the model parameters and the "noise" of the model are calculated using Bayesian regression analysis. Application to the Skeena River Sockeye fishing with 67 years of data showed that parameter uncertainty was much less than model noise, thus putting in question the applicability of the Ricker model for management decisions. The analysis is extended to experimentation on the "optimal" spawning level, non-stationarity of the fishery, and model uncertainty.

Actions (login required)

View Item