ANALYSIS OF CONTRADICTORY DATA SOURCES IN FISH STOCK ASSESSMENT

Fisheries stock assessments sometimes prove, in retrospect, to be wron g. Errors may be due to poor model assumptions or to data that do not reflect the biological process of interest. We develop a method that f ormally admits the possibility of such errors. Likelihood functions de rived from this method indicate greater uncertainty in parameter value s than conventional likelihoods, whose derivations presume that models correctly describe the observed data. The problem of uncertainty is p articularly acute when more than one data source is available and diff erent data sets provide contradictory parameter estimates. Traditional methods of stock assessment involve weighted averages of the contradi ctory data, and these generally produce parameter estimates intermedia te to those obtained from the data sets individually. We demonstrate t hat, when model or data errors are considered, the most likely paramet er values are not intermediary to conflicting values; instead, they oc cur al one of the apparent extremes. We provide an example using contr adictory trends in catch-per-unit-effort data for the Canadian norther n cod stock (1978-88).