PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Detecting and correcting underreported catches in fish stock assessment
Nicolas Bousquet, Noel Cadigan, Thierry Duchesne and Louis-Paul Rivest
CFJAS 2009.


In many cases landings from fisheries are under-reported; that is, the true landings are greater than those reported. Despite this bias, reported landings are still used in fish stock assessments which might lead to overoptimistic exploitation strategies. Among others, the cod stock (Gadus morhua) from the southern Gulf of St. Lawrence (Canada), which has declined severely in recent years, is known to have unreported overfishing during 1985-1992. In this article, we use the statistical theory of censoring, which was recently introduced by Hammond and Trenkel (2005) for surplus production models, to correct estimates of catches. We show in simulations that this approach can detect (or confirm) and correct the problematic landings, reducing significantly the error implied by considering landings as true values. In typical frameworks such corrections are found nearly insensitive to subjective boundaries placed on real catches, and robust to modifications imposed to the variance structure of simulated catches

EPrint Type:Article
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Learning/Statistics & Optimisation
ID Code:5370
Deposited By:Nicolas Bousquet
Deposited On:31 March 2009