Generic modelling tools for the assessment of marine populations

Mormede S., Dunn A., Pinkerton M.

Date de parution: septembre 2011
Volume: 35
Number: SP
Pagination: 257-262

How to cite: Mormede, S., Dunn, A., & Pinkerton, M. (2011). Generic modelling tools for the assessment of marine populations. Cybium, 35(SP): 257-262.


In the New Zealand EEZ, most fish stocks are managed using TAC limits. Much of the scientific advice for the management of these stocks and associated species was derived from analyses using generic or standardised software packages and modelling techniques. These modelling techniques have also been applied to several Antarctic fish populations through various CCAMLR working groups, including the setting for TAC of toothfish fishery. CASAL is an age- (or size-) structured fish stock assessment model. Developed about ten years ago, it was mostly used in NZ EEZ stock assessments and in CCAMLR toothfish stock assessments. SeaBird is a population modelling package developed to undertake assessments of sea birds and help evaluate the impact of fisheries-related mortality on sea bird populations. The Spatial Population Model (SPM) is a spatially explicit integrated statistical catch-at-age stock assessment model that is similar in functionality aspects to CASAL. Implemented as a generalised package, it investigates the dynamics of more spatially complex populations, in particular, to test assumptions surrounding tag mixing in toothfish populations. All three models share similar design and structural characteristics: they generate integrated analyses, where there is flexibility in specifying the population dynamics, parameters, and model outputs; flexibility in the choice of observations; and are structured to allow easy interpretation of input and output files. Each of these models was designed to have a high level of documentation, quality assurance procedures, and scientific transparency to allow for robust peer review. The use of a documented, robust, and validated software tool in assessments, benefits both scientists and science managers. For scientists, it allows easy documentation of methods through the input and output files, improved efficiency in producing outputs, reduction of errors in underlying software code and its implementation, and easy transfer of methodology between scientists (regardless of stocks). For managers, benefits include greater efficiency in obtaining scientific advice, common methods used in assessments between stocks, and a standardised terminology. In addition to the standardised software packages, we also briefly describe two other modelling methods relevant to developing an ecosystem approach to fisheries: a parameter adjustment technique for mass-balance food-web models, and a boosted regression technique for developing relationships between species distributions and environmental conditions.

Mots-clés: Bayesian - Fisheries - Integrated analyses - Population models - Software - Stock
PDF visible pour tous