%0 Conference Proceedings
%9 ACTI : Communications avec actes dans un congrès international
%A Katara, I.
%A Gaertner, Daniel
%T Some news approaches for standardizing tropical purse seiners CPUEs
%S 16ème groupe de travail sur les thons tropicaux
%C Victoria
%D 2014
%L fdi:010063285
%G ENG
%I CTOI
%K PECHE THONIERE ; SENNEUR ; CAPTURE ; EFFORT DE PECHE ; ESTIMATION DE STOCK ; MODELISATION ; ANALYSE STATISTIQUE
%K OCEAN INDIEN
%N IOTC-2014-WPTT16-16
%P 16
%U https://www.documentation.ird.fr/hor/fdi:010063285
%> https://horizon.documentation.ird.fr/exl-doc/pleins_textes/divers15-01/010063285.pdf
%W Horizon (IRD)
%X Indices of CPUE are presented for skipjack tuna (Katsuwonus pelamis) for the Indian Ocean over the period 1980-2013. The analysis was initiated under the framework of the E.U. research project CECOFAD, whose objectives are to enhance our understanding of fishing effort units and improve the accuracy of purse-seine CPUE estimates. Skipjack stock assessments mainly depend on abundance indices derived from purse-seine fleets, thus standardisation of purse seiner CPUEs for this species is a priority. We follow a CPUE definition framework where three different types of CPUE are calculated to describe trends in number of schools, school detectability and catchability and school size. CPUE standardization is typically achieved with the development of GLM models; these estimates are often biased because the range of the fishing area changes from one year to the next. We compare CPUE trends derived from conventional GLMs with trends based on GLMMs, where the spatial explanatory variables are treated as a random effect. According to a definition, a variable is random when it constitutes only a part of the entire population: spatial variables can be considered random since each year fishing vessels sample only a part of the historically fished area. Exploratory analyses revealed issues of heterogeneity and autocorrelation, supporting the use of mixed models and the shift towards a Bayesian approach to CPUE standardization modeling. Our analysis uses fishing sets targeting free swimming tuna schools for the French purse seiners operating in the western Indian Ocean. Future work will address the problem of accounting for the major changes in purse seine fishing strategies and the implementation of new technology on board (mainly the use of FAD technology) under a Bayesian framework.
%B GTTT : Groupe de Travail sur les Thons Tropicaux
%8 2014/11/15-19
%$ 040PECHE03 ; 040INFSTA01