Environmental and stock effects on the recruitment of anchovy in the Bay of Biscay : a multivariate analysis
|Allain Gwenhael1, Petitgas Pierre1, Lazure Pascal2
|1 : Ifremer, Centre de Nantes, BP 12105, F-44311 Nantes, France
2 : Ifremer, Centre Bretagne, ZI de la Pointe du Diable - CS 10070, F-29280 Plouzané, France
|International Council for the Exploration of the Sea 1999 Conference
|International Council for the Exploration of the Sea Conference. ICES CM 1999/Y:22
|Bio-physical coupling, Recruitment, Anchovy, Hdrodynamics, Step by step regression
Variability in anchovy recruitment in Biscay is temptatively explained using variables from the stock and from the environment during the ichtyoplankton phase. Anchovy recruitment depends on many different processes : weight of adults and proportion of multiple spawners (age 2), river discharges, upwelling events at the coast and at the shelf break. Meteorological variables (wind, rain, tempe rature ) are forcing effects on the sea but they are not the effective processes that govern the production in the sea. So they do not relate directly to the survival of larvae and to recruitment. In Biscay, many different physical processes are active and used by anchovy which spawns in different physical systems (river plumes, open ocean enrichment systems). Therefore we use a hydrodynamical physical model to construct environmental variables that relate directly to the physical pro cesses that occur in the sea. Such environmental variables are thought to have a more powerfull explanatory power than general meteorological variables alone. Stock effects are parametrised by the importance in the population of year class 2 and by individual average weight. Environmental effects are physical indices derived from the physical model: vertical turbulence, stratification, upwelling, river plumes. As we have a small amount of years and a lot of variables, a step by step regression analysis (Linear Model) is used to analyse hierarchy in the explanatory power of the factors. Upwelling is the first explanatory variable with a very strong positive effect on recruitment, then cornes breakdown of stratification events with a negative effect. All other variables have a smaller explanatory power on recruitment and the stock effect is found to be not significant. A model with 2 physical variables explains 80% of the recruitment variability observed in the 1987-1996.