We examine the spatial scale of variability in recruitment for 11 marine, three anadromous, and five freshwater species. Generally the spatial scale of recruitment correlations for marine species is approximately 500 km, compared with less than 50 km for freshwater; anadromous species fall between these two scales. The scale for marine species is comparable with (but less than) that of the largest-scale environmental variables (and is compatible with the idea that large-scale environmental agents influence recruitment). Our results are consistent with the hypothesis that predation is a more important factor in determining recruitment in freshwater than it is in the marine environment.

We examine the spatial scale of variability in recruitment for 11 marine, three anadromous, and five freshwater species. Generally the spatial scale of recruitment correlations for marine species is approximately 500 km, compared with less than 50 km for freshwater; anadromous species fall between these two scales. The scale for marine species is comparable with (but less than) that of the largest-scale environmental variables (and is compatible with the idea that large-scale environmental agents influence recruitment). Our results are consistent with the hypothesis that predation is a more important factor in determining recruitment in freshwater than it is in the marine environment.Résumé : Nous avons examiné l'échelle spatiale de la variabilité du recrutement chez 11 espèces marines, trois espèces anadromes et cinq espèces d'eau douce. Généralement, l'échelle spatiale des corrélations de recrutement pour les espèces marines est d'environ 500 km, comparativement à moins de 50 km pour les espèces d'eau douce; pour les espèces anadromes, l'échelle se trouvent entre les deux. L'échelle des espèces marines est comparable (mais inférieure) à celle des variables environnementales de la plus grande échelle (et concorde avec l'idée que des agents environnementaux à grande échelle influent sur le recrutement). Nos résultats concordent avec l'hypothèse que la prédation est un facteur plus important pour la détermination du recrutement en eau douce que dans l'environnement marin. [Traduit par la Rédaction]

A variety of mechanisms have been proposed to explain the renowned British Columbia sockeye salmon (Oncorhynchus nerka) cycles, most of which invoke between-brood interactions (delayed density dependence) or depensatory harvest practices. We examine the dynamics of the Ricker model with realistic parameter values and suggest that the cycles could result from a stable mode excited by stochastic forcing. The previously proposed mechanisms are not required to generate cyclic patterns, although they could play a role in reinforcing cycles. Our results suggest that relaxing harvest rates will increase yields and decrease variability in sockeye abundances.

Delayed density-dependent mortality can be a cause of the cyclic patterns in abundance observed in many populations of sockeye salmon (Oncorhynchus nerka). We used a meta-analytical approach to test for delayed density dependence using 34 time series of sockeye data. We found no consistent evidence for delayed density-dependent mortality using spawner - spring fry or spawner-recruit data. We did find evidence for delayed density-dependent mortality at a 1 year lag for the spawner - fall fry and the spawner-smolt data; however, effects at lags greater than 1 year were weak and not statistically significant. Between-cohort interactions at a 1-year lag are not sufficient to cause the observed cycles in adult abundance, although they will tend to reinforce patterns established by other mechanisms. Our results imply that reductions in fishing mortality in off-peak years should result in an increase in abundance.

A variety of mechanisms have been proposed to explain the renowned British Columbia sockeye salmon (Oncorhynchus nerka) cycles, most of which invoke between-brood interactions (delayed density dependence) or depensatory harvest practices. We examine the dynamics of the Ricker model with realistic parameter values and suggest that the cycles could result from a stable mode excited by stochastic forcing. The previously proposed mechanisms are not required to generate cyclic patterns, although they could play a role in reinforcing cycles. Our results suggest that relaxing harvest rates will increase yields and decrease variability in sockeye abundances.Résumé : On a envisagé toutes sortes de mécanismes pour expliquer les cycles bien connus qu'on observe chez le saumon rouge (Oncorhynchus nerka) de la Colombie-Britannique : dans la plupart des cas, on suppose des interactions entre les pontes annuelles (effet retardé dépendant de la densité) ou des pratiques de pêche anticompensatoires. Nous avons étudié la dynamique du modèle de Ricker en attribuant aux paramètres des valeurs réalistes et nous pensons que les cycles du saumon rouge pourraient refléter un mode stable stimulé par un forçage stochastique. Les mécanismes envisagés antérieurement ne devaient pas nécessairement provoquer un effet cyclique, même s'ils pouvaient avoir un effet de renforcement. D'après les résultats que nous avons obtenus, la diminution de la pêche devrait se traduire par une augmentation des rendements et une baisse de la variabilité des effectifs du saumon rouge.[Traduit par la Rédaction]