2010 Volume 74 Issue 2 Pages 88-97
We investigated the relationships between environmental conditions and fluctuations in the recruitment of Japanese sardine, Sardinops melanostictus, in the northwestern Pacific. Stock-recruitment relationships were estimated based on Ricker, Beverton–Holt, power, linear, and proportional models. The error terms of the stock-recruitment relationship were assumed to follow normal or log-normal distributions. Furthermore, we calculated the correlation coefficients between environmental factors and residuals by subtracting estimated values from observed values. The correlation coefficients were calculated using two different datasets: the entire dataset (1976–2004), and the entire dataset minus data from 1988 to 1991, as crucial failures in recruitment occurred during this period. Analysis of the full dataset yielded a greater number of environmental factors with significant correlation coefficients than did analysis of the partial dataset. This finding suggests that the crucial failures in recruitment can be explained by environmental factors (e.g., Arctic Oscillation, Pacific Decadal Oscillation, sea surface temperature of southern area of the Kuroshio Extension), especially those factors that occurred during the spawning season, which are critical in terms of population fluctuations. These environmental factors should be taken into account in recruitment forecasting models.
