2019 Volume 83 Issue 3 Pages 181-190
We proposed a new forecasting method of annual (May–December) landing of immature fresh skipjack tuna at Kesennuma port (northeastern Japan) in May, whose annual landing of fresh skipjack tuna has been the largest in Japan over the past 22 years. Monthly landing of fresh skipjack tuna by commercial size categories at Kesennuma during May–December and Boso Katsuura (middle-eastern Japan, hereafter called Katsuura) during January–April, 2002–2018 were converted into putative seasonal cohorts: large-sized (XA group), middle-sized (BC group), and small-sized (DE group), by using estimated growth curves for these groups. In order to examine relationships among annual/seasonal landing of these groups, considering an observed declining trend in fishing effort of Japanese skipjack pole-and-line fishery, which is the major fishery for fresh skipjack tuna, original landing data for each group was detrended as year-specific residuals from a fitted linear regression line. The BC group, the most abundant among the three groups, migrates from subtropical waters to northeastern waters off Japan, i.e., Kuroshio-Oyashio Transition Zone (KOTZ), during early summers. There was a statistically significant positive relation between detrended annual landing of BC group at Kesennuma and detrended landing during January–April at Katsuura. The XA group is supposed to represent DE groups in a previous year according to recently estimated growth curves, and it migrates to KOTZ in summers. There was also a statistically significant positive correlation between detrended annual landing of XA group and that of DE group in a previous year at Kesennuma. The best multiple regression model suggested that annual landing of the BC group at Kesennuma in a given year can be forecasted with two explanatory variables: landing of BC group at Katsuura during January–April and mean March sea surface temperature around the southern Izu Islands, south of Katsuura, in the same year.