Relationship between the recruitment of the Pacific-cohort of chub mackerel (Scomber japonicus) and the influence factors on the spawning ground based on GLM and GAM

Chub mackerel (Scomber japonicus) is an important pelagic fish in the Northwest Pacific Ocean. It is necessary for us to find the relationship between the abundance and influence factors, which are beneficial to exploit and utilize this resource. In this study, based on the recruitment data and the spawning stock biomass (SSB) data of the Pacific-cohort of S.japonicus during 1980—2016 obtained from Japan fisheries institution, the normality test of natural logarithm of recruitment was finished and the time period that passed the normality test was during 1980—1999, with the environmental data of spawning ground, we analyzed the relationship between the sea surface height (SSH), sea surface salinity (SSS), sea surface temperature (SST) and the natural logarithm of SSB ln(SSB) and the recruitment during 1980—1999 with generalized linear model (GLM) and generalized additive model (GAM). The GLM results revealed the order of importance of variables ranked by decreasing magnitude was ln(SSB) ×Year, ln(SSB), SSS ×Year and SSS, which were significant (P<0.05) and considered the combined effects of factors. Considering the single factor in GLM models affecting the recruitment, order of the importance of variables ranked by decreasing magnitude was SST, SSS, Year, ln(SSB) and SSS. The GAM results indicated that the model which contained Year, SST and SSH was the optimal model based on Akaike’s Information Criterion (AIC), the importance ranked by decreasing magnitude was Year, SST and SSH. However, considering the single factor in GAM models affecting the recruitment, the importance of variables ranked by decreasing magnitude was Year, SSS, ln(SSB), SST and SSH. The suitable range of SSH was 62—65 cm, the suitable range of SSS was 34.72—34.74 and 34.78—34.83 and the suitable range of SST was 20.2—20.6 °C. When ln(SSB)>6.0, the recruitment was at a high level, based on GAM analysis.