Time-varying relationships between early growth rate and recruitment in Japanese sardine

•Effects of growth rate on recruitment changed over time.•Baseline recruitment level also changed over time.•Time-varying relationship model was more statistically plausible than time-constant models.•Growth–recruitment relationship is more dynamic than previously recognized. Growth is an important factor that regulates survival during early life stages, and growth rates and recruitment are often thought to be positively related. However, environmental conditions generally fluctuate over time, which could lead to temporal changes in growth–recruitment relationships. Here, we investigate whether growth–recruitment relationships are time-varying by applying dynamic linear models, which allow coefficients to vary over time, to a 23-yr data set of recruitment and early life-stage growth rate in Japanese sardine Sardinops melanostictus. We found that the impact of growth rate on recruitment (i.e., the slope coefficient) was relatively high during the late 1990s, then gradually weakened, and has remained low in recent years. The baseline recruitment level (i.e., the intercept) also shifted over time, being relatively low in the late 1990s, gradually increasing, and remaining high in recent years. Furthermore, a model with time-varying intercept and slope was more statistically plausible than models with constant coefficients. These results indicate that the growth–recruitment relationship varied dynamically over time. Our findings highlight the need to consider the possibility of time-varying growth–recruitment relationships when investigating such relationships and predicting recruitment.