Abstract
Environment‐recruitment relationships often break down over time. This can occur because of spurious results from testing numerous environmental variables in the original analyses or from relative changes in the importance of specific environmental drivers over time. An environment‐recruitment model was developed for the Newfoundland and Labrador capelin stock, but the relationship first weakened with the addition of new data and then recently broke down. We hypothesized that the thermal and wind parameters in the original environment‐recruitment model were now less important for recruitment dynamics following a stock collapse in 1991 due to the observed change in stock demographics, while other mechanisms like timing of spawning and annual climate conditions may be more important. Furthermore, three different measures of age‐2 recruitment were used in the original model and subsequent re‐testing of the environment‐recruitment relationships, which may have instigated a spurious breakdown of this relationship. Using a consistent empirical measure of age‐2 recruitment for the 1983–2022 year‐classes, an earlier timing in peak spawning predicted higher age‐2 recruitment, although only a moderate level of variability ( R 2 < 40%) was explained. The timing of spawning has been persistently delayed since the stock collapsed in 1991, which has had a profound impact on stock dynamics. This study demonstrates the importance of renewed and frequent reassessment of prior knowledge of stock dynamics and environment‐recruitment relationships, especially when dramatic changes in stock dynamics or ecosystem regime shifts are observed.