Unravelling the Gordian knot! Key processes impacting overwintering larval survival and growth: A North Sea herring case study

•A 30-yr time series reveals spatiotemporal changes in larval herring survival and growth processes.•Biophysical modelling connected post winter larval abundance and length to pre-winter spawning locations.•Modelled and observed larval length agreed well in specific prey, mortality and behaviour scenarios.•Field and model results suggest altered larval growth environments after 1990 in the southern North Sea.•Size-based mortality and prey field estimates are needed to understand overwinter survival dynamics.

Unraveling the key processes affecting marine fish recruitment will ultimately require a combination of field, laboratory and modelling studies. We combined analyzes of long-term (30-year) field data on larval fish abundance, distribution and length, and biophysical model simulations of different levels of complexity to identify processes impacting the survival and growth of autumn- and winter-spawned Atlantic herring (Clupea harengus) larvae. Field survey data revealed interannual changes in intensity of utilization of the five major spawning grounds (Orkney/Shetland, Buchan, Banks north, Banks south, and Downs) as well as spatio-temporal variability in the length and abundance of overwintered larvae. The mean length of larvae captured in post-winter surveys was negatively correlated to the proportion of larvae from the southern-most (Downs) winter-spawning component. Furthermore, the mean length of larvae originating from all spawning components has decreased since 1990 suggesting ecosystem-wide changes impacting larval growth potential, most likely due to changes in prey fields. A simple biophysical model assuming temperature-dependent growth and constant mortality underestimated larval growth rates suggesting that larval mortality rates steeply declined with increasing size and/or age during winter as no match with field data could be obtained. In contrast better agreement was found between observed and modelled post-winter abundance for larvae originating from four spawning components when a more complex, physiological-based foraging and growth model was employed using a suite of potential prey field and size-based mortality scenarios. Nonetheless, agreement between field and model-derived estimates was poor for larvae originating from the winter-spawned Downs component. In North Sea herring, the dominant processes impacting larval growth and survival appear to have shifted in time and space highlighting how environmental forcing, ecosystem state and other factors can form a Gordian knot of marine fish recruitment processes. We highlight gaps in process knowledge and recommend specific field, laboratory and modelling studies which, in our opinion, are most likely to unravel the dominant processes and advance predictive capacity of the environmental regulation of recruitment in autumn and winter-spawned fishes in temperate areas such as herring in the North Sea.