Sensitivity to life-history parameters in larval fish drift modelling predictions for contrasting climatic conditions

Abstract

Listen

Translate article

Many marine fish species display a long larval drift between offshore spawning grounds and coastal nursery areas. This drift, whose efficiency drive the renewal of the population, critically depends on many environmental factors and on the spawning behaviour of species. Given the complexity to collect field data on fishes early life stages, modelling has proved to be one of the most valuable tool to explore such questions. However, these complex models require many parameters, many of which are uncertain, making a comprehensive sensitivity analysis a crucial step before drawing conclusions on the effects of environmental drivers. This study focused on the larval drift of five species archetypes mimicking five commercially important fish species with contrasted spawning strategies and larval ecology, in five different sub regions over Western Europe, from the Iberian coast to the North Sea. Using an hydrodynamic model (MARS3D) coupled with an individual-based model, and by building a simulation design suitable for such sensitivity analysis, we assessed the relative influence of extrinsic (years and region’s hydrodynamics) and intrinsic (spawning grounds’ depth, spawning period, vertical migration, developmental rate, daily mortality and tolerance to temperature) factors and their two-ways interactions on recruitment success to nurseries. The thermal tolerance of species was one of the leading factors, suggesting that this is a critical parameter, especially when modelling species recruitment success at the edge of their distribution range. On the other hand, other parameters such as the nycthemeral vertical migration, that are often discussed in such larval drift simulation exercise, have proved to have a more limited influence for an analysis at such a large spatial scale. Overall, when studying larval drift on a large scale, local hydrodynamics and inter-annual variations hold most of the explained variance in the recruitment success, confirming the importance of repeating simulations over multiple years before exploring the connectivity among spawning and nursery areas.