Inter-Research > MEPS > Prepress Abstract

MEPS prepress abstract   -  DOI: https://doi.org/10.3354/meps14536

Lessons from the calibration and sensitivity analysis of a fish larval transport model

Léo Barbut*, Sigrid Lehuta, Filip A. M. Volckaert, Geneviève Lacroix

*Corresponding author:

ABSTRACT: Numerous fish populations show strong year-to-year variations in recruitment. The early life stages play a crucial role in determining recruitment and dispersal patterns. A helpful tool to understand recruitment and dispersal are simulations with a Lagrangian transport model, coupling a hydrodynamic model with an individual-based model. Larval transport models require a sound knowledge of the biological processes governing larval dispersal, while they may be highly sensitive to the parameters selected. Various assumptions about larval traits, behaviour and other model parameters can be tested by comparing simulation results with field data to identify the most sensitive parameters and to improve model calibration. This study shows that biological parameterization is more important than inter-annual variability to explain year-to-year variability of larval recruitment of common sole in the North Sea and Eastern English Channel. In contrast, year-to-year variability of connectivity leads to higher variability than changes of the biological parameters. The most influential parameters are pelagic larval duration, spawning period and mortality. Calibration over a 12-year recruitment survey shows that a scenario with a low mortality associated with a long larval duration, and a behaviour involving nycthemeral and tidal migration, reproduces best the observations. The research provides insights in factors influencing fish dispersal and recruitment, suggesting a strategy for enhancing the accuracy of models in upcoming studies. The study supports the improvement of larval dispersal modelling by incorporating an easily applicable sensitivity analysis for both calibration and validation. Incorporating sensitivity analyses enhances larval dispersal models, aiding informed fisheries management and understanding recruitment variability.