Fish recruitment is the result of the integration of small-scale processes affecting larval survival over a season and large oceanic areas. A hydrodynamic model was used to explore and model these physical-biological interaction mechanisms and then to perform the integration from individual to population scales in order to provide recruitment predictions for fisheries management. This method was applied to the case of anchovy (Engraulis encrasicolus) in the Bay of Biscay (NE Atlantic). The main data available to investigate survival mechanisms were past growth (otolith) records of larvae and juveniles sampled at sea. The drift history of these individuals was reconstructed by a backtracking procedure using hydrodynamic simulations. The relationships between (real) growth variation and variations in physical parameters (estimated by hydrodynamic simulations) were explored along the individual trajectories obtained. These relationships were then used to build and adjust individual-based growth and survival models. Thousands of virtual buoys were released in the hydrodynamic model in order to reproduce the space-time spawning dynamics. Along the buoy trajectories (representative of sub-cohorts), the biophysical model was run to simulate growth and survival as a function of the environment encountered. The survival rate after 3 months of drift was estimated for each sub-cohort. The sum of all these survival rates over the season constituted an annual recruitment index. This index was validated over a series of recruitment estimations. The modelling choices, model results and the potential use of the recruitment index for fisheries management are discussed.