Résumé: The investigation of larval dispersal and retention, their variability and dependence on wind conditions, has become a major topic in fisheries research owing to potential effects on stock recruitment and stock structuring. The present study quantifies the wind-induced variability of larval retention of herring in a highly productive coastal lagoon of the Western Baltic Sea. This lagoon, the Greifswalder Bodden, represents the main spawning area of Western Baltic Spring-Spawning Herring, a stock that has recently undergone a continuous decline in recruitment. The study tests whether this decline was related to changes in larval retention, more precisely to changes in wind conditions, the main forcing of the lagoon's circulation. To answer this, a model approach was applied. Larvae were tracked as Lagrangian drifters under constant and variable wind conditions, examining the main drift patterns and reconstructing the incidents during the period of recruitment decline. For the latter, weekly cohorts of virtual larvae were released in the lagoon over the entire spawning period (April–June; \textgreater16 weeks). The fraction of retained larvae per cohort was related to observed larval abundances. On this basis, a new retention index was defined to evaluate the annual larval retention. The results presented cannot explain the observed recruitment decline but characterize the lagoon as an important larval retention area by virtue of unsteady wind conditions that prevent a steady outflow of larvae.

Résumé: AimAssessing the spatial structure and dynamics of marine populations is still a major challenge in ecology. The need to manage marine resources from ecosystem and large-scale perspectives is recognized, but our partial understanding of oceanic connectivity limits the implementation of globally pertinent conservation planning. Based on a biophysical model for the entire Mediterranean Sea, this study takes an ecosystem approach to connectivity and provides a systematic characterization of broad-scale larval dispersal patterns. It builds on our knowledge of population dynamics and discusses the ecological and management implications. LocationThe semi-enclosed Mediterranean Sea and its marine ecosystems are used as a case study to investigate broad-scale connectivity patterns and to relate them to oceanography and population dynamics. MethodsA flow network is constructed by evenly subdividing the basin into sub-regions which are interconnected through the transport of larvae by ocean currents. It allows for the computation of various connectivity metrics required to evaluate larval retention and exchange. ResultsOur basin-scale model predicts that retention processes are weak in the open ocean while they are significant in the coastal ocean and are favoured along certain coastlines due to specific oceanographic features. Moreover, we show that wind-driven divergent (convergent, respectively) oceanic regions are systematically characterized by larval sources (sinks, respectively). Finally, although these connectivity metrics have often been studied separately in the literature, we demonstrate they are interrelated under particular conditions. Their integrated analysis facilitates the appraisal of population dynamics, informing both genetic and demographic connectivities. Main conclusionsThis modelling framework helps ecologists and geneticists to formulate improved hypotheses of population structures and gene flow patterns and to design their sampling strategy accordingly. It is also useful in the implementation and assessment of future protection strategies, such as coastal and offshore marine reserves, by accounting for large-scale dispersal patterns, a missing component of current ecosystem management.