Agent-based modeling and genetics reveal the Limfjorden, Denmark, as a well-connected system for mussel larvae

Fishery of blue mussels Mytilus edulis constitutes a very important economic activity in Denmark, whereas mussel farming on long-lines or nets is a new, growing sector. Spawning from natural mussel beds takes place during early summer, and larvae disperse via water currents before settling on the bottom or on spat collectors in the farms. In the present study, we coupled a 3D physical model system (FlexSem) with an agent-based model in order to examine the connectivity of this marine system in terms of mussel larval dispersal and settling potential. To address this question, we (1) estimated the dispersal and connectivity between 17 areas in the Limfjorden, (2) identified the main donor and receiver areas of mussel larvae and (3) identified possible dispersal barriers. The results show that the central narrow strait in the Limfjorden was the main donor area in all the studied years, and that the adjacent eastern areas were the main receiver areas. Towards the inner basins of the Limfjorden, isolation increased and limited connectivity was observed. The results from the cluster analysis grouped the Limfjorden into 3 to 5 clusters, but there was still some exchange of simulated larvae observed among these clusters. Analysis of molecular markers revealed no genetic differentiation between areas and supports the model results, indicating that despite distinguishable hydrographic boundaries, the mussel populations in the Limfjorden are well connected. This study demonstrates how connectivity modeling can be used to support site selection processes in aquaculture.

Automated summary

The study analyzed the connectivity of mussel larvae in Denmark and found that the central strait in Limfjorden was the main donor area, while the eastern areas were the main receiver areas. There was increased isolation towards the inner basins of Limfjorden, but some exchange of simulated larvae was still observed among different clusters.