Modeling passive dispersal through a large estuarine system to evaluate marine reserve network connections

The importance of larval dispersal to the performance of marine reserve networks is widely recognized. We characterized patterns of passive dispersal in the eastern basin of the Strait of Juan, de Fuca, Washington, and interpreted the results in the context of marine reserve network connectivity. We used a surface current model to describe the dispersal of passive particles released from 16 sites over periods of 4 and 30 d in the spring of three consecutive years. We then used this approximation to infer the extent to which existing marine reseves and protected areas established on an ad hoc basis are likely to function as a network connected via larval exchange. Dispersal patterns varied substantially between release sites. Release site location accounted for the greatest amount of variation in dispersal distance, exceeding variation due to year, month, or tidal phase. After 30 d, dispersal distance and variance combined to describe three groups of release sites: those characterized by short distance, low variance dispersal; those characterized by long distance, low variance dispersal; and those characterized by intermediate to long distance, high variance dispersal. We suggest that sites within this third group are likely to make the strongest contributions to network connectivity in this system. Our findings underscore the importance of using both dispersal distance and variance estimates to model connectivity between sites and suggest that the performance of ad hoc collections of protected sites can be enhanced through the establishment of additional protected sites chosen to fill critical gaps in existing networks.