Résumé
Along the coast of Chile, fisheries targeting the marine gastropod Concholepas concholepas, commonly named "loco", were highly valuable until the end of the 80s when catches declined significantly. Since the late 90s, a management plan based on territorial-user-rights areas has been implemented, with limited effect on stock recovery. More effective loco conservation and management is impeded by lack of information regarding connectivity via larval dispersal between these individually-managed areas. To develop a regional view of loco connectivity, we integrate loco life history information into a biophysical, individual-based larval dispersal model. This model is used to evaluate scales of loco connectivity and seasonality in connectivity patterns, as well as to partition the coast into largely disconnected subpopulations using a recently developed connectivity-matrix clustering algorithm. We find mean dispersal distances ranging from 170 to 220 km depending on release depth of larvae and planktonic larval duration. Settlement success levels depend quantitatively on the physical and biological processes included in the model, but connectivity patterns remain qualitatively similar. Model estimates of settlement success peak for larval release dates in late austral autumn, consistent with field results and with favorable conditions for larval coastal retention due to weak upwelling during austral autumn. Despite the relatively homogeneous Chilean coastline, distinct subpopulations with minimal connectivity between them are readily identifiable. Barriers to connectivity that are robust to changes in model configuration exist at 23 degrees S and 29 degrees S latitudes. These zones are all associated with important headlands and embayments of the Chilean coast. (C) 2014 Published by Elsevier Ltd.