The potential risk for fisheries imposed by the proposed petroleum activities in the Lofoten area is currently a subject of intense political as well as scientific debate (Box I). This area is important for the earliest life stages of several fish species spawning along the west and north coasts of Norway, including the commercially and ecologically important Northeast Arctic cod and Norwegian spring-spawning herring stocks. The controversy is to a large degree based on lack of quantitative information ab out how survival and growth of fish larvae varies in space and time. It therefore remains unknown how likely it is that a minor proportion of a fish cohort dying from oil as eggs or larvae may include the majority of the eggs and larvae that would otherwi se have survived to recruitment, thus contributing to population persistence and growth. In the proposed project newly available data on fish eggs and larvae and zooplankton are analysed by a combination of probabilistic drift modelling, statistical analy sis of drift-model output and biological data, and coupled biological-oceanographic modelling. A method used in environmental impact assessment severely underestimates the probabilistic variability in pathways. We will establish new methods that are consi stent with the stochastic nature of the ocean's meso-scale. For this purpose, an existing ensemble simulation will be examined, based on a validation with newly available ocean current data. The proposed work will be an extension of ongoing work (project LEO) and closely linked to a planned Joint Industry Project (project EOP) being developed in dialogue with the oil industry by a consortium of Norwegian marine research institutions. The project is expected to generate new insight into the spatiotemporal dynamics of growth and survival of fish larvae through their first months. This insight will provide a considerably firmer basis for predicting the possible population effects of oil-caused mortality of fish larvae.