The Barents Sea-Norwegian Sea are used for oil production and thus susceptible to oil accident. This calls for providing an improved understanding of long-term impacts of oil accident on this pelagic ecosystem. The Barents Sea-Norwegian Sea ecosystem is c haracterized by strong interactions among fish species; the Arcto-Norwegian cod, the capelin, and the Norwegian spring-spawning herring. These three large and economically valuable fish stocks are particularly important for the ecosystem as a whole as foo d for seabirds and mammals, and the state of these three stocks may therefore in sum represent the state of the pelagic ecosystem as a whole in these seas. To predict possible long-term ecosystem consequences of oil spills, we therefore need to consider t he entire process from the spill itself to the cascading effects within the ecosystem. However, even in a pristine state without human intervention, the Barents Sea-Norwegian Sea ecosystem is a highly dynamics system: the ecosystem is inherently stochasti c. For instance, recruitment (and larval mortality) of the herring is extremely variable from year to year, in part linked to variations in sea climate. Thus, even with perfect knowledge of the system processes and the lethal and non-lethal effects of an oil spill, we can only predict the ecosystem effects of the spill on a probabilistic basis, and then only for a few years ahead. In addition, we have incomplete knowledge about ecosystem processes in this system. Therefore, one cannot expect a large degre e of certainty in the conclusions about long-term effects. We therefore aim to provide information of the range of possible outcomes of oil spills in an ecosystem context. Further we aim to provide a qualitative indication as to whether an oil spill at a given place and time can be expected to have a relatively large or small long-term effect, and under which climatic regimes the effect of spills will be most harmful.