Long records of geophysical forcing have been used in numerous studies to estimate a statistical distribution of oil spill scenarios. The resulting set of spill scenarios is then used as a basis for planning a robust response capability that should be able to handle all likely real spills. For model developers to be able to support these expectations there are a number of criteria that must be satisfied: (1) Models must develop and retain the data necessary to answer key response questions; (2) developers must understand the limitations in resolution imposed by the specific algorithms they use; and (3) the cardinality of the long geophysical records (with respect to modeled spill behavior) should be determined and the final collection of spill scenarios must span this set. This paper considers these specific constraints and discusses methods that can be used to quantify some aspects of the uncertainty in the output. Published by Elsevier Science Ltd.