ABSTRACT Estimating damages is both a spatial and temporal problem; spatial distribution and intensity of pollution depends on when a spill occurs, which way the wind is blowing, currents, and water temperature among other factors [4]. This results in a range of outcomes ranging from minor consequences, when winds blow persistently offshore, through to a worst case where oil is brought ashore in larger quantities and in many locations. Often modellers resort to identifying a “worst case” which might be the run when oil reaches shore soonest, or when the most oil washes ashore [5], [6]. However, it is well known that damages and costs of clean-up vary spatially by shoreline type and activity [7], [8]. So how can decision makers be confident that the so-called “worst case” selected by these methods is in fact a worst case. And in any case, is the “worst case” an appropriate basis for setting financial assurance amounts. The researchers explicitly addressed these uncertainties in a novel way for oil spill damages assessments, by providing a cumulative probability distribution of outcomes, with each outcome representing the total damages from a particular spill event. An automated method using oil pollution damage models was developed and applied to enable this approach.