Robustness against different implementation of uncertainty in a contingency DSS: the Prestige oil spill revised

The dramatic consequences of oil or chemical spills for coastal ecosystems and economic uses make contingency management a politically sensitive task. We here propose a new Decision Support System (DSS) consisting of a combination of modeling and evaluation methods with which different containment strategies can be compared in the light of a variety of potential ecological and economic damages. In this study, the DSS is tested by hindcasting the Prestige accident off the coast of Spain in 2002. In particular, an oil spill contingency simulation model coupled with multi-criteria analysis techniques was used to assess oil spill impacts for weather and current forecasts available at that time. Five different towing directions resulted in many different hypothetical sinking sites which all were simulated separately. Two procedures of including the array of pollution scenarios into the evaluation were considered: the first method uses expected pollution intensities for each towing route, in the second procedure all hypothetical sinking scenarios are rank–ordered so that towing directions are only indirectly compared. Additional uncertainty resulting from incomplete or imprecise information about pollution impact thresholds or different weighting schemes representing major stakeholder groups is also taken into account. We identified clearly one worst option and one or two almost equally well performing routes. For reasonable range limits of evaluation parameters or sinking probabilities, rankings turn out to be robust against the many uncertainties included so far. Even the two different methods of implementing various sinking scenarios did only differ with respect to confidence limits. Robustness as well as transparency of the coupled approach carry a large potential for enhancing the efficiency of decision making even in politically sensitive situations.