Mysterious oil spill on the Brazilian coast – Part 2: A probabilistic approach to fill gaps of uncertainties

Over 5000 tons of spilled oil reached the northeast coast of Brazil in 2019. The Laboratory for Computational Methods in Engineering (LAMCE/COPPE/UFRJ) employed time-reverse modeling and identify multiple potential source areas. As time-reverse modeling has many uncertainties, this article carried out a methodology study to mitigate them. A probabilistic modeling using Monte Carlo approach was developed to test these source areas with the Spill, Transport, and Fate Model (STFM) and a scenario tree methodology was used to select possible spill scenarios. To estimate the performance of Lagrangian models, two new model performance evaluations were added to Chang and Hanna (2004). The combination of probabilistic simulations, scenario tree analysis, and model performance evaluation proved to be a powerful tool for mitigating the uncertainties of time-reverse modeling, yielding good results and simple implementation. [Display omitted] •A method to assess the oil modeling uncertainties using Monte Carlo approach•The importance of subsurface oil drift in the 2019 spill has been confirmed.•Development of investigative oil spill modeling based on scenario tree approach