摘要：Sunken oil transport processes in rivers differ from those in oceans, and currently available models may not be generally applicable to sunken oil in river settings. The open-source Subsurface Oil Simulator (SOSim) model has been expanded to handle spills of sunken oil in navigable rivers, utilizing Bayesian inference to integrate field concentration data with bathymetric data to predict the location and movement of sunken oil. A novel prior likelihood function incorporates bathymetric input, with sampling grid and default parameters adapted appropriately for rivers. SOSim v2 was demonstrated versus field observations taken following the M/T (Motor Tanker) Athos I oil spill. The model was also modified to operate in 1-D, to assess the longitudinal distribution of sunken oil in a non-navigable river using available poling data collected following the Enbridge Kalamazoo River oil spill in 2010. Results of both case studies were consistent with observed data and local bathymetry in 2-D and 1-D, and the model is suggested as a complement to deterministic models for oil spill emergency response in rivers.
