Development of the SOSIM model for inferential tracking of subsurface oil

Mary Jacketti, Chao Ji, James D. Englehardt, C. J. Beegle-Krause

Research output: Contribution to conferencePaperpeer-review

7 Scopus citations


Detection and modeling of subsurface oil is currently difficult due to limited visibility, limited availability of wide-range remote sensing technology, and the effects of changes in temperature (oil density), salinity, weathering, and wave-induced sediment entrainment. This paper summarizes the work done and currently in progress for the development of the SOSim (Subsurface Oil Simulator) model, which uses Bayesian inference and limited available field reconnaissance data on oil concentrations to assess the location of the subsurface oil and project its movement in time. The original multimodal Gaussian SOSim model was designed for assessment only of sunken oil on bay bottoms and continental shelves following instantaneous spills. The focus of the current 2018-2019 project is to expand capability to allow tracking of water column and sunken oil, and oil released continuously over a period of time such as from well blowouts, based on available 2-D and 3-D field data, output from other models, and bathymetric data. In particular, for submerged oil, the spatiotemporal concentration information from other oil trajectory models, such as SINTEF's Oil Spill Contingency and Response (OSCAR) model, can now be accepted as SOSim input to create a prior likelihood function, and temporally-varying advection and dispersion are inferred from such prior data and available field data. Alternatively, for sunken oil, bathymetry can be incorporated as prior information, with the bounded and scaled bathymetric profile treated as an inverse prior distribution, and 2-D river modeling capability is being developed. In addition, the new model implements tracking of continuous oil releases by numerical temporal integration, as well as 95% confidence bounds on the 15% oil concentration contour by comparison of the difference in its -2 log-likelihood function from that of the maximum likelihood contour with a chi-squared random variate. These updates will allow the model to better integrate available field data with other input information to predict the location of submerged oil following a well blowout, vessel release, or pipeline break, to complement existing emergency response tools and better assist recovery efforts.

Original languageEnglish (US)
Number of pages17
StatePublished - Jan 1 2019
Event42nd Arctic and Marine Oilspill Program - Technical Seminar on Environmental Contamination and Response, AMOP 2019 - Halifax, Canada
Duration: Jun 4 2019Jun 6 2019


Conference42nd Arctic and Marine Oilspill Program - Technical Seminar on Environmental Contamination and Response, AMOP 2019

ASJC Scopus subject areas

  • Water Science and Technology
  • Waste Management and Disposal
  • Geochemistry and Petrology
  • Pollution
  • Nature and Landscape Conservation


Dive into the research topics of 'Development of the SOSIM model for inferential tracking of subsurface oil'. Together they form a unique fingerprint.

Cite this