Scientific Methodology

The oil spill hazard mapping method applied in GLAMOR is based on the work developed by Sepp Neves et al., (2015) and later improved by Sepp Neves et al., (2016). The methodology, named IT-OSRA, is aligned with modern and standardized concepts of risk (ISO 31000 – Risk management principles) critically adapted to the oil spill case. The user can find the theoretical IT-OSRA basis in Sepp Neves et al., (2015).

The oil spill hazard estimation in IT-OSRA relies on outputs generated by ensemble oil spill modelling experiments. As many as necessary spill scenarios are simulated with different release sites, meteo-oceanographic conditions and oil spill model configurations addressing uncertainties about where, when and how (i.e. spilled volume, type of oil, etc.) a spill might happen. The magnitude and frequency of the beaching events in the simulated spills are evaluated at pre-defined coastal areas (e.g. municipality, country) and used to compute the oil spill hazard.

Some important simplifications were applied to IT-OSRA for the GLAMOR-AtlantOS experiment in order to reduce the computational time involved. To address uncertainties regarding “where” the spill might happen, about 10,000 potential release points were defined covering areas between 30 and 130km away from the coastline. A fixed distance of approximately 25km between release points was defined based on the Lagrangian spatial-decorrelation for surface currents in the Atlantic. To address uncertainties regarding “when” the spill will happen or, in other words, the meteo-oceanographic conditions during the spill event,  simulations were repeated for every release point using sampled current and wind fields for the year of 2013. Finally, uncertainties regarding the spill characteristics (i.e. spill volume, oil type, spill duration) were not addressed in the AtlantOS experiment. More on the reasoning behind the ensemble members’ configuration is presented in Sepp Neves et al., (2016), where the user will also find the results obtained with a small-scale case study carried out off the Portuguese coast.



Atlantic release points
First attempt to estimate the oIl spill hazard for the Algarve area, Southern Portugal, and its counties (assigned with their respective names).
Hazard values range from 0 (very low/absent hazard) to 1 (very high). Mapping relative to the year of 2013. Extracted from Sepp Neves et al., 2016
Histogram of oil concentrations found at the coasts generated by the Algarve ensemble experiment. The oil spill hazard and its uncertainties are estimated in GLAMOR based on the concentration distribution

The AtlantOS experiment setup

Here the details for the model setup for the oil spill hazard bulletin

  • Release Point grid: 1/4deg resolution
  • Spatial coverage of RP grid: coastal band (30 – 130 km offshore) covering ice-free areas in the Atlantic
  • Number of parcels released in the simulations: 25,000 representing 10,000 tons (accidental scenarios)
  • Oil API: 38
  • Current fields: CMEMS Global ocean daily analyses (1/12deg spatial resolution)
  • Wind fields: ECMWF ERA-Interim Reanalysis 6-hourly wind fields (80km spatial resolution)
  • Wave fields: Stokes drift disregarded
  • Temporal coverage: only 2013 for now
  • Number of simulations to be performed: ~600,000

Further information can also be found in our online tutorial.