How many runs should we run? – Stochastic oil spill modelling for environmental impact assessment over large operational areas

---

Daniel Machado, Murray Burling

Thursday 2 july 2015

12:39 - 12:42h at Europe 1 (level 0)

Themes: (T) Special session, (ST) Oil spill modelling, Poster pitches

Parallel session: Poster pitches: 11J. Special session: Oil Spill Modelling

---

Quantifying the potential exposure from hypothetical oil spill scenarios is critical for preparing environmental impact assessments (EIA), emergency contingency response plans and financial risk analysis for oil and gas operations in offshore waters. Stochastic oil spill trajectory and fate modelling has become the industry standard and best practice for assessing risks of exposure. This intensive approach involves running numerous simulations to track the trajectory and fate of oil at sea due to variable metocean conditions, with the aim of accounting for the natural variability of conditions in a study region. This allows the determination of, for example, the zone of potential impact (ZPI) and probability statistics for risk assessment. Such approach is based on the assumption that the specific location where a spill scenario may occur is clearly defined, which is the case of some operations e.g. a drilling of a well or operations at a specific site. However, several oil and gas related operations involve large operational areas instead of a single location, e.g. seismic surveys, laying of pipelines on the seafloor, shipment of hydrocarbons, etc. Adaptation of the conventional, single location, approach is required to achieve comprehensive assessment of those cases where a spill event could potentially occur anywhere within an operational area. Basic practical questions in this regard arise, such as: how many potential spill locations and simulations are required for assessments over large operational areas? Here we present a conceptual framework to contribute to answering these questions. The framework was tested for a hypothetical case study, with 1,000 stochastic oil spill simulations completed at randomly selected sites within an operational area of approximately 8,000 km2. The simulation outputs were analysed to calculate the area of the ZPI at a series of threshold concentrations of oil floating on the surface and entrained in the water column. For this case study, 80% and 95% of the predicted ZPI area were reached when including 600 and 900 simulations, respectively. A detailed description of the procedure and insights gained during the assessment are presented, with the aim of contributing towards better representation of the risks when considering exploration and development of our ocean resources.