Use of a probabilistic particle tracking model to simulate the Prestige oil spill

A probabilistic particle tracking model is used to simulate the oil dispersion after the Prestige wreckage. This oil spill constitutes a suitable benchmark to analyze the capabilities of a probabilistic model, since the time elapsed from wreckage to oil landing (12 days) is much longer than the reliability time associated with forecast winds, usually on the order of 3–4 days. The particle model can be run in two different modes: real time mode (when existing reliable wind fields for the event under scope) and in probabilistic mode (in absence of reliable wind fields but with historical fields corresponding to a similar period). The validity of the particle model is first evaluated in a hindcast way, running the Prestige case with the wind fields corresponding to the period November 19 to November 30, 2002, which were not available at the moment of the wreckage. Calculations show the accuracy of the model to provide the right impact point and timing. The probabilistic model is then used to simulate the same event by means of historical data. The region where the oil landed is shown to be the area with the highest probability to be impacted.     

Operational Oceanography System applied to the Prestige oil-spillage event

This contribution describes the procedure used during the Prestige oil-spillage event, by means of an Operational Oceanography System, and the behaviour of the present prediction tools (hydrodynamic and dispersion models) applied to it. The accuracy of these tools is estimated by a reanalysis of field data transmitted by a sea surface drifting buoy, released at the time of the oil spill. The numerical models applied were the Regional Ocean Modeling System (ROMS), fed by the available six-hourly NCEP atmospheric information, together with a Lagrangian Particle-Tracking Model (LPTM). ROMS has been used to estimate the current fields for the Bay of Biscay, whilst the LPTM has provided the oil spill trajectories. The results demonstrate that the accuracy of the numerical models depends upon the quality of the meteorological input data. In this case, the current fields at the sea surface, derived by ROMS, have been underestimated by the wind fields of the NCEP reanalysis data. An efficient calibration of these wind fields, with data provided by the Gascony buoy (fixed oceanic and atmospheric station), achieves more realistic looking results; this is reflected in the comparison between the buoy trajectory predicted numerically and the tracked movements of the drifting buoy.