GEOPHYSICAL MONOGRAPH SERIES, VOL. 195, PP. 179-195, 2011
Surface Drift Predictions of the Deepwater Horizon Spill: The Lagrangian Perspective
Forecasting the movement of the oil after a massive spill like that from the Deepwater Horizon drilling platform in the Gulf of Mexico is critical for pollution containment and mitigation. Assessing such forecasts a
posteriori is also relevant for future drilling risk assessments. This study considers surface drift simulations based on
a regional circulation model. Initialization and assessment is performed with two distinct satellite-derived observational
products. Two metrics are introduced for measuring forecast success: percent of the predicted spill area contained in the
observation and percent of the observed spill area contained in the forecast. Simple passive tracer advection for 2–5 days
yields about 30%–75% of the forecast in the observation. The second metric scores around 30% for almost all cases studied.
The unquantified continuing leakage of oil from the ruptured well is modeled at two rates. At the lower rate, it has barely
any effect on the forecast, while the higher rate leads to a doubling or greater of the captured observation (second metric),
accompanied by a significantly smaller deterioration in the first metric. The role of wind is also explored and found to be
negligible away from the coastal areas. Lagrangian coherent structures, in the form of direct Lyapunov exponents and mesohyperbolicity,
are tested for their ability to describe the oil advection characteristics without the sensitivity to observational errors
in the initialization and without the details that are most error-prone. They are found to have some skill in capturing the
deformation patterns exhibited by the observed oil slick.
Citation: Huntley, H. S.,