Gulf of Mexico - Invisible oil beyond the Deepwater Horizon satellite footprint

Major oil spills are catastrophic events that immensely affect the environment and society, yet determining their spatial extent is a highly complex task. During the Deepwater Horizon (DWH) blowout, ~149,000 km2 of the Gulf of Mexico (GoM) was covered by oil slicks and vast areas of the Gulf were closed for fishing. Yet, the satellite footprint does not necessarily capture the entire oil spill extent. Here, we use in situ observations and oil spill transport modeling to examine the full extent of the DWH spill, focusing on toxic-to-biota (i.e., marine organisms) oil concentration ranges. We demonstrate that large areas of the GoM were exposed to invisible and toxic oil that extended beyond the boundaries of the satellite footprint and the fishery closures. With a global increase in petroleum production–related activities, a careful assessment of oil spills’ full extent is necessary to maximize environmental and public safety.


The Deepwater Horizon (DWH) blowout was a mega oil spill, introducing ~795 million liters of live oil into the Gulf of Mexico (GoM) with oil slicks covering an estimated area of 149,000 km2 (1). As a result, vast areas of the GoM were closed for fishing, at one point reflecting a maximum of more than a third of the U.S. national exclusive economic zone (2). The application of the fishery closures was based on satellite and areal imagery by the National Environmental Satellite, Data, and Information Service (NESDIS), using mainly synthetic aperture radar (SAR), and tracking of the NESDIS footprints forward in time, predicting their approximated locations for up to 72 hours (3). The closures were reopened on the basis of a systematic seafood sampling, performing chemical and sensory testing in various seafood specimens (2).

The cumulative satellite oil slick footprint was largely accepted as the DWH oil spill extent from scientific, public, and management perspectives (2, 4, 5). Yet, accumulating field data support a much wider extent of the DWH spill beyond the satellite footprint, reaching the West Florida Shelf (WFS), the Texas Shores (TXS), the Loop Current (LC) system, and the Florida Keys (FK) (see regional map in fig. S1). Specifically, in the WFS, studies indicate high concentrations of oil, including toxic and mutagenic levels, in the water (6), in sediments (7), in sand patties (8), and on the coast (9). Furthermore, high levels of polycyclic aromatic hydrocarbons (PAHs) from DWH oil were found in red snappers’ livers, which co-occurred with a high frequency of skin lesions in bottom-dwelling fish (10). Last, satellite imagery and particle tracking showed that an oil slick present east of Pensacola (north WFS) was transported southeast along the WFS, reaching Tampa and the Dry Tortugas (FK) within a few weeks (11). In TXS, high and toxic levels of oil were found in the water (12) and in sediments (12, 13). In the LC system, the European Space Agency reported the presence of DWH oil during mid-May (14). Later, during early June, NESDIS satellite imagery revealed the presence of 12 oil slicks in the LC system, stretching from FK to the GoM interior. High concentrations of oil were reported in the LC between late June and mid-July west of the fishery closures (15). Furthermore, a deep intrusion (~1000 to 1300 m), which was documented (16) and represented in oil transport simulations (17), included toxic PAH concentrations within 13 km from the well and above-background concentrations extending southwest ~300 km beyond the satellite footprint and closure areas. Overall, these observations indicate that DWH oil extended beyond the satellite footprint and the fishery closures (18).

Oil transport models are effective tools for predicting and reconstructing marine oil spills. Numerous modeling efforts were previously carried out in an attempt to reconstruct spatiotemporal dynamics of the DWH oil spill. The surface oil movement was modeled using Lagrangian Coherent Structure (LCS) core analysis, successfully reconstructing the oil footprint from the satellite imagery (19). Similarly, the three-dimensional (3D) oil application of the Connectivity Modeling System (oil-CMS) (20) and the General NOAA (National Oceanic and Atmospheric Administration) Operating Modeling Environment (GNOME) (21) simulated the oiled coastline areas, both successfully reconstructing the observed beaching patterns (9). The GNOME was also used to conduct probabilistic Monte Carlo simulations, running hundreds of oil transport scenarios based on historical wind and currents data (22). While some of these models provided possible scenarios of the DWH spill, a robust spatiotemporal examination of the spill’s extent, resolving toxic and nontoxic concentrations, does not exist for the DWH. Similarly, the relationship between satellite detection and in situ oil concentrations was rarely examined. This is a key point because satellite detection is the main source of information used for determining the oil spill extent and the corresponding management actions, e.g., fishery closures (2). Having the capacity to quantify the toxic-to-biota extent, which is not necessarily equal to the satellite-detected extent, would improve the understanding and, therefore, the management of oil spills.

Read the full story here.