Deepwater Horizon - Estimating surface oil volume distribution in real time
Abstract
Spill responders to the Deepwater Horizon (DWH) oil spill required both the relative spatial distribution and total oil volume of the surface oil. The former was needed on a daily basis to plan and direct local surface recovery and treatment operations. The latter was needed less frequently to provide information for strategic response planning. Unfortunately, the standard spill observation methods were inadequate for an oil spill this size, and new, experimental, methods, were not ready to meet the operational demands of near real-time results. Traditional surface oil estimation tools for large spills include satellite-based sensors to define the spatial extent (but not thickness) of the oil, complemented with trained observers in small aircraft, sometimes supplemented by active or passive remote sensing equipment, to determine surface percent coverage of the 'thick' part of the slick, where the vast majority of the surface oil exists. These tools were also applied to DWH in the early days of the spill but the shear size of the spill prevented synoptic information of the surface slick through the use small aircraft. Also, satellite images of the spill, while large in number, varied considerably in image quality, requiring skilled interpretation of them to identify oil and eliminate false positives. Qualified staff to perform this task were soon in short supply. However, large spills are often events that overcome organizational inertia to the use of new technology. Two prime examples in DWH were the application of hyper-spectral scans from a high-altitude aircraft and more traditional fixed-wing aircraft using multi-spectral scans processed by use of a neural network to determine, respectively, absolute or relative oil thickness. But, with new technology, come new challenges. The hyper-spectral instrument required special viewing conditions that were not present on a daily basis and analysis infrastructure to process the data that was not available at the command post. Very few days provided sufficient observation quality and spatial coverage. Future application of this method will require solving both the observational and analysis challenges demonstrated at DWH. Similarly, the multi-spectral scanner results could only be interpreted by a handful of individuals, causing some logistical problems incorporating the observational results with the incident command decisions. This roadblock may go away as the spill response community becomes more familiar with the technology.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2011
- Bibcode:
- 2011AGUFMIN14A..03L
- Keywords:
-
- 4262 OCEANOGRAPHY: GENERAL / Ocean observing systems;
- 4337 NATURAL HAZARDS / Remote sensing and disasters