Oil Spill Detection and Forecasting using SAR imagery

Europe is the world's largest market in crude oil imports, representing about one third of the world total. Ninety percent of oil and refined products are transported to and from Europe by sea. Some of this oil make its final way into the sea. Of the oil released by ships, seventy five percent is estimated to have come from operational discharges and only twenty five percent from accidental spills. Furthermore, while past statistical assessments identified tankers as the main marine polluters with crude oil, recent ones switch the emphasis to fuel oil sludge, bilge water and engine room waste, which are produced by all types of ships.

Oil spills can be detected in SAR imagery through supervised algorithms, and the spread and dispersion of spills forecasted by simulation models coupled with oceanographic models.

Example 3: Forecasting of oil spill transport in the Mediterranean Sea

MARSAIS investigates the operational requirements of oil spill transport forecasting. The goal is to explore the potential use of SAR derived information on oil slicks by operational forecasting systems that simulate their dispersion and transformations. An oil spill model is being set up for the whole Eastern Mediterranean Sea where many records of small oil spills (vessel discharges) exist since 1999. This oil spill model is coupled with a hydrodynamic model and an offshore wave model, the provides the oceanographic field that drives the oil spill transport model.

The area covered by the oil spill model is shown below, along with an example of the forecasted transport of a spill originally positioned north of Cyprus (32°E, 35.5°N). These two figures represent the initial position of the oil spill on December 1st, 2001 and its fate after 17 days on the sea. In this scenario, realistic meteorological forcing was provided by POSEIDON-ETA limited area model.

Coverage of the MARSAIS oil spill model

Map of the Oil Spill Model area. © 2003 NCMR

 
Initial location of oil spill

Initial location of oil spill. © 2003 NCMR

Forecast of oil spill transport from the MARSAIS model

Forecasted oil drift after 17 days. © 2003 NCMR

 

More examples:

(1) Oil spill detection in the North Sea - ERS SAR image November 25, 1994 (2) Oil spill detection in the North Sea - X-SAR image April 11, 1994
 

Algorithms and models:

The ERS SAR supervised slick algorithm is used to detect potiential oil spills in ERS SAR images. Other algorithms for detection of oil spills in SAR imagery, include the OSDWS and JRC's interactive algorithm for oil spill detection. The Oil Spill Model is used to forecast transport of oil spills in the Mediterranean Sea.

Operational application:

Slick Fact Sheet (ERS SAR) (5 pages, 287kB) Slick Fact Sheet (X-band SAR) (6 pages, 1.33MB) Oil Spill Model Fact Sheet (8 pages, 500kB)