The confidence level of oil spill detections in satellite Synthetic Aperture Radar (SAR) imagery requires the analysis of
many different factors. Unfortunately, oil slicks are not the only phenomena which can appear as a dark feature in a SAR
image. These include a number of parameters like wind speed, currents, internal waves, upwelling sea areas, algae
bloom, mixing water areas, et cetera. These phenomena are called look-alikes. The largest challenge in detecting oil
spills in SAR images remains in the accurate discrimination between oil spills and look-alikes.
This study introduces the vantages of using geospatial analysis of various metocean data (e.g. wind speed and direction,
sea surface temperature, wave direction, ocean colour data) and environmental ancillary data (e.g. vessel traffic, port
locations) as a supplementary information source for the oil spill probability assessment in SAR imagery. The analysed
data exists in different formats with different value scales. In addition, the parameters of the metocean data analysis are
not equally important for a reliability of oil spill detection. The weight of metocean parameters depends on the impact of
natural phenomena on SAR systems (e.g. wind and currents have pro rata more influence on the probability than sea
surface temperature and chlorophyll-a) and the area of interest (e.g. chlorophyll-a is a more important value for the
Baltic Sea than for the Mediterranean Sea).
The derived oil spill probability categorisation based on the weighted analysis of metocean environmental ancillary data
could be a useful tool for authorities for an efficient planning of cost-intensive verification flights.