2004
borne
near-
solar
ncans
e oils
varies
urface
lected
wave
of sea
ontrast
at any
urface
are is
y and
kinetic
LU The
ide oil
'htness
1d 6 at
an oil
unding
ved in
f solar
id of
irface is
| smooth
5 the sea
enerated
y appear
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
A large archive of historical satellite radar imagery exists for
the Arabian Gulf region stretching back to 1984. However, not
necessarily all of these images are useful for showing oil slicks.
The suitability of radar imagery for slick detection is largely a
function of the local weather and sea-state conditions at the
time of image acquisition. Detection hinges on a significant
difference in sea-surface texture between the area covered by
oil and the surrounding clean water. This occurs when a
uniform wind of sufficient strength blows across the surface,
raising small capillary wavelets in the absence of oil. However,
if the wind blows too strongly I overwhelms the surface-tension
effects of a film of oil, and the oil covered area becomes as
rough as the surrounding seas. Precipitation can also break apart
or obscure thin oil slicks.
The image data were geographically transformed to fit the base
map image using corner point locations provided with the
images. Images brightness and contract were manipulated to
optimise the discrimination of slicks. After the radiometric
range correction and the georefrencing of each image, the
procedure starts with the definition of a target area. For each
ERS SAR image, we obtained handcart wind speed data every 6
hours interval spanning the 24-hour prior to the time of
satellite’s overpass. Wind speed at the time of image acquisition
was between 3 and 5 metres per second. Ideal condition were
considered to be a wind speed of 3-4 metres per second at
acquisition time, with a maximum speed well below 10 metres
per second and average speed well below 5 meters per second
during the prior 24 hours. As a minimum, several hours of fair
conditions are required for new slick to accumulate to a
detectable size. If at most likely have been dispersal. If the wind
speed exceeded 5 meters per second during the previous 24
hours, it will most likely have prevented the formation of a thin
oil slick. The ERS-1/2 images used in this study were observed
during suitable wind condition for slick analysis.
Analytical efforts were focused on overall assessments of
image quality and suitability for slick detection, identification
of possible oil slicks in the area, and comparison of slick
patterns between the different dates of imagery and also
different types of sensors .the interpretation results indicate that
certain coastal areas of the UAE face frequent oil spills.
Satellite images show striking examples, one offshore Fujairah
(centred at the coordinates 25030' N/56025'E). Considerable
spill concentrations were found within successive JERS-1 OPS,
Landsat-7 ETM+ images and ERS-1/2 SAR browse images.
Other ‘eventful’ but less critically affected regions appear to be
the congested waterway approaches, such as the Strait of
Hormuz, and zones with major oíl production and transport
routes offshore Dubai, Sharjah, and Ajman. The images indicate
that the areas in the offshore Fujairah, the coast lineof the Gulf
of Oman, face a higher incidence of oil spills than the other of
areas. Here considerable spill concentrations were found within
the continuously observed JERS-1 OPS, Landsat-7 ETM+
images and ERS-1/2 SAR browse images. The heavily polluted
regions lie off the coast between Fujairah and Khor Fakkan
associated with anchorage of tankers prior to entering the
Arabian Gulf. The pollution in this area is probably due to
deliberate discharges of ballast water containing oil from ships.
Oil discharging from both anchored and moving vessels can be
observed in each image. Immediately after discharging flush
ballast water, the simmering water surface can be seen as bright
silver to gray colour patches on the surrounding water. Based
on the size of the image pixel, most of vessels are super tankers
661
whose size of hull is more than 300 m in length. Although
sequential sampling of the ERS-1/2 SAR browse images can be
considered, some firs conclusions can already be drawn about
this potential threat to the Gulf of Oman environment, such as
the generally higher abundance of spills along the coast of
Fujairah.
Shipping routes with relatively frequent incidences of oil spills
are the offshore area running parallel to the coast of Abu Dhabi,
Dubai, Shajah and Ajman where intensive oil production
activities exist, and the routes through the Strait of Hormuz.
This study is probably the first time that spatial distribution of
oil pollution is mapped over a quite wide sea area in the region
using high spatial resolution satellite images. Attention was
focused on the area of the offshore from Abu Dhabi to Ajman
with analysis using seven ERS-1/2 SAR images including 2
tandem mode 24-hour interval image observed on 24 April, 29,
May and 30 May 1996 (Figure 3). Numerous small oil slicks
caused by oil spilled from ships leaked from oil platforms, and
natural seeps from the seafloor are visible throughout the
images.
A partial problem is the effect of extensive dark, low wind areas
across the north-eastern parts of the 29 and 30 May images.
This weather pattern is not uncommon in the Arabian Gulf
region. Scattered oil spill are marked by dark patches with a
variety of sizes and shapes on the sea surface. Despite these
localized problems, both images were acquired under generally
acceptable conditions for slick detection. It is seen clearly that
there are two polluted regions, one in the west side of the image
and one in the southwest corner of the image. If we put this map
over the layer representing the oil fields and the shipping
routes, a close relationship between spills and both oil fields
and the shipping routes can be found visually.
Pollutions
Continuous Oil Leaking from Platform ;
{RADARSAT
(13 February 2002)
Figure 3. ERS-1/2 SAR images including 2 tandem mode 24-
hour interval image observed on 24 April, 29, May and 30 May
1996.
Near the oil field north-western quadrant of the image in many
distinct elongated and dog-leg slicks appear with very similar
morphology to natural seepage slicks. Hence very bright spots
can be seen at one end of the slicks as origin points. Observed
slicks and bright spots in the 3 images acquired on different
dates are confirmed as leakage oil slicks from same oil
production platforms. In addition, known well location
correspond to the leaking pints as well. The clear discrimination
of these slicks on these images provides strong evidence that
I
|
p
I
4
1
