Oil Spill Detection Using Remote Sensing and GIS in Eastern Coast

of United Arab Emirates

Afra Al Teneiji

1,2

, Aaesha Al Mesafri

1,2

and Rami Al-Ruzouq

1,2

Department of Applied Physics and Astronomy, University of Sharjah, Sharjah 27272, U.A.E.

Civil and Environmental Engineering Department, University of Sharjah, Sharjah 27272, U.A.E.

Keywords: Remote Sensing, Geographic Information System, Oil Spill, United Arab Emirates.

Abstract: One of the most dangerous oceanic pollutions nowadays is marine oil spills, which occur when crude oil is

released into the oceanic water and must be contained quickly since they can extend to large areas and result

in serious ecological, economic, and health consequences. Places like the Gulf of Oman are highly vulnerable

to oil spill accidents due to the high marine activity there. Remote sensing and geographic information systems

(GIS) have proven their capabilities in countless fields, and detecting oil spills is one of them. This study

explores the potential of combining remote sensing and Geographic Information Systems (GIS) for oil spill

detection on the Eastern Coast of the United Arab Emirates. Leveraging Sentinel-1 SAR and Sentinel-2 optical

data, we develop and evaluate a methodology to identify oil spills. While specific accuracy assessments await

further testing, initial visual analysis indicates promising results. The study contributes to advancements in

oil spill detection by demonstrating the potential of using these remote sensing techniques in this region.

Additionally, it highlights the value of GIS integration for data analysis and visualization. This research holds

promise for improved oil spill monitoring and environmental protection efforts on the Eastern Coast of the

United Arab Emirates.

1 INTRODUCTION

Marine oil spills are a hot topic that concerns many

governments due to their comprehensive impacts.

This kind of pollution can not only disrupt wildlife

habitats, but it can also harm people's health and

affect the fishing and tourism industries as well. The

oil spill is harmful to ocean and shoreline

environments, especially in places like the Arabian

Gulf and Gulf of Oman that are exposed to such

threats due to the marine activities that take place

there. Desalination plants, fish farms, and tourism

suffer from economic drawbacks whenever an oil

spill incident occurs. To reduce the consequences of

this issue, quick and strategic solutions must be

obtained. Remote sensing and GIS are some of the

most reliable tools that are used in such

circumstances. Remote sensing is a very powerful

tool that can be employed in protecting the marine

environment and monitoring the oil spills that change

the physical and chemical qualities of the sea surface

due to its ability to cover large areas in a short amount

of time as well as functioning day and night during

different weather conditions (synthetic aperture

radars). Oil slicks form fewer rough surfaces than the

surrounding water; therefore, it is less likely for a

radar pulse to bounce back to the sensor, creating dark

spots in the images that represent the oil spill. But

these dark spots can also be caused by natural events

in the ocean, like areas with low wind speeds, weed

beds, or algae blooms. It is common to call these dark

areas "look-alikes."

Using satellite imagery to detect and monitor the

oil spill is not a new topic. Countless studies have

proven the role of remote sensing in this field. For

instance, Dhavalikar and Choudhari (2022) have

applied remote sensing techniques by using synthetic

aperture radar (SAR) images to capture, quantify, and

classify oil spills and lookalikes. They have found

that oil spills caused by moving platforms (ships or

rigs) over the Eastern Arabian Sea are greater than the

oil spill detected near the Bombay High Oil

Platforms. Evaluating the quality of the images

enables the locating of slicks and categorizing them

according to wind speed, known oil infrastructure,

and natural occurrences. Through this, Issa (2005)

was able to map marine oil pollution in the Arabian

Gulf and Gulf of Oman, create an oil spill atlas

Al Teneiji, A., Al Mesafri, A. and Al-Ruzouq, R.

Oil Spill Detection Using Remote Sensing and GIS in Eastern Coast of United Arab Emirates.

DOI: 10.5220/0012561300003696

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 10th International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2024), pages 105-111

ISBN: 978-989-758-694-1; ISSN: 2184-500X

105

offshore the United Arab Emirates, and identify over

six hundred potential spills between 1992 and 2003

using SAR and optical data. Chaturvedi et al. (2019)

demonstrated that Sentinel-1 VV polarization serves

as the most effective tool for oil spill detection.

Furthermore, Gafoor and Al Shehhi (2022) used

Sentinel-1 SAR and Sentinel-2 optical data, which

were pre-processed and analyzed using SNAP and

ArcGIS Pro software. The results of the study showed

that remote sensing and GIS are effective tools for

detecting and monitoring oil spills. The combination

of SAR and optical data was particularly effective,

and the use of band ratios from Sentinel-2 data was

useful in distinguishing oil spills from other

lookalikes. Also, Grimaldi et al. (2010) evaluated

AVHRR TIR channels 4 and 5 data and were able to

detect thin and old oil films with high sensitivity and

dependability. Kolokoussis & Karathanassi (2018)

use it in both known natural outflows and light oil

spill events. Researchers developed and evaluated

two object-based image analysis (OBIA) methods for

detecting oil spills from Sentinel-2 imagery: spectral

matching and texture analysis. Both methods were

able to detect oil spills in the Sentinel-2 imagery, with

the spectral matching method being more effective at

detecting thicker and more concentrated oil spills, and

the texture analysis method being more effective at

detecting thinner and more dispersed oil spills. Fahim

Abdul Gafoor and Maryam R. Al Shehhi (2022)

processed and analyzed Sentinel-1 SAR and Sentinel-

2 optical data using SNAP and ArcGIS Pro software.

The results of the study showed that remote sensing

and GIS are effective tools for detecting and

monitoring oil spills. The combination of SAR and

optical data was particularly effective, and the use of

band ratios from Sentinel-2 data was useful in

distinguishing oil spills from other lookalikes. Also,

Grimaldi and others (2010) evaluated AVHRR TIR

channels 4 and 5 data and were able to detect thin and

old oil films with high sensitivity and dependability.

Researchers Polychronis Kand Vassilia

Karathanassi (2018) utilized in both known natural

outflows and light oil spill events. Researchers

developed and evaluated two object-based image

analysis (OBIA) methods for detecting oil spills from

Sentinel-2 imagery: spectral matching and texture

analysis. Both methods were able to detect oil spills

in the Sentinel-2 imagery, with the spectral matching

method being more effective at detecting thicker and

more concentrated oil spills, and the texture analysis

method being more effective at detecting thinner and

more dispersed oil spills.

The eastern coast of the United Arab Emirates is an

active marine shipping area that overlooks the Gulf of

Oman, which has recorded several oil spills over the

past years; some of them were reported and others

were not. These frequent accidents must be taken into

consideration to reduce or avoid any drawbacks from

the situation. In this paper, we utilized some remote

sensing and GIS technologies that can significantly

contribute if such accidents occur. The main

objectives of this study are summarized in the

following points:

• Detection and identification of the oil spill using

Sentinel-1 SAR images.

• The oil spills were detected using Sentinel-2

band ratios.

• Indicating the size of the oil slicks.

2 STUDY AREA AND

MATERIALS

The study area covers the eastern coast of the UAE,

which has an extent of approximately 1747.6 km2,

with centroid coordinates of 56.4831371 Eo and

25.28068759 No (Figure 1). This area is known for

the marine transportation traffic that occurs there, and

this is due to the Fujairah and Khor Fakkan ports that

are located along the eastern coast. Fujairah Port is

considered the largest port in the Gulf of Oman and

the second largest bunkering hub globally (Port of

Fujairah, n.d.). In addition, Khor Fakkan Port, which

lies beyond the Strait of Hormuz on the UAE's Indian

Ocean shore, is a world leader in container

transshipment. Its advantageous location gives it the

perfect center for transhipping goods to locations in

East Africa, the Red Sea, and the upper Gulf (Ports &

Terminals, Sharjah Ports Authority, 2023). Due to

these conditions, this region is exposed to continuous

oil spill incidents that might occur from ship

accidents or intentional or unintentional oil leakage.

For instance, a large oil spill was reported on March

30, 1994, which resulted from a collision between

tanker Baynunah and tanker Seki and resulted in

16,000 metric tons spilling into the Gulf of Oman.

Then oil washed ashore about thirty kilometers of

shoreline north of the UAE port of Khor Fakkan due

to the wind and currents. The impact affected several

economically and ecologically delicate regions.

Moreover, in 2005, Fujairah Port was prohibited, and

in 2007, the UAE issued violations for pollution

caused by ships.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

106

Figure 1: Study area.

3 DATA AND METHODOLOGY

3.1 Data Used

In this study, two different data sets were considered

to detect the oil spills, which are synthetic aperture

radar (SAR) and optical images from Sentinel-1 and

Sentinel-2. The data from the sites was acquired from

the Copernicus Open Access Hub and the Alaska

Satellite Facility (ASF) for different dates when the

oil spill incidents occurred. The SAR Level 1 Ground

Range data of the sites was acquired from Copernicus

Open Access Hub and Alaska Satellite Facility (ASF)

for different dates when the oil spill incidents

occurred. Detected products were used in this study

with medium resolution in IW (interferometric width)

mode, 10x10 m pixel resolution, 5.5 cm wavelength,

and 251.8 Km ground range coverage. The data was

provided in VV and VH polarization. Furthermore,

the Sentinel-2 data used was level-2A, which is

atmospherically corrected and provided in thirteen

spectral bands and different spatial resolutions (10m,

20m, and 60m), with 60m resolution bands used to

perform the analysis.

The images were collected on different dates,

depending on the date of the incident and their

availability. Table 1 shows the date when the oil

incident took place, according to the local news, and

the dates at which the available data was captured.

These incidents were caused by distinct reasons, such

as pipe leakage, ship collisions, and terrorist attacks.

Table 1: Data availability, dates, and types.

Reported Captured Type

2019-05-12 2019-05-14 Sentinel-2A

2019-10-28 2019-10-31

Sentinel-2A

Sentinel-1

2022-05-12 2022-05-13 Sentinel-2A

2022-08-22

2022-08-12 Sentinel-1

2022-08-16 Sentinel-1

Furthermore, digital elevation models of the study

area were collected from the USGS Earth Explorer,

which were used to apply the terrain corrections for

the SAR data. The DEM images are Shuttle Radar

Topography Mission Void. Filled data has a

resolution of 90 m. The images were enhanced to

complete missing data, thus providing a more

comprehensive dataset of the digital elevation model.

3.2 Sentinel-1 SAR Data Processing

The collected SAR images were pre-processed using

the Synthetic Aperture Radar toolset in ArcGIS Pro

software, which provides correction and processing

tools for the SAR data. The pre-processing stage

started with downloading the orbit file and applying

the orbit correction, which uses an orbit state vector

(OSV) file that is more precise for modifying the

orbital data in the SAR dataset. Then the thermal

noise will be removed to eliminate distortions caused

by noise in SAR imagery, creating a smoother and

more visually appealing image. Radiometric

calibration will be applied, where SAR reflectivity

will be transformed into calibrated normalized

backscatter values using a reference plane for

normalization. Radiometric terrain flattening will be

implemented to adjust SAR data to compensate for

radiometric variations caused by terrain variations.

Then speckle filtering will be applied using the

Despeckle tool to reduce speckle noise in SAR

imagery by eliminating the grainy or salt-and-pepper

appearance caused by coherent illumination. Next,

geometric terrain correction will be used that

transforms the distorted geometry of SAR imagery

into a map-projected coordinate system using a range-

Doppler backgeocoding algorithm. Finally, the SAR

data will be converted from linear to decibels (db).

The dB scale is the ideal choice for displaying SAR

Oil Spill Detection Using Remote Sensing and GIS in Eastern Coast of United Arab Emirates

107

images because of its logarithmic structure, which

allows it to manage big numerical values and wide

dynamic ranges with efficiency. Since SAR data

reduces the range of amplitude or intensity values,

converting it to dB units makes image interpretation

easier and improves the image's visual representation.

After the preprocessing stage is done, the data is

ready to be analysed. A threshold was implemented

on the VV band, in which the decibel values that

represent the oil slicks will be selected, which are

lower than -22 db. The pixels with these values will

be represented by a class, which will be converted

into a polygon, and further analysis will be executed,

such as measuring the surface area and the relative

variation in the backscatter intensity across the oil

slicks.

3.3 Sentinel-2A Data Processing

Next, Sentinel-2A band ratios help detect oil spills in

remote sensing imagery. We may divide two spectral

bands to show their respective intensities. This may

assist in identifying image features. First, create a

model builder to run two ratios. The first one is R:

B3/B2, G: (B11+B12)/B8, B: (B3+B4)/B2, and the

second one is R: b3/b2; G: (b3 + b4)/b2; B: (b6 +

b7)/b5. Separately, for each ratio, we will composite

three bands to get the false color or ratio band. And

for unsupervised classification, group pixels in a

picture by spectral or spatial similarity. It may extract

oil pixels from a picture by finding pixels with

comparable spectral properties to previous oil spills.

After that, convert the raster to a polygon to estimate

the size of the oil spill. The flowchart shows the

fundamental procedures used to identify oil spills

using satellite imagery.

4 RESULTS AND DISCUSSION

The oil spill incidents that were reported in the local

news were captured by the satellites either after a

number of days since the incidents occurred or before

they occurred. Some of the incidents were captured

by Sentinel 1, and others by Sentinel 2. Most of the

Sentinel-2 images contained clouds, which degraded

the quality of the image and affected the analysis.

4.1 SAR Products

The resulting SAR images were able to detect the oil

slicks very clearly in the image due to the low

backscattering intensity of the oil compared with the

features surrounding it, such as the water. Oil slicks

Figure 2: Methodology Framework.

appear as dark patches in radar imagery due to the

suppression of capillary waves, resulting in a flatter

ocean surface and reduced radar reflection (Figures

3–5) (Moreira et al., 2013). Figure 3 shows the oil

spill that resulted from the collision of two ships in

the Gulf of Oman on August 28, 2022. This accident

resulted in intensive pollution of the seawater in the

region. The resulting image shows how the oil slick

extends widely along the coast. Figure 4 shows the oil

leakage that occurred in Khor Fakkan Port; however,

there is another oil spill caused by a ship in the middle

right side of the image. Figure 5 shows the same

location as the previous image, but after four days. It

is noticeable how the extent of the oil spill increased

significantly, which creates doubts that this oil spill

was not caused by a simple leakage only, but there is

a probability that there are multiple agents in this

situation.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

108

Figure 3: Oil spill detection and extraction on 2019-10-31.

Figure 4: Oil Spill detection and extraction on 2022-08-12.

The circle marks Khor Fakkan port location.

Figure 5: Oil Spill detection and extraction on 2022-08-16.

The circle marks Khor Fakkan port location.

After detecting the oil spill in the images, we

applied a binary threshold classification to extract the

oil spill features. This allowed us to estimate the

overall area of each oil slick, in which the largest

extent of oil spill was recorded on August 16, 2022,

with approximately 165 km2 of oil that extends in

front of the eastern coast of the UAE (Table 2).

Determining the extent of the oil spill can aid the

responsible authorities in creating strategies to clean

up the spills. Furthermore, the overall images show

that the oil slicks have unique shapes and textures, in

which they exhibit smooth, uniform, and continuous

texture with regular shapes (Al-Ruzouq et al., 2020).

Table 2: Overall area of each oil slick.

Incidents Area (Km

)

2019-10-31 54.8

2022-08-12 13.1

2022-08-16 164.9

Moreover, if the oil slick feature were extracted solely

in VV polarization, variation in the backscatter values

would be illustrated within the oil spills themselves

(Figure 6). This can pinpoint a very crucial fact: many

factors can affect the backscattering of the oil, such

as the type of oil, thickness, and weathering degree

(Garcia-Pineda et al., 2020).

Figure 6: The image shows the backscatter intensity

variation within the oil slicks.

4.2 Optical Products

According to Figure 7, the band ratio of the oil spill

on Fujairah's eastern coast shows that an oil spill was

there on October 31, 2019, May 14, 2019, and May

13, 2022. The ratio of B3/B2, (B11+B12)/B8, and

(B3+B4)/B2 shows the oil spill more clearly and

Oil Spill Detection Using Remote Sensing and GIS in Eastern Coast of United Arab Emirates

109

darker. These band ratios are more sensitive to oil

spills than others. It is good at reflecting near-infrared

light. Since water does not reflect much in the NIR

range, it is easy to tell the difference between oil and

water. Algae has a high reflection in the blue, green,

and red spectra, while oil has a low reflectivity in

those ranges. Also, it reflects more SWIR light in the

B11 and B12 bands than in the visible wavelengths,

and SWIR bands are less scattering-prone, making

them more vulnerable to oil spills (Diaz, 2023).

Figure 7: False Color Subset of Sentinel-2 Imagery.

While the band ratio in b3/b2; (b3 + b4)/b2; (b6 +

b7)/b5 was pale and less vulnerable to oil spills since

they are based on visible light reflectance at the same

dates, Cloud cover obstructed the determination of

the oil spill presence in the image captured on May

13, 2022. Water and suspended particles scatter

visible light, making oil spills more difficult to see.

The spectral response of the oil spill over the picture

is enhanced by using the ratio (b3 + b4)/b2, while the

ratio (b6 + b7)/b5 is used to display vegetation-related

information. Furthermore, an oil spill may be mapped

using a false-color composite of sentinel-2 (MSI), as

shown in Figure 8 (F. A. Gafoor et al., 2022).

Figure 8: False Color Subset of Sentinel-2 Imagery.

Then classify the oil spill by using the unsupervised

classification technique to identify and extract the

composite of oil pixels, which can then be utilized to

estimate the extent of surface area that has been

contaminated by the oil. The use of object-based

classifications and segmentation algorithms has the

potential to enhance the accuracy of oil slick

detection by mitigating the occurrence of false

negatives. The estimated extent of the oil leak spans

around 1.215676 square kilometres as of May 14,

2019, as presented in Figure 9.

Figure 9: Unsupervised classification technique.

5 CONCLUSIONS

Detecting oil spills on the Eastern Coast of the UAE

is a challenging task due to the intricate water

conditions, limited depth, and frequent occurrence of

cloudy weather. The careful processing and

segmentation of satellite photos are imperative to

enhance the detection of oil spills in each area. The

utilization of both synthetic aperture radar (SAR) and

optical data has proven to be effective in identifying

oil spills, with the application of band ratios derived

from Sentinel-2 data being particularly valuable in

distinguishing oil spills from similar phenomena.

According to the research findings, Sentinel-1 VV

polarization and Sentinel-2 band ratios, namely

B3/B2, (B11+B12)/B8, and (B3+B4)/B2, proved to

be very efficient in detecting oil spills. The use of

remote sensing and Geographic Information Systems

(GIS) in the realm of oil spill identification and

monitoring is gaining significance, given the

escalating occurrence of oil spills. The use of remote

sensing and Geographic Information Systems (GIS)

may provide significant insights to decision-makers

tasked with addressing oil spills and minimizing their

ecological ramifications. As a future plan, we believe

that the use of artificial intelligence and machine

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

110

learning techniques can significantly enhance the

results, generating more precise and accurate

outcomes. Furthermore, this technology can be used

to distinguish the oil spill from other lookalikes.

REFERENCES

Al-Ruzouq, R., Gibril, M. B., Shanableh, A., Kais, A.,

Hamed, O., Al-Mansoori, S., & Khalil, M. A. (2020).

Sensors, features, and machine learning for oil spill

detection and monitoring: A Review. Remote Sensing,

12(20), 3338. https://doi.org/10.3390/rs12203338

Chaturvedi, S. K., Banerjee, S., & Lele, S. (2019). An

assessment of oil spill detection using Sentinel 1 SAR-

C Images. Journal of Ocean Engineering and Science,

5(2), 116–135. https://doi.org/10.1016/j.joes.2019.09.0

Dhavalikar, A. S., & Choudhari, P. C. (2022). Detection

and quantification of daily marine oil pollution using

remote sensing. Water, Air, & Soil Pollution,

233(8). https://doi.org/10.1007/s11270-022-05752-0

Diaz, A. (2023, September 21). Visualizing a marine oil

spill with sentinel-2 MSI imagery. Medium. https://

medium.com/@anapau_diazg/visualizing-a-marine-

oil-spill-with-sentinel-2-msi-imagery-cb0aaa2045b

Gafoor, F. A., & Al Shehhi, M. R. (2022). Oil spill detection

and mapping using sentinel-1 and sentinel-2 in the

Arabian Gulf Coastal Waters. IGARSS 2022 - 2022

IEEE International Geoscience and Remote Sensing

Symposium. https://doi.org/10.1109/igarss46834.20

22.9883723

Garcia-Pineda, O., Staples, G., Jones, C. E., Hu, C., Holt,

B., Kourafalou, V., Graettinger, G., DiPinto, L.,

Ramirez, E., Streett, D., Cho, J., Swayze, G. A., Sun,

S., Garcia, D., & Haces-Garcia, F. (2020).

Classification of oil spill by thicknesses using multiple

remote sensors. Remote Sensing of Environment, 236,

111421. https://doi.org/10.1016/j.rse.2019.111421

Grimaldi, C. S. L., Casciello, D., Coviello, I., Lacava, T.,

Pergola, N., & Tramutoli, V. (2010). Satellite oil spill

detection and monitoring in the Optical Range. 2010

IEEE International Geoscience and Remote Sensing

Symposium. https://doi.org/10.1109/igarss.2010.5651

967

Issa, S. (2005, January). Monitoring oil spills offshore the

United Arab Emirates using multi ... Semantic Scholar.

https://www.semanticscholar.org/paper/Monitoring-Oi

l-Spills-Offshore-the-United-Arab-Data-Issa/0b17b82

d9fe8c41a4d57d6cb62d1bd3ac293c8c9

Kolokoussis, P., & Karathanassi, V. (2018). Oil spill

detection and mapping using Sentinel 2 imagery.

Journal of Marine Science and Engineering, 6(1), 4.

https://doi.org/10.3390/jmse6010004

Moreira, A., Prats-Iraola, P., Younis, M., Krieger, G.,

Hajnsek, I., & Papathanassiou, K. P. (2013, March). A

tutorial on Synthetic Aperture Radar | IEEE Journals

& Magazine IEEE Xplore. https://ieeexplore.ieee.org/

document/6504845

Port of Fujairah. (n.d.). https://fujairahport.ae/

Ports & Terminals - Sharjah Ports Authority. SPA -. (2023,

August 4). https://sharjahports.gov.ae/ports-terminals/

Oil Spill Detection Using Remote Sensing and GIS in Eastern Coast of United Arab Emirates

111