Available online at www.sciencedirect.

com

ScienceDirect
Aquatic Procedia 3 (2015) 151 – 156

International Oil Spill Response Technical Seminar

Observation of Oil Spills through Landsat Thermal Infrared Imagery:

A Case of Deepwater Horizon
Qianguo Xinga,*, Lin Lia, Mingjing Loua, Lei Binga,b,c, Ruxiang ZhaocˈZhenbo Lic
a
Yantai Institute of Coastal Zone Research, Chinese Academy of Sciences, Yantai, Shandong, China
b
University of Chinese Academy of Sciences, Beijing 100049 China
c
Yantai Oil Spill Response Technical Center of Yantai MSA, Yantai 264000 China

Abstract

In the accidental oil spills of Deepwater Horizon at the Gulf of Mexico in 2010, cold plume of initial oil cover with
the apparent surface temperature lower than the surrounding sea surface temperature by 0.6 K, was detected; and
away from the initial leakage location, the apparent temperature of oil film was found higher than the surrounding
sea surface water temperature with a maximum difference of 3.2 K. Both the cold and hot oil patches had relatively
thicker film, but the cold patches were due to the initial low temperature during the crude oil upwelling from deep
water while the hot ones were caused by sun heating. This suggests that thermal infrared imagery has the potential in
locating the leakage place of crude oil spill upwelling especially from deep water and identifying thick oil
aggregations.
© 2014 The Authors. Published by Elsevier B.V.
© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of International Oil Spill Response Technical Seminar.
Peer-review under responsibility of China Offshore Environmental Services Ltd

Keywords: Crude oil spill; deep sea oil rig; sea surface temperature; thermal infrared imagery; Landsat; deepwater horizon; the Gulf of Mexico

*
Corresponding author. Tel.: +86-535-2109125; fax: +86-535-2109000
E-mail address: qgxing@yic.ac.cn

2214-241X © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of China Offshore Environmental Services Ltd
doi:10.1016/j.aqpro.2015.02.205
152 Qianguo Xing et al. / Aquatic Procedia 3 (2015) 151 – 156

1. Introduction

Due to the increasing activities of marine oil resource exploitation, oil spill has become very dangerous to marine
ecosystem, e.g., the Deepwater Horizon oil gush (Klemas, 2010) at the Gulf of Mexico in April-July, 2010, and the
oil leakage of Penglai oil rig platform at the Bohai Sea China since June, 2011. Remote sensing techniques can
provide information on the location of ocean oil spills, the rate and direction of oil patches movement, and input to
prediction model and may help to the management of clean-up and control.

Optical and microwave remote sensing techniques are mostly used to monitor marine oil spill (Jha et al., 2008).
Optical detection techniques include passive and active visible remote sensing, e.g., LANDSAT, MODIS visible
imagery, and laser fluorescence. Passive optical sensors can not work at night when there is no sunlight; active ones,
e.g., laser fluorescence, can be used to monitor oil spill in daytime and night. However, clouds, fogs and haze may
have negative impacts on the passive optical remote sensing as well as the laser fluorescence in the application of oil
spill remote sensing. Microwave can see through clouds, and work at daytime and night; SAR sensors on the basis of
microwave can image oil spills. However, application of SAR is highly dependent on sea surface wind speed: the
low wind speed area may be mistaken as oil spill, and oil spills in the very high wind speed area may be not detected.

Oil film differs from sea water in thermal characteristics and absorption of sun ray energy. Thermal infrared
radiation can see though haze and thin fog, and can work in daytime and night. Thermal infrared remote sensing
technique is a very important tool for oil spill remote sensing. Using AVHRR data, Tseng and Chiu (1994) observed
hot spots of oil spill with the apparent temperature 2-4 oC higher than the surrounding sea surface temperature, SST
in daytime during the 1991 Persian Gulf War, and these hot spots weakened or disappeared at night. Cai et al (2007)
used high resolution thermal infrared imagery of ASTER to detect the possible oil regions in the Yinggehai area of
Hainan Province, China. As the modeling by Shih and Andrews (2008a,b), differential heating of the oil film and
water during daytime can lead to the infrared radiance contrast reversal.

For monitoring oil spill, our work aims to study the thermal dynamics on the basis of Landsat thermal infrared
imagery.

2. Data and Methods

The largest accidental marine oil spill in the history of the petroleum industry happened on April 20, 2010 at the
Gulf of Mexico, Fig.1. The spill stemmed from a sea-floor oil gusher with a water depth 4992 feet, 1521.6 meters,
and the location is more than 100 Km away from the nearest land, BP Exploration and Production Inc., 2009. This
accident was used as a case study. Landat-7 ETM+ and Landsat-5 TM data were acquired at about 10:18 AM, local
time in the morning to map the oil spill. Radiance, and apparent temperature retrievals were based on the work of
Xing et al (2006a, b): converting digital number, DN, dimensionless to at-satellite radiance, LȜ, W m-2 sr-1 ȝm-1 as
the following equation:

LȜ=gainhDNˇoffset , 1

where the value of gain and offset can be found or calculated from parameters provided in the header file of data;
then, we can get the at-satellite brightness temperature as Eq. 2 which is an approximation to Planck’s radiance
function:
K2
TB
ln(k1 / LO  1) , 2
where k1 and K2 are calibration constants.
 Qianguo Xing et al. / Aquatic Procedia 3 (2015) 151 – 156 153

To get the real temperature of target, Weng et al (2004) gave an approximate method as Eq. 3:
TB
T
1  ( OTB / U ) ln H
,3
Where Ȝ is the wavelength for TM and ETM+ thermal infrared bands, which was set as 11.5×10-6 m; U= h u c /ıˈ
1.438×10-2 m K, ı is the Boltzman constant, 1.38×10-23J K-1, h is the Plank’s constant, 6.626×10-34 J s, c is the light
velocity, 2.998×108m s-1, and H is the emissivity of ground object: for seawater, it can be set as 0.985 (Xing et al,
2006a); crude oil has a lower emissivity which usually ranges in 0.94-0.97.

Landsat-5/TM and Lansat-7/ETM+ has similar configurations in thermal band. ETM+ (60 m) has finer spatial
resolution than TM (120 m) in thermal infrared imagery, and more sensitive to thermal radiation. Landsat-7 ETM+
has two modes: Band 6L, B61 provides an expanded dynamic range and lower radiometric resolution, sensitivity,
with less saturation at high Digital Number ,DN values; Band 6H, B62 has higher radiometric resolution, sensitivity
although it has a more restricted dynamic range. Since 2003, a technical malfunction has caused the ETM+ imagery,
including thermal imagery to have a series of horizontal stripes, but which doesn’t influence the radiometric
performance (Chander et al., 2009).

Fig. 1. Location of oil rig platform, red * mark and the oil spills nearby, shown by grey patches. The background image of Landat-7 ETM+ band
1 was acquired on 01-May-2010.

3. Result and Discussion

The imagery shows both cold and hot plumes during the process of oil spill and their drifting in the case of
Deepwater Horizon. On the morning of 1-May-2010, cold plumes of oil spill were found near the location of oil rig
platform, Fig.2, and the apparent temperature of oil cover was about 293.1 K, 0.6 K lower than the nearby SST, see
Table 1. Oil patches with higher apparent temperature than the surrounding apparent SST were also identified, Fig.3
on the imagery acquired in the morning of 25-May-2010, and the maximum temperature of the hot spots was about
296.5 K, 3.2 K higher than the nearby SST. As the sensors of Landsat/TM and Landsat/ETM+ have been in
operation for many years, there may be calibration errors especially in the thermal bands (Barsi et al., 2007), which
may bring errors in the calculated surface temperature. However, we don’t think that the calibration errors have
significant impacts on the contrasts in the retrieved SST imagery.

Oil usually absorbs heat in daytime and becomes warmer than the surrounding sea water, which makes it show as
hot slicks on the thermal images, as Fig. 3; during the night, the oil cover loses heat faster than the surrounding
water, which leads to a cooler region for oil. Even with enough sun-heating, a certain-thickness thin film can lead to
a cooler oil film (Shih and Andrews, 2008a, b), but it is not the case in this work. From the concurrent visible
154 Qianguo Xing et al. / Aquatic Procedia 3 (2015) 151 – 156

imagery of Landsat (Fig.2 and 3), we found that the cold patches and hot patches both had relatively higher
thickness of film, which is far from the threshold of oil film thickness.

(a) (b)
Fig.2. The cold plume of oil spill: the reddish plume in the false-color composite image ,RGB: band741 ,a and the dark broom-like plume in the
thermal infrared band are the oil leak. The imagery was acquired by Landsat 7 ,ETM+ at 16:18GMT on 01-May-2010.

(a) (b)
Fig.3. The hot slicks of oil spill. The green slicks in the false-color composite image ,RGB: Band741 ,a and the white slicks in the thermal
infrared band ,b are the oil patches. The imagery was acquired by Landsat 5, TM at 16:17GMT on 25-May-2010.

The temperature of sea water near the sea bed is usually much lower than the surface. In-situ survey results at the
Mexico Bay during May 8, July 24, 2010 demonstrated that the bottom temperature ranged from 1.2ć to 17.0ć
with water depth of 220 m - 1100 m, while surface temperature ranged from 28ć to 36ć*. Due to the heat
exchange between oil and sea water, crude oil near the sea bed also has relatively lower temperature. When oil
initially rises to the sea surface and the duration of sun heating is not long enough, its temperature may be lower
than the surrounding sea water temperature although the heat exchange always exists between crude oil and sea
water during the rising process. As shown by Fig. 4, with the decrease of thermal emissivity, the bright temperature,
TB will decrease accordingly. So, even though the oil temperature is very close to that of sea water, oil cover has a
lower bright temperature and appears a cooler region.
 Qianguo Xing et al. / Aquatic Procedia 3 (2015) 151 – 156 155

375 0.98
365
355 0.94
345
0.95

T, K
335
325 0.96
315
0.97
305
295
285 290 295 300 305 310 315
TB, K
Fig. 4. Relationship between bright temperature, TB and the temperature of sea surface, sea water and oil with the change of thermal emissivity,
Landsat TM and ETM+ thermal band.

With the drifting of oil patches, the oil film might become thinner while the covering area would be expanded,
and due to the larger surface area of heat-exchanging with sea water, its temperature would be more close to the
water temperature. Oil on the surface sea water absorbs solar radiation and can be heated by the sun at daytime, and
the drifting oil with different size of film thickness has different sun-heating effects (Shih and Andrews, 2008a). Oil
cover with larger film thickness can absorb and store more radiation energy and thus has higher temperature. These
different thermal dynamics of oil spill can be caught with the approach of remote sensing, as the case in this paper.

Table 1. Satellite-derived temperatures of sea water and oil cover

Satellite/sensor, Acquisition date, Bright Temperature ,TB, K
Surface type Band
time Min Max Mean Stdev
B62 292.8 293.7 293.1 0.20
Crude oil
Landsat-7/ETM+, 01-May-2010, B61 292.4 293.9 293.1 0.18
16:18GMT B62 293.4 294.3 293.7 0.18
Sea water
B61 293.4 294.4 293.7 0.26
Crude oil Landsat-5/TM, 25-May-2010, B6 296.4 301.1 299.2 1.01
Sea water 16:17GMT B6 296.0 297.3 296.5 0.20

4. Conclusions

In the case of oil spill at the Gulf of Mexico in 2010, both cold patches and hot spots under different situations
were remotely detected from the images of Lansat-5/TM and Landsat-7/ETM+: the cold oil patches were found near
the deep sea oil rig platform where the oil had relatively lower temperature when it rose to the sea surface initially;
oil patches with larger size of film thickness were also found away from the original leaking site, and they had
higher temperature than surroundings, which was caused by the sun heating. These results suggest that thermal
infrared remote sensing techniques may be used to identify the initial place of deep sea crude oil leakage and thick
oil aggregations.

Acknowledgements

This work was co-supported by the open fund of Key Laboratory of Marine Spill Oil Identification and Damage
Assessment Technology, SOA, No.201208 and Yantai Science and Technology Bureau, No.2013ZH081. Landsat
image data from USGS is also acknowledged.

References

Barsi, JA, Hook, SJ, Schott, JR, Raqueno, NG, and Markham, BL, 2007. Landsat-5 Thematic Mapper Thermal Band Calibration update, IEEE
156 Qianguo Xing et al. / Aquatic Procedia 3 (2015) 151 – 156

Transactions on Geoscience and Remote Sensing, 44, 552í555.

BP Exploration and Production Inc, 2009. Initial Exploration Plan Mississippi Canyon Block 252 ,OCS-G 32306.
Cai, G, Wu, J, Xue, Y, Wan, W and Huang, X ,2007. Oil Spill Detection from Thermal Anomaly using ASTER Data in Yinggehai of Hainan,
China, Geoscience and Remote Sensing Symposium, IGARSS, 23-28.
Chander, G, Markham, BL, and Helder, DL, 2009. Summary of Current Radiometric Calibration Coefficients for Landsat MSS, TM, ETM+, and
EO-1 ALI sensors, Remote Sensing of Environment, 113, 893ˀ903.
Jha, MN, Levy, J, and Gao, Y , 2008. Advances in Remote Sensing of Oil Spill Disaster Management: State-of-the-Art Sensor Technology for
Oil Spill Surveillance, Sensors, 8, 236–255.
Klemas, V , 2010. Tracking Oil Slicks and Predicting their Trajectories using Remote Sensors and Models: Case Studies of the Sea Princess and
Deepwater Horizon Oil Spills, Journal of Coastal Research, 26, 789–797.
Shih, WC and Andrews, AB , 2008a. Infrared Contrast of Crude-Oil-Covered Water Surfaces, Optics Letters, 33, 3019-3021.
Shih, WC and Andrews, AB ,2008b. Modeling of Thickness Dependent Infrared Radiance Contrast of Native and Crude Oil Covered Water
Surfaces, Optics Express, 16, 10535-10542.
Tseng, WY and Chiu, LS, 1994. AVHRR Observations of Persian Gulf Oil Spills, Geoscience and Remote Sensing Symposium, IGARSS, 2,
779 - 782.
Weng, Q, Lu, D, and Schubring, J, 2004. Estimation of Land Surface Temperature–Vegetation Abundance Relationship for Urban Heat Island
Studies, Remote Sensing of Environment, 89, 467-483.
Xing, Q, Chen C, Shi, P, and Yang, J, 2006a. Atmospheric Correction of Landsat Data for the Retrieval of Sea Surface Temperature in Coastal
waters, Acta Oceanologica Sinica, 25, 25-34.
Xing, Q, Chen, C, and Shi, P, 2006b. Method of Integrating Landsat-5 and Landsat-7 Data to Retrieve Sea Surface Temperature in Coastal waters
on the Basis of Local Empirical Algorithm, Ocean Science Journal, 41, 97-104.