DEMOSSTitle: Development of Marine Oil Spills/slicks Satellite monitoring System elements for the Black Sea, Caspian Sea and /Kara/Barents Seas

INTAS Thematic Call on Earth Sciences and Environment in co-operation with with ESA, 2006

byStein Sandven1, Vladimir Kudriavtsev2 and Vladimir Malinovsky3

1NERSC, Bergen, Norway2NIERSC, St. Petersburg, Russia

3MHI, Sevastopol, UkraineWith contribution form the other DEMOSSS partners

Partners1. Nansen Environmental and Remote Sensing Center (NERSC),

Norway2. BOOST Technologies, Brest, France3. University of Hamburg, Hamburg, Germany4. Nansen International Environmental and Remote Sensing Center

(NIERSC), St.Petersburg, Russia5. Institute of Applied Physics Russian Academy of Sciences (IAP),

Nizhny Novgorod, Russia6. Marine Hydrophysical Institute of the Ukrainian National Academy

of Sciences, Sevastopol, Ukraine7. Arctic and Antarctic Research Institute (AARI), St.Petersburg,

Russia8. Research Center for Earth Operative Monitoring (NTs OMZ),

Moscow, Russia

Project Objectives

• To develop and demonstrate components of a marine oil spill detection and prediction system based on satellite SAR and other space data in combination with models for oil slick/spill monitoring and prediction

Overview of Tasks

Satellite monitoring of selected areas:* Barents/Kara Sea, NIERSC* Black Sea, MHI* Caspian Sea, NTsOMZand validation of oil slick detection, BOOST

Task 4

Radar Imaging Model Development, NIERSCTask 2Algorithm for detection & quantification of oil spills and look-alikes, NIERSC

Task 3

Task 5

Task 1

Oil spill modelling and drift forecasting, AARI

Field experiment with oil slicks in the Black Sea, IAP/MHI

SAR acquisition of the study areas1 2

34

The Barents Sea (1) and Kara Sea (2) are relatively clean areas with little ship traffic and offshore exploration has just started. The areas are expected to become much more exposed to oil pollution in the future.

The Black Sea (3) and the Caspian Sea (4) have already significant tanker traffic and offshore exploitation has started from several platforms

WSM: 38APM: 1IMM: 52

91Caspian

WSM: 58APM: 38IMM: 86

182Black

WSM: 276APM: 3IMM: 168

447Kara

WSM: 131APM: 12IMM: 240

383Barents

DetailsTotal*Region

* Number of image obtained from ESA rolling archive from May to December 2007. In addition, archived data from earlier years are available for the studies.

Radar scattering modelling

DEMOSSS develops an improved model of radar scattering from a sea surface covered by oil and biogenic films to be used in detection and classification of surface film in SAR images

Flow diagram of the radar scatter model for simulation of a given surface condition

10−1

100

101

102

103

104

0

10

20

30

40

k, rad/m

Con

tras

t, dB

10−1

100

101

102

103

104

0

10

20

30

40

k, rad/m

Con

tras

t, dB

10−1

100

101

102

103

104

10−6

10−4

10−2

k, rad/m

B(k

)

10−1

100

101

102

103

104

10−6

10−4

10−2

k, rad/m

B(k

,00 )

u10=6m/sOLEE=0.022

OmniDirectional Up−Wind Direction

Wind Waves Spectrum and Effect of Thin OLE Film

Oleic adic (OLE): monomolecular film

Backscatter from clean and film-covered water in tank experiments (inc. angle 30°)

• Blue dots: observed backscatter from clean water• Blue circles: observed backscatter from oil films (also triangles and crosses)• Blue line: modelled backscatter from clean surface• Green line: modelled backscatter from surface film

(Ref. Gade at el.,JGR 1998, Kudriavtsev et al, JGR, 2005)

Spectral Contrasts for different surface films: Comparison of models with data

Wavenumber k, rad/cm

Data from tank experiments (Ermakov et al.)

OLE

Vegetable oil

Crude oil

Diesel oil

101

102

103

100

101

102

k, rad/m

Spe

ctra

l Con

tras

t

u10=7m/s

OLE E=0.022

VO E=0.012

CO E=0.004

Contrast between wind wave spectrum for clean water and different surface films

Model simulations

Effective Oil Film ViscosityExperimental estimates by Ermakov et al. vs. Jacobs and Jenkins (1997) model

Wave damping coefficients as function of film thickness

Oil thickness in mm Oil thickness in mm

15 Hz waves 25 Hz waves

10−6

10−5

10−4

10−3

10−2

0

5

10

15

20

Film Thickness [m]

VV

&H

H c

ontr

ast i

n dB

10−6

10−5

10−4

10−3

10−2

0

5

10

15

20

Film Thickness [m]

VV

&H

H c

ontr

ast i

n dB

10−6

10−5

10−4

10−3

10−2

0

5

10

15

20

Film Thickness [m]

VV

&H

H c

ontr

ast i

n dB

10−6

10−5

10−4

10−3

10−2

0

5

10

15

20

Film Thickness [m]

VV

&H

H c

ontr

ast i

n dB

C−bandθ=200

u10

=5m/s

X−bandθ=200

u10

=5m/s

C−bandθ=200

u10

=10m/s

X−bandθ=200

u10

=10m/s

Up-wind Radar contrasts vs. oil film thicknessat C- and X-band

Analysis of oil spill signatures in SAR images

NRCS

Incidence angle

direction

Wind speed

0 2 4 6 8 100

1

2

3

4

5

6

7

8

9

10

Wnd Speed, m/s

C−

band

Oil

Slic

ks C

ontr

ast,

dB

400

300

200

Comparison of observed (from SAR archive) and modelled C-band backscatter contrasts in oil spill signatures

Simulated NRCS field (in dB) for an eddy current field in presence of surfactants. Wind speed (a) 5 m/s and (b) 15 m/s. Radar geometry is for ERS SAR.

Modelled backscatter of surfactants in an eddy

Field experiment from an offshore tower in the Black Sea

Optical system to measure short wave spectrum and surface mean slope

Video system to measure wave breaking characteristics

Optical Spectrum Analyzer

X-band radar

Ka-band radar

OLO (11:55)Periodic Natural slicks

250º2 - 4 m/s?70º06.10.07, 10:02

VO (12:22)Natural slick (12:30)

OLO (12:44) DA (13:00)VO (13:39) VO (15:39)VO (15:50) DA (16:09)

OLO (16:23) VO (16:23)

270º2-4 m/s?0º - 5º05.10.07, 10:28

Natural slick (17:17)330º1-3 m/s150º04.10.07, 15:46

Natural slick (12:37)300º0-2 m/s120º04.10.07, 10:34

Natural slick (15:50)Natural slick (16:03)

OLO (16:09)Natural slick (16:14)

280º1-2 m/s110º03.10.07, 15:46

Natural slick (12:00)Natural slick (12:42)Natural slick (12:54)

Natural slick (12:59) + VONatural slick (13:13)

OLO (13:43)

270º0 m/s1-2 m/s2 -3 m/s

No wind107º

(11:43)90º

(12:28)

03.10.07, 10:27

Natural slick (17:12)DF (17:14)

270º2.5m/s?90º (E)02.10.07, 16:47

VO (13:03)VO (14:15)

OLO (15:07)

260º0 m/s<0.5 m/s (14:07)

No wind145º(SE)

02.10.07 12:30

Slick observationsWaveVector dir

Wind speedZ=4m

Wind directionDate, start time

Experiments with slick observations from the towerArtificial slicks: Veg.Oil, Olive oil, Dodecyl Alcohol, Diesel fuel.

Total: 16

Natural slicks periodic and single bands

Total number of slicks: 13

ASAR image

Contrasts in slicks observed on 05 Oct 2007

Dodecyl alcohol slick (film elasticity E=50-70 mN/m)

Vegetable oil slick (film elasticity E=12-15 mN/m)

Wind velocity 2-4 m/s

0 1 10Wavenumber, rad/cm

1

10

100

1000

Con

trast

OSA

X-band

Ka-band

RadarsPhoto

0 1 10Wavenumber, rad/cm

1

10

Con

trast

OSA

SAR observation of experimental oil spills

NRCS profile across a slick observed in ASAR APP data

Results from SAR analysis of AP images form 2003:

The largest slick has a contrast of about 15 dB compared to the surrounding clean water

Subset of ENVISAT ASAR AP image on 23 August 2003 off Novorossisk coast:

(a) VV-pol

(b) HH-pol

(c) Pol ratio (PR)

For clean seas PR is defined by contribution of bragg scattering and wave breaking, with typical value of 5 for this inc angle. In slicks bragg waves disappear and PR becomes close to 1

Another slick observation in SAR APP image

Distribution of oil spills in the Black Sea derived from 68 SAR images

SAR images from ERS-2 and ENVISAT were analyzed for a period of three years (2001 - 2004), resulting in 68 images with 424 likely oil spill events.

The distribution of the spills are concentrated along the main shipping lanes and in the offshore drilling area in the western Black Sea

Example from the Caspian Sea

• The ASAR Wideswath image from 04 July 2007 covers most of the Caspian Sea (left figure). A subset of the image (above) was analyzed for the area off Baku (red circle) where a spill event could be detected. The SARTool provided by BOOST Technologies was used to detect and quantify the oil spill area.

Oil spill event Kerch Strait 11 November 2007

C-band: RADARSAT X-band: TerraSAR L-band: ALOS PALSAR

Images obtained 16 November - > case study for model comparison

Courtesy: Scanex Courtesy: DLR

Comparison with previous SIR-C/X SAR dataand field experiments

Oil drift modellingoil spill locationoil spill volume and spill rateoil propertiesfractional composition of the oil

Oil spill input

currentswindwind wavesthermohaline structurebathymetryice conditions

advectionturbulent diffusionevaporationemulsificationdispersionphoto-oxidationbio-oxidation

Oil slick spatial distributionOil mass balance

spreading of the spillets

advectionevaporation

Oil spill simulation

Oil spill output

Sea state input

AARI is developing an oil drift model, OilMARS, based on the components shown in the diagram. The model has been tested in the Barents and Kara Seas.

Oil spill modelling in Kara Sea: open water• AARI uses its oil

spill model OilMARS to simulate oil drift in the Kara Sea. The figures show the spreading of a spill over a period of 20 days. The red area indicate where oil reached the coast and caused pollution at the beach.

Oil spill modelling in Kara Sea: sea ice waters• AARI uses its oil spill

model Oilmars to simulate oil drift in the Kara Sea. The figures show the spreading of a spill over a period of 20 days in winter when the sea ice ice-covered. Black indicates oil spill in open water, blue indicate oil spill on top of th eice and red is oil spill under the ice.

Summary and further work• Radar scattering modelling tools is ready for use• Field experiments with artificial oil spills at the tower in the Black Sea were

performed in 2007, more experiments are planned in 2008• Lab experiments with radar observation of wave damping by various oil

types have been conducted• Build-up of SAR data base for the study regions have started, primarily

with ASAR data. Will be supplemented by other SAR data (X- and L-band)• Analysis of SAR data for slick and other ocean surface features, including

comparison with models has started• Verify hypothesis that PR can be used to identify oil and natural slicks and

discriminate them from look-alikes• Establish monitoring scheme using satellite data in combination with

models and in situ data for validation

Acknowledgement

• The SAR data for the study is provided ESA through AOBE-2780)

• The research is supported by INTAS (contract no. 06-1000025-9264), EU

FP6 (contract no. 031001- MONRUK), and national projects