Rdc Submerged Oil Detection Report June 2009

Report No. CG-D-08-09

Heavy Oil Detection
(Prototypes) Final Report
Distribution:
Approved for public release; distribution is unlimited. This document is available
to the U.S. public through the National Technical Information Service, Springfield,
VA 22161.

June 2009

UNCLASSIFIED | CG-926 RDC | Hansen, et al. | CG-5332 | June 2009
1

Heavy Oil Detection (Prototypes) – Final Report

N O T I C E
This document is disseminated under the sponsorship of the Department of
Homeland Security in the interest of information exchange. The United States
Government assumes no liability for its contents or use thereof.

The United States Government does not endorse products or manufacturers.
Trade or manufacturers’ names appear herein solely because they are
considered essential to the object of this report.

This report does not constitute a standard, specification, or regulation.

Timothy R. Girton
Technical Director
United States Coast Guard
Research & Development Center
1 Chelsea Street
New London, CT 06320-5506

ii


Technical Report Documentation Page
1. Report No. 2. Government Accession Number 3. Recipient’s Catalog No.

CG-D-08-09
4. Title and Subtitle 5. Report Date

Heavy Oil Detection (Prototypes) - Final Report June 2009
6. Performing Organization Code

Project No. 4153
7. Author(s) 8. Performing Report No.

Kurt A. Hansen; Michele Fitzpatrick; Penny R. Herring; Mark VanHaverbeke R&DC UDI # 827
9. Performing Organization Name and Address U.S. Coast Guard 10. Work Unit No. (TRAIS)

Potomac Management Group, Inc. Research and Development Center
a Division of ATSC 1 Chelsea Street 11. Contract or Grant No.

7925 Jones Branch Drive New London, CT 06320-5506 HSCG32-04-D-R00005
McLean, VA 22102 HSCG32-09-J-000085
12. Sponsoring Organization Name and Address 13. Type of Report & Period Covered

U.S. Department of Homeland Security Interim
United States Coast Guard
14. Sponsoring Agency Code
Commandant (CG-5332)
Commandant (CG-5332)
Washington, DC 20593-0001 U.S. Coast Guard Headquarters
Washington, DC 20593-0001
15. Supplementary Notes

The R&D Center’s technical point of contact is Mr. Kurt A. Hansen, 860-271-2865, email: Kurt.A.Hansen@uscg.mil.
16. Abstract (MAXIMUM 200 WORDS)

Current methods for locating and recovering submerged oil spills are inadequate. Detection methods are
often improvised on-scene, and recovery techniques are labor intensive and not always successful. The
U.S. Coast Guard Research and Development Center has embarked on a multi-year project to develop a
complete approach for dealing with spills of submerged oils. This report describes the assessment of
detection techniques using sonar, laser fluorometry, real-time mass spectrometry, and in-situ fluorometry
to locate oil sitting on the sea floor. Evaluation of four proof-of-concept devices was conducted at the Oil
and Hazardous Material Simulated Environmental Test Tank (OHMSETT), now called The National Oil
Spill Response Test Facility, in Leonardo, NJ, between November 2007 and February 2008. Further
testing of two of these prototype devices, plus three additional detection systems, was conducted at
OHMSETT in January 2009. This report contains the results of these tests and recommendations for
Federal On-scene Coordinators when responding to spills of heavy oil (contained in Appendix E).

17. Key Words 18. Distribution Statement

Heavy oil detection, multi-beam sonar, fluorescence polarization Approved for public release; distribution is
unlimited. This document is available to the U.S.
public through the National Technical
Information Service, Springfield, VA 22161.
19. Security Class (This Report) 20. Security Class (This Page) 21. No of Pages 22. Price

UNCLASSIFIED UNCLASSIFIED 74

iii


(This page intentionally left blank.)

iv


EXECUTIVE SUMMARY

Even though heavy (sinking) oils have historically accounted for a small percentage of spills, environmental
and economic consequences resulting from such a spill can be high. Heavy oils can sink and affect shellfish
and other marine life in addition to causing closure of water intakes at water treatment facilities and power
plants. Regardless of whether the heavy oil is near the surface, neutrally buoyant in the water column, or on
the bottom, its recovery is difficult. The underwater environment poses major problems including poor
visibility, difficulty in tracking oil spill movement, and colder temperatures, complicating containment and
recovery. Developing effective methods and technologies suitable for this environment is a major
challenge.

Current methods are inadequate to find and recover spills of submerged oil. Many of the detection
approaches are ad-hoc and the recovery techniques very labor intensive. The U.S. Coast Guard (USCG)
Research and Development Center (RDC) has embarked on a multi-year project to develop a complete
approach for spills of submerged oils, including detecting and mapping the spilled oil (Stage I) and
containing and recovering it (Stage II). Each Stage is itself a multi-phase effort. This report discusses the
process and results for Stage I.

Phase I – Detection and Mapping Proof-of-concept
A Broad Agency Announcement (BAA) was used to identify several potential submerged oil detection and
mapping technologies in the proof-of-concept phase of development. Four companies were chosen to
develop proof-of-concept instruments, which were then tested for the ability to locate and identify test
patches of three types of heavy oil in sediment trays deployed in the Oil and Hazardous Material Simulated
Environmental Test Tank (OHMSETT), now called The National Oil Spill Response Test Facility. The
technologies included sonar, laser fluorometry and real-time mass spectroscopy. Based on test results, two
were selected for prototype development using technical requirements and the risks associated with further
development. The companies selected were EIC of Norwood, MA (fluorescence polarization) and RESON,
Inc. of Goleta, CA (multi-beam sonar combined with data-processing software).

Phase II – Prototype Test Results
The configuration of the prototype tests, conducted at OHMSETT, included four types of heavy oil, four
types of sea bottom, and intermittently placed rocks and seaweed. The two companies selected in Phase I,
EIC and RESON, returned to test their equipment. Three additional vendors, Biosonics, CodaOctupus, and
SRI International, tested their detection equipment at OHMSETT as well (at no cost to the USCG) on the
same test configuration.

EIC returned to OHMSETT with a compact unit but bright sunlight during these tests saturated the
instrument with strong backscatter. On-scene modifications allowed EIC to continue the tests and achieve
usable results. EIC later developed a successful method to reduce the external light by modulating the laser
and looking for the returned fluorescence that was also modulated. Due to the fluctuations of the GPS input,
a direct mapping of the results was not possible.

The RESON prototype included a detection algorithm that uses backscattering strength to estimate oil patch
location and size. While not done in real-time, the data transfer and calculations were completed for the
entire test section in less than one day for the 400 kHz runs. While it was relatively easy to discriminate oil

v


from the bottom, the probability of detection can be increased as more information is known about specific
oils and their properties and entered into the model. The system detected 87 percent of the target areas but
had 24 percent false alarms.

BioSonics, Inc. (Seattle, WA) tested a unit equipped with two single-beam sonar transducers (200 kHz and
420 kHz) that are usually used to classify substrate (sub-bottom) or submerged vegetation. The system was
successful in classifying the oil as a different kind of material in real-time. It was also able to differentiate
the four types of bottom material that were used.

The CodaOctopus (New York, NY) EchoScope4D Imaging sonar operating at 375 kHz was tested. This is
the same system that the USCG is evaluating for other uses. Like the RESON system, it uses return signal
to differentiate between rocks, bottom, and oil. At almost all angles and frequencies, the contrast between
oil and sand is about 15 dB. It also records and can display the bathymetry.

SRI International (Menlo Park, CA) was externally funded to evaluate a real-time mass spectrometer that
has been used to map the field of a sewage outfall among other things. No oil was detected during tests in
the large tank, but the system was able to detect low-level components in a set-up similar to the Phase I
Woods Hole Oceanographic Institution (WHOI) configuration.

Conclusions and Recommendations
The technologies discussed in this report represent an improvement over the existing sunken oil detection
methods. Although these systems have not been tested in the difficult harsh environment of low visibility,
they may be useful immediately in some situations. This use could reduce the amount of effort currently
required and increase reliability of oil detection on the bottom or in the water column.

The multi-beam and imaging sonars appear to be the best sensors to conduct wide area detection surveys.
Some of the signal return issues, which can cause false positive detections for the low grazing angles of
common side-scan sonar, are reduced in the tested systems. Most of these types of systems should be able
to automatically detect large clumps of oil, but the resolution for widely dispersed product is still not
complete. The sooner that a system is deployed before the oil breaks up, the better chance that detection
will occur. Spill responders should ensure that detection equipment has some type of processing software to
interpret raw sensor data.

The laser systems and narrower beam sonars may be better suited as a follow-up to the areas scanned by the
wide scan sonars. These should provide better resolution and should be able to calculate general thickness,
which could provide some information about the amount of oil. The narrow areas covered could introduce
resolution issues, especially for widely scattered oil. On the other hand, the narrow area covered could be
advantageous for guiding recovery efforts.

The real-time mass spectrometry systems should be evaluated for neutrally buoyant oil detection in the
water column. For some spills, especially those in rough waves or fast moving currents, these instruments
may be useful as mounted sensors in a fixed place. This would be especially useful for municipalities and
power plants that use the water for cooling.

The use of this equipment by a Federal On-scene Coordinator (FOSC) is limited at this time due to the level
of development. Guidance is contained in the Appendixes that provide information about the specific
technologies tested. A decision-tool and recommendations for FOSC use are contained in Appendix E.

vi


TABLE OF CONTENTS

EXECUTIVE SUMMARY ............................................................................................................................ v
LIST OF FIGURES ....................................................................................................................................... ix
LIST OF TABLES .......................................................................................................................................... x
LIST OF ACRONYMS, ABBREVIATIONS, AND SYMBOLS............................................................... xi
1 INTRODUCTION................................................................................................................................... 1
1.1 Background ....................................................................................................................................... 1
1.2 Purpose/Objective ............................................................................................................................. 2
1.3 Technical Approach .......................................................................................................................... 2
1.3.1 Performance Requirements ....................................................................................................... 3
1.3.2 Phase I Proof-of-concept Testing.............................................................................................. 4
1.3.3 Phase II Prototype Testing ........................................................................................................ 4
2 PHASE I PROOF-OF-CONCEPT TESTING ..................................................................................... 4
2.1 Overview ........................................................................................................................................... 4
2.1.1 Test Set Up................................................................................................................................ 5
2.1.2 Oil Selection and Sinking Oil ................................................................................................... 5
2.1.3 Test Trays.................................................................................................................................. 6
2.2 RESON, Inc. 7125 SeaBat System ................................................................................................... 7
2.2.1 SeaBat System Description ....................................................................................................... 7
2.2.2 SeaBat Test Description............................................................................................................ 7
2.2.3 SeaBat Results .......................................................................................................................... 9
2.2.4 SeaBat Next Steps ................................................................................................................... 10
2.3 SAIC Laser Line Scan System (LLSS) ........................................................................................... 10
2.3.1 LLSS Description.................................................................................................................... 10
2.3.2 LLSS Test Description............................................................................................................ 11
2.3.3 LLSS Results .......................................................................................................................... 11
2.3.4 LLSS Next Steps ..................................................................................................................... 12
2.4 EIC Laboratories Fluorescence Polarization (FP) .......................................................................... 13
2.4.1 FP System Description............................................................................................................ 13
2.4.2 FP Test Description................................................................................................................. 14
2.4.3 FP Results ............................................................................................................................... 14
2.4.4 FP Next Steps.......................................................................................................................... 15
2.5 Woods Hole Oceanographic Institution (WHOI) Detection and Identification System................. 15
2.5.1 WHOI System Description ..................................................................................................... 15
2.5.2 WHOI Test Description .......................................................................................................... 16
2.5.3 WHOI System Results ............................................................................................................ 17
2.5.4 WHOI System Next Steps ...................................................................................................... 18
2.6 Phase I Summary of Results ........................................................................................................... 19
2.6.1 Test Set-up Considerations ..................................................................................................... 19
2.6.2 POC Test Results .................................................................................................................... 19

vii


TABLE OF CONTENTS (Continued)

3 PHASE II PROTOTYPE TESTING .................................................................................................. 21
3.1 Overview ......................................................................................................................................... 21
3.1.1 Test Set-up .............................................................................................................................. 21
3.1.2 Test Trays................................................................................................................................ 21
3.2 RESON, Inc. 7125 SeaBat Sonar System ....................................................................................... 23
3.2.1 SeaBat System Modifications ................................................................................................. 23
3.2.2 SeaBat Results ........................................................................................................................ 24
3.2.3 Other Considerations .............................................................................................................. 25
3.3 EIC Laboratories Fluorescence Polarization (FP) .......................................................................... 26
3.3.1 EIC System Modifications ...................................................................................................... 26
3.3.2 EIC Results ............................................................................................................................. 27
3.3.3 Other Considerations .............................................................................................................. 29
3.4 Tests of Opportunity ....................................................................................................................... 30
3.4.1 BioSonics ................................................................................................................................ 30
3.4.2 CodaOctopus ........................................................................................................................... 32
3.4.3 SRI International ..................................................................................................................... 35
3.5 Phase II Summary ........................................................................................................................... 35
3.5.1 Test Set-up Considerations ..................................................................................................... 35
3.5.2 Prototype Test Results ............................................................................................................ 35
4 CONCLUSIONS ................................................................................................................................... 37
5 RECOMMENDATION ........................................................................................................................ 38
6 REFERENCES ...................................................................................................................................... 39
APPENDIX A. OHMSETT TEST FACILITY................................................................................... A-1
APPENDIX B. ACOUSTIC DETECTION OF HEAVY OIL .......................................................... B-1
APPENDIX C. OPTICAL/FLUORESCENT DETECTION OF HEAVY OIL .............................. C-1
APPENDIX D. CHEMICAL DETECTION AND IDENTIFICATION OF HEAVY OIL ............ D-1
APPENDIX E. RECOMMENDATIONS FOR FEDERAL ON-SCENE COORDINATORS
FOR OIL SUSPECTED TO BE ON THE SEA BOTTOM .....................................E-1
APPENDIX F. VENDOR CONTACT INFORMATION ..................................................................F-1

viii


LIST OF FIGURES

Figure 1. OHMSETT test facility. ....................................................................................................................5
Figure 2. Chart for calculating oil/barite density. ............................................................................................ 6
Figure 3. Test trays (a) before placing into tank and (b) on tank bottom. ....................................................... 6
Figure 4. RESON 7125 SeaBat system............................................................................................................ 7
Figure 5. RESON 7125 SeaBat system mounted on the OHMSETT tank. ..................................................... 8
Figure 6. Set-up of SeaBat in OHMSETT tank. .............................................................................................. 8
Figure 7. RESON SeaBat sample results. ........................................................................................................ 9
Figure 8. SAIC LLSS (a) being lowered over the side of a vessel and (b) mounted in the OHMSETT
tank. ................................................................................................................................................ 11
Figure 9. SAIC LLSS sample results for (a) visual and (b) fluorescent wavelengths. .................................. 12
Figure 10. EIC FP system (a) on table top, (b) inside probe case, and (c) mounted on OHMSETT tank..... 13
Figure 11. Sample results from EIS FP showing (a) single line and (b) summary of all lines scanned. ....... 14
Figure 12. EIC FP contour plot. ..................................................................................................................... 15
Figure 13. WHOI system (a) MS & fluorometer in waterproof housing, (b) suction hose and transducer,
and (c) navigation system with three transponders. ...................................................................... 16
Figure 14. Results of WHOI lab sensitivity study. ........................................................................................ 17
Figure 15. Results of WHOI sampling in OHMSETT tank (a) before container of oil was added and (b)
after oil was placed. ...................................................................................................................... 18
Figure 16. Phase II test tray configuration (not to scale). .............................................................................. 22
Figure 17. Details of tray configurations. ...................................................................................................... 23
Figure 18. RESON (a) layout of trays and (b) sample results. Dotted red line in (a) is instrument
centerline. ...................................................................................................................................... 25
Figure 19. Estimated survey and processing time for a one square mile survey as a function of depth. ...... 26
Figure 20. EIC FP probe (a) close-up and (b) in the test tank at OHMSETT. .............................................. 27
Figure 21. Two views of EIC FP sample results. .......................................................................................... 28
Figure 22. Superimposed images of tray set-up and FP results. .................................................................... 29
Figure 23. Results of modulation test conducted in bright sunlight showing output and return signals. ...... 29
Figure 24. BioSonics DT-X Digital Scientific Echosounder sensor.............................................................. 30
Figure 25. BioSonics sample echogram......................................................................................................... 31
Figure 26. BioSonics sample analysis results. ............................................................................................... 32
Figure 27. CodaOctopus EchoScope 4D transponder.................................................................................... 33
Figure 28. CodaOctopus sample results. (a) shaded mosaic showing bathymetry (color palette wraps
around so red to red spans 20 cm elevation); (b) raw data of area containing oil, rock, and
seaweed; and (c) data after UIS software processing (blue is oil). ............................................... 34
Figure 29. SRI International mass spectrometer. ........................................................................................... 35
Figure B-1. Acoustic backscattering from the sea floor. ............................................................................. B-1
Figure C-1. Graphical representation of light transmission in water. Water color, turbidity, and other
factors impact the actual attenuation of transmitted light, as well as any corresponding
reflectance or fluorescence. ..................................................................................................... C-1
Figure E-1. Detection decision tree. .............................................................................................................E-4
Figure E-2. The V-SORS used to search for and recover submerged oil. ....................................................E-6

ix


LIST OF TABLES

Table 1. Phase I test results. ........................................................................................................................... 20
Table 2. Phase II oil types and properties. ..................................................................................................... 21
Table 3. Detections and missing detections for RESON position #3. ........................................................... 25
Table 4. Phase II test results. ......................................................................................................................... 36
Table E-1. Advantages and disadvantages of submerged oil detection technologies...................................E-3

x


LIST OF ACRONYMS, ABBREVIATIONS, AND SYMBOLS

AUV Autonomous underwater vehicle
BAA Broad Agency Announcement
BS Backscattering strength
cm Centimeter(s) (10-2 meters)
cP Centipoise
dB Decibel(s)
EIC EIC Laboratories, Inc.
FOSC Federal On-scene Coordinator
FP Fluorescence Polarization
g/ml Grams per milliliter
GPS Global Positioning System
HFO Heavy fuel oil
IMO International Maritime Organization
kHz Kilohertz (1000 cycles/second)
LLSS Laser Line Scan System
m Meter(s)
mm Millimeter(s) (10-3 meters)
MMS Minerals Management Service
MS Mass spectrometer
g/l Micrograms (10-6 grams) per liter
nm Nanometer(s) (10-9 meters)
No. Number
Oil and Hazardous Material Simulated Environmental Test Tank, now called The
OHMSETT
National Oil Spill Response Test Facility
Oil Pollution Preparedness, Response and Co-operation to Pollution Incidents by
OPRC-HNS
Hazardous and Noxious Substances
PAH Polyaromatic hydrocarbons
PMT Photomultiplier tube
POC Proof-of-concept
RDC USCG Research and Development Center
RFI Request for Information
ROV Remotely operated vehicle
SAIC Science Applications International, Inc.
TETHYS TETHered Yearlong Spectrometer
TS Target strength
UIS Underwater Inspection System
USCG U.S. Coast Guard
UV Ultraviolet
V-SORS Vessel-Submerged Oil Recovery System
WHOI Woods Hole Oceanographic Institution

xi



xii


1 INTRODUCTION

1.1 Background
Increased consumption and transportation of oil and its products increase the problem of pollution in lakes
and oceans. Even with strict rules and regulations on oil transportation, accidents leading to oil spills still
occur frequently, with thousands of tons of oil being spilled into seas. This results in the contamination of
marine environments and endangers marine ecology. It is evident that oil spills in coastal waters, harbors,
and oil terminals are especially dangerous and necessitate fast response in order to prevent contamination of
marine habitats when such accidents occur. Therefore, reliable response systems for all types of oil spills
and all environments are needed.

Locating and identifying heavy oil is a problem of growing concern as the use of heavy oil and related
slurry products becomes more prevalent. In addition, even though heavy (sinking) oils have historically
accounted for a small percentage of spills, environmental and economic consequences resulting from such
spills can be high. Heavy oils can sink and affect shellfish and other bottom dwelling marine organisms, in
addition to causing closure of water intakes at industrial facilities and power plants in both salt and fresh
water. The underwater environment poses major problems for spill detection and response, including poor
visibility, difficulty in tracking oil spill movement, and cold temperatures, all of which complicate
containment and recovery. Developing effective methods and technologies suitable for this environment is
a major challenge.

Numerous papers have been written about techniques to detect and recover oil sitting on the sea bottom. In
early papers, authors focused on what conditions are required for oil to sink (Michel and Galt, 1995) or what
should not be done (Castle et al., 1995). Others addressed specific recovery processes (Elliott, 2005 and
Schnitz and Wolf, 2001). The International Maritime Organization (IMO) sponsored a forum in 2002
during which monitoring, modeling, and recovery of heavy oils were addressed (Brown, et al., 2002,
Parthiot, Cabioc’h, 2002). The USCG RDC attempted to build on the efforts of Environment Canada
(Brown et al, 2004 and 2006) by investigating an airborne laser fluorosensor to detect submerged oil (Fant
and Hansen, 2005 and 2006). At least one underwater laser fluorometer system had previously been
deployed (Barbini et al, 2000), although it had not been designed for detecting submerged oil. For the
purpose of this report and following the Coastal Response Research Center (2007) definition, “submerged
oil” describes any oil that is not floating at or near the surface. “Sunken oil” describes the accumulation of
bulk oil on the seafloor.

After the major spills in the U.S. of the M/T Athos in 2004 in the Delaware River and T/B DBL-152 in 2005
in the Gulf of Mexico, the RDC decided to re-examine heavy oil response efforts (Michel 2006). At least
one other commercial effort was pursuing a recovery method at that time (Usher, 2006). Parallel efforts by
international organizations (Parthiot, 2004) were also ongoing. A workshop, co-sponsored by RDC in
December of 2006, also reemphasized research needs (CRRC, 2007).

There is also an ongoing effort within the IMO Oil Pollution Preparedness, Response and Co-operation to
Pollution Incidents by Hazardous and Noxious Substances (OPRC-HNS) Working Group for Submerged
Oils, headed by Italy.

1


1.2 Purpose/Objective
The U.S. Coast Guard (USCG) and industry do not have a consistent and reliable method to recover
submerged oil on the sea floor, a task that includes the multiple phases of detecting, tracking, containing,
and ultimately recovering submerged oil. Response to spills of heavy oils is often ad-hoc, with detection
and removal strategies developed at the time of the spill. These methods are generally inadequate to find
and recover the oil, and the recovery techniques are very labor intensive. The USCG needs to develop a
blueprint for method(s) within the oil response industry to recover heavy oil located on the sea floor. To
assist in this effort, the USCG Research and Development Center (RDC) has embarked on a multi-year
project to develop a complete approach for response to spills of submerged oils, including detecting and
mapping the spilled oil (Stage I) and containing and recovering it (Stage II).

The objective of the first stage is to identify and develop technologies capable of detecting heavy oil on the
sea floor. The ability to detect, track, delineate, and quantify heavy oil on the sea floor will permit
operational decisions to be made regarding feasibility of and best strategy for recovery. This report covers
the results of the heavy oil detection and mapping research.

The long-term objective for the second stage of this effort is to create a system that will recover heavy oil
from the sea floor. Such a system will have to accomplish a variety of tasks to be successful. These include
detecting the oil, possibly concentrating/corralling the oil for collection, and collecting the oil into a
containment vessel for proper disposal. The proofs-of-concept and prototypes that are developed will not be
USCG-owned. They will be owned by industry and available for use if directed by the USCG to recover
heavy oil. These are expected to be incorporated into response plans in the future.

1.3 Technical Approach
The identification of potential technologies was accomplished using a Request for Information (RFI) and a
Broad Agency Announcement (BAA). The RDC released an RFI in the summer of 2006 asking vendors to
provide potential approaches for the detection and recovery of oil on the sea floor. A summary of past
experiences, especially with respect to the two latest spills (Michel, 2006), was provided as part of the RFI.
RDC received responses to the RFI from 15 organizations, some of which addressed several topic areas.
The five major topics addressed in the responses to the RFI were:

• Detection of Oil in the Water Column,
• Detection of Oil on the Bottom,
• Containment of Suspended Oil/Protection of Water Intakes,
• Containment of Submerged Oil on the Bottom, and
• Recovery of Submerged Oil on the Bottom.
The range of costs in the responses indicated that the project would need to proceed in stages. If a reliable
detection technique can not be developed, then a major research effort should not be mounted for the
recovery part of the process. As a result of the information submitted, it was decided to divide the research
effort into detection (Stage I) and then recovery (Stage II).

2


In April of 2007, RDC published a BAA that requested approaches for detection only. The objective of the
specification in the BAA was that the sensors should provide enough information for decision-makers to
determine if an amount of oil sufficient to merit recovering could be identified. The approach was to divide
the BAA process into a proof-of-concept phase (Phase I) where three to five vendors would be awarded
contracts, and then a prototype development phase (Phase II) where two to three vendors would be awarded
contracts. Two sets of performance requirements were listed, one for immediate verification for the
concepts and one for the prototypes. Four technologies were selected for proof-of-concept testing in Phase I.
From these, two proof-of-concept technologies were later selected for further prototype development and
testing in Phase II. That selection was based on the technical requirements as well as the risks associated
with further development.

1.3.1 Performance Requirements

For a successful proof-of-concept, the BAA requested the following capabilities for the detection
technology:

1) Able to identify the presence of heavy oil on the sea floor with 80 percent certainty.
2) Able to detect oil on the bottom from at least 1 meter (m) (3.28 feet (ft)) away.
3) Oil location shall be geo-referenced to 1 m (3.28 ft) in accuracy.
4) Ideally, will provide real-time data, but at a minimum shall produce results and data interpretation
hourly.
5) Able to provide data for all sea floor conditions (i.e., silty, rocky, and gravel bottom types;
vegetation and shellfish-covered bottoms; and over flat and sloped areas and areas with rapid
substrate changes). Phase I testing will be of simple sea floor conditions (i.e., flat and scattered
protrusions).
6) Operate in fresh and sea water conditions equally well.
7) Operate in water depths of up to 33.3 m (100 ft).
8) Have minimal maintenance requirements (easy to maintain and calibrate).
9) Easy to operate and involve minimal training.
10) Easily de-contaminated and durable.
11) Equipment operation not adversely affected by exposure to oil.

Once the proof-of-concept has been demonstrated, the prototype device (or combination of devices) should
be able to operate in the following conditions:

1) Able to search a one square mile area in a 12-hour shift.
2) Operate in water current of up to 1.5 knots.
3) Operate in up to 5-foot seas.
4) Operable during the day and night.
5) Able to be set up within 6 hours of arriving on site.
6) Easily deployable and transportable. Capable of being deployed from a vessel of opportunity and a
variety of other platforms (i.e., towed bodies, remotely operated vehicles (ROVs), autonomous
underwater vehicles (AUVs), and manned submersibles).

3


1.3.2 Phase I Proof-of-concept Testing

The proof-of-concept (POC) testing permitted RDC to determine if the proposed technologies that are being
adapted from other areas can actually be used to find oil. For the Phase I proof-of-concept evaluation, four
vendors were selected:

• RESON: Multi-beam Sonar.
• Science Applications International Corp. (SAIC): Laser Line Scan System (LLSS) adapted for
fluorescence.
• EIC Laboratories: Fluorescence Polarization.
• Woods Hole Oceanographic Institution (WHOI): In-Situ Mass Spectrometry and In-Situ
Fluorometry.
The systems were tested at the Department of Interior Minerals Management Service (MMS) Oil and
Hazardous Material Simulated Environmental Test Tank (OHMSETT), now called The National Oil Spill
Response Test Facility, in Leonardo, NJ between November 2007 and February 2008. Section 2 of this
report contains the discussion and results of the Phase I proof-of-concept testing.

1.3.3 Phase II Prototype Testing

Two of the systems from Phase I were selected for prototype testing at OHMSETT. They were: RESON
Multi-beam Sonar and EIC Laboratories Fluorescence Polarization. Three additional vendors, Biosonics,
CodaOctupus, and SRI International, tested their detection equipment at OHMSETT as well (at no expense
to the USCG). Section 3 contains the discussion and results of the Phase II prototype testing. The
Appendices contain supporting information.

2 PHASE I PROOF-OF-CONCEPT TESTING

2.1 Overview
The facility chosen for the heavy oil detection POC testing was MMS OHMSETT in Leonardo, NJ. The
RDC believed that OHMSETT could provide a somewhat realistic environment while providing the ability
to create targets and sufficient area to address the multiple aspects of each type of approach. Of the 11
capabilities for a successful POC listed in the BAA (see Section 1.3.1), the following could potentially be
demonstrated at OHMSETT:

• Able to identify the presence of heavy oil on the sea floor with 80 percent certainty.
• Able to detect oil on the bottom from at least 1 m (3.28 ft) away.
• Oil location shall be geo-referenced to 1 m (3.28 ft) in accuracy.
• Ideally, will provide real-time data, but at a minimum shall produce results and data interpretation
hourly.
• Able to provide data for all sea floor conditions (i.e., silty, rocky, and gravel bottom types;
vegetation and shellfish-covered bottoms; and over flat and sloped areas and areas with rapid
substrate changes). Phase I testing will be of simple sea floor conditions (i.e., flat and scattered
protrusions).

4


Vendor ability to meet the other POC performance requirements was determined by RDC discussions with
vendors and review of materials they provided.

2.1.1 Test Set Up

Three of the four systems were tested in OHMSETT’s large outdoor tank (see Figure 1). Because of the
time of year of the testing (winter) and the nature of the sensors, the system from WHOI was tested in an
inside tank (described in Section 2.5). The OHMSETT facility provides an environmentally safe place to
conduct objective testing and to develop devices and techniques for the control of oil and hazardous material
spills. Appendix A gives more details about the OHMSETT facility.

Figure 1. OHMSETT test facility.

The large test tank at OHMSETT is an above-ground concrete tank measuring 203 m (665 ft) long by 20 m
(66 ft) wide and 3.4 m (11 ft) deep. At the time of testing, however, the water depth was reported as 2.4 m
(8 ft). The test tank is equipped with a tow bridge that spans the width of the tank and moves along the tow
length at speeds of up to 6.5 knots. The bridge is outfitted with a climate-controlled laboratory space to
accommodate personnel, computers, bridge motion controls, and other components sensitive to the
elements. The salinity in the tank during the tests was 26 parts per thousand.

2.1.2 Oil Selection and Sinking Oil

The first major challenge in the proof-of-concept testing was to determine how to create stable oil targets
under water. The two oils selected were Sundex 8600, a heavy oil used by OHMSETT, and Number (No.) 6
fuel oil (also known as Bunker C or heavy fuel oil (HFO)). Due to the specific gravity of the test oils
relative to the tank water, barite (BaSO4) was incorporated in the oil samples to increase their bulk density
to ensure the samples would be heavier than the tank water and that the samples would remain deposited
within the test trays throughout the study period. Figure 2 shows the chart used to determine the amount of
barite needed to ensure that the densities of the oils were heavier than the water, about 15 percent by weight.

5


Barite Mixing Chart

50
45

Weight Percent Barite
y = 100.77Ln(x) + 2.49
40
35
HFO
30 Sundex
25
20
y = 99.404Ln(x) + 6.28
15
10
5
0
0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.60 1.70
Mixture Density (g/cm 3 at 20°C)

Figure 2. Chart for calculating oil/barite density.

2.1.3 Test Trays

Two test trays were constructed to hold the oil at the bottom of the OHMSETT tank. Each was fabricated
from aluminum and measured 2.4 m by 2.4 m (8 ft by 8 ft). One tray had a 15.24 cm (6 inches) lip and the
other a 20.32 cm (8 inches) lip. The trays were filled with construction sand and then depressions were
made for false targets and oil-filled locations (see Figure 3a). Both types of oil and a piece of roofing tar
were placed in their respective locations along with some false depressions. The targets were 1-3 inches
thick and 2-3 ft in diameter. The trays were filled with water to saturate the sand and moved to the bottom
of the OHMSETT main tank (see Figure 3b). Over time, sediment settled on some of the targets.

(a) (b)
Figure 3. Test trays (a) before placing into tank and (b) on tank bottom.

6


2.2 RESON, Inc. 7125 SeaBat System
2.2.1 SeaBat System Description

The SeaBat 7125 system is a dual-frequency multi-beam echo sounder (sonar) designed to measure relative
water depths over a wide swath perpendicular to the towing vehicle’s track. It can operate at either 200 kHz
or 400 kHz. For the POC test, the sonar was connected to RESON PDS2000, a software package designed
for hydrographic survey and dredging operations. Figure 4 shows the components of the SeaBat system.
According to the manufacturer, the SeaBat 7125 ensonifies a 128 degree sector below the sonar head
assembly and is suitable for mounting on a surface vessel, ROV, or AUV.

Figure 4. RESON 7125 SeaBat system.

The theory for the acoustic detection of oil is that the oil would be less reflective than the sand. On the
sonar image, the less reflective material is represented as darker shades on the display and the more
reflective material is displayed in lighter shades. Appendix B gives a more detailed discussion of the
acoustic detection of heavy oil.

2.2.2 SeaBat Test Description

The Phase I POC demonstration focused on the sonar capability to detect the oil sample by visual inspection
of the acoustic image and simple thresholding. The SeaBat POC test was carried out at the OHMSETT test
facility on 28 November, 2007. The sonar head was installed rigidly on a pole that could be moved across
the bridge spanning the width of the tank (see Figure 5). The sonar was placed just below the water surface,
2 m (6.6 ft) above the bottom of the tank (see Figure 6). The bridge moved along the tank at speeds ranging
from 0.1 to 6.0 knots. The test conditions simulated a very calm sea. The wind did not affect the test, and
no rain or other source of acoustic noise was present.

7


Figure 5. RESON 7125 SeaBat system mounted on the OHMSETT tank.

Figure 6. Set-up of SeaBat in OHMSETT tank.

The sonar was set to operate in a normal survey mode and the bridge simulated a vessel. The sonar was
moved laterally across the width of the bridge to test the effective swath of the system. The system was
primarily operated at 400 kHz; however, two test runs at 200 kHz were also conducted.

8


2.2.3 SeaBat Results

The sonar was moved over the targets several times, with most runs performed at a frequency of 400 kHz.
The very high ping rate of the system and the short range dictated by the test set-up provided plenty of data
for detection. All targets (Sundex, No. 6 oil, and asphalt) were positively detected, as each of the samples
appeared clearly on the monitor as dark areas against a brighter background. This was the expected result,
given the low acoustic reflectivity of oil compared to that of typical seabed sediment. Figure 7 gives an
example of the SeaBat results.

(b) (c)

(a) (d) (e)

Figure 7. RESON SeaBat sample results.

The left portion of Figure 7 (a) shows the raw intensity data collected while the sonar was traveling over the
tray. The top center image (b) is a zoom on the data being processed (bottom tray in figure on left, area
inside white rectangle). The top right image (c) shows color-coded bathymetry results provided by the
multi-beam echo sounder. The bottom center image (d) shows the automatic detection results and the
bottom right image (e) shows the outline of the detected objects overlaid on the raw intensity data.

All targets present in the sonar swath were positively detected with a probability exceeding 80 percent, with
the exception of one target. That target was properly segmented as oil, but the detection process merged it
with the tray walls. On this target, the detection rate was 75 percent. The same target oil in the other tray
was detected with a rate of 85 percent. The processing parameters were kept constant for all the runs. The
process generated one false alarm on a blank target. This represents an overall false alarm rate of 8 percent.

9


2.2.4 SeaBat Next Steps

The POC test showed that a trained operator can visually detect the patches representing heavy oil on flat,
coarse, sandy sea floor with a high degree of certainty. A series of further experiments will be required to
gather enough data for the development of a more robust solution involving software that can call suspicious
areas to the attention of the operator.

The data collected during this exercise, as well as data collected by RESON at OHMSETT independent of
this exercise, were used to develop an automated detection system that does not rely solely on the operator
to visually detect the oil. The prototype system was expected to include an advanced image processing
solution and a model inversion solution. These solutions would be based on measuring the backscattering
strength as a function of multiple parameters, including the physical characteristics of the seafloor.

2.3 SAIC Laser Line Scan System (LLSS)
2.3.1 LLSS Description

The SAIC SM-2000 LLSS was originally developed as a reconnaissance tool for sea floor characterization
and underwater search and recovery operations. This high-resolution survey tool was designed for the
identification of substrate types, assessment of biological resources, and detection of hard targets on the
seafloor. The optically-based system was originally developed to bridge the gap between underwater video
and side-scan sonar by emitting a high-power, blue-green (532 nanometer (nm) wavelength) laser and
reading the intensity of light reflected back to an internal receiver at that wavelength.

Although the system provides an excellent light source for imaging the seafloor and gathering data based on
light reflectance, the blue-green laser was not considered the optimal light source for exciting deposits of oil
on the seafloor. As a result, SAIC incorporated a lower wavelength, higher energy laser light source that
approached the ultraviolet-A (UV-A) band (405 nm) to exploit the fluorescent properties of heavy oil and
develop an accurate submerged oil detection tool. Appendix C discusses the optical and fluorescent
detection of heavy oil in more detail. Although just above the UV-A band, the 405 nm laser was selected as
the best option to elicit fluorescence underwater due to its ability to deliver a high amount of energy per
photon while offering increased resistance to attenuation as it is transmitted through the water column.

Further modifications to the LLSS included the incorporation of precision optical filters to control the
intensity and wavelength of light that entered the receiver unit. The filtering scheme was specifically
designed to target the fluorescent response of oil-based compounds centered at 480 nm, which was
determined to be the most advantageous starting point for the Phase I testing. Previous research had shown
that crude oil responds to laser excitation from a low-wavelength light source over a broad range (400 to
650 nm), with peak intensity recorded in proximity to 480 nm. In addition, the filtering scheme was
designed to eliminate the effects of light reflection associated with the 405 nm laser, as well as the potential
of false positives based on the fluorescence of dissolved organic matter (420 nm) and chlorophyll-a (685
nm) in the water.

10


2.3.2 LLSS Test Description

Prior to the proof-of-concept testing at OHMSETT, SAIC conducted a dockside wet test of the LLSS in
local waters (see Figure 8a). This test was conducted to better facilitate efficient POC testing, as well as to
provide insight into possible interference or noise associated with fluorescence from other compounds in the
natural environment. Representatives of RDC observed much of the wet test so they could see how the
modified LLSS worked in a harbor environment.

(a) (b)
Figure 8. SAIC LLSS (a) being lowered over the side of a vessel and (b) mounted in the OHMSETT tank.

During the POC testing at the OHMSETT tank, the LLSS was suspended in water by a four-point harness
system, which was secured beneath the moving bridge (Figure 8b). The bridge was then maneuvered to
make multiple passes over the two test target trays during daylight, dusk, and darkness. Due to the
relatively shallow water depth within the tank, the LLSS was positioned to scan the trays at an angle
(approximately 30 degrees). This increased the focal distance between the LLSS and the test trays on the
tank bottom to the minimum operating distance.

2.3.3 LLSS Results

During daylight conditions, the sunlight saturated the test area with the wavelength of light that the filtered
receiver was designed to capture. As a result, the modified LLSS provided accurate imagery data (Figure
9a) but failed to elicit and/or detect any fluorescent signal over the background light.

The LLSS imagery acquired over the test trays during the night runs was essentially a monochromatic image
with dark areas indicating zero to weak fluorescent return (Figure 9b). The brighter areas in the imagery,
representing relatively intense return within the preferred bandwidth, were indicative of a response to the
excitation laser at sufficient strength to pass through the filter. The scan angle of 30 degrees required by the
test configuration resulted in a narrower swath than would be expected in field conditions. The outermost
returns were out of sync, resulting in a darker image on one side of the swath. The intensity of the return
signal suggests that the 10 nm band pass, 480 nm filter was adequate to capture the light emitted by the
weathered Sundex 8600 oil deposits that were embedded within the sediment matrix (this is not obvious in
Figure 9b due to the Sundex 8600 being in the area of no data in that image). In contrast, the fluorescent
response of the No. 6 fuel oil and roofing tar deposits were present and detectable by the modified LLSS,
but recorded at a much lower intensity.

11


(a) (b)
Figure 9. SAIC LLSS sample results for (a) visual and (b) fluorescent wavelengths.

2.3.4 LLSS Next Steps

In general, the data obtained as part of the POC testing indicated that the modified LLSS could be an
effective tool for the detection and mapping of heavy oils on the seafloor through fluorescence. However,
the results also indicated that further refinement and experimentation with the laser technology would be
required to fully satisfy the objectives of the research and development effort.

Due to the depth requirement and the light reflection off of the tank sides and bottom, this system cannot be
fairly tested at OHMSETT again. The system needs a better way to improve signal/noise ratio to detect
fluorescence. Possibilities include a more powerful laser and/or better processing. The following is a list of
key elements of any future modifications and testing of the LLSS as a submerged oil deposit detection and
mapping system.

1) Limit ambient light – The clear water and white epoxy paint in the OHMSETT test tank did not
represent optimal conditions to conduct the testing. Future testing should be performed in an
environment that better mimics conditions in a coastal harbor and/or port facility where sizable spills
of heavy oil are more likely to occur.
2) Increase the power of the laser light source – Major considerations in the design of this optical tool
are compensating for the attenuation of the 405 nm laser excitation light, as well as maximizing the
intensity of the fluorescent response of an oil deposit. Even if dissipated somewhat by turbidity or
dissolved organics, a higher intensity laser should increase the operational range of the LLSS and
reduce the effects of suspended particulates and water color.
3) Filtering – The initial testing indicated some apparent differences in the intensity of the returns
associated with each type of test oil, suggesting the current configuration of the laser light source
and filtering scheme may be better suited to low molecular weight polyaromatic hydrocarbons
(PAHs) relative to high molecular weight PAHs. Continued refinement of the modified LLSS
through the alteration of the return light signal filtering scheme to allow the passage of a broader
spectrum of visible light, inclusive of the green and red color range, is likely to improve the
capability to detect a wider variety of heavy oil deposits on the seafloor.
4) System Dimensions – The results of the tank testing indicated that the minimum distance between
sensor and target of 2.5 m and the sheer size and weight of the existing LLSS unit are limitations that
need to be addressed as part of the future refinement of the LLSS as a oil detection tool.

12


2.4 EIC Laboratories Fluorescence Polarization (FP)
2.4.1 FP System Description

Fluorescence spectroscopy has been shown to be an effective tool for monitoring oil contaminants in water.
Because the main constituents of oils are aromatic compounds, illumination of oil samples with ultraviolet
or visible light causes the oil samples to emit fluorescence. Fluorescence-based methods have several
advantages, including: they are non-contact, have high-sensitivity to the presence of aromatic hydrocarbons,
and are easily miniaturized. There are other fluorescing species in marine environments however, such as
humic compounds and chlorophyll, that may interfere with direct fluorescence measurements. In addition,
ambient light interferes with fluorescence measurements and limits the use of fluorescence to night unless
costly pulse excitation/detection schemes are used. One way to mitigate these problems and to enhance the
selectivity of fluorescence is to incorporate polarization into the measurement technique. FP measurements
are based on the assessment of the rotational motions of species. In particular heavy oils, which are very
viscous, will show significant fluorescence polarization when excited with polarized light. Appendix C
discusses the optical and fluorescent detection of heavy oil.

In EIC’s system, a compact, continuous wave, green (532 nm), diode-pumped, solid-state laser is used for
fluorescence excitation. Figure 10 shows the EIC FP system. The main components of the FP probe are the
fiber optic fluorescence polarizer and a telescopic focusing/collection optic. The fiber optic fluorescence
polarizer consists of three miniature optical trains (laser excitation, perpendicular FP collection, and parallel
FP collection) arranged in a backscattering collection probe configuration. The probe telescope is a simple
refractor telescope consisting of a 50 mm diameter, 100 mm focal length objective lens and a 9 mm
diameter, 11 mm focal length eyepiece. The telescope is used to focus the laser beam into the sample and
also to collect the fluorescence emitted by the sample. With the telescope as the front optics of the FP
probe, the probe can detect fluorescence signals from fluorescent samples less than 1 m away from the
probe to several meters away. The telescope focus is adjusted by moving the eyepiece between the polarizer
and the objective lens. In the current POC probe, the eyepiece is moved manually; however, the eyepiece
can be mounted into a linear actuator where the linear movement can be controlled via software and thus
allow active focusing of the FP probe.

(a)

(b) (c)

Figure 10. EIC FP system (a) on table top, (b) inside probe case, and (c) mounted on OHMSETT tank.

13


2.4.2 FP Test Description

After the FP probe and instrument were deployed from the tow bridge at OHMSETT, the first test was to
determine whether the FP probe could detect the individual oil targets in the test trays. The FP probe was
slowly scanned (0.5 knot speed) through each of the oil targets while the FP signal was continuously
recorded. In some oil targets, the probe was stopped for a short time. Several strong polarization signals
(>0.25) were observed during the scan, and these signals correspond to areas when the probe focus was on
oil targets. In several of the targets, the oil samples were partially covered with sand. Even with these
samples however, an FP signal was still detected.

FP grid scans of the test trays were also performed. To allow geo-referencing of the detected FP signals, a
differential Global Positioning System (GPS) antenna was attached to the FP probe during the grid scans to
record GPS coordinates that were then paired with the FP signals. Grid scans were started on one corner of
the tray so that the first scan is about 0.67 m (2 ft) away from the edge of the tray. The tow bridge was
moved from along the length of the tray, and at the end of each scan the probe was translated by about 15.24
cm (6 inches). The last scan was about one meter away from the other edge of the tray. The tow speed was
kept constant during the scans.

2.4.3 FP Results

Test results of the POC fluorescence polarization instrument at OHMSETT indicate that the FP probe is
capable of accurately detecting heavy oil in real time. Oil targets in the test trays showed significant FP
signals that can be easily distinguished from ambient backgrounds such as sunlight or background
fluorescence. Figure 11a shows the linear plot of a grid scan that was performed at a 1-knot speed. In this
plot, it can be seen that several strong fluorescence polarization peaks were observed in the middle section
of the graph, corresponding to the area when the probe was over oil targets. Figure 11b shows the GPS plot
of the 1-knot speed grid scan and clearly shows the FP peaks in the center of the grid. Figure 12 shows the
EIC FP results as a contour plot. All testing was done during daylight hours on cloudy to overcast days. No
interference from sunlight was observed. Furthermore, it was determined during testing that the test tank
surface paint fluoresces, but did not give a strong FP signal.

(a) (b)

Figure 11. Sample results from EIS FP showing (a) single line and (b) summary of all lines scanned.

14


Figure 12. EIC FP contour plot.

2.4.4 FP Next Steps

The ultimate goal of this project is to develop an autonomous submersible fluorescence polarization detector
for heavy oil that can be integrated with different types of deployment vehicles. To achieve this goal, the
plan for Phase II was to miniaturize the components of the POC FP instrument, assemble them into a
compact instrument, and encase them in a sealed tubular housing. The FP instrument should incorporate an
embedded computer to allow the system to operate autonomously and communicate with the host vehicle.

2.5 Woods Hole Oceanographic Institution (WHOI) Detection and Identification
System
2.5.1 WHOI System Description

The WHOI detection system relies on two complementary modes of hydrocarbon sensing: a TETHered
Yearlong Spectrometer (TETHYS) mass spectrometer (MS) in combination with an off-the-shelf UV
fluorometer. See Figure 13 for the WHOI system: (a) MS & fluorometer in waterproof housing, (b) suction
hose and transducer, and (c) navigation system with three transponders. The TETHYS instrument is an
underwater in-situ MS developed through a partnership between WHOI and Monitor Instruments LLC. The
UV fluorometer (Chelsea Instruments Ltd., Surrey, England) is sensitive to aromatic hydrocarbons
fluorescing at 360 nm. The TETHYS instrument is capable of identifying and describing hydrocarbon
composition across a broad spectrum ranging from methane to tridecane, as well as halogenated
hydrocarbons and many other toxic industrial chemicals. The instrument utilizes a proprietary non-evaporable
getter ion pump and mass analyzer developed by Monitor, called the Miniature Mass Analyzer (MMA). One
suction pump pulls the water into the instrument for sampling. Another pump and hose combination is used to
spray the surface of the oil to get oil particles into the water column which the other system can sample.

15


(a) (c)

(b)

Figure 13. WHOI system (a) MS & fluorometer in waterproof housing, (b) suction hose and transducer, and
(c) navigation system with three transponders.

2.5.2 WHOI Test Description

2.5.2.1 Laboratory Sensitivity Study
Prior to the OHMSETT tests, WHOI conducted lab sensitivity tests with the two test oils used at
OHMSETT. Based on these heavy oil samples, a series of molecular “fingerprints” of hydrocarbon
constituents was developed with the TETHYS MS and UV fluorometer. These data were then used to
develop mission scripts to monitor for specific hydrocarbon compounds with high signal-to-noise ratios.
This resulting classification system optimized the MS’s operational parameters to track only relevant ion
peaks, thereby improving the operational response of the MS. The aromatic UV fluorometer was operated
in parallel during this sensitivity analysis to develop a composite limit of detection and response metrics for
detection of heavy end members from these heavy oil types. Figure 14 shows the results of the sensitivity
analysis (the vertical scale for the blue bars is MS ion counts and for the red bars is carbazole equivalent
micrograms per liter (μg/l)). The system can detect trace amounts of specific compounds but it is not clear
if the levels during testing would reach minimum values needed.

16


Figure 14. Results of WHOI lab sensitivity study.

2.5.2.2 OHMSETT Testing
The WHOI system was tested inside (in February) in a portable FasTank (3.05 m (10 ft) diameter, water
about 0.965 m (3.2 ft) deep). WHOI used a short baseline transponder system that included a transducer at
the end of the snorkel pole to position the system. Trial operations were conducted as three surveys, with
TETHYS MS and UV fluorometer system operating from an aluminum platform directly above the test
tank. An intake snorkel was moved through the water in a pattern consisting of four parallel tracklines,
spaced with approximately 0.5 m separation and at a vertical distance of approximately 0.5 m above the
tank bottom. During each grid survey, the snorkel was moved to and held at discrete waypoints (3 to 4
waypoints per trackline) while the MS and UV fluorometer measured the hydrocarbon levels at that
position. The first survey was conducted as a negative control, without any hydrocarbons. The second
survey, of similar geometry, was conducted with a hydrocarbon sample in the tank. The third survey was
conducted after the hydrocarbon sample was repositioned in order to characterize the oil diffusion rate.

2.5.3 WHOI System Results

Analysis of the Sundex 8600 and No. 6 fuel oil samples indicate that the TETHYS MS is well suited to
detect trace levels of volatile short-chain hydrocarbons (e.g., methane through octane), while the UV
fluorometer is able to detect water-soluble aromatic hydrocarbon components (e.g., benzene, toluene,
xylene, and naphthalene). Most heavy oil spills contain small but significant fractions of these volatile or
water soluble petroleum fractions, making the MS and fluorometer combination highly useful for detecting
heavy oil hydrocarbon contamination. Gas chromatographic analysis of the short-chain hydrocarbons from
samples taken in parallel with TETHYS MS and UV fluorometer data suggest that even when undispersed
(i.e., settled on the bottom and not disturbed), Sundex 8600 and No. 6 fuel oil both emit small but detectible

17


amounts of these light hydrocarbons into the water column. Furthermore, these sensitivity data suggest that
because the flux rates are extremely low, plumes of these heavy oil tracers may persist at detectable levels in
the water column for weeks to months in calm water.

Figure 15 shows sample results from the OHMSETT testing. Results from this research program suggest
that common heavy petroleum product spills, including sinking fuel oils such as No. 6, can be located and
identified through the use of light fractions as tracers using MS techniques. Low molecular weight aromatic
compounds possessing high water solubility may also serve as tracers using UV fluorometry, although the
addition of barite to oil samples appears to render this technique ineffective. It is unclear if turbulent mixing
of the oil will improve UV fluorometer sensitivity.

(a) (b)
Figure 15. Results of WHOI sampling in OHMSETT tank (a) before container of oil was added and (b)
after oil was placed.

POC tank tests at the OHMSETT facility have demonstrated that plumes of light tracer compounds
(methane through butane) readily diffuse from sunken oil sources and persist with sufficient intensity over
time to be detected using the TETHYS mass spectrometer. This technique is valid at distances greater than
1 m with better than 80 percent accuracy under the conditions tested. Furthermore, by combining this
sensory technique with high precision acoustic navigation, intensity contour plots can be constructed that
accurately characterize the source location and spatial extent. It is not clear, however, if the oil will actually
release components into the water column and what their movement will be in even a small amount of
current. In addition, the sensor may have to be deployed very close to the bottom, which could be
problematic, especially for rough bottoms.

2.5.4 WHOI System Next Steps

To improve the system, the TETHYS components could be optimized to improve their spectral resolution
and sensitivity to the oil fractions identified in the fuel oil. Information regarding the behavior of other
submerged oils, whether through models or experimentation, would be needed to further refine the system.

This system is already being used in the Gulf of Mexico. A recent hurricane caused an underwater
avalanche that buried some oil pipeline. To relocate the pipeline, a water jetting tool was used to disturb the
upper layer of the silt, and the WHOI system was then used to sample the water column for hydrocarbons.
The process was successful in mapping the bottom that was saturated with oil.

18


2.6 Phase I Summary of Results
2.6.1 Test Set-up Considerations

As discussed in Section 2.1, OHMSETT was considered to be the best facility for conducting the POC tests.
Not all of the BAA performance requirements could be tested at this facility. Some capabilities, such as the
ability to operate equally well in fresh and sea water, needed to be determined independent of the
OHMSETT tests.

There were some other issues resulting from the limitations of the test set-up. The relatively small tray
areas as compared to the swath width of the instruments caused problems for some of the instruments during
this round of evaluations. In addition, it was not clear in the beginning if the silt in the tank would influence
results.

2.6.2 POC Test Results

The testing objective was met and the proof-of-concept evaluation was successful. All four of the systems
located oil under the conditions that were given; that is, clear water with a limited amount of turbidity or
sand covering the oil. Table 1 gives a summary of the test results compared to the BAA performance
requirements listed in Section 2.1.

RESON is adapting an existing system. Although sonar systems have been used in the past to locate
submerged oil, the issue of concern is the turn-around time of the interpretation, and RESON appears to be
addressing that issue. It is not clear how this system will perform in muddy bottoms where the difference in
density between the oil and bottom is closer than the conditions documented in this test.

The SAIC system is adapted from an existing system and appears to work in low light conditions – again
given reasonable clarity. Additional refinements were recommended for any additional efforts. Any future
tests should take place in a more realistic environment so that the light levels and focal length are in line
with the system performance, as these conditions cannot be met in a controlled tank environment.

The EIC equipment is a new approach and while it may have more risk than the other systems, it also may
have the most applicability. The small size of the equipment may lend its applications to multiple uses,
including mounting in small ROVs or AUVs. It also may be small enough to be mounted on a suction head
during recovery operations.

The WHOI system has already been used to detect some oils in a calm water column and it appears the
approach could be refined depending upon the oil spilled. The sinking mechanism for the spill must be such
that the lighter components of the oil are still available, such as fuel oil that mixes with sand. But it is not
clear how much dissolved or particulate oil would be in the water column under more realistic
circumstances, especially after several days or weeks or with current flow. Predictive models of heavy oil
are not currently available for that scenario.

19


Table 1. Phase I test results.
Requirement RESON SAIC EIC WHOI
Identification The detection rate was Poor response in Yes [certainty not Both oils tested were
of heavy oil on 100% on most of the daylight detection of listed in report]. detected in the water
sea floor (80% targets. It was at least ambient light. Better column in the limited
certainty) 80% on all targets. results at night. configuration.
Ability to The detection range was Focal length of the Can detect signals No, uses very close-
detect oil on only limited by the tank laser was longer from fluorescent in sampling
the sea floor geometry. The longest than the depth of the samples <1 m away technique.
from at least detection range was 4 m. tank. Should be from the probe to
1 meter away able to meet. several meters
away.
Georeference If the tank facility is located Done with previous Software for the data Real-time position
oil locations outdoors and a valid system. acquisition board estimate of the
Differential GPS track is allows a GPS signal hydrocarbon sample
obtainable then the oil to be recorded along was accomplished
targets could possibly be with the FP signals using a 150 kHz
geo-referenced as part of for geo-referencing. long baseline
the tank test. navigation system.

Real time data The POC system had a Can view from Detection display Generally yes, but
real-time display, but needs screen with available in real- not clear if covering
work to produce real-time additional time. Contour plots a large area.
analysis. coordination with require a grid scan
locations needed. and data processing.
Operate in Tested in sea water only Tested in sea water Tested in sea water Tested in fresh
fresh and sea but no impediments noted. only but no only but no water only but no
water impediments noted. impediments noted. impediments noted.
conditions
equally well

The technologies represented here are an improvement over the existing ad-hoc methods. Although these
systems have not been tested in the difficult harsh environment of low visibility, they may have immediate
use in some situations, for which they could reduce the amount of effort and increase reliability of oil
detection on the bottom or in the water column. T

The amount of available funding limited selection to two choices for Phase II prototype development.
Since the WHOI system was not able to reliably detect the oil from 1 meter away, it was eliminated from
future testing. The SAIC system was large and could not be adequately evaluated in the shallow confines of
the Ohmsett tank. In addition, unacceptably high risks are associated with the large amount of work needed
for further development for the SAIC system; so it was also eliminated from further testing. The other two
systems showed more capabilities when used in combination; so they were chosen for further evaluation.;.
The next step was to complete prototype development and evaluate the RESON sonar and EIC fluorosensor
at OHMSETT in a more realistic environment.

20


3 PHASE II PROTOTYPE TESTING

3.1 Overview
As discussed in Section 1.3, the BAA requires that the prototype device (or combination of devices) shall be
able to operate in the following conditions:

1) Able to search a 1 square mile area in a 12-hour shift.
2) Operate in water current of up to 1.5 knots.
3) Operate in up to 5-foot seas.
4) Operable during the day and night.
5) Able to be set up within 6 hours of arriving on site.
6) Easily deployable and transportable.
7) Capable of being deployed from a vessel of opportunity and a variety of other platforms (i.e., towed
bodies, ROVs, AUVs, and manned submersibles).

3.1.1 Test Set-up

The two types of oil (No. 6 fuel and Sundex) and asphalt from the first test were used again, as well as a
new slurry oil with a high enough density so that addition of barite was not required. During the tests, the
water temperature was about -1° C and the salinity was 23 parts per thousand. Table 2 gives the densities
(in grams per milliliter (g/ml)) and the estimated viscosities (in centipoise (cP)) for the oils used in Phase II.

Table 2. Phase II oil types and properties.
No. 6 Fuel Oil Sundex 8600 Tesoro Slurry
o
Density (g/ml @ 1 C) 1.083 1.071 1.0626
Viscosity (cP @ 30.5o F (-0.8oC)) 700,000 550,000 80,000

3.1.2 Test Trays

A new test configuration was designed using ten trays, each 2.4 meters by 6.1 meters (8 ft by 20 ft),
resulting in a 12 meters by 12 meters (40 ft by 40 ft) test area. The trays had 4-inch sides and were filled
two to four inches with four types of bottom or substrate (stone chip/sand mix, river silt, pea gravel, and
#100 sand). Each tray had a different combination of oil types, oil deposit configurations (approximately
one inch deep), and substrates. Rocks and seaweed were placed intermittently. The layout and contents of
the trays are shown in Figure 16. Figure 17 shows the details of the target configurations.

21


Rocks/oil
Rocks rocks 1 ft.
seaweed
2 ft Dia 12” Wide
6” Wide
2
1
Rake oil
6” Wide into gravel
3 ft Wide 12” Wide
12” Wide
3
4
Oil around 2- 3 ft Wide Rocks/oil
foot rock rocks 1 ft.

5 2 ft Dia 2 ft Dia
6

12” Wide 12” Wide

Rocks 8
7 seaweed 3 ft Dia
6” Wide

3 ft Dia
12” Wide
2 ft Dia 10
9 2 ft Wide
3 ft Dia 12” Wide

Substrates: Original Sand Oils: Sundex

Silt from Delaware Bay No. 6

Pea Sized Gravel Asphalt Roofing

#70 Sand (210 microns) Tesoro Slurry Oil

Rocks/Cement Sand Wave (3-6 inches high) # Tray Number

Figure 16. Phase II test tray configuration (not to scale).

22


Figure 17. Details of tray configurations.

3.2 RESON, Inc. 7125 SeaBat Sonar System
3.2.1 SeaBat System Modifications

RESON modified their SeaBat system with an algorithm to interpret the data from their multi-beam echo
sounder. For input, the detection algorithm uses calibrated backscatter levels from which the backscattering
strength (BS) is derived. The detection algorithm functions in two steps. First, it estimates bottom
topography and uses it to set a zero boundary. Then signals of BS below that boundary are evaluated
graphically as “oil” to produce “bins.” The BS of the bins is then compared to a reference angular response
curve for black oil. (The reference curve is based on measurements carried out in independent tests by
RESON in 2008, also at the OHMSETT facility.) If the average difference between the reference and
measured backscatter is below a pre-set threshold, the response is classified as oil.

Calibration of the system is crucial. The best method is to use a smooth hard (metal) sphere directly below
the sonar. The correction factor calculated is then applied to all data. A calibration can also be done using a
known type, such as in this case, the bottom of the tank.

23


The sonar was mounted on OHMSETT’s moveable bridge and positioned approximately 1.9 m (6.2 ft)
above the bottom of the test tank. The depth to the sediment of the trays was approximately 1.65 m (5.4 ft).
The swath width applied was 110°, consequently the swath width on the ground was approximately 4.9 m
(16.1 ft). Five survey passes were conducted to cover the test area with an overlap of approximately 2 m
(6.6 ft). A sound velocity probe was mounted beside the sonar to provide sound velocities in real time to the
beam former. The detection algorithm was set up in a software package, MATLAB, which is not embedded
in the sonar software. A GPS device was installed above the sonar for real-time positioning data
acquisition. Processing was done off-line and the input from the GPS units was correlated with the data.

3.2.2 SeaBat Results

Evaluation of the detections was based on the areas of the detected objects rather than on the number of
detected objects. Evaluation based on number of detected objects would discriminate against the limited
number of large detected areas in favor of the high number of small areas. The detection software estimated
the area of each detected patch. For each detected patch, it was decided qualitatively whether a detected
patch belonged to a real patch. For each survey, the detection rate and the false alarm rate were derived.

Figure 18 and Table 3 show the results for sonar position #3 (near the centerline of the array). The other
four positions showed similar results. It appears that the software algorithm can learn what is most likely oil
versus bottom and automatically outline these areas (Figure 18(b)). This includes complex geometries with
oil near rocks and seaweed. While it is relatively easy for the model to distinguish oil from the bottom, the
probability of detection can be increased as more information is known about specific oils and their
properties and entered into the model.

The five surveys yielded an average detection rate equal to 87 percent and an average false alarm rate equal
to 24 percent. False alarms resulted from seaweed, fine sand, and small inaccuracies in the positions and
dimensions of the objects in the trays.

Data processing was not done in real time but was done later on a separate computer. Because the rate of
data acquisition outstripped the rate of data processing, the lag time to produce the results accumulated. It
was estimated that the earliest processing results would lag real time by 8 minutes. Total processing time
for one square mile is a function of depth (or sonar altitude), which impacts sweep width. Total processing
time goes up exponentially as depth decreases. Total processing time would be 12 hours in 30 meters of
water and increase to 22 hours in 10 meters of water. (These figures assume a ping rate of 15s-1 and a vessel
speed of 6 knots.) While not done in real-time, the data transfer and calculations were completed for the
entire test section in less than one day for the 400 kHz runs. Additional tests were done at 200 kHz. In an
attempt to demonstrate coverage capability, a slow-ping run at 400 kHz that used 1 ping/second at a tow
speed of 0.5 knots was conducted. This is equivalent to using 10 pings/second at 5 knots. The additional
data are not included in this report but were used to estimate coverage capability.

RESON estimates that the detection processing time for a 1 square mile survey at a depth of 30 m can be
made in 12 hours. At shallower depths the swath will be smaller, thus requiring more runs over the area. At
deeper depths, the swath will cover a larger area, and the processing time for a 1 square mile survey will be
reduced at the cost of poorer resolution. The detection processing software currently uses raw beam-formed
signals as input. This has been done in order to ensure full control over all the stages of the computations.
When the input signals are replaced by “snippet” data, which are the type of data that only originate from

24


the region near the seabed, the computation time is expected to decrease significantly. Lag time may also be
decreased by increasing the number of data processors or computer speed.

There appear to be some discrepancies for left and right beams when observing the same targets because of
the geometry differences. This may require additional overlaps when performing actual searches.

(a) (b)
Figure 18. RESON (a) layout of trays and (b) sample results. Dotted red line in (a) is instrument centerline.
Table 3. Detections and missing detections for RESON position #3.
True area Patch Estimated area Missing area
Oil Patch (m2) No. (m2) (m2)
T1C (Tesoro) 0.28 11 0.28 0.01
T2A (Tesoro) 0.37 23 0.24 0.13
T3C (Scattered oil) 0.07 1,2,3 0.4 -0.33
T3D (Tesoro) 0.86 5 1.29 -0.43
T4A (#6 Oil mixed into stone ) 0.73 22 0.08 0.66
T5B (#6 Oil) 0.29 4 0.21 0.08
T6A (#6 Oil) 1.67 19 1.61 0.06
T8A (Tesoro) 0.66 20 0.76 -0.1
T9D (# 6 Oil) 0.93 7 0.46 0.47
T10A (# 6 Oil) 0.29 21 0.15 0.14
Total 6.15 5.47 .68

3.2.3 Other Considerations

3.2.3.1 Areal Coverage, Processing Time
There are trade-offs to consider for bottom coverage, speed of tow, and depth. To ensure some overlap of
coverage of the bottom, the altitude of the transponder above the sea bed must be greater than 10 meters at a
tow speed of 6 knots. Slower speeds or deeper water will increase the overlap of each ping. The amount of
time needed to perform the processing is about double the data acquisition time at 100 meters but almost
four times at 20 meters (see Figure 19). The requirement for surveying and processing for 1 square mile in

25


12 hours is met at a transponder altitude of 30 meters. The operators must understand the size of a patch of
oil that is of interest (amount recoverable, amount that could be toxic to environment, etc.) and make
decisions about the coverage, processing time, and resolution.

Figure 19. Estimated survey and processing time for a one square mile survey as a function of depth.

3.2.3.2 Summary
Methods are needed to reduce the processing time. Several options are available including reducing overlap
or using a range-gate type of analysis to reduce the amount of data collected. Creation of a library of oil
response characteristics and typical returns for various bottom types could enhance oil detection. Likewise
additional “seaweed” detection algorithms could reduce false alarms. Data from the 200 kHz tests need to
be analyzed to determine if this is useful in finding buried oil.

3.3 EIC Laboratories Fluorescence Polarization (FP)
3.3.1 EIC System Modifications

Based on the results of the Phase I development and testing, an autonomous, compact, underwater FP
prototype instrument was designed and fabricated in Phase II. This instrument is about 20 inches (0.51 m)
long and weighs about 16 pounds (Figure 20(a)). The altimeter sonar is the black rectangular piece attached
on the outside of the cylinder. The FP instrument prototype was designed to be compact, remotely operated,
and housed in a waterproof cylindrical housing so that it could be easily configured and deployed with
different types of platforms such as towed bodies, ROVs, AUVs, and manned submersibles.

26


(a) (b)
Figure 20. EIC FP probe (a) close-up and (b) in the test tank at OHMSETT.

The main components of the probe are as described in Section 2.4.1 and did not change from Phase I.
Detection of the two FP components is done with two fiber optically coupled photomultiplier tubes (PMT)
incorporating bandpass filters (15 nm bandwidth) centered at 589 nm. Data acquisition is performed
through the embedded computer via software developed by EIC. The software in the embedded computer
allows the FP instrument to be controlled remotely or to perform the detection in an automated fashion,
including GPS tagging of the data acquired. A remote computer aboard the survey vessel can be used to
control the operation of the FP instrument via serial communication. Instrument functions such as data
acquisition, data display with GPS coordinates, instrument parameter inputs, and data logging can be
performed remotely. The remote software both records and displays the raw fluorescence signals from the
two PMT channels and also calculates and displays the polarization values. In addition, the software allows
a GPS signal to be recorded along with the FP signals to allow georeferencing of FP data. The software
allows the operator to change the PMTs data acquisition time, balance the response of the two signals from
the two emission legs of the FP probe, and obtain the bias of the two PMTs.

3.3.2 EIC Results

During testing of the FP probe in the test tank at OHMSETT, the FP instrument was attached to one end of a
6-foot long, 1-inch diameter aluminum extension rod (Figure 20(b)), which allowed positioning of the FP
instrument at a given depth in the test tank. The extension rod was attached to a metal flange that bolted to
the bottom of the bridge tow bar. A GPS unit was mounted at the above-water end of the extension rod.

Grid scans of the test trays containing the oil targets were performed. Grid scans were started at one corner
of the test bed and ended at the opposite corner, so that the first line and last line ends at the edges of the test
bed. The tow bridge was moved from north to south, and at the end of each line the probe was translated by
a set distance. The tow speed was kept constant during each of the scans.

Although the individual scan lines were straight and parallel, the detection results were scattered because of
the accuracy of the GPS readings (see Figure 21 for sample results). The readings were taken about
0.3 meters (1 foot) apart, but the accuracy of the GPS was about 1 meter. The scatter indicates that some of
the track lines crossed when in reality they did not. The GPS direction of travel also indicated that the
instrument doubled back when it did not.

27


The results indicate that the FP instrument is capable of accurately detecting heavy oil in real time. Each of
the oil targets in the test platforms showed significant fluorescence polarization signals and could easily be
distinguished from the FP signal of the surrounding background. Figure 22 shows the FP results stretched
and superimposed on the complete tray layout. Some of the problems with plotting the location of the oil
are GPS position errors, but the reasons for missing some of the targets are not known. It is probably a
combination of the GPS movement and trying to stretch the data to match the overall size of the schematic
of the tray layout.

Figure 21. Two views of EIC FP sample results.

During the Phase II tests, bright sunlight, which did not occur during the Phase I testing, caused problems
for the FP detection. Although some fluorescence was detected, sunlight saturated the input. It appears that
there are several ways to reduce the external light. Solutions to minimizing the solar background
interference such as spatial filtering and modulation detection schemes were investigated. The most
promising is to modulate the laser and look for the returned fluorescence that will also be modulated.

Option 1, using a pinhole aperture to perform spatial filtering, could only reduce the background light by a
factor of four, which is still not enough to minimize the solar effect. The other approach was to modulate
the laser excitation at a specific frequency and set the detection for that frequency. When a unit so modified
was evaluated in bright sunlight with Tesoro oil in a parking lot, the return fluorescent signal was also
modulated (see Figure 23). The modulation technique still needs to be verified in an actual deployment

28


Figure 22. Superimposed images of tray set-up and FP results.

Figure 23. Results of modulation test conducted in bright sunlight showing output and return signals.

3.3.3 Other Considerations

The use of fluorescence polarization increases the usefulness of lasers over standard fluorescence. With the
addition of the modulation process, the amount of false alarms is greatly reduced. The problem of turbidity
may still limit the use of this system.

29


The areal coverage of a single sensor is limited. Indications are that several units could be mounted in an
array, but the footprint would still be limited by the small footprint of each. 100 percent coverage will be
virtually impossible. This system could probably not be used to find a small amount of oil. The user must
consider tradeoffs between coverage and resolution.

3.4 Tests of Opportunity
Three detection vendors came to OHMSETT using their own funds in order to take advantage of the test
setup before it was dismantled. The results are discussed below.

3.4.1 BioSonics

3.4.1.1 BioSonics Description
This company tested their DT-X Digital Scientific Echosounder (Figure 24), a unit equipped with two single
beam transducers (200 kHZ and 420 kHZ) that is usually used to classify substrate (sub-bottom) or
submerged vegetation. It has a very narrow, 6o beam width and weighs about 20 pounds. It is normally
connected to a GPS system but was not for this test.

Figure 24. BioSonics DT-X Digital Scientific Echosounder sensor.

Approximately 93 acoustic datasets were collected from the test site over a two-day period. These were
primarily collected as linear transects across the test site, the same as the scan lines of the other systems.
The transects were made at known locations and at measured speeds, providing the ability to estimate
position based on the timing of the acoustic samples.

3.4.1.2 BioSonics Test Results
The system was successful in classifying the oil as a different kind of material in real-time (see

Figure 25 and Figure 26 for sample results of one pass along the length of the tray). It was also able to
differentiate the four types of bottom material that were used. This differentiation was made possible by
collecting sufficient data to develop an on-site reference library so that the same bottom material could be
recognized and designated as not of interest during a search for oil. BioSonics estimates that when working
with an unknown bottom, it would take approximately 30 minutes to characterize the bottom type(s), prior
to beginning the search for the submerged oil.

30


3.4.1.3 Other Considerations
It is unlikely that the methods used here will reliably detect and discriminate small quantities of oil in
association with vegetation and complex rock environments. Oil patches thicker than those tested
(1-2 inches) would probably be easier to detect. Further, there is a need for additional development of the
reference library to include collection and analysis of acoustic data from other oil samples, in larger
quantities, to continue development of a reliable tool.

This technology is already being used in many different environments by natural resource managers to
classify substrate and measure submerged vegetation. These capabilities were developed with extensive
testing and ground-truth trials over time. Presumably the same approach could be used for submerged oil.

Figure 25. BioSonics sample echogram.

31


Figure 26. BioSonics sample analysis results.

3.4.2 CodaOctopus

3.4.2.1 CodaOctopus Description
The CodaOctopus EchoScope4D Imaging sonar, operating at 375 kHz, was used for these tests (Figure 27).
This is the same system that the USCG is evaluating for other uses. It generates 128 by 128 beams in a 50
by 50 degree grid. It weighs about 45 pounds. The range of the sonar is from 1 meter to about 100 meters,
depending on target strength (TS). The range resolution of the standard unit is 4 cm. It is typically
deployed with a navigation system so that position and orientation are known. Like the RESON system, it
uses return signal strength to differentiate between rocks, bottom, and oil.

32


Figure 27. CodaOctopus EchoScope 4D transponder.

3.4.2.2 CodaOctopus Test Results
At almost all angles and frequencies, the contrast between oil and sand was about 15 dB. Sample results
are shown in Figure 28: (a) shaded mosaic showing height (bathymetry); (b) raw data for area containing
oil, rock, and seaweed; and (c) data after Underwater Inspection System (UIS) software processing. The
company recommends that typical seabed types and heavy oil types be analyzed and their respective
intensity returns (TS) measured as functions of frequency and angle of incidence. The TS database can then
be used by a calibrated EchoScope to determine the class of seabed. A calibrated EchoScope used for this
application will give calibrated TS data compensated for projector beam pattern.

With online navigation included, an EchoScope survey will typically be carried out at 3 to 5 knots and
therefore cover large areas efficiently. (This equates to 5 to 8 hours for complete coverage of 1 nautical
mile by 1 nautical mile area at 100 meter swath width and 20 percent overlap.) Mosaics will be built on-
the-fly using the UIS software. The UIS mosaic software uses averaging in geo-referenced cells – a
technique that improves the signal-to-noise ratio significantly. As the data are instantaneous 3D, a vessel
can move to an object of interest and obtain better data and even use it to position tools or divers in zero
visibility water.

3.4.2.3 Other Considerations
CodaOctopus recommends that a calibrated instrument be used in any further tests. This will compensate
for combined receiver and projector sensitivity and result in fully normalized image intensity. The intensity
of returns is the most important parameter in distinguishing between sand and oil. None of the fully
calibrated echoscope heads were available at the time of the OHMSETT tests.

33


(a)

(b)

(c)

Figure 28. CodaOctopus sample results. (a) shaded mosaic showing bathymetry (color palette wraps around
so red to red spans 20 cm elevation); (b) raw data of area containing oil, rock, and seaweed; and
(c) data after UIS software processing (blue is oil).

34


3.4.3 SRI International

This company was funded by MMS to evaluate real-time mass spectrometry that has been used to map the
field of a sewage outfall among other things. This system uses a different approach than the WHOI system.
The cylindrical vessel is normally mounted in a remotely operated vehicle or autonomous vehicle. As seen
through the windows of the OHMSETT tank (Figure 29), it was strapped into its maintenance stand for this
use. No oil was detected in the tank but the system was later able to detect low-level components in a barrel
in a high bay area at OHMSETT. This is similar to the experience that WHOI had with their system. SRI
also indicated that the development of a pressurized system might allow the detection of heavier, less
volatile compounds.

Figure 29. SRI International mass spectrometer.

3.5 Phase II Summary
3.5.1 Test Set-up Considerations

The test setup was as realistic as it could be given that the oil was contained. The bright backscatter did
affect systems. The water was significantly cold (30-31 oF and under ice) but temperature did not seem to
affect systems once they had warmed.

3.5.2 Prototype Test Results

Table 4 provides a summary of the prototype test results compared to the BAA performance requirements
listed in Section 3.1.

35


Table 4. Phase II test results.

SRI
Requirement RESON EIC BioSonics CodaOctopus
International
Identification of Yes Yes Yes with some Yes with No
heavy oil on sea limitations additional work
floor (80%
certainty)
Ability to detect oil Yes Yes yes Yes No
on the sea floor
from at least 1
meter away
Real time data No Yes Yes Yes Yes

Able to provide Yes (some false Yes, although Yes Yes Yes
data for all sea alarms from turbidity results
floor conditions seaweed; oil in reduced
under gravel not instrument
detected) sensitivity
Search a one At depths over Yes, but limited Yes Yes Undetermined
square mile area 100 ft coverage
in a 12-hour shift
Water currents of Yes, if mounted Yes, if mounted Yes, if mounted Yes, if mounted Yes, if mounted
up to 1.5 knots on boat/ROV on boat/ROV on boat/ROV on boat/ROV on boat/ROV
that can handle that can handle that can handle that can handle that can handle
this this this this this
Operate in up to 5 Yes, if mounted Yes, if mounted Yes, if mounted Yes, if mounted Yes, if mounted
foot seas on boat/ROV on boat/ROV on boat/ROV on boat/ROV on boat/ROV
that can handle that can handle that can handle that can handle that can handle
this this this this this
Operable during Yes Strong solar Yes Yes Yes
the day and night background
originally
reduced
performance, but
modifications
eliminated the
problem.
Able to be set up Yes Yes Yes Yes Yes
within 6 hours
Easily deployable Yes Yes Yes Yes Yes
and transportable
Capable of being Yes Yes Yes Yes Not clear
deployed from a
vessel of
opportunity and a
variety of other
platforms

36


The evaluation of prototype systems was accomplished and several additional systems were tested. All
systems need further development to make them into practical tools. In addition, the complexity of the
environment was limited and the instruments should be evaluated further in the field.

In General:

• The methods were successful in detecting oil in a benign environment.
• There is no single method that can cover 100 percent of the area with no false alarms.
• Resolution of results is still and issue.
o Easier if oil stays together.
o Random hits need to be correlated.
• Use of techniques in turbid water and very soft bottom (such as rivers and harbors) is also an issue.
• Additional research needed for real-time mass spectrometry systems.

4 CONCLUSIONS
The technologies presented here represent an improvement over the existing ad-hoc methods. Although
these systems have not been tested in the difficult harsh environments of low visibility, currents, and
complex bottoms, they may be immediately useful in some situations, which could reduce the amount of
effort and increase reliability of oil detection on the bottom or in the water column. Additional work needs
to be done for all systems before they can be considered fully operational and it is hoped that the vendors
can find additional funding sources.

The multi-beam and imaging sonars appear to be the best sensors to conduct wide area detection. Some of
the signal return issues, which cause false positive detections for the low grazing angles of common side-
scan sonar, are reduced in the systems tested. Most systems should be able to automatically detect large
clumps of oil, but the resolution for widely dispersed product is still not complete. Spill responders should
ensure that detection equipment has some type of processing software to interpret raw sensor data. This will
ensure timely processing and require minimal training for response personnel. The sooner that a system is
deployed before the oil breaks up, the better will be the chance that detection will occur.

The laser systems and smaller beam sonars may be better suited as a follow-up to the wide scan areas.
These should provide better resolution and should be able to calculate general thickness which could
provide some information about the amount of oil. The narrow areas covered could introduce resolution
issues especially for widely scattered oil. On the other hand, the narrow area covered could be
advantageous for guiding recovery efforts.

The real-time mass spectrometry systems should be evaluated for neutrally buoyant oil detection in the
water column. For some spills, especially those in rough waves or fast moving currents, these instruments
may be useful in tracking plumes. This would be especially useful for municipalities and power plants that
use the water for cooling.

Positioning of the systems should be evaluated according to needs. In good visibility, the oil can be located
within 5-10 meters that will permit divers or other operators to find it for recovery. For limited visibility
and under special circumstances, underwater navigation systems, similar to the WHOI system, should be
utilized for better accuracy.

37


5 RECOMMENDATION
The companies that had systems evaluated here and other related technologies should be encouraged to
further develop the systems. Regulations require that vessel and facility plans contain response techniques
and that the response organization has the equipment to respond to submerged oils. As progress is made on
their development, regulations should also ensure that this capability is available for heavy oils (Type V)
and those that could sink if exposed to the environment. Research funded by the Coastal Response
Research Center (CRRC) at the University of New Hampshire is being done in Canada to further document
what conditions are needed to cause oil to sink. These research results should provide responders with
better information about oil behavior. Better models for submerged oil should also be developed that can be
used to predict behavior and help further define detection and recovery techniques.

The use of this equipment by a Federal On-scene Coordinator (FOSC) is limited at this time due to the level
of development. Guidance is contained in the Appendixes that provide information about the specific
technologies tested. A decision-tool and recommendations for FOSC use is contained in Appendix E.

These types of systems should be integrated into recovery systems along with visual detection methods for
clearer water. The USCG RDC has begun a project to develop full recovery systems that should be
completed by 2012. It is hoped that companies that are in the field of detection will combine with other
hardware manufacturers to develop systems that can be:

• Easily deployed in response to sinking oil.
• Be readily available, either through taking off the shelf or having a clear plan of integration.
• Be able to clearly find the oil and immediately recover it before it has a chance of being disturbed.

38


6 REFERENCES
Barbini, R. F., Colao, R. Fantoni, C. Ferrante, A. Palucci and S. Ribezzo. (2000). “Development of
LIDAR Fluorosensor Payload for Submarine Operation,” in Proceedings of EARSeL-SIG
Workshop LIDAR, Dresden, FRG, pages 39-45, June 2000.

Bello, J.M. (2009). “Fluorescence Polarization Detection of Heavy Oil on the Sea Floor –
OHMSETT Testing Results Conducted on 1/19/2009 to 1/21/2009,” EIC Laboratories, Inc.,
submitted to USCG R&DC, February 19, 2009.

Bello, J.M. & Clauson, S. (2008). “Fluorescence Polarization Detection of Heavy Oil on the Sea
Floor – OHMSETT Testing Results,” EIC Laboratories, Inc., submitted to USCG R&DC,
January 2, 2008.

BioSonics, Inc. (2009). “BioSonics Report on OHMSETT Test Opportunity; Dates 1/26/2009 and
1/27/2009,” submitted to USCG R&DC, February, 2009.

Brown, C. E., Fingas, M. F., Gamble, R. L. & Mysliski, G. E. (2002). “The Remote Detection of
Submerged Oil”, 3rd R&D Forum on High Density Oil Spill Response, March 11-13, 2002.

Brown, C.E., Fingas, M.F. & Marois, R. (2006). “Oil Spill Remote Sensing: Flights Around
Vancouver Island”, in Proceedings of the Twenty-ninth Arctic Marine Oilspill Program
Technical Seminar, Environment Canada, Ottawa, ON, pp. 921-929, 2006

Brown, C.E., Fingas, M.F., & Marois, R. (2004). “Oil Spill Remote Sensing: Laser Fluorosensor
Demonstration Flights off the East Coast of Canada”, in Proceedings of the Twenty-Seventh
Arctic Marine Oilspill Program Technical Seminar, Environment Canada, Ottawa, ON, pp.
317-334, 2004.

Cabioc’h, F. (2002). “Sunken Hydrocarbons and Chemical Products: Possible Responses and
Available Methods”, 3rd R&D Forum on High Density Oil Spill Response, Brest, France,
March 11-13, 2002.

Camilli, R. & Bingham, B. (2008). “Detection of Heavy Oil on the Seafloor – Phase I Final
Report,” Woods Hole Oceanographic Istitution, submitted to USCG R&DC, March 11, 2008.

Castle, R.W., Wehrenberg, F., Bartlett, J. &. Nuckols, J. (1995). “Heavy Oil Spills: Out of Sight,
Out of Mind”, 1995 International Oil Spill Conference, pages 565-571, 1995.

Coastal Response Research Center (CRRC). (2007). “Submerged Oil – State of the Practice and
Research Needs,” Durham, NH, July, 2007.

CodaOctopus R&D. (2009). “ EchoScope Trials at OHMSETT January 28-29, 2009.” Submitted
to USCG R&D Center, February 9, 2009.

Committee on Marine Transportation of Heavy Oils, National Research Council. (1999). “Spills of
Nonfloating Oils: Risk and Response,” National Academy of Sciences, 1999.

39


Elliott, J. (2005). “An Analysis of Underwater Oil Recovery Techniques, 2005 International Oil
Spill Conference, 11 pages, 2005.

Eriksen, M.T. (2009). “Prototype Demonstration Report – HSCG32-07-C-R00030, Deliverable 11,”
RESON, Inc., submitted to USCG R&DC, May 21, 2009.

Eriksen, M.T. & Maillard, E. (2007). “Proof of Concept Demonstration Report – HSCG32-07-C-
R00030, Deliverable 5,” RESON, Inc., submitted to USCG R&DC, December 21, 2007.

Fant, J.W. & Hansen, K.A. (2005). “U.S. Coast Guard Oil Spill Remote Sensing: Preliminary Laser
Fluorosensor Studies”, in Proceedings of the Twenty-eighth Arctic Marine Oilspill Program
Technical Seminar, Environment Canada, Ottawa, ON, pp. 861-872, 2005.

Fant, J.W. & Hansen, K.A. (2006). ”U.S. Coast Guard Laser Fluorosensor Testing”, in Proceedings
of the Twenty-ninth Arctic Marine Oilspill Program Technical Seminar, Environment
Canada, Ottawa, ON, pp. 951-964, 2006.

Hansen, K.A. (2009). “Research Efforts for Detection and Recovery of Submerged Oils,” in
Proceedings of the Thirty-second Arctic Marine Oilspill Program Technical Seminar,
Environment Canada, Ottawa, ON, 2009 (in preparation).

Hansen, K.A, Bello, J., Clauson, S., Eriksen, M.T., Maillard, E., Camilli, R., Bingham, B., Morris, J.
& Luey, P.J. (2008). “Preliminary Results for Oil on the Bottom Detection Technologies,”
in Proceedings of the Thirty-first Arctic Marine Oilspill Program Technical Seminar,
Environment Canada, Ottawa, ON, 2008.

Michel, J. (2006). “Assessment and Recovery of Submerged Oil: current State Analysis,” prepared
for US Coast Guard Research and Development Center, June, 2006.

Michel, J. & Galt, J.A. (1995). “Conditions Under Which Floating Slicks Can Sink in Marine
Settings,” 1995 International Oil Spill Conference, API Publication No. 4620, American
Petroleum Institute, Washington, D.C., pp 573-576, 1995.

Morris, J. & Luey, P.J. (2008). “Modification and Use of the Laser Line Scan System as an Optical
Tool for the Detection of Heavy Oil on the Seafloor – Proof of Concept Demonstration
Report,” Science Applications International Corporation, submitted to USCG R&DC,
February 1, 2008.

Parthiot, F., De Nanteuil, E., Merlin, F., Zerr, B., Guedes, Y., Lurtin, X., Augustin J.M.,
Cervenka, P., Marchal, J., Sessarego, J.P., & Hansen, R.K. (2004). “Sonar Detection and
Monitoring of Sunken Heavy Fuel Oil on the Seafloor,” Proceedings of the Interspill 2004
Conference, Trondheim, Norway, June 14-17, 2004.

Parthiot, F. (2002). “Monitoring of Sunken Fuel Oils,” 3rd R&D Forum on High Density Oil Spill
Response, March 11-13, 2002.

Schnitz, P. R. & Wolf, M. A. (2001). “Nonfloating Oil Spill Response Planning,” 2001
International Oil Spill Conference, pp 1307-1311, 2001.

40


Short, R.T. (2009). “Demonstration of Detection and Location of Submerged Oil Using In Situ
Mass Spectrometry,” SRI International, submitted to USCG R&DC, February, 2009.

Usher, D. (2008). “The Use of Manned Submersible Units to Accomplish Submerged Oil
Recovery,” International Oil Spill Conference, Savannah, GA, May, 2008, pages 1289-1291.

Wendelboe, G., Fonseca, L., Eriksen, M., Hvidbak, F., and Mutschler, M. (2009). “Detection of
Heavy Oil on the Seabed by Application of a 400 kHz Multi-beam Echo Sounder,” in
Proceedings of the Thirty-second Arctic Marine Oilspill Program Technical Seminar,
Environment Canada, Ottawa, ON, 2009 (in preparation).

41



42


APPENDIX A. OHMSETT TEST FACILITY
The Oil and Hazardous Material Simulated Environmental Test Tank (OHMSETT), now called The
National Oil Spill Response Test Tank Facility (www.ohmsett.com), is the only facility where full-scale oil
spill response equipment testing, research, and training can be conducted in a marine environment with oil
under controlled environmental conditions (i.e., waves, temperature, oil types). The facility provides an
environmentally safe place to conduct objective testing and to develop devices and techniques for the
control of oil and hazardous material spills. OHMSETT’s mission is to increase oil spill response capability
through independent and objective performance testing of equipment, providing realistic training to
response personnel, and improving technologies through research and development.

OHMSETT is located at the Naval Weapons Station Earle Waterfront in Leonardo, New Jersey
(approximately one hour south of New York City). It is maintained and operated by the Department of
Interior Minerals Management Service (MMS) through a contract with MAR, Incorporated of Rockville,
Maryland. OHMSETT’s above ground concrete test tank is one of the largest of its kind, measuring 203 m
long by 20 m wide by 3.4 m deep. The tank is filled with 2.6 million gallons of crystal clear saltwater.

OHMSETT has a mechanically operated control bridge that spans the width of the tank and traverses the
tank’s length; two stand-alone work bridges can be stationary or rigidly attached to the mobile control
bridge. The OHMSETT test tank allows testing of full-scale equipment. The tank’s wave generator creates
realistic sea environments, while state-of-the-art data collection and video systems record test results. The
facility has proven to be ideal for testing equipment, evaluating acquisition options, and validating research
findings.

Public and private sector entities are invited to contract the use of OHMSETT as a research center to test oil
spill containment/clean-up equipment and techniques, to test new designs in response equipment, and to
conduct training with actual oil spill response technologies.

Features & Capabilities

• A main towing bridge capable of towing test equipment at speeds up to 6.5 knots
• An auxiliary bridge oil recovery system to quantify skimmer recovery rates
• A wave generator capable of simulating regular waves up to one meter in height, as well as a simulated
harbor chop
• A movable, wave-damping artificial beach
• An oil distribution and recovery system that can handle heavy, viscous oils and emulsions
• A control tower with a fully-computerized 32-channel data collection system as well as above-and
below-water video
• A centrifuge system to recover and recycle test oil
• Blending tanks with a water and oil distribution system to produce custom oil/water emulsions for
testing
• A filtration and oil/water separator system
• An electrolytic chlorinator to control biological activity
• Permanent and mobile storage tanks that can hold over 227,000 liters of test fluids
• A vacuum bridge to clean the bottom of the tank
• Staging and shop area for special fabrication

UNCLASSIFIED | CG-926 RDC | K. Hansen, et al. | CG-5332 | June 2009
A-1



A-2


APPENDIX B. ACOUSTIC DETECTION OF HEAVY OIL
Acoustic techniques for sea floor mapping are widespread and they could potentially provide relatively
rapid coverage for heavy oil detection. Side-scan sonar mapping systems are normally interfaced with a
Global Positioning System (GPS) and hydrographic mapping software to generate maps of seafloor features.
These systems can provide relatively rapid coverage, and are primarily useful for identifying areas of natural
collection for the sunken oil. Multi-beam sonar systems can potentially be used to differentiate the oil from
the sea bottom by sensing the contrast in roughness.

B.1 Acoustic Detection Mechanism
Sonar mapping techniques rely on acoustic sounding principles, specifically on the differential density and
sound speeds of water compared to those of sediments and the seafloor. Oil and oil-sediment mixtures will
differ from sediments in a similar manner and thus should be recognizable.

A sonar device contains a transducer that converts the electrical signal from a transmitter within the
transducer into an acoustic pulse, and transmits that energy into the water. In reciprocal fashion, the
transducer receives acoustic echoes (from targets on the bottom) and converts them to electrical signals.
The pulse of energy travels through the water at a speed of approximately 1500 m/sec, and depends on
pressure (therefore depth), temperature (a change of 1 °C ~ 4 m/s), and salinity (a change of 1% ~ 1 m/s).
When the acoustic pulse encounters an object, some of the energy (i.e. an echo) is reflected back to the
transducer (this reflected energy is also called backscatter (see Figure B-1)) and some continues forward.

Figure B-1. Acoustic backscattering from the sea floor.

The amount of energy that is forward versus backscattered from the seabed is a result of impedance changes
at the water/sediment interface, the roughness of the water/sediment interface, and the sediment volume
heterogeneities. These heterogeneities include objects buried in the sediment that reflect energy that is
originally forward scattered rather than backscattered. The acoustic impedance of a material depends on its
density, viscosity, and the speed of sound in the material. Impedance contrast and roughness governs the
scattering mechanisms at the water/sediment interface. Sediment heterogeneities govern the scattering

B-1


mechanisms within the sediment volume. The intensity of the backscattered signal also depends on the
acoustic frequency, and on the grazing angle θg (with respect to the seabed plane) of the incoming field.
The return echo signal also depends on the equipment parameters (i.e. frequency, transducer’s beamwidth,
and others).

B.2 Data Processing
In order for a sonar system to serve as an oil detection tool, the data received from the sonar must be
processed and interpreted. There are different types of processing used, usually based on the specific sonar
equipment and target strength differences between the oil and the bottom. Vendors generally design
software specifically to work with their systems.

Returns come in and an image is generated based on user-defined thresholds. That image then undergoes
image processing to determine what is oil and what is not. If the signal level exceeds a user-selected
threshold level, a mark appears on the echogram. The distance from the centerline to the mark is
proportional to the travel time for the pulse to travel from the transducer to the target and back. Since the
velocity of sound in water is known, range (distance from the transducer) can be calculated from this travel
time. By collecting the echoes from many consecutive transmissions, the time in the acoustic beam, the
change in range and the direction of travel of targets can be determined.

B.3 Advantages
There are some advantages to using an acoustic seafloor classification system. Appropriate systems are
commonly available at relatively low cost. They are portable, so they can be deployed on boats of
opportunity, and they have minimal power requirements. Due to their ping rates, they are also capable of
collecting data quickly.

Multi-beam sonars can be designed to operate at different frequencies. Higher frequencies give better
angular resolution. Lower frequencies provide lower resolution but offer additional range.

One advantage is that these systems are relatively common and thus their properties and peculiarities are
well known.

B.4 Limitations
Sonar images need to be interpreted. Since oil spills do not occur frequently, vendors have not designed
specific software to provide rapid delineation of oil patches on the bottom. Analysis has taken more than a
day to identify oil on the bottom and this usually has to be confirmed. This is also usually complicated by
biological interference (plants (particularly kelp) and animals) that can confound the signals. This time
delay is not useful for many spills when the oil is still on the move.

Someone trained to interpret the image can generally do so rapidly. A couple of sediment samples, which
could be taken during the survey, should be enough to allow accurate interpretation of the sonar data. They
could also be augmented with video or imagery collection. The building of a library of oil-on-sediment
returns would probably allow for more rapid interpretation. Again, this would require someone trained in
oil recognition.

B-2


Systems with narrow swath width make continuous coverage of the seafloor difficult, and their acoustic
“footprint” is relatively small and dependent on depth. These types of systems would be useful for
confirming the presence of oil once its general location is suspected.

Some sonar systems may be unable to distinguish between oiled sediments and underlying sediments
because of their acoustic similarity. This is especially true in rivers and harbors. Therefore, sampling or in
situ observations are necessary to confirm the maps. Because the sonar is reflecting the roughness or
smoothness of the seafloor, oil that is covered by sediments may be missed in a sonar survey. Changes in
salinity of the water will have a direct effect on the propagation of the Sonar’s acoustic signal in the water.

B-3



B-4


APPENDIX C. OPTICAL/FLUORESCENT DETECTION OF HEAVY OIL
Visual observations (by aircraft, ship, diver, or camera/television) have been the principal methods of
locating and tracking submerged oil. Airborne photography and visual-based systems, which are widely
available and can rapidly survey large areas, are frequently used to locate submerged oil. The performance
of these systems is limited by water clarity and depth, the quantity of oil, and the characteristics of bottom
sediment. Figure C-1 shows the transmission distance for the visible spectrum for various water types.
Given the possibility of misidentifying natural materials (seaweed, seagrass beds) as oil, in situ observations
are always required to validate airborne assessments. Direct observations can also be performed by divers
within safe depth restrictions and visibility limits. Observations by underwater cameras, either operated by
divers or deployed from ships, can also be used to locate submerged oil. These visual methods must
generally be confirmed by sampling and have relatively limited coverage. During R&D Research, Hansen
and Fant (2006) detected fluorescence from a target about 40 ft away in clear water using an airborne laser.

Figure C-1. Graphical representation of light transmission in water. Water color, turbidity, and other factors
impact the actual attenuation of transmitted light, as well as any corresponding reflectance or
fluorescence.

C.1 Laser Fluorescence Mechanism
An alternative detection technique using the visible light spectrum is fluorescence spectroscopy. Laser
fluorosensors are active sensors that rely on the fact that certain compounds in petroleum oils absorb
ultraviolet light and become electronically excited. This excitation is removed by the process of
fluorescence emission, primarily in the visible region of the spectrum.

C-1


Crude and refined oil products are primarily composed of saturated and aromatic hydrocarbons, resins, and
asphaltenes. Polyaromatic hydrocarbons (PAHs) are chemical compounds comprised of fused rings
containing strong unsaturated bonds. Due to the structural arrangement of PAHs, they tend to fluoresce in
response to light energy. Through the process of fluorescence, the light energy that was absorbed by the oil-
based compound is released back to the ambient environment, returning the molecules to their original
ground state. Despite a difference in molecular structure, alkanes (saturated hydrocarbons) will also
fluoresce when exposed to a focused light source.

Various intensities and wavelengths of light can be used to excite PAH and alkane molecules into a state of
fluorescence. However, numerous studies have shown that high-energy ultraviolet (UV) light in the
wavelength range 200 nm to 400 nm is the most effective source of excitation, yielding the strongest
fluorescent emission. PAH compounds responding to UV tend to fluoresce quite distinctly, emitting
photons in the visible light wavelength range (400–600 nm: violet to orange), with specific wavelengths of
emission serving to identify the types of PAH compounds present. Similarly, alkane molecules will
fluoresce in response to UV, but they do so at a lower wavelength outside the visible light spectrum (UV-A
bandwidth; 320-400 nm), making them less viable indicators of oil.

C.2 Fluorescence Polarization
In addition to oils there are other potential fluorophors (compounds that fluoresce) in the marine
environment, such as chlorophyll from algae and seaweed. One technique developed to distinguish between
oil and other fluorophors is fluorescence polarization (FP). FP measurements are based on the assessment
of the rotational motions of species. FP can be considered a competition between the molecular motion and
the lifetime of fluorophors in solution. If linear polarized light is used to excite an ensemble of fluorophors,
only those fluorophors aligned with the plane of polarization will be excited. The FP depends on the
fluorescence lifetime and the rotational correlation time (θ). The rotational correlation time is given by
θ = ηV/kT, where k is the Boltzman constant, T is the absolute temperature, η is the viscosity, and V is the
molecular volume. Thus, for viscous compounds, fluorescence polarization will be observed. Other
fluorophors in the marine environment will not exhibit fluorescence polarization since they are a less
viscous medium and will not be conducive to fluorescence polarization.

C.3 Advantages
Fluorescence based methods have several advantages, including they are non-contact (e.g., can be deployed
with fiber optic probes for remote sensing), have high sensitivity to the presence of aromatic hydrocarbons,
and can be easily miniaturized.

Fluorescence polarization (FP) enhances the selectivity of fluorescence by incorporating polarization into
the measurement technique. FP measurements are based on the assessment of the rotational motions of
species. In particular heavy oils, which are very viscous, will show significant fluorescence polarization
when excited with polarized light.

C-2


C.4 Limitations
In addition to naturally occurring fluorescence that can interfere with fluorescence based methods, ambient
or reflected light can affect results. For example during dockside tests of SAIC’s Laser Line Scan System
(LLSS), white test rays and white writing on polyethylene bags were visible as well as the oil. In the LLSS
OHMSETT tests, the paint of the tank reflected the laser light. It is not known whether elements in the
natural environment would also cause the same problem, but one would expect that a highly reflective clean
sand could result in similar reflection of the laser light. In addition, turbid water could have the opposite
effect by interrupting the signal such that oil cannot be detected.

C-3



C-4


APPENDIX D. CHEMICAL DETECTION AND IDENTIFICATION OF HEAVY
OIL
Direct sampling of the water column or seabed may be used to locate and map the movement of oil.
Sampling can be done by a vessel, a remote vehicle, or a diver (in shallow water). Sampling generally
becomes more difficult and time consuming as the water depth, current speed, and wave height increase. A
variety of sampling techniques are available, including grab sampling of water or sediments with subsequent
visual or chemical analysis, sorbent materials deployed on weighted lines or in traps, and core sampling of
the seabed sediments. Sampling is typically limited in scope and may not provide representative
observations of the impact area. Water-column and bottom trawls may be useful for selected spills because
they can cover larger areas. The effectiveness of sampling methods is strongly dependent on the
composition of the oil and oiled sediment and on environmental factors, such as current speed, water depth,
and substrate type.

Some companies have developed instruments designed to conduct in situ sampling and analysis of the water
column. These instruments perform a chemical analysis of the water to determine the presence of oil and
identify its components. In situ chemical analysis techniques include mass spectroscopy and ultraviolet
fluorometry.

D.1 Mass Spectroscopy
Mass spectrometry (MS) is an analytical technique that is used to identify unknown compounds, to quantify
known compounds, and to determine the structure and chemical properties of molecules. It does this by
ionizing the components to generate charged molecules and molecule fragments, and then measuring their
mass-to-charge ratio (m/z). In an MS procedure, a sample is introduced into the MS instrument and its
components undergo ionization through one of a variety of mechanisms (e.g., by impacting them with an
electron beam), resulting in the formation of charged particles (ions). The m/z of the particles can then be
calculated based on behavior of the ions as they pass through electric and magnetic fields generated by the
MS instrument.

D.2 Ultraviolet Fluorometry
Ultraviolet (UV) fluorometry employs a flow-through, fixed-wavelength UV fluorometer to measure and
map components of oil that can be induced to fluoresce. These are generally aromatic hydrocarbons. In situ
and towed fluorometric detection devices are widely available and routinely used to detect and map
petroleum leaks and spills. These systems may be mounted on buoys, boats, or remotely operated vehicles.
When mounted on boats and coordinated with GPS, they can provide maps of the subsurface oil
concentration field. They are restricted to making oil concentration measurements in the water column and
have a detection range from parts per billion to parts per million, depending on environmental conditions
and oil type. Given the three-dimensional nature of submerged oil plumes, mapping of subsurface oil
requires an extensive effort.

D-1


D.3 Advantages
Mass spectrometers and UV fluorometers are able to detect wide range of components and distinguish
between different chemical species. They are also able to provide real-time data.

D.4 Limitations
Detection and characterization of oil in the water column depends on the solubility of the oil in sea water. It
is not clear whether oil that has been submerged for more than 1-2 days will emit volatile compounds and
create a signature trail. It also means that the sensor may have to be very close to the bottom and be tightly
controlled which will limit the speed of the sensor through the water. Bottom type and organic growth may
further restrict the applicability of these systems.

D-2


APPENDIX E. RECOMMENDATIONS FOR FEDERAL ON-SCENE
COORDINATORS FOR OIL SUSPECTED TO BE ON THE SEA
BOTTOM
In responding to any oil spill, it is essential that the Federal On-scene Coordinator (FOSC) knows the
location, area coverage, and general physical condition of the oil to effectively deploy cleanup resources and
protect environmentally sensitive areas. Detecting and tracking oil beneath the surface is a particularly
challenging problem. For the purpose of this report and following Coastal Response Research Center
(2007), “submerged oil” describes any oil that is not floating at or near the surface. “Sunken oil” describes
the accumulation of bulk oil on the seafloor.

E.1 Fate of Spilled Oil and Oil-based Compounds
When oils are initially released into the marine or aquatic environment, a number of processes can affect the
slick. These include spreading, evaporation and oxidation, dispersion, dissolution, emulsification,
biodegradation, and sedimentation. Chemical make-up, density, and viscosity of the oil will have a large
impact on the resultant behavior of the spilled oil. Oil products with a specific gravity less than the
surrounding seawater at the time of release will tend to form a surface slick, while oils with a specific
gravity greater than that of the seawater will likely sink to the seafloor or suspend in the water column;
under both conditions, the oil can be transported by currents. Deposits of sunken oil are challenging to
detect, map, and recover following an oil spill. Methods of detection and mapping using existing techniques
are often inefficient and time consuming, involve labor intensive searches, and thus contribute to low
recovery volumes for these kinds of spills.

Oils and chemicals with similar physical properties and low solubility can make their way to the seabed
through a number of different mechanisms:

• The pollutant has an initial specific gravity already greater than that of seawater.
• The specific gravity of the pollutant becomes greater than seawater through the incorporation of
sediments either as a result of being stranded on sand shorelines and washed back into near-shore
waters or becoming entrained with high levels of suspended sand in breaking waves (either on the
beach or offshore bars).
• The oil sinks following a fire that not only consumes the lighter components but also results in
heavier pyrogenic products as a consequence of the high temperatures associated with the fire.
• The pollutant is injected directly into the seabed and sticks to it through mechanical adhesion.

Since 1991, there have been at least nine major spills that involved submerged oil. All of the past spills
where the oil submerged initially (without picking up sediment) were heavy, refined oil products or coal tar
oil that were denser than the receiving water. Most of the past spills where the oil initially floated then sank
were spills of heavy crude oils or heavy refined oil products that sank after picking up sand.

Regardless of whether the spilled oil exists as a surface slick or as a deposit at the sediment/water interface,
natural physical processes within the water column (surface waves, tidal currents, etc.), evaporation, and
dissolution will cause the spilled oil-product to weather and properties to change over time. Submerged oils
however, weather at much slower rates than floating or stranded oil. Higher density oil deposits or tarballs
on the seafloor are also affected by bottom current action and the incorporation of sediment grains into the

E-1


oil matrix. Exposure to near-bottom currents of significant magnitude may result in transport of the oil
along the bottom by tidal, river, or storm wave currents and continued incorporation of native ambient
sediment grains, as well as widespread dispersion from the original point of origin.

Submerged and sunken oil may move uncontrolled in the water column due to temperature changes,
currents, gain or loss of sediments, and wave action. The result of a spill of heavy oil that sinks to the sea
floor may therefore cause significant damage to the marine environment, recreational areas, sensitive
industrial installations, and property such as boats and docks.

E.2 Tracking and Mapping Submerged Oil
The appropriate method for tracking and mapping a particular spill depends on whether the oil is suspended
in the water column or deposited on the seabed and on the water depth and clarity. In general, visual and
photobathymetric techniques are restricted to water depths of 20 m or less and are suitable for both
suspended and deposited oil. Diver-based visual observations can only be used in low-current and small
wave areas with moderately clear water. Acoustic techniques, video observations, water-column and
bottom sampling, in situ detectors, and nets and trawls typically have no depth restrictions except that the
water must be deep enough for the instrument to be deployed and operated safely. They become more
difficult to operate, however, as the current speed and wave height increase. Measurements near the seabed
become more challenging as the topographic relief of the bottom increases and the bottom surface becomes
rougher. Fouling of instruments can be a serious issue.

Locating and identifying heavy oil are problems of growing concern as the use of heavy oil and related
slurry products becomes more prevalent. Despite the technological improvements that have been made in
identifying oils spills through surface slick detection, heavy oils with limited or no surface slick expression
remain challenging. In recent years, a number of spills such as the M/T Athos 1 and DBL-152 (Michel
2006) have been difficult to remediate because of poor estimates of subsurface spill volume and the inability
to track petroleum product migration (advection and dispersion on the seafloor and within the water
column). This inability to provide clear estimates of subsurface spill extent and movement persists because
of inadequate sensing technology. Experimental technologies such as airborne laser fluorescence show
promise in detecting aromatic hydrocarbons at water depths to a couple of meters. However, the
effectiveness of this technique rapidly deteriorates with increasing depth or water turbidity (Fant and
Hansen 2006). Other methodologies such as side-scan sonar have been periodically employed but proven
unreliable in detecting sunken oil (Michel 2006).

Present state-of-the-art techniques are generally slow, labor intensive, and expensive. Systems such as the
Vessel-Submerged Oil Recovery System (V-SORS – an array of heavy chain and sorbent pom-poms
dragged across the bottom) have proven effective in localizing the general areas of pooled and mobile spills,
but are unable to determine precise locations or actual amounts of oil. Furthermore, because V-SORS and
other technologies (such as sorbent drops and sediment cores) are used in contact with the seafloor, these
systems pose significant risk to snagging on or otherwise damaging benthic marine life and structures (e.g.,
reefs, cables, and pipelines). Other non-contact seafloor survey techniques such as ROV video surveys pose
the additional problems of only being operational in high visibility water and low sea states and generally
being un-navigated, or if navigated, then requiring large and costly dynamically positioned ships. Table E-1
(modified from Michel (2006)) lists the advantages and disadvantages of a variety of submerged oil
detection technologies and Figure E-1 shows a detection decision tree (Castle et al., 2005).

E-2


Table E-1. Advantages and disadvantages of submerged oil detection technologies (modified from Michel
(2006)).
ADVANTAGES DISADVANTAGES
Visual
- Can cover large areas quickly using standard - Only effective in areas with high water clarity
resources available at spills - Sediment cover will prevent detection over time
- Ground truthing required
Manual (V-SORS, Net Trawls, Snare Sentinels)
- Could detect both pooled and mobile oil moving above - Time and labor intensive for deployment, inspection,
the bottom and replacement
- Relatively efficient in that large areas could surveyed - Susceptible to snagging on the bottom
- Provided spatial data on extent of submerged oil - Cannot determine where along the trawl the oil
- Can vary the length of the trawl to refine spatial extent occurred
- Could be used in vessel traffic lanes - Difficult to calibrate the effectiveness of oil recovery
- Good positioning capability with onboard GPS and - Requires a vessel with a boom/pulley and adequate
navigation system deck space on the stern for handling, inspection, and
replacement
- Requires use of white snare, which has to be special
ordered
Side Scan Sonar
- Good spatial coverage - Once the oil spreads out, has reduced success at oil
- Not affected by poor visibility identification
- Good visualization of large oil accumulations and other - Slow turnaround (days) for useful product
bottom features (e.g., debris piles, pipelines) - Needs validation of targets as oil
- Less accuracy in muddy substrates
Multi-beam Sonar
- Some systems can generate high-quality data with - Data processing can be slow
track lines - Requires extensive ground truthing
- Good locational accuracy - Requires skilled operators
- Software detection algorithms can increase search
efficiency
Laser
- Almost no false positives - Of limited use in turbid waters
- Can use systems close to bottom
- Data output easy to interpret
Bottom Sampling
- Can be effective in small areas for rapid definition of a - Samples a very small area, which may not be
known patch of oil representative
- Low tech option - Too slow to be effective over a large area
- Has been proven effective for certain spills - Does not indicate quantity of oil on bottom
Real-Time Mass Spectrometry
- Able to detect wide range of components - Droplets of oil or soluble oil must be in the water
- Able to provide real-time data column
- Oil on the bottom cannot be solid (as in low
temperatures)

E-3


Tracking methods
Oil Density Water Depth
• Visual (aircraft)
0–20 m ± • Photobathymetric
techniques

Near neutral
buoyancy • Visual (diver)
0–30 m ±
(oil is suspended
in water column) • Sonar
• Visual (television,
No depth remotely operated
restriction vehicle)
• Water-column sampling
– water samples
– midwater trawls
• In situ detectors

Water Depth
• Visual (aircraft)
0–20 m ± • Photobathymetric
techniques
Negative
buoyancy
• Visual (diver)
(oil sinks to bottom) 0–0 m ±

• Geophysical
• Sonar
No depth • Side-scan sonar
restriction • Enhanced acoustic
• Grab samples
• Bottom trawls
• Visual (television,
remotely operated
vehicle)
• In situ detectors

Figure E-1. Detection decision tree.

E.3 Recommendations for Detection
The technology and approaches have not changed since the National Research Council (NRC) report
(Committee on Marine Transportation of Heavy Oils, National Research Council 1999). Experiences
during spills for the period since the report have contributed to some better understanding. Decision-makers
should still refer to the chart from Castle et al. (1995) and referenced in the NRC study (Figure E-1). Use of
the V-SORS was refined during the spills of 2004 (Delaware River) and 2006 (Gulf of Mexico). Additional
guidance includes:

1) Determine amount of impacted oil (oil that may contact or effect water inputs, sensitive areas, etc.)
or recoverable oil.
• Collectable amount is a function of time to reach the oil (including transit and mooring),
capability of cleanup technique, weather and amount of storage available.
2) Try most simple method first that addresses amount of oil being detected.
3) Use sophisticated methods for deeper and larger amounts of oil. Use models if available to
determine search area and potential amount of oil that may be recovered.

E-4


4) Make decisions based on the minimum amount of actionable oil (oil that may contact or effect water
intakes, sensitive areas, etc.) or recoverable oil. This helps to define the resolution of the detection
method needed.
• Recoverable amount is a function of time to reach the oil (especially if offshore and mooring
arrangements are made), capability of cleanup technique, weather and amount of storage
available. For example, the amount of “recoverable oil” for the DBL-152 spill was 500 barrels?
This was based on the cost for the recovery system to transit to the site, set up, and perform the
recovery. It was also partially based on the amount of oil that was perceived to harm the
environment or ultimately end up on shore.
5) Sonar can search a wide area but processing must be timely and resolution sufficient. Have vendors
determine resolution (i.e., the size of the patch of oil that can be detected), amount of time to search
any area, and the amount of time to process the data.
6) Operators of laser systems also need to define the area covered, estimated patch size, and the time to
process the data.
7) Utilize differential GPS systems for finer search grids if available.
8) Minimize the amount of time between the detection and collection phases of the response.

E.4 Manual Detection Methods
E.4.1 Snare Sentinels

“Snare sentinels” can consist of any combination of the following: a single length of snare on a rope
attached to a float and an anchor, one or more crab or lobster pots on the bottom that are stuffed with snare,
or a minnow trap or eel pot stuffed with snare and deployed at selected water depths. The configuration
depends on the water depth and where the oil is in the water column.

E.4.2 Vessel-Submerged Oil Recovery System (V-SORS)

The V-SORS consists of an 8 to10-foot pipe, 6 to 8 inches in diameter, rigged in a bridle fashion, attached
with several 6 to 8 foot lengths of 3/8-inch or larger chain (Figure E-2). Around the chains, snare is tied.
The system is towed behind a vessel and dragged along the bottom and somewhat angled through the water
column. It is pulled up regularly and inspected for oil. The oil coverage on the snares is roughly estimated.
The V-SORS Light system consists of a single chain with snare. This lighter system samples a smaller area
but requires less logistics.

E-5


Figure E-2. The V-SORS used to search for and recover submerged oil.

E-6


APPENDIX F. VENDOR CONTACT INFORMATION

BioSonics, Inc.: Robert McClure, MSc, FP-C
Business Development
4027 Leary Way NW
Seattle, WA 98107
bmcclure@biosonicsinc.com
Phone: 206-782-2211
www.biosonicsinc.com

CodaOctopus Group, Inc.: Anthony Davis
President US Operations
Bluewater House, 1601 3rd Street South,
St. Petersburg, FL 33701
anthony.davis@codaoctopus.com
Phone: 727-822-1565
www.codaoctopus.com

EIC Laboratories, Inc.: Job Bello
111 Downey Street
Norwood, MA 02062
bello@eiclabs.com
Phone: 781-769-9450
www.eiclabs.com

RESON Inc.: Mette T. Eriksen
Consulting Project Manager
100 Lopez Road
Goleta, CA93117
mette.T.Eriksen@reson.com
Phone: 805-964-6260
www.reson.com

Science Applications International Corp.: John T. Morris
Marine Survey Manager
221 Third Street
Newport, RI 02840
JOHN.T.MORRIS@saic.com
Phone: 401-847-4210
www.saic.com

F-1


SRI International: R. Timothy Short
Chemical Sensors Group, Manager
Marine Technology Program
333 Ravenswood Avenue
Menlo Park, CA 94025
timothy.short@sri.com
Phone: 727-553-3990
www.sri.com

Woods Hole Oceanographic Institution: Dr. Richard Camilli
Assistant Scientist
Deep Submergence Lab
Woods Hole, MA 02543
rcamilli@whoi.edu
Phone: 508-289-3796
www.whoi.edu

F-2

Rdc Submerged Oil Detection Report June 2009

More Related Content

Viewers also liked (7)

Similar to Rdc Submerged Oil Detection Report June 2009 (20)

Recently uploaded (20)

Rdc Submerged Oil Detection Report June 2009