Accepted Manuscript

Assessing the Performance and Cost of Oil Spill Remediation Technologies

Daniel P. Prendergast , Philip M. Gschwend
PII: S0959-6526(14)00409-0
DOI: 10.1016/j.jclepro.2014.04.054
Reference: JCLP 4260
To appear in: Journal of Cleaner Production
Received Date: 16 September 2013
Revised Date: 17 April 2014
Accepted Date: 21 April 2014
Please cite this article as: Prendergast DP, Gschwend PM, Assessing the Performance and
Cost of Oil Spill Remediation Technologies, Journal of Cleaner Production (2014), doi: 10.1016/
j.jclepro.2014.04.054.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
10,406 Words
Assessing the Performance and Cost of Oil Spill Remediation 1
Technologies 2
Daniel P. Prendergast
a
and Philip M. Gschwend
b,
* 3
a
Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, 4
MIT Bldg. 48-123, 77 Massachusetts Ave, Cambridge, MA 02139, USA 5
dprender@mit.edu 6
b
Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, 7
MIT Bldg. 48-413, 77 Massachusetts Ave, Cambridge, MA 02139, USA 8
pmgschwe@mit.edu 9
10
* To whom all correspondence should be addressed: 11
Philip Gschwend [tel: 617-253-1638 email: pmgschwe@mit.edu] 12

13
14
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
2

Abstract 1
Oil spills are an especially challenging chemical contamination event to remediate. Predicting the 2
fate and effects of spilled oil is a formidable task, complicated by its complex chemical composition and 3
the potential for catastrophically large discharge volumes. The proper choice of cleanup technique is 4
equally complex, and depends on a host of factors, including oil type, spill location, spill size, weather, 5
and local regulations and standards. This paper aims to provide a broad review of the current technologies 6
used to remediate oil spills, and the context in which they operate. The chemical characteristics of an oil 7
spill are discussed, including implications for transport modeling, and impacts that arise from short-term 8
and chronic toxicity. The most common remediation technologies (mechanical recovery, dispersants, and 9
in-situ burning) are reviewed, as are emerging technologies (hydrophobic meshes and magnetic sorbents). 10
A comparative analysis is performed on these methods by calculating a maximum oil encounter rate for 11
each device, which is an under-reported performance characteristic critical to planning a response effort. 12
Finally, a review of cleanup cost estimation techniques is used to assess the cost-effectiveness of 13
remediation methods. Analysis shows that waiving the legal penalty for recovered oil can result in 14
significant cost savings for the liable party, and may drive improvements in recovery-focused technology. 15
The authors suggest continued research into improving oil spill recovery methods and understanding the 16
fate of individual compounds in the spilled oil. This will both minimize potential environmental damages, 17
and reduce the uncertainty of their impacts. 18
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Highlights
Assessing the Performance and Cost of Oil Spill Remediation Technologies:
Prendergast, Gschwend

Review of the composition and fate of spilled oil, including modeling approaches.
Calculation of the maximum oil encounter rate for various remediation techniques.
Estimation of cleanup costs including cost averted by recovering spilled oil.
Net negative cost of cleanup can be achieved, promoting removal of contamination.
Recommend increased development of oil spill recovery methods.
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
10,406 Words
Assessing the Performance and Cost of Oil Spill Remediation 1
Technologies 2
Daniel P. Prendergast
a
and Philip M. Gschwend
b,
* 3
a
Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, 4
MIT Bldg. 48-123, 77 Massachusetts Ave, Cambridge, MA 02139, USA 5
dprender@mit.edu 6
b
Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, 7
MIT Bldg. 48-413, 77 Massachusetts Ave, Cambridge, MA 02139, USA 8
pmgschwe@mit.edu 9
10
* To whom all correspondence should be addressed: 11
Philip Gschwend [tel: 617-253-1638 email: pmgschwe@mit.edu] 12

13
14
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
2

Abstract 1
Oil spills are an especially challenging chemical contamination event to remediate. Predicting the 2
fate and effects of spilled oil is a formidable task, complicated by its complex chemical composition and 3
the potential for catastrophically large discharge volumes. The proper choice of cleanup technique is 4
equally complex, and depends on a host of factors, including oil type, spill location, spill size, weather, 5
and local regulations and standards. This paper aims to provide a broad review of the current technologies 6
used to remediate oil spills, and the context in which they operate. The chemical characteristics of an oil 7
spill are discussed, including implications for transport modeling, and impacts that arise from short-term 8
and chronic toxicity. The most common remediation technologies (mechanical recovery, dispersants, and 9
in-situ burning) are reviewed, as are emerging technologies (hydrophobic meshes and magnetic sorbents). 10
A comparative analysis is performed on these methods by calculating a maximum oil encounter rate for 11
each device, which is an under-reported performance characteristic critical to planning a response effort. 12
Finally, a review of cleanup cost estimation techniques is used to assess the cost-effectiveness of 13
remediation methods. When recovering spilled oil averts a fine, and then the fine is subtracted from the 14
cost of the spill, mechanical recovery methods are found to have a negative cost per unit recovered for 15
offshore spills with a variety of oil types and sizes. Waiving the legal penalty for spilled oil that is 16
recovered can result in significant cost savings for the liable party, and may drive improvements in 17
recovery-focused technology. The authors suggest continued research into improving oil spill recovery 18
methods and understanding the fate of individual compounds in the spilled oil. This will both minimize 19
potential environmental damages, and reduce the uncertainty of their impacts. 20
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
3

1. Introduction 1
The worldwide use and distribution of crude oil and its derivatives continues to impose a 2
potential threat to aquatic environments. Accidental releases can occur from a variety of sources including 3
tankers, pipelines, storage tanks, refineries, drilling rigs, wells, and platforms (Vanem et al., 2008). 4
Fortunately spill frequency and volume from all international sources have decreased since the 1970s 5
(Burgherr, 2006) due to the identification of management-based risk factors (Bergh et al., 2013), 6
increasing implementation of preventative regulations, and the development of corporate social 7
responsibility practices by the oil production and transportation industries (Rauffleta et al., 2014). Despite 8
these global improvements, there may be an increased risk of spills on a local level due to increased 9
industrial activities in countries with high economic growth, e.g. in the South China Sea (Woolgar, 2008). 10
Additionally, catastrophic spills remain a possibility from all sources. Noteworthy examples include: the 11
1989 sinking of the Exxon Valdez oil tanker off the coast of Alaska (Peterson et al., 2003), the subsea 12
blowout in the Gulf of Mexico of the Deepwater Horizon drilling rig in 2010 (Camilli et al., 2011), and 13
the 2010 pipeline spill of diluted bitumen in Michigan (EPA, 2011). The inability of responders to 14
prevent the spilled oil from reaching sensitive areas led to economic, social, and environmental damages. 15
These large-scale spills in highly mobile aquatic environments highlight the need for remediation 16
technologies that can respond swiftly to mitigate potential damages. 17
Oil spill clean-up technology has expanded to include a variety of approaches in the past 50 18
years. Spill response techniques are typically classified as mechanical/physical, chemical, and biological 19
(Dave and Ghaly, 2011). While only briefly described below, detailed reviews of these techniques have 20
been published, including their operational limitations (Ventikos et al, 2004) and a qualitative assessment 21
of their strengths and weaknesses (Dave and Ghaly, 2011). The mechanical/physical class includes 22
deployment of oil booms, which are floating barriers designed to control the movement of surface oil 23
slicks. Skimmers are a broad category of stationary or mobile mechanical devices specifically designed to 24
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
4

recover oil from the waters surface (Schulze, 1998). To separate the water and oil, they typically take 1
advantage in the difference in density or adhesive properties of water and oil. An example of a chemical 2
technique is the application of dispersants, which are surfactants sprayed on the oil slick from aircraft or 3
boats in order to reduce the water/oil interfacial tension and cause the oil to break-up into smaller drops. 4
This promotes dissolution and biodegradation while limiting movements of large volumes of oils to 5
sensitive receptors such as coastal wetlands. Bioremediation consists of the addition of nutrients and/or 6
oxygen to stimulate the growth of indigenous microbes that can utilize oil as a carbon source. Microbes 7
designed to degrade the oil can also be added, if it is felt that natural oil-degrading strains are not present 8
in sufficient numbers. Most recent research has focused on chemical surfactants or bioremediation 9
applications, in order to improve their efficiency and/or the impact of their addition on the environment 10
(Dave and Ghaly, 2011). 11
Newer techniques are also becoming well-known and applied in the oil spill response community. 12
One such technique, in-situ burning, consists of using specially designed high-temperature boom to corral 13
oil slicks into a smaller area, where it is ignited in a controlled burn (Allen and Ferek, 1993). This 14
technique was widely used in the Deepwater Horizon response (Allen et al., 2011). Absorbents also see 15
widespread use, especially when cleanup goals demand the complete removal of oil. However, due to the 16
difficulty in handling oil-soaked materials, this technique is typically confined to small areas, and is not 17
examined in this paper. 18
Material science offers the potential for innovation beyond current techniques. Skimmers have 19
been modified to have oleophilic surfaces, and this advance has seen widespread implementation in the 20
oil spill response industry (Broje, 2006). Magnetic particles offer many advantages over traditional 21
absorbent techniques. Their high hydrophobicity and oleophilicity makes them extremely efficient 22
separators, and their uptake capacity can match or exceed current absorbents (Chun and Park, 2001). In 23
addition, their inherent magnetic properties provide a facile method of recovering and handing oil-sorbent 24
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
5

amalgam. Another technique originating from materials research involves hydrophobic meshes, which 1
can separate oil and water in-situ without additional energy input (Deng et al., 2013). While these 2
techniques remain largely untested under field conditions, their potential to improve the rate and 3
efficiency of cleanup operations is worth investigating. 4
This paper has three primary objectives. The first is to review the inherent complexity in 5
predicting the fate and impact of spilled oil in the marine environment. Crude oil and its derivatives are 6
extremely complex mixtures of organic chemicals. Recent advances in fate modelling are reviewed, while 7
highlighting the uncertainties and gaps in current knowledge. Ideally, perfect knowledge leads to an 8
optimal response to an oil spill, defined as one that balances response costs with environmental damages. 9
However, the literature shows that quantifying the damage to social, economic, and environmental 10
resources from oil spills is an uncertain endeavor. Thus, the second objective is to review and reanalyze 11
the performance of the major classes of oil spill cleanup techniques in order to assess the current 12
technological capabilities for responding to a large-scale oil spill. Emphasis will be placed the encounter 13
rate of each technique, a common limiting factor for large spills. The third objective is to review how the 14
costs of response efforts are currently estimated. These methods are then used to establish a financial 15
incentive to recover oil, under the hypothetical scenario whereby the responsible party is not fined for oil 16
that is recovered from the environment at or very near the point of discharge. This scenario will highlight 17
the financial benefits of recovering, rather than dispersing or destroying, spilled oil, and show how it 18
complements the mitigation of environmental damage. 19
20
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
6

2. Method of Review 1
This study relies solely on peer-reviewed scientific papers, publically available government 2
reports, and published results of remediation technique performances. Preference was given to recent 3
reviews of a given topic, and papers that first identified a phenomenon. Using this literature, the most 4
common approaches to oil spill remediation are identified, and their performance metrics are compiled. In 5
a parallel effort, the factors controlling the fate and impact of oil spills are identified, as are the 6
quantitative models that predict these factors. Only models that had published the scientific basis and 7
validation of their algorithms are included in this review. These models are assessed for their adherence to 8
physical mechanisms, and ability to predict the transport and impact of oil spills. We then introduce the 9
idea of a theoretical maximum oil encounter rate, and show how the underlying formulation is consistent 10
with published predictive tools currently used by oil spill response community. After identifying the 11
major classes of oil spill response technology, published methodologies for estimating the cost of each 12
response are used to assess the economic implications of oil spill recovery. 13
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
7

3. The Composition and Fate of Spilled Oil 1
3.1 Oil Composition and Characterization 2
Oil is not a single-component substance with well-defined physical properties and behavior. For 3
example, crude oil is a mixture of individual chemicals with more than 10,000 unique elemental 4
compositions (Marshall and Rodgers, 2008). Each unique elemental composition, in turn, potentially 5
represents thousands of unique chemical structures. The total number of individual compounds is 6
estimated to be in the billions (Beens and Brinkman, 2000). The compounds mostly consist of 7
hydrocarbons, but also include organic compounds with various heteroatom substituents, notably oxygen, 8
nitrogen, sulfur, and trace metals (Shi et al., 2010). More processed forms of oil, such as diesel fuel, 9
lubricating oils, or diluted bitumen, represent a subset of the composition of crude oil which has been 10
separated or modified to produce desired physical or chemical properties. The origin of biofuels is distinct 11
from that of crude oil, and as such is considered separately in performance and analysis (although they 12
may have many components in common) (Brynolf et al., 2014). Desired properties of any fuel depend on 13
individual chemical composition, which has been limited by the overwhelming complexity of the mixture. 14
Gas chromatography coupled with mass spectrometry (GC-MS) is the conventional method used 15
to elucidate oil composition. The columns used to separate oil components largely do so based on their 16
London dispersive interactions as reflected by their boiling points. In order to resolve compounds with 17
similar boiling points, two-dimensional gas chromatography (GC x GC) can be used (developed by Liu 18
and Phillips (1991), with recent applications by Ventura et al. (2008) and Reddy et al. (2012)), which 19
separates by both the London interactions with the first stationary phase as well as polar interactions with 20
the second stationary phase. Unfortunately, GC is only effective at separating compounds with a boiling 21
point of less than about 400C. For crude oil, this can represent a significant blind spot: only about half of 22
the mass in the Macondo oil well was resolvable by conventional gas chromatography (McKenna et al., 23
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
8

2013). Only very recently have these high-boiling point compounds been separated by mass using ultra- 1
high resolution mass spectrometry, but the structure of the individual compounds for a significant fraction 2
of many crude oils remains unknown. 3
Additional characterization methods have been developed in order to empirically link oil 4
composition with bulk behavior. The most widespread technique is SARA fractionation, which separates 5
oil into saturates, aromatics, resins, and asphaltenes (Jewell et al., 1972). Resins and asphaltenes are more 6
polar and differentiated empirically. Asphaltenes are insoluble in heptane or pentane. Saturates are 7
nonpolar saturated hydrocarbons, while aromatic compounds are more polarizable. Heavier oils tend to be 8
enriched in resins while lighter oils tend to be enriched in saturates. Depending on the analytical 9
technique used, different methods may yield different mass fractions for each class (Fan and Buckley, 10
2002). 11
Nevertheless, numerous correlations have been developed that link the SARA fractions to the 12
bulk properties of spilled oil. For example, emulsification has been found to be related to SARA 13
information (Fingas and Fieldhouse, 2012). Emulsification of the oil, where water droplets become 14
entrained in the oil phase due to wind and wave action, increases the oil's viscosity, thereby making the 15
oil slick more resistant to skimming and dispersion. Fingas and Fieldhouse (2012) found that the resin and 16
asphaltene fractions of more than 300 crude oils correlate with the ease of such emulsification. In 17
addition, the specific gravity of oil has often been used to predict its un-emulsified viscosity, with more 18
recent correlations utilizing resin and asphaltene content (Hossein et al., 2005). Notably, in both of the 19
above studies, higher asphaltene content was shown to significantly increase the viscosity of the oil and 20
stability of the emulsion. Spilled oil becomes more viscous and emulsified as time passes. Field trials 21
(Figure 1) have shown that in less than 24 h, the water content of emulsified oil can reach up to 80% by 22
volume, and its viscosity can increase by a factor of 100 (Daling et al., 1997). An increase in 23
emulsification or viscosity causes any remediation technique to be less effective. Understanding the 24
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
9

propensity of spilled oil to alter its physical properties under environmental conditions is one key to 1
understanding its fate and choosing the optimal remediation response. 2
3
Figure 1: Field data of the viscosity and water content of crude oil released on the open sea. Water 4
temperature was 15C in the summer and 10C in the winter. Lines are added for visualization purposes 5
only. Data taken from Daling et al. (1997), who note that dispersant effectiveness is dramatically reduced 6
for this oil when its viscosity becomes greater than 4000 cP. 7
8
3.2 Modeling Oil Distribution and Fate 9
After release to an aquatic environment, oil undergoes numerous processes collectively known as 10
weathering that alter its composition and fate (Blumer et al., 1973). An oil slick both moves with the 11
underlying water currents and spreads relative to the water surface (Fay, 1971). Spreading broadens and 12
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
10

thins the oil slick at a rapid rate controlled by gravity and the oils viscosity. As it is exposed to wind and 1
wave action, lighter oil components are transferred to the air via evaporation (Mackay and Matsugu, 2
1973). Simultaneously, lighter polar constituents dissolve into the water column (Arey et al., 2007). Both 3
of these processes cause the remaining oil to become enriched with heavier, apolar compounds, which 4
increases the viscosity of the slick. Depending on the composition of the oil, the depletion of lighter 5
compounds can enhance the transformation of a slick into a water-in-oil emulsion through wind and wave 6
action (Fingas, 1995). Eventually, the oil can entrain so much water (up to 80% by volume) that it breaks 7
up into nearly neutrally buoyant tar balls, causing it to sink and be transported hundreds of miles with 8
prevailing currents (Reed et al., 1999). Oil on the surface can undergo photodegradation where it is 9
broken down into smaller components or transformed into oxygenated compounds (Aeppli et al, 2012). 10
Oil can also be broken down by native microorganisms, especially when entrained in the water column 11
(Atlas, 1995). Models that couple all of the transport and weathering process have been developed by 12
various governmental and industry groups. One goal of such models is to predict the physical extent and 13
distribution of a spill in order to plan or implement response operations. 14
The National Oceanographic and Atmospheric Administration (NOAA) has developed two 15
models to this end: the General NOAA Operational Modeling Environment (GNOME) and the 16
Automated Data Inquiry for Oil Spills 2 (ADIOS2). 17
GNOME focuses on predicting the trajectory of an oil spill using a 2D Lagrangian approach 18
(Beegle-Krause, 2001). The model accounts for wind, current, oil spreading, and beaching along 19
shorelines (Zelenk et al., 2012). It also takes into account evaporation by idealizing the oil as a mixture of 20
three pseudocomponents, each assigned an independent degradation half-life. Users may select one of six 21
included oils with predetermined component fractions and half-lives, or specify oil with custom 22
characteristics. Other inputs include local wind velocities, currents, bathymetry, maps, and extent of the 23
oil spill. The program models the development of the spills spatial extent over time. The model does not 24
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
11

take into account any losses besides evaporation, and thus likely overestimates the total slick volume at 1
any given time. NOAA also acknowledges that the trajectory model is simplistic and sensitive to the 2
accuracy of the local current data, but that it includes tools to estimate the uncertainty of a predicted 3
trajectory (NOAA, 2013). 4
ADIOS2 (Lehr et al., 2002) focuses on the weathering process of spilled oil, rather than its spatial 5
distribution. Using a database of over 100 oils, the program tracks the macroscopic properties of the oil 6
(specifically density, viscosity, and water content) as it undergoes evaporation, spreading (to determine 7
thickness), emulsification, dispersion (entrainment of oil droplets into the water column), and beaching. It 8
also simulates the effects of clean-up techniques, including skimming, in-situ burning, and application of 9
dispersants. Inputs to the program include wind speed, a uniform current, oil type, and the details 10
regarding the cleanup operations. The main program outputs are the evolving properties of the spilled oil 11
and remaining volume. Predictions do not extend for more than five days, and so they do not consider 12
effects of biodegradation or phototransformations. ADIOS2 also does not model the formation of tar 13
balls, which is the final fate of much of the heavier, nonpolar components of the oil. In addition, many of 14
the algorithms are based on empirical studies where the properties of interest were simulated in a 15
laboratory with a limited set of oils, rather than a field trial. Lehr et al. (2002) note that validation of the 16
model has been limited to basic observations of 40 small oil spills, and they recommend that it be used as 17
a rough guide for cleanup planning. 18
Oil spill models have also been developed by nongovernmental agencies. The Spill Impact Model 19
Application Package (SIMAP) was developed by Applied Science Associates, Inc. as part of an 20
ecological risk assessment process required by the Department of the Interior (French-McCay, 2004). As 21
a result of this developmental goal, SIMAP is unique insofar as it combines a physical fate model with a 22
biological effects model. The physical model predicts slick transport, evaporation, component 23
concentrations in the water column, sorption to sediments, and shoreline fouling. The model still 24
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
12

simplifies the oil composition by treating it as an eight component mixture, but the pseudo-components 1
have been chosen to emphasize the toxic, volatile, and residual natures of a given oil. The algorithms used 2
to model the physical degradation resemble those in the AIDOS2 model, and so suffer from the same 3
limitations. The biological effects model uses the output from the physical fate model to determine the 4
extent of exposure of local wildlife (birds, fish, marine mammals, and reptiles), invertebrates, and plant 5
communities (French-McCay, 2009). A toxicity model (Di Toro et al., 2000) is then used to determine the 6
percent mortality, which is in turn fed into a food web model to determine overall ecological impact. 7
SIMAP does not take into account chronic toxicity or changes in the behavior of birds and wildlife. 8
Validation has been undertaken with at least 20 spill events (French-McCay, 2009), but most cases were 9
limited to observation of wildlife. 10
Modeling the outcome of an oil spill faces many challenges. Many of the algorithms used to 11
predict physical processes rely on parameters derived from empirical studies. These parameters are based 12
on a limited dataset, and it is unclear if or when the models extrapolate from that dataset. Another major 13
challenge to developing accurate fate models is the difficulty of validation. Oil spills can cover a large 14
area, with complicated local ecosystems and physical environments. Collecting appropriately detailed 15
spatial and temporal field data is a challenging, time-consuming, and expensive endeavor. Future 16
modeling efforts should focus on validation, particularly by hindcasting the effects of a past spill where 17
extensive fieldwork was undertaken. 18
3.3 Environmental Impact of Spilled Oil 19
The environmental damage caused by spilled oil is probably its least understood aspect, owing to 20
the combined complexity of both the oil and the environment. The toxicity of the individual compounds 21
can vary greatly. Quantitative structure-activity relationships (QSARs) have been developed to correlate 22
the n-octanol-water partition coefficient (a measure of a chemicals tendency to partition into nonpolar 23
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
13

media from aqueous solution) to the nonspecific toxicity of hydrocarbons, which is caused by the build- 1
up of oil components in the membranes of cells (Di Toro et al., 2000). The nonspecific toxicity can also 2
be determined solely through GC x GC analysis of the oil, without the need to identify individual 3
components (Tcaciuc et al., 2012). In addition, compound-specific QSARs often contain correction 4
factors for chemicals with unique toxicity mechanisms, such as polycyclic aromatic hydrocarbons 5
(McGrath et al. 2005). 6
However, our knowledge of toxicity mechanisms is incomplete. First, byproducts can form after a 7
spill is released into the environment. For example, as first noted by Hansen (1975) and recently reviewed 8
by Lee (2003), compounds produced by photodegradation can be toxic to aquatic animals and 9
microorganisms, yet this aspect is not typically tracked as part of an oil fate model. Second, there is some 10
evidence that the bioassays used to test for chemical toxicity do not fully expose the test organisms to 11
fraction of the oil that is not resolved by GC (Hong et al., 2012). Weathered oil is enriched in high 12
molecular weight compounds, and its toxicity relative to fresh oil is still being debated. Finally, most 13
experiments can only determine the acute toxicity of chemicals, since exposure times rarely last longer 14
than 96 hours (Weber, 1993). This has led to instances where chronic exposure to low levels of pollutants 15
caused lasting adverse effects on the local ecosystem, such as in Prince William Sound following the 16
Exxon Valdez oil spill (Peterson et al., 2003). In general, the complexity, variability, and incomplete 17
characterization of oil compounds magnify the uncertainty in predicting the impact of a spill. 18
4. Technical Assessment of Oil Spill Response Technologies 19
4.1 Overview of Response Techniques 20
A broad review of the operational limitations of various countermeasure techniques has been 21
reported by Ventikos et al. (2004), including mechanical barriers/booms, skimmers, skimmer vessels, 22
sorbent materials, and chemical dispersants. They provide quantitative performance limits for all 23
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
14

techniques based on environmental conditions: wind speed, current velocity, wave height, and sea state. 1
For recovery techniques (skimmers and sorbents) additional data are provided for oil viscosity, recovery 2
efficiency (ratio of oil encountered to oil recovered), sensitivity to debris, recovery capacity, and nominal 3
recovery rate (range of possible volumetric recovery rates). For dispersants, rougher numbers are given 4
for the application ratio (volume of dispersant/oil) and the effectiveness. These values will be used as the 5
basis of subsequent analysis here. 6
A more comprehensive, but qualitative, review has been completed by Dave and Ghaly (2011). In 7
addition to the techniques reviewed in Ventikos et al. (2004), Dave and Ghaly (2011) included chemical 8
solidifiers, in-situ burning, and bioremediation. The authors listed qualitative advantages and 9
disadvantages for each technique, then assigned a weighted score based on a variety of criteria, including: 10
efficiency, time, cost, impact on marine life, level of difficulty (to operate), weather, reliability, oil 11
recovery, effect on oil characteristics, and the need for post-remediation treatment. Their work provides 12
insight into the less quantifiable aspects of implementing each technique. 13
4.2 Maximum Oil Encounter Rate 14
Currently, the recovery capacity of oil spill remediation techniques is reported as the Effective 15
Daily Recovery Capacity (EDRC) (USCG, 1997). This value is used in the design of oil spill response 16
plans, and is required by the Oil Spill Protection Act. The calculation is as follows: 17
EDRC = T 24 h E (1)
18
where EDRC is in bbl/d, T is the nameplate recovery capacity in bbl/h as defined by ASTM F2709 19
(2008), and E is an efficiency factor, which must be at most 20%. The nameplate recovery capacity is the 20
maximum rate that a collection system can recover oil, given optimal conditions. Thus, the EDRC 21
represents a corrected estimation of a days worth of recovery efforts. Unfortunately, the correction factor 22
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
15

is the only parameter than can be adjusted for irregularities in spill conditions, and it does not explicitly 1
take into account variable oil viscosity and thickness, emulsified oil, dispersant addition, or other 2
complicating factors (Lee, 1993). In terms of assessing the potential of spill technology, it 3
overemphasizes the importance of the nameplate recovery capacity. An important missing parameter is 4
the oil encounter rate. A framework for assessing this parameter is developed below. 5
In simple terms, a device cannot recover more oil than it encounters. Thus, the volumetric 6
recovery rate of any system will be the lower of either its EDRC or the encounter rate. The encounter rate 7
for a given spill is dependent on the physical distribution of the spill and the capacity of the remediation 8
technique. In order to focus on the potential of a technology, it is assumed that the recovery system is 9
always in contact with an oil slick, making this framework an estimation of the maximum oil encounter 10
rate (MOER). Furthermore, the slick is characterized by its average thickness, which accounts for both its 11
volume and spatial distribution. Thus, the MOER is described in units of area/time. This approach 12
quantifies the limiting factor for large spills spread over a wide area. In contrast, for confined spills the 13
rate of recovery is more likely to be limited by the EDRC. The methodology of this approach is consistent 14
with previously developed spill response operations planning tools, such as the Response Operations 15
Calculator (Dale et al., 2011) and the Estimated Recovery System Potential (ERSP) Calculator (Allen et 16
al., 2012). 17
The MOER is designed as a complementary metric for assessing oil spill technologies, not an 18
improvement of the EDRC. Below (sections 4.2.1 to 4.2.4), we develop a MOER expression for major 19
classes of remediation technologies, and estimate representative values. Unless otherwise noted, these 20
calculations do not take into account the preparation time, transport time, or post-operational processing. 21
The MOER is oil encounter rate under optimal conditions. In-situ burning is an exception because it relies 22
on a two-step technique of corralling the oil, followed by a stationary burn. Sorbents were not included in 23
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
16

this analysis, since they are generally more suited for confined spills where complete oil removal is 1
necessary (Dave and Ghaly, 2011). 2
It is important to emphasize that a MOER estimation assumes the technique is in constant use, 3
which is never the case. For example, aircraft applying dispersants are often only over the spill for about 4
20% of operational time (Fingas, 1999). Thus, by focusing on the maximum encounter rate possible, the 5
MOER is an overestimation of actual performance. The calculated MOER values are intended to be 6
representative of the techniques potential; the unique circumstances of a given spill will degrade the 7
actual encounter rate. 8
4.2.1 Skimmers 9
For skimmers, the MOER is simply the product of the maximum operational water velocity 10
(relative to the skimmer) and the width of the skimmer opening. For mobile skimmers that operate in 11
conjunction with a boom, the boom is assumed to funnel the encountered oil with perfect efficiency, 12
giving the skimmer an effective swath equal to the width of the boom opening. The maximum feasible 13
swath of a boom is about 1000 ft, or 300 m (Allen et al., 2012), and the maximum speed a boom can be 14
towed under optimal sea conditions without losing oil is about 0.5 m/s (Amini et al., 2008). Thus, the 15
MOER for a boom-skimmer system is 150 m
2
/s, or 54 ha/h. 16
4.2.2 Dispersants 17
Dispersants are typically applied from fixed-wing or rotary-wing aircraft, covering a wide area 18
using specially designed application equipment (ASTM, 2013). The MOER can be found directly from 19
previous operational reports of areal coverage rate (Fingas, 2011). At a 1:20 dispersant-to- oil ratio, 20
typical values for rotary- and fixed-wing aircraft are 300 ha/h and 800 ha/h, respectively. 21
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
17

4.2.3 In-situ Burning 1
In-situ burning is a two-step process. The first step involves collecting oil within a boom, which 2
would have the same encounter rate as the skimming system described above. The second is the 3
controlled burn, which requires the boom to remain stationary. In calculating the MOER for in-situ 4
burning, it is assumed that the time required to seal the boom and ignite the collected oil is negligible 5
compared to the total collection and burn time. In addition, the boom assumes the shape of a catenary 6
curve (Wicks, 1969), and a collection is ceased when oil occupies one third of the total enclosed area. The 7
oil burns downward from the surface of the enclosed slick at a typical rate of 3 mm/min (Fingas, 2011), 8
although this value may be slightly lower for emulsified oil (Evans et al., 1990). Given a capacity, C (m
3
), 9
of a boom configuration, the time required to collect the oil (T
1
) and to burn the oil (T
2
) is given by: 10
T
1
= C / S ; T
2
= C / (A B) (2)
11
where S is the encounter rate (m
3
/s) when moving, A is the area occupied when the boom is at 1/3 12
capacity (as recommended by Allen (1991)), and B is the burn rate, noted above. The total encounter rate 13
is given by dividing the capacity of the boom by the total time required to fill and burn it: 14
EnR = C / (T
1
+T
2
) = C / [(C / S ) + (C/AB)] = [(1 / S) + (1 / AB)]
-1
(3)
15
Normalizing by the spill thickness h and introducing parameters for tow speed u and swath width w gives: 16
MOER = EnR / h = [(h / wuh) + (h / AB)]
-1
= [( 1 / wu) + (h / AB)]
-1
(4)
17
Booms used for in-situ burning are intentionally towed such that the opening width is about 30% of the 18
boom length (Allen, 1991). As outlined by Wicks (1969), a catenary boom with a known length and 19
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
18

width-to-length ratio has a fixed area, A (see Appendix). Using a 300 m long boom with a 90 m wide 1
opening (Allen, 1991), the area is found to be approximately 3000 m
2
. 2
Equation 4 shows that, due to the time required to remain stationary and to burn the oil, the 3
MOER for in-situ burning must depend on the thickness of the oil encountered. In order to show the range 4
of possible values, two MOER values are calculated: one for a relatively thick oil slick of 5mm, and one 5
for a relatively thin slick of 50 m, Using a tow speed of 0.5 m/s, those values are 18 ha/h and 45 ha/h, 6
respectively. 7
4.2.4 Hydrophobic Mesh 8
Hydrophobic mesh acts much like a filter: it allows oil to pass through the mesh while rejecting 9
water. This process is passive, utilizing the difference in interfacial energy to drive the separation step. 10
However, oil must still be collected and removed from the interior of the mesh to perpetuate the process. 11
Deng et al. (2013), using bench-scale tests, found that the rate of oil recovery was faster than the 12
spreading rate of relatively inviscid oils. Thus, much like a skimming system, field-scale devices 13
incorporating hydrophobic mesh will need a way to store recovered oil, as well as continuously move the 14
mesh to unrecovered oil. Although such devices have been yet to be developed, this assessment envisions 15
a device that is handled like a skimmer, with a wide booming system in place to tow the mesh across 16
floating oil slicks, maximizing the oil encountered. In this case, hydrophobic meshes would have a 17
MOER equivalent to those of skimmers. They would also likely be classified as a mechanical recovery 18
technique. 19
4.3 Discussion 20
Table 1 contains a summary of the calculated MOER values. Dispersants applied with aircraft 21
clearly have the largest MOER. This is not unexpected, since aircraft can move much faster than marine 22
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
19

vessels towing boom, and do not need to handle the oil after encountering it. Burning and recovery 1
methods cannot encounter oil at the same rate dispersants can. However, due to the time required to reach 2
and return from a slick, typical application rates are 20% of this maximum potential (Fingas, 1999). Even 3
with this practical limitation, if time is the only consideration, dispersants will likely continue to be the 4
fastest way to protect vulnerable resources. 5
Technology Maximum Oil Encounter
Rate [ha/h]
Dispersants: Fixed-wing aircraft 160
Dispersants: Rotary Aircraft 60
Skimming 54
Hydrophobic Meshes 54
In-situ Burning; 5 mm oil 18
In-situ Burning; 5 m oil 45
Table 1: Oil encounter rates normalized by spill thickness for various oil spill remediation technologies 6
operating under optimal operational conditions except applying a 20% factor to dispersant use. 7
8
In this analysis, in-situ burning has a lower MOER than mechanical recovery.This is the opposite 9
of what is typically observed in practice (e.g. Allen et al., 2011). Both utilize a boom to collect oil, but 10
only in-situ burning must collection be (in theory) periodically stopped to burn it off. Most common 11
skimmer systems cannot recover continuously, and must also stop when onboard oil capacity is reached 12
(Schulze, 1998). In addition, the MOER increased when the slick thickness decreased. This result 13
indicates that it is more efficient for the fire-resistant boom to remain mobile for as long as possible. 14
Additionally, according to Equation 3, only the area of the boom affects the encounter rate, not the 15
booms volumetric capacity. A deeper draft is a tradeoff: it allows a responder to collect more oil, but it 16
also means the boom must remain stationary for longer during the burning step. The only way to increase 17
the area of the boom is to use longer boom, or drag it in a wider configuration (see Equation A.3 in the 18
Appendix). 19
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
20

Unlike the EDRC, a MOER calculation can compare recovery and non-recovery remediation 1
techniques. The previous calculations compare the implementation of a new technique (hydrophobic 2
meshes) to well-established methods, under the assumption that it could match the recovery capacity of a 3
skimmer system. Unfortunately, there is a lack of quantitative studies on the separation capabilities of 4
hydrophobic meshes, especially under the harsh conditions of a marine environment (Liu and Jiang, 5
2011). Matching the EDRC of a boom/skimmer system may require unmanageably large areas of the 6
mesh to be brought into contact with the oil slick. 7
Nonetheless, there is reason to believe that hydrophobic meshes have a higher potential MOER. 8
A skimming apparatus is relatively small, and relies on a wide boom swath to encounter oil at a useful 9
rate. In contrast, large areas of hydrophobic mesh may be fabricated, and may potentially be incorporated 10
along the entire length of a boom. Additionally, boom failure usually occurs when the oil slick is pushed 11
at a high velocity relative to the underlying water (Amini et al., 2008). A boom that recovers oil along its 12
entire length reduces that relative velocity, and enables the system to be towed at a faster rate. 13
Incorporating hydrophobic mesh could increase both the MOER and the EDRC of a towed-boom system. 14
An improved determination of the MOER will allow this new technology, or any other, to be more 15
thoroughly assessed for remediation potential than by an EDRC calculation alone. 16
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
21

1
5. Cost of Response 2
Cost is important in determining the effectiveness of a remediation technique. A full analysis 3
would consider both the cost of implementing a remediation technique, as well as the cost averted by 4
preventing the oil from damaging the environment. While these two factors can be difficult to estimate, 5
the balance between the two provides an optimization objective. A spill should not be cleaned in the least 6
expensive manner; rather, an optimal spill response minimizes both environmental damages and cost of 7
operations. This section reviews methods by which the cost of oil spills is typically calculated, and 8
provides additional analysis regarding the choice of response techniques. 9
A variety of cost analyses for oil spills have been performed in the last 30 years (Kontovas and 10
Psaraftis, 2008). These studies were undertaken to quantify the full cost of pollution incidents in a manner 11
that was valuable to risk management decisions and conflict mitigations (Yang et al., 2014). In general, 12
the studies found that the cost of the spill, on a per-weight-spilled basis, depended largely on oil type, 13
weather, location (both in terms of geography and national jurisdiction), extent of shoreline oiling, and 14
cleanup technique. In addition, the types of cost tended to vary, but could be largely classified as cleanup 15
(removal, research, etc.), socioeconomic losses (tourism, marketable resources lost), and environmental 16
(deaths of flora and fauna, ecosystem impacts) costs. According to Kontovas and Psaraftis (2008), the 17
determination of total cost for a spill always uses historical data, but has taken different approaches: 18
1. Estimating and adding up all relevant cost components (cleanup, socioeconomic, and 19
environmental). 20
2. Estimating cleanup costs, and then estimating the environmental and socioeconomic costs 21
through a modeled comparison ratio. 22
3. Estimating total costs directly, but controlling for various influencing factors. 23
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
22

4. Estimating total costs based on compensation eventually paid to claimants. 1
Additional efforts have been made to develop a single relationship between the total cost and volume of 2
oil spilled, despite the often-acknowledged complexity inherent to these approaches. As a result, the 3
regression analysis of Kontovas et al. (2010) showed significant variability, even when outliers are 4
removed and the form of the fitting equation was varied. The cost variability often exceeded an order of 5
magnitude for a given spill size. 6
The model of Etkin (2000) (a Type 3 approach) handles this variability by accounting for 7
different spill scenarios. This model was derived from historical cost data of worldwide oil spills and 8
validated with multiple case scenarios. Cost data were initially obtained from the OSIR International Oil 9
Spill Database (ITOPF, 2012), which also tracks the characteristics of the oil spills (size, oil type, etc.). 10
The model takes into account oil type, extent of shoreline oiling, size of the oil spill, location type, 11
country where the spill occurs, and primary cleanup method. The model predicts that the per-weight cost 12
of oil spill cleanup increases with: decreasing spill size, proximity to shore, extent of shoreline oiling, and 13
oil viscosity (reflecting a larger composition of non-volatile, sparingly-soluble components). The trend in 14
costs for shore proximity and spill size is counter-intuitive. However, Etkin (2000) notes that the high 15
per-volume cost of small spills is likely due to a variety of fixed costs related to response resources that 16
are required by law (environmental monitors, stand-by crews), but not always fully implemented for a 17
small spill. In any case, by statistically controlling for influencing variables, this model allows a 18
straightforward comparison of a variety of oil spill scenarios. 19
A simplified form of the main equation of the model is: 20
C
li
= C
n
t
i
s
i
m
i
(5)
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
23

where C
li
is the response cost per unit spilled for scenario i, C
n
is the average response cost per unit 1
spilled in nation n, t
i
is the oil type cost modifier for scenario i, s
i
is the spill size cost modifier for 2
scenario i, and m
i
is the cleanup method cost modifier for scenario i. The cost modifiers represent the 3
median percentage difference in cost for a given scenario with that characteristic from the overall median 4
spill cost. 5
Overall, Etkin (2000) showed that the most expensive cleanup techniques are (in increasing 6
order) natural cleansing, in-situ burning, dispersants, mechanical recovery, and manual cleanup. The 7
analysis in this work only compares dispersants, in-situ burning, and mechanical recovery, which are the 8
three most likely options for a large, offshore spill. Accordingly, the spill location type is limited to 9
offshore, and the extent of shoreline oiling to zero kilometers, since this factor is a reflection of the cost of 10
shoreline remediation techniques, not on-water cleanup techniques. A final summary of the cost modifiers 11
considered in this analysis is shown in Table 2 below. 12
Oil Type Oil Type
Modifier
(ti)
Spill Size Spill Size
Modifier
(si)
Primary Cleanup
Method
Method
Modifier
(mi)
No. 2 Fuel (Diesel) 0.18 < 34 t 2.00 In-Situ Burning 0.25
Light Crude 0.32 34-340 t 0.65 Dispersants 0.46
Crude 0.55 340-1,700 t 0.27 Mechanical Recovery 0.92
Heavy Crude 0.65 1,700-3,400 t 0.15
No. 6 Fuel 0.71 3,400-34,000 t 0.05
No. 4/5 Fuel 1.82 > 34,000 t 0.01
Table 2: Cost modifiers (t
i
, s
i,
m
i
) used in this analysis. Taken from Etkin (2000). 13
Etkin (2000) found that the average cost per unit of oil spilled varied greatly depending on the 14
nation in which it was spilled. For example, in 1999 USD, the average cost of a spill in the United States 15
was approximately $25,600/t, while in Singapore it was only $390/t. This large variation was attributed to 16
a variety of factors, including spiller liability, cleanup standards, labor costs, and the scarcity of data for 17
some regions. Assuming the cost of the average spill has only been affected by inflation, the current value 18
(2013 USD) for the United States is $35,800/t, calculated from the consumer price index (BLS, 2013). 19
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
24

The cost of the spill usually falls upon the party responsible for the incident. This cost includes 1
any fines imposed by regional laws. Although the exact amount of the fine and method of calculation 2
varies between regions, in the United States the Oil Pollution Act of 1990 imposes a civil penalty of 3
$1,100 per barrel spilled (about $8,070/t). This fine increases to $4,300 per barrel spilled (about 4
$31,600/t) if the liable party was found to be grossly negligent in its actions (33 U.S.C. 40 2716a). 5
This system ensures the responsible party has an incentive to minimize any damages resulting from the 6
spill. The Etkin model includes these fines within the listed cost of the spill. 7
A hypothetical scenario is now proposed. If the liable party were fined for each barrel spilled, 8
minus each barrel recovered during cleanup operations, this could greatly influence the choice of 9
cleanup method. In other words, the liable party is fined only for each unit of oil remaining in the 10
environment. The reasoning for this suggestion is straightforward: if the oil does not remain in the 11
environment, and is confined to an offshore location near where it was spilled, much of the potential 12
damage is mitigated, and the liable party should receive a lesser penalty. 13
Under this legal scenario, one can analyze the cost of cleanup as follows. For remediation efforts 14
utilizing dispersants or in-situ burning, the oil components are merely transformed or moved to a different 15
phase, and still have the potential to damage the environment. The cost associated with these methods will 16
be assessed using Equation 5. In contrast, mechanical recovery of the oil removes it from the 17
environment, and thus would not be fined per unit spilled. This cost would be determined by 18
C
li
= C
n
t
i
s
i
m
i
F (6)
where F is the fine per unit of oil spilled (but not recovered). C
n
was taken as the overall average cost for 19
the United States ($35,600) multiplied by the cost factors for an offshore spill (0.46) and 0-1 km of 20
shoreline oiling (0.47). Figure 2 shows the results of comparing the prices of the three original types of 21
cleanup methods (mechanical recovery, dispersants, and in-situ burning) with mechanical recovery with 22
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
25

two types of fines: $1,100 per barrel, and $4,300 per barrel. The costs have been sorted by spill scenario 1
(oil type and spill volume, 6 each) and ordered from least to most expensive per unit oil spilled (top-to- 2
bottom and left-to-right). A complete list of the costs can be found in the Supplementary material. The 3
cost could be further reduced by the value of the recovered oil. However, even if the costs of processing 4
this recovered oil were not included, the market price of oil, about $100/bbl at the time of this writing, is 5
an order of magnitude less than the fine associated with spilling it. Thus, our cost estimate is slightly 6
conservative by not taking this value into account. 7
8
Figure 2: Graphical representation of the estimated cost per tonne (t) to remediate oil spilled in the open 9
ocean. Each square in the 6x6 grids represents the cost of an oil spill for a given type of oil and spill size, 10
as indicated in the blank Scenario Grid (upper left). For example, the bottom-left corner of every grid 11
estimates the cost, per weight of oil spilled, to remediate a spill of No. 4 or 5 fuel oil that was less than 34 12
t in size. Results were derived using a model developed by Etkin (2000). Included are two hypothetical 13
scenarios whereby the value of an averted fine is taken into account, reducing the cost (lower right). 14
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
26

The most apparent result is that all mechanical recovery operations with the larger fine averted 1
operate at a negative cost. Equivalently stated, this model always estimates that the cost of recovering oil 2
from an offshore spill with no shoreline oiling is less than the value of the fine associated with leaving 3
that spill in the environment. No non-recovery technique has a negative cost per unit oil cleaned, 4
including a no action alternative, which requires monitoring, modeling, and decision-making. Thus, this 5
cost model indicates that mechanical recovery with an averted fine is the most cost effective technique, 6
even without directly taking into account the socioeconomic and environmental damages prevented. 7
When the fine is at the lower value of $1,100 per barrel, this model also estimates negative costs 8
in 31 out of the 36 scenarios. These five remaining scenarios were under the most costly conditions, 9
having both small spill sizes and heavy oil types. However, even in these remaining scenarios, mechanical 10
recovery was still less expensive than dispersants in four instances, and in-situ burning twice. In addition, 11
these scenarios are examples of where the most cost-efficient technique determined by the model may not 12
be the one chosen by responders. Dispersants and in-situ burning are the least effective at remediating 13
small spills of heavy oils; the oils tend to emulsify and resist burning or dispersion. Thus, mechanical 14
recovery would likely be the method of choice based on effectiveness, which is not directly accounted for 15
in this cost model. 16
These rough cost estimations showcase the value of only imposing a fine for unrecovered spilled 17
oil. Such flexible regulations, in combination with a competitive industry, have been shown to foster 18
technological innovation (Ford et al., 2014). In this case, by providing a specific financial incentive for 19
liable parties to remove as much oil from the environment as possible, in a manner that is cost effective in 20
a directly observable way, two goals are accomplished. First, potential damages from the spill are 21
minimized. The more oil that is removed from the environment, the less damage it causes. Second, oil 22
recovery decreases the uncertainty associated with impacts from the accident. As outlined in Section 1, 23
the exact numerical value of socioeconomic and environmental damages it not well known, and varies 24
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
27

greatly based on many factors unique to a given spill. This uncertain value often becomes the focus of 1
debates that decide how much compensation should be rewarded, and to whom (Yang et al., 2014). By 2
reducing the number of impacted parties, the uncertainty in assigning just compensation is diminished. 3
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
28

1
5. Conclusions 2
A review of the oil spill literature showcased the challenges in fully understanding an oil spill. 3
Oils are complex mixtures with varied bulk properties that substantially affect our remediation 4
approaches. Viscosity and tendency to emulsify (water content) are the most important properties to 5
identify for spill remediation efforts. In order to better assess the potential of different cleanup 6
technologies in large spills, the concept of a maximum oil encounter rate (MOER) was developed, which 7
reflects the ability to remediate oil under perfectly optimal conditions. Representative MOER values for 8
common spill technologies were calculated, showing that dispersants had a significantly higher MOER, 9
followed by skimming and in-situ burning. Determining the MOER for hydrophobic meshes, a new oil 10
recovery technology, highlighted the inability of the EDRC to assess the benefits gained from improving 11
the encounter rate of systems. 12
As a complementary assessment, a cost model was implemented to understand the cost- 13
effectiveness of a given remediation technique under various spill conditions. Mechanical recovery 14
methods were the most costly under various spill sizes and volume. However, if the penalty imposed per 15
unit of spilled oil were waived for any recovered oil, the cost of recovering the oil would be smaller than 16
the averted penalty. Care should be taken in waiving this penalty, since the components in oil vary in their 17
fate and impact. But, creating this financial incentive would encourage removal of the spilled oil from the 18
environment, reduce the uncertainty of its impact, and may spur the development of better recovery 19
technologies. To this end, the authors recommend a heightened effort to develop oil spill recovery 20
techniques, as well as additional research into understanding the fate of individual oil components. 21
22
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
29

1
Appendix Additional MOER Calculations for In-Situ Burning 2
Wicks (1969) derived a series of equations describing the geometry of towed boom. For a 3
catenary shaped section of boom, its length L and opening width W are related by 4
L = 2a sinh(wL / 2a) (A.1)
where a is a constant that reflects the steepness of the curve. If the ratio of boom length to opening width 5
is specified, the value of a can be found numerically. 6
The sag d of towed boom is the perpendicular distance from its apex to the open end. A boom 7
with a smaller opening width to length ratio will have a deeper curve, and thus a larger value of d. The 8
sag is given by: 9
d = a(cosh(L/2a) 1) (A.2)
The area enclosed by the boom can then be found from the sag, boom length, and opening width: 10
A = W(a + d) a L (A.3)
Note that the area A used in Section 4.2.3 is one third of the area calculated with Equation A.3. 11
12
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
30

1
References 2
Aeppli, C.; Carmichael, C.A.; Nelson, R.K.; Lemkau, K.L.; Graham, W.M.; Redmond, M.C.; Valentine, 3
D.L.; Reddy, C.M. 2012. Oil Weathering after the Deepwater Horizon Disaster Led to the Formation 4
of Oxygenated Residues. Environ. Sci. Technol. 46, 87998807. doi: 10.1021/es3015138. 5
Alan, A.A., 1991. Controlled Burning of Crude Oil on Water Following the Grounding Of the Exxon 6
Valdez. International Oil Spill Conference Proceedings: March 1991, Vol. 1991, No. 1, pp. 213-216. 7
doi: 10.7901/2169-3358-1991-1-213 8
Allen, A.A.; Ferek, R.J. 1993. Advantages and disadvantages of burning spilled oil. In Proceedings of the 9
1993 International Oil Spill Conference. 1, pp. 765-772. doi: 10.7901/2169-3358-1993-1-765 10
Allen, A.A.; Mabile, N.J.; Jaeger, D.; Costanzo, D. 2011. The Use of Controlled Burning during the Gulf 11
of Mexico Deepwater Horizon MC-252 Oil Spill Response. In Proceedings of the 2011 International 12
Oil Spill Conference. 1, pp. abs 194. doi: 10.7901/2169-3358-2011-1-194 13
Allen, A.A.; Dale, D.H.; Galt, J.A.; Murphy, J.A. 2012. EDRC Project Final Report (Under GSA 14
Contract GS-00F-0002W; BSEE Order # E12-PD-00012). www.bsee.gov/Research-and- 15
Training/Technology-Assessment-and-Research/Project-673.aspx (accessed November 1, 2013). 16
Amini, A.; Bollaert, E.; Boillat, J.L.; Schleiss, A.J. 2008. Dynamics of low viscosity oils retained by rigid 17
and flexible barriers. Ocean Eng. 35(14-15), 1479-1491. doi: 10.1016/j.oceaneng.2008.06.010 18
Arey, J.S.; Nelson, R.K.; Reddy, C.M. 2007. Disentangling Oil Weathering Using GCxGC 1. 19
Chromatogram Analysis. Environ. Sci. Technol. 41, 5738-5746. doi: 10.1021/es070005x 20
ASTM Standard F1413, 2013. Standard Guide for Oil Spill Dispersant Application Equipment: Boom 21
and Nozzle Systems," ASTM International, West Conshohocken, PA, USA, 2013, DOI: 22
10.1520/F1413 23
Atlas, R.M. 1995. Bioremediation of petroleum pollutants. Mar. Pollut. Bull. 31(4), 178-182. doi: 24
10.1016/0964-8305(95)00030-9 25
Beegle-Krause, C.J. 2001. General NOAA Oil Modeling Environment (GNOME): A New Spill 26
Trajectory Model. International Oil Spill Conference Proceedings Mar 2001, Vol. 2001, No. 2 (March 27
2001) pp. 865-871. doi: 10.7901/2169-3358-2001-2-865 28
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
31

Beens, J.; Brinkman, U.A.Th.; 2000. The role of gas chromatography in compositional analysis in the 1
petroleum industry. TrAC, Trends Anal. Chem. 19 (4), 260-275. doi: 10.1016/S0165-9936(99)00205- 2
8 3
Bergh, L. I. V.; Ringstad, A. J.; Leka, S.; Zwetsloot, G. I. J. M. Psychosocial risks and hydrocarbon leaks: 4
an exploration of their relationship in the Norwegian oil and gas industry, Journal of Cleaner 5
Production (2013), doi: 10.1016/j.jclepro.2013.09.040. 6
BLS, 2013. Consumer Price Index. www.bls.gov/cpi (accessed August 27, 2013). 7
Blumer, M.; Ehrhardt, M. Jones, J.H. 1973. Environmental fate of stranded crude oil. Deep-Sea Res. 8
20(3) 239-259. doi: 10.1016/0011-7471(73)90014-4. 9
Broje, V.; Keller, A.A. 2006. Improved Mechanical Oil Spill Recovery Using an Optimized Geometry for 10
the Skimmer Surface. Environ. Sci. Technol., 40 (24), 79147918. DOI: 10.1021/es061842m 11
Brynolf, S.; Fridell, E.; Andersson, K. Environmental assessment of marine fuels: Liquefied natural gas, 12
liquefied biogas, methanol and bio-methanol. J. Cleaner Prod. (2014), doi: 13
10.1016/j.jclepro.2014.03.052. 14
Burgherr, P. J. 2006. In-depth analysis of accidental oil spils from tankers in the context of global spill 15
trends from all sources. Hazard. Mater. 140, 245-256. 16
Camilli, R.; Di Iorio, D.; Bowen, A.; Reddy, C. M.; Techet, A. H.; Yoerger, D. R.; Whitcomb, L. L.; 17
Seewald, J. S.; Sylva, S. P.; Fenwick, J. 2011. Proc. Natl. Acad. Sci. U. S. A. 109 (50), 20260-20267. 18
Chun, C. and Park, J. 2001. Oil Spill Remediation Using Magnetic Separation. J. Environ. Eng., 127(5), 19
443449. doi: 10.1061/(ASCE)0733-9372(2001)127:5(443). 20
Dale, D.H., Allen, A., Broje, V. 2011. The Response Options Calculator (ROC). International Oil Spill 21
Conference Proceedings Mar 2011, Vol. 2011, No. 1 (March 2011) pp. abs179 doi: 10.7901/2169- 22
3358-2011-1-179 23
Daling, P.S.; Aamo, O.M.; Lewis, A.; Strom-Kristiansen, T. 1997. SINTEF/IKU Oil-weathering model: 24
predicting oils properties at sea. In International Oil Spill Conference Proceedings Apr 1997, Vol. 25
1997, No. 1 (April 1997) pp. 297-307. doi: 10.7901/2169-3358-1997-1-297. 26
Dave, D.; Ghaly, A.E.; 2011. Remediation Technologies for Marine Oil Spills: A Critical Review and 27
Comparative Analysis. Am. J. Env. Sci. 7 (5), 423-440 28
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
32

Deng, D.; Prendergast, D.P.; MacFarlane, J.; Bagatin, R.; Stellacci, F.; Gschwend, P.M. 2013. 1
Hydrophobic Meshes for Oil Spill Recovery Devices. ACS Appl. Mater. Interfaces. 5 (3), 774781. 2
doi: 10.1021/am302338x. 3
Di Toro, D.M.; McGrath, J.A.; Hansen, D.J. 2000. Technical basis for narcotic chemicals and polycyclic 4
aromatic hydrocarbon criteria. I. Water and Tissue. Environ. Toxicol. Chem. 19(8), 1951-1970. DOI: 5
10.1002/etc.5620190803 6
EPA, 2013. EPA Response to Enbridge Spill in Michigan. www.epa.gov/enbridgespill (accessed August 7
20, 2013). 8
Etkin, D. S., 2000. Worldwide analysis of marine oil spill cleanup cost factors. In Arctic and marine 9
oilspill program technical seminar Vol. 1, pp. 161-174. Environment Canada; 1999. 10
Evans, D.A., Walton, W., Baum H., Lawson, R., Rehm, R., Harris, R., Ghoniem, A., Holland, J. 1990. 11
Measurement of Large Scale Oil Spill Burns. Arctic and Marine Oil Spill Program Technical 12
Seminar, 13
th
. June 6-8, 1990, Edmonton. Alberta, Canda. pp-38 13
http://fire.nist.gov/bfrlpubs/fire03/PDF/f03157.pdf (accessed November 1, 2013). 14
Fan, T.; Buckley, J.S. 2002. Rapid and Accurate SARA analysis of medium gravity crude oils. Energy 15
Fuels. 16, 1571-1575. doi: 10.1021/ef0201228 16
Fay, J.A. 1971. Physical processes in the spread of oil on a water surface. In Proceedings, International 17
Oil Spill Conference Proceedings: June 1971, Vol. 1971, No. 1, pp. 463-467. doi: 10.7901/2169- 18
3358-1971-1-463 19
Fingas, M. 1995. Water-in-oil emulsion formation: A review of physics and mathematical modeling. Spill 20
Sci. Technol. Bull. 2, 5559. doi: 10.1016/1353-2561(95)94483-Z 21
Fingas, M., 2011. Oil Spill Science and Technology, first ed. Elsevier, Amsterdam. 22
Fingas, M.; Fieldhouse, B. 2012. Studies on water-in-oil products from crude oils and petroleum 23
products. Mar. Pollut. Bull. 64, 272-283. doi: 10.1016/j.marpolbul.2011.11.019. 24
Ford, J. A.; Steen, J.; Verreynne, M.-L. How environmental regulations affect innovation in the 25
Australian oil and gas industry: going beyond the Porter Hypothesis. J. Cleaner Prod. (2014), doi: 26
10.1016/j.jclepro.2013.12.062. 27
French-McCay, D.P.; 2004. Oil Spill Impact Modeling: Development and Validation. Environ. Toxicol. 28
Chem. 23(10), 2441-2456. DOI: 10.1897/03-382 29
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
33

French-McCay, D.; 2009. State-of-the-art and research needs for oil spill impact assessment modeling. In 1
Proceedings of the 32nd AMOP Technical Seminar on Environmental Contamination and Response, 2
Emergencies Science Division, Environment Canada, Ottawa, ON, Canada (pp. 601-653). 3
Hansen, HP. 1975. Photochemical degradation of petroleum hydrocarbon surface films on seawater. Mar. 4
Chem. 3, 183-195. DOI: 10.1016/0304-4203(75)90001-8. 5
Hong, S., Khim, J. S., Ryu, J., Park, J., Song, S. J., Kwon, B. O., Choi, K., Ji, K., Seo, J., Lee, S., Park, J., 6
Lee, W., Choi, Y., Lee, K.T., Kim, C.K., Shim, W.J., Naile, J.E., Giesy, J.P. 2012. Two Years after 7
the Hebei Spirit Oil Spill: Residual Crude-Derived Hydrocarbons and Potential AhR-Mediated 8
Activities in Coastal Sediments. Environ. Sci. Technol., 46(3), 14061414 DOI: 10.1021/es203491b. 9
Hossain, M.S.; Sarica, C.; Zhang, H.-Q.; Rhyne, L.; Greenhill, K.L. 2005. Assessment and Development 10
of Heavy-Oil Viscosity Correlations. In SPE/PS-CIM/CHOA International Thermal Operations and 11
Heavy Oil Symposium, Lecture 97907-MS. doi: 10.2118/97907-MS. 12
ITOPF, 2012. Oil spill tanker statistics: 2004, 2005. www.itopf.com/information-services/data-and- 13
statistics/statistics (accessed July 14, 2013). 14
Jewell, D. M.; Weber, J. H.; Bunger, J. W.; Plancher, H.; Latham, D. R. 1972. Ion-exchange, 15
coordination, and adsorption chromatographic separation of heavy-end petroleum distillates . Anal. 16
Chem. 44, 1391-1395. doi: 10.1021/ac60316a003. 17
Kontovas, C.A.; Psaraftis, H.N. 2008. Marine Environment Risk Assessment: A Survey on the Disutility 18
Cost of Oil Spills. In Society of Naval Architects and Marine Engineers (SNAME) Greek Sections 19
2nd International Symposium on Ship Operations, Management and Economics (pp. 18-19). 20
www.martrans.org:8093/docs/ws2009/Kontovas%20Psaraftis%20%20Disutility%20Cost%20of%20o 21
oi%20spills.pdf (accessed November 1, 2013). 22
Kontovas, C. A., Psaraftis, H. N., Ventikos, N. P. 2010. An empirical analysis of IOPCF oil spill cost 23
data. Mar. Pollut. Bull. 60(9), 1455-1466. doi: 10.1016/j.marpolbul.2010.05.010 24
Lee, R.F. 2003. Photo-oxidation and photo-toxicity of crude and refined oils. Spill Sci. Tech. Bull. 8(2), 25
157-162. DOI: 10.1016/S1353-2561(03)00015-X 26
Lees, J.E. 1993. Contingency planning, contractor requirements, and Oil Pollution Act of 1990 27
implementation. International Oil Spill Conference Proceedings Mar 1993, Vol. 1993, No. 1 (March 28
1993) pp. 51-56 doi: 10.7901/2169-3358-1993-1-51. 29
Lehr, W.; Jones, R.; Evans, M.; Simecek-Beatty, D.; Overstreet, R. 2002. Revisions of the ADIOS oil 30
spill model. Environ. Model. Softw. 17, 191-199. doi: 10.1016/S1364-8152(01)00064-0. 31
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
34

Liu, K.; Jiang, L. 2011. Metallic surfaces with special wettability. Nanoscale, 3, 825838. DOI: 1
10.1039/C0NR00642D 2
Liu, Z.; Phillips, J.B. 1991. Comprehensive Two-Dimensional Gas Chromatography using an On-Column 3
Thermal Modulator Interface. J Chromatogr. Sci. 29 (6), 227-231. doi:10.1093/chromsci/29.6.227 4
Mackay, D.; Matsugu, R.S. 1973. Evaporation rates of liquid hydrocarbon spills on land and water. Can. 5
J. Chem. Eng. 51, 434439. doi: 10.1002/cjce.5450510407 6
Marshall, A.G.; Rodgers, R.P. 2008. Petroleomics: Chemistry of the Underworld. Proc. Natl. Acad. Sci. 7
U. S. A. 105 (47), 18090-18095. doi: 10.1073/pnas.0805069105 8
McGrath, J. A.; Parkerton, T. F.; Hellweger, F. L.; Di Toro, D.M. 2005. Validation of the narcosis target 9
lipid model for petroleum products: gasoline as a case study. Environ. Toxicol. Chem. 24(9), 2382- 10
2394. DOI: 10.1897/04-387R.1. 11
McKenna, A.M.; Nelson, R.K.; Reddy, C.M.; Savory, J.J.; Kaiser, N.K.; Fitzsimmons, J.E.; Marshall, 12
A.G.; Rodgers, R.P. 2013. Expansion of the Analytical Window for Oil Spill Characterization by 13
Ultrahigh Resolution Mass Spectrometry: Beyond Gas Chromatography. Environ. Sci. Technol. 47 14
(13), 75307539. doi: 10.1021/es305284t. 15
NOAA, 2013. Response Tools: GNOME. response.restoration.noaa.gov/oil-and-chemical-spills/oil- 16
spills/response-tools/gnome.html (accessed August 15, 2013). 17
Peterson, C.H.; Rice, S.D.; Short, J.W.; Esler, D.; Bodkin, J.L.; Ballachey, B.E.; Irons, D.B. 2003. Long- 18
term ecosystem response to the Exxon Valdez Oil Spill. Science 302, 2082. DOI: 19
10.1126/science.1084282 20
Rauffleta, E.; Cruzb, L. B.; Bres, L. An assessment of corporate social responsibility practices in the 21
mining and oil and gas industries, J. Cleaner Prod. (2014), doi: 10.1016/j.jclepro.2014.01.077. 22
Reddy, C.M.; Arey, J.S.; Seewald, J.S.; Sylva, S.P.; Lemkau, K.L.; Nelson, R.K.; Carmichael, C.A.; 23
McIntyre, C.P.; Fenwick, J.; Ventura, G.T.; Van Mooy, B.A.S.; Camilli, R. 2012. Composition and 24
fate of gas and oil released to the water column during the Deepwater Horizon oil spill. Proc. Natl. 25
Acad. Sci. U. S. A. 109 (50), 20229-20234. doi: 10.1073/pnas.1101242108. 26
Reed, M.; Johansen, ; Brandvik, P.J.; Daling, P.; Lewis, A.; Fiocco, R.; Mackay, D.; Prentki, R. 1999. 27
Oil spill modeling towards the close of the 20th century: overview of the state of the art. Spill Sci. 28
Technol. Bull. 5(1), 3-16. doi: 10.1016/S1353-2561(98)00029-2 29
Schulze, R. 1998. Oil spill response performance review of skimmers. ASTM Manual Series MNL34, 30
Scranton, PA, USA. DOI: 10.1520/MNL34-EB 31
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
35

Shi, Q.; Hou, D.; Chung, K.H.; Xu, C.; Zhao, S.; Zhang, Y. 2010. Characterization of Heteroatom 1
Compounds in a Crude Oil and Its Saturates, Aromatics, Resins, and Asphaltenes (SARA) and Non- 2
basic Nitrogen Fractions Analyzed by Negative-Ion Electrospray Ionization Fourier Transform Ion 3
Cyclotron Resonance Mass Spectrometry. Energy Fuels. 24, 2545-2553. doi: 10.1021/ef901564e 4
Tcaciuc, A.P.; Nelson, R.K.; Reddy, C.M.; Gschwend, P.M. 2012. Estimating Phospholipid Membrane 5
Water Partition Coefficients Using Comprehensive Two-Dimensional Gas Chromatography. Environ. 6
Sci. Technol.46 (6), 34493456. DOI: 10.1021/es203792e 7
USCG, 1997. Determining and Evaluating Required Response Resources for Vessel Response Plans, 8
33CFR154 and 33CFR155, Washington, D.C., U.S. Government Printing Office, 1997. 9
Vanem, E.; Endresen, .; Skjong, R. 2008. Cost-effectiveness criteria for marine oil spill preventive 10
measures. Reliab. Eng. Syst. Saf. 93, 1354-1368. doi:10.1016/j.ress.2007.07.008. 11
Ventura, G.T.; Kenig, F.; Reddy, C.M.; Frysinger, G.S.; Nelson, R.K.; Van Mooy, B.; Gaines, R.B. 2008. 12
Analysis of unresolved complex mixtures of hydrocarbons extracted from Late Archean sediments by 13
comprehensive two-dimensional gas chromatography (GCxGC). Organic Geochemistry. 39(7), 846- 14
867. doi: 10.1016/j.orggeochem.2008.03.006. 15
Ventikos, N. P.; Vergetis, E.; Psaraftis, H. N.; Triantafyllou, G. J. 2004. A high-level synthesis of oil spill 16
response equipment and countermeasures. J. Hazard. Mater. 107, 5158. 17
doi:10.1016/j.jhazmat.2003.11.009. 18
Weber, C.I., ed. 1993. Methods for measuring the acute toxicity of effluents and receiving waters to 19
freshwater and marine organisms, fourth ed. Environmental Monitoring Systems Laboratory, Office 20
of Research and Development, US Environmental Protection Agency, Cincinnati, OH, USA. 21
Wicks, M. 1969. Fluid Dynamics Of Floating Oil Containment By Mechanical Barriers In The Presence 22
Of Water Currents. International Oil Spill Conference Proceedings: December 1969, Vol. 1969, No. 23
1, pp. 55-106. doi: 10.7901/2169-3358-1969-1-55 24
Woolgar, L. 2008. Assessing the increasing risk of marine oil pollution spills in China. In Proceedings of 25
the 2008 International Oil Spill Conference. 1, pp. 711-715. doi: http://dx.doi.org/10.7901/2169- 26
3358-2008-1-711 27
Yang, D.; Kao, W. T. M.; Huang, N.; Wang, R.; Zhang, X.; Zhou, W. Process-based environmental 28
communication and conflict mitigation during sudden pollution accidents. J. Cleaner Prod. (2014), 29
doi: 10.1016/j.jclepro.2013.11.023. 30
Zelenke, B., C. O'Connor, C. Barker, C.J. Beegle-Krause, and L. Eclipse (Eds.). 2012. General NOAA 31
Operational Modeling Environment (GNOME) Technical Documentation. U.S. Dept. of Commerce, 32
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
36

NOAA Technical Memorandum NOS OR&R 40. Seattle, WA, USA: Emergency Response Division, 1
NOAA. 105 pp.Figure Captions 2
Figure 1: Field data of the viscosity and water content of crude oil released on the open sea. Water 3
temperature was 15C in the summer and 10C in the winter. Lines are added for visualization purposes 4
only. Data taken from Daling et al. (1997), who note that dispersant effectiveness is dramatically reduced 5
for this oil when its viscosity becomes greater than 4000 cP. 6
Figure 2: Graphical representation of the estimated cost per tonne (t) to remediate oil spilled in the open 7
ocean. Each square in the 6x6 grids represents the cost of an oil spill for a given type of oil and spill size, 8
as indicated in the blank Scenario Grid (upper left). For example, the bottom-left corner of every grid 9
estimates the cost, per volume of oil spilled, to remediate a spill of No. 4 or 5 fuel oil that was less than 10
34 t in size. Results were derived using a model developed by Etkin (2000). Included are two hypothetical 11
scenarios whereby the value of an averted fine is taken into account, reducing the cost (lower right). 12
13
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT

M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Supplementary Material
Assessing the Performance and Cost of Oil Spill Remediation Technologies:
Daniel P. Prendergast, Philip M. Gschwend

Results of cost estimation model, corresponding to Figure 2 in the text.

Oil Type (6) Spill Size (t) (6) Cleanup Method (3) Total Coefficient $$/t (US) $$/bbl (US)
Crude <34 Dispersants 0.109 $ 3,920.00 $ 534.00
Crude <34 In-situ Burning 0.0595 $ 2,130.00 $ 290.00
Crude <34 Mechanical 0.219 $ 7,830.00 $ 1,070.00
Crude <34 Mechanical F 0.251 $ 931.00 $ 127.00
Crude <34 Mechanical GF 0.251 $ (22,500.00) $ (3,070.00)
Crude > 34,000 Dispersants 0.000547 $ 19.60 $ 2.67
Crude > 34,000 In-situ Burning 0.000297 $ 10.60 $ 1.45
Crude > 34,000 Mechanical 0.00109 $ 39.20 $ 5.34
Crude > 34,000 Mechanical F 0.00126 $ (7,990.00) $ (1,090.00)
Crude > 34,000 Mechanical GF 0.00126 $ (31,500.00) $ (4,290.00)
Crude 1,700-3,400 Dispersants 0.00820 $ 294.00 $ 40.00
Crude 1,700-3,400 In-situ Burning 0.00446 $ 160.00 $ 21.80
Crude 1,700-3,400 Mechanical 0.0164 $ 587.00 $ 80.10
Crude 1,700-3,400 Mechanical F 0.0189 $ (7,410.00) $ (1,010.00)
Crude 1,700-3,400 Mechanical GF 0.0189 $ (30,900.00) $ (4,210.00)
Crude 3,400-34,000 Dispersants 0.00273 $ 97.90 $ 13.30
Crude 3,400-34,000 In-situ Burning 0.00149 $ 53.20 $ 7.25
Crude 3,400-34,000 Mechanical 0.00547 $ 196.00 $ 26.70
Crude 3,400-34,000 Mechanical F 0.00628 $ (7,850.00) $ (1,070.00)
Crude 3,400-34,000 Mechanical GF 0.00628 $ (31,300.00) $ (4,270.00)
Crude 340-1,700 Dispersants 0.0148 $ 529.00 $ 72.10
Crude 340-1,700 In-situ Burning 0.00803 $ 287.00 $ 39.20
Crude 340-1,700 Mechanical 0.0295 $ 1,060.00 $ 144.00
Crude 340-1,700 Mechanical F 0.0339 $ (6,850.00) $ (934.00)
Crude 340-1,700 Mechanical GF 0.0339 $ (30,300.00) $ (4,130.00)
Crude 34-340 Dispersants 0.0356 $ 1,270.00 $ 174.00
Crude 34-340 In-situ Burning 0.0193 $ 692.00 $ 94.30
Crude 34-340 Mechanical 0.0711 $ 2,540.00 $ 347.00
Crude 34-340 Mechanical F 0.0817 $ (5,140.00) $ (701.00)
Crude 34-340 Mechanical GF 0.0817 $ (28,600.00) $ (3,900.00)
Heavy Crude <34 Dispersants 0.129 $ 4,630.00 $ 631.00
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Heavy Crude <34 In-situ Burning 0.0703 $ 2,510.00 $ 343.00
Heavy Crude <34 Mechanical 0.259 $ 9,250.00 $ 1,260.00
Heavy Crude <34 Mechanical F 0.297 $ 2,570.00 $ 350.00
Heavy Crude <34 Mechanical GF 0.297 $ (20,900.00) $ (2,850.00)
Heavy Crude > 34,000 Dispersants 0.000646 $ 23.10 $ 3.15
Heavy Crude > 34,000 In-situ Burning 0.000351 $ 12.60 $ 1.71
Heavy Crude > 34,000 Mechanical 0.00129 $ 46.30 $ 6.31
Heavy Crude > 34,000 Mechanical F 0.00149 $ (7,990.00) $ (1,090.00)
Heavy Crude > 34,000 Mechanical GF 0.00149 $ (31,500.00) $ (4,290.00)
Heavy Crude 1,700-3,400 Dispersants 0.00970 $ 347.00 $ 47.30
Heavy Crude 1,700-3,400 In-situ Burning 0.00527 $ 189.00 $ 25.70
Heavy Crude 1,700-3,400 Mechanical 0.0194 $ 694.00 $ 94.60
Heavy Crude 1,700-3,400 Mechanical F 0.0223 $ (7,270.00) $ (991.00)
Heavy Crude 1,700-3,400 Mechanical GF 0.0223 $ (30,700.00) $ (4,190.00)
Heavy Crude 3,400-34,000 Dispersants 0.00323 $ 116.00 $ 15.80
Heavy Crude 3,400-34,000 In-situ Burning 0.00176 $ 62.90 $ 8.57
Heavy Crude 3,400-34,000 Mechanical 0.00646 $ 231.00 $ 31.50
Heavy Crude 3,400-34,000 Mechanical F 0.00743 $ (7,770.00) $ (1,060.00)
Heavy Crude 3,400-34,000 Mechanical GF 0.00743 $ (31,200.00) $ (4,260.00)
Heavy Crude 340-1,700 Dispersants 0.0175 $ 625.00 $ 85.20
Heavy Crude 340-1,700 In-situ Burning 0.00949 $ 339.00 $ 46.30
Heavy Crude 340-1,700 Mechanical 0.0349 $ 1,250.00 $ 170.00
Heavy Crude 340-1,700 Mechanical F 0.0401 $ (6,630.00) $ (904.00)
Heavy Crude 340-1,700 Mechanical GF 0.0401 $ (30,100.00) $ (4,100.00)
Heavy Crude 34-340 Dispersants 0.0420 $ 1,500.00 $ 205.00
Heavy Crude 34-340 In-situ Burning 0.0228 $ 817.00 $ 111.00
Heavy Crude 34-340 Mechanical 0.0840 $ 3,010.00 $ 410.00
Heavy Crude 34-340 Mechanical F 0.0966 $ (4,610.00) $ (629.00)
Heavy Crude 34-340 Mechanical GF 0.0966 $ (28,100.00) $ (3,830.00)
Light Crude <34 Dispersants 0.0636 $ 2,280.00 $ 311.00
Light Crude <34 In-situ Burning 0.0346 $ 1,240.00 $ 169.00
Light Crude <34 Mechanical 0.127 $ 4,560.00 $ 621.00
Light Crude <34 Mechanical F 0.146 $ (2,830.00) $ (386.00)
Light Crude <34 Mechanical GF 0.146 $ (26,300.00) $ (3,590.00)
Light Crude > 34,000 Dispersants 0.000318 $ 11.40 $ 1.55
Light Crude > 34,000 In-situ Burning 0.000173 $ 6.19 $ 0.84
Light Crude > 34,000 Mechanical 0.000636 $ 22.80 $ 3.11
Light Crude > 34,000 Mechanical F 0.000731 $ (8,070.00) $ (1,100.00)
Light Crude > 34,000 Mechanical GF 0.000731 $ (31,500.00) $ (4,300.00)
Light Crude 1,700-3,400 Dispersants 0.00477 $ 171.00 $ 23.30
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Light Crude 1,700-3,400 In-situ Burning 0.00259 $ 92.90 $ 12.70
Light Crude 1,700-3,400 Mechanical 0.00955 $ 342.00 $ 46.60
Light Crude 1,700-3,400 Mechanical F 0.0110 $ (7,700.00) $ (1,050.00)
Light Crude 1,700-3,400 Mechanical GF 0.0110 $ (31,200.00) $ (4,250.00)
Light Crude 3,400-34,000 Dispersants 0.00159 $ 56.90 $ 7.77
Light Crude 3,400-34,000 In-situ Burning 0.000865 $ 31.00 $ 4.22
Light Crude 3,400-34,000 Mechanical 0.00318 $ 114.00 $ 15.50
Light Crude 3,400-34,000 Mechanical F 0.00366 $ (7,920.00) $ (1,080.00)
Light Crude 3,400-34,000 Mechanical GF 0.00366 $ (31,400.00) $ (4,280.00)
Light Crude 340-1,700 Dispersants 0.00859 $ 308.00 $ 41.90
Light Crude 340-1,700 In-situ Burning 0.00467 $ 167.00 $ 22.80
Light Crude 340-1,700 Mechanical 0.0172 $ 615.00 $ 83.90
Light Crude 340-1,700 Mechanical F 0.0197 $ (7,330.00) $ (1,000.00)
Light Crude 340-1,700 Mechanical GF 0.0197 $ (30,800.00) $ (4,200.00)
Light Crude 34-340 Dispersants 0.0207 $ 740.00 $ 101.00
Light Crude 34-340 In-situ Burning 0.0112 $ 402.00 $ 54.90
Light Crude 34-340 Mechanical 0.0414 $ 1,480.00 $ 202.00
Light Crude 34-340 Mechanical F 0.0475 $ (6,370.00) $ (868.00)
Light Crude 34-340 Mechanical GF 0.0475 $ (29,800.00) $ (4,070.00)
No. 2 Fuel (Diesel) <34 Dispersants 0.0358 $ 1,280.00 $ 175.00
No. 2 Fuel (Diesel) <34 In-situ Burning 0.0195 $ 696.00 $ 95.00
No. 2 Fuel (Diesel) <34 Mechanical 0.0716 $ 2,560.00 $ 349.00
No. 2 Fuel (Diesel) <34 Mechanical F 0.0823 $ (5,120.00) $ (698.00)
No. 2 Fuel (Diesel) <34 Mechanical GF 0.0823 $ (28,600.00) $ (3,900.00)
No. 2 Fuel (Diesel) > 34,000 Dispersants 0.000179 $ 6.41 $ 0.87
No. 2 Fuel (Diesel) > 34,000 In-situ Burning 0.0000973 $ 3.48 $ 0.47
No. 2 Fuel (Diesel) > 34,000 Mechanical 0.000358 $ 12.80 $ 1.75
No. 2 Fuel (Diesel) > 34,000 Mechanical F 0.000411 $ (8,070.00) $ (1,100.00)
No. 2 Fuel (Diesel) > 34,000 Mechanical GF 0.000411 $ (31,500.00) $ (4,300.00)
No. 2 Fuel (Diesel) 1,700-3,400 Dispersants 0.00269 $ 96.10 $ 13.10
No. 2 Fuel (Diesel) 1,700-3,400 In-situ Burning 0.00146 $ 52.20 $ 7.12
No. 2 Fuel (Diesel) 1,700-3,400 Mechanical 0.00537 $ 192.00 $ 26.20
No. 2 Fuel (Diesel) 1,700-3,400 Mechanical F 0.00617 $ (7,850.00) $ (1,070.00)
No. 2 Fuel (Diesel) 1,700-3,400 Mechanical GF 0.00617 $ (31,300.00) $ (4,270.00)
No. 2 Fuel (Diesel) 3,400-34,000 Dispersants 0.000895 $ 32.00 $ 4.37
No. 2 Fuel (Diesel) 3,400-34,000 In-situ Burning 0.000486 $ 17.40 $ 2.37
No. 2 Fuel (Diesel) 3,400-34,000 Mechanical 0.00179 $ 64.10 $ 8.74
No. 2 Fuel (Diesel) 3,400-34,000 Mechanical F 0.00206 $ (7,990.00) $ (1,090.00)
No. 2 Fuel (Diesel) 3,400-34,000 Mechanical GF 0.00206 $ (31,500.00) $ (4,290.00)
No. 2 Fuel (Diesel) 340-1,700 Dispersants 0.00483 $ 173.00 $ 23.60
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
No. 2 Fuel (Diesel) 340-1,700 In-situ Burning 0.00263 $ 94.00 $ 12.80
No. 2 Fuel (Diesel) 340-1,700 Mechanical 0.00967 $ 346.00 $ 47.20
No. 2 Fuel (Diesel) 340-1,700 Mechanical F 0.0111 $ (7,700.00) $ (1,050.00)
No. 2 Fuel (Diesel) 340-1,700 Mechanical GF 0.0111 $ (31,200.00) $ (4,250.00)
No. 2 Fuel (Diesel) 34-340 Dispersants 0.0116 $ 416.00 $ 56.80
No. 2 Fuel (Diesel) 34-340 In-situ Burning 0.00632 $ 226.00 $ 30.90
No. 2 Fuel (Diesel) 34-340 Mechanical 0.0233 $ 833.00 $ 114.00
No. 2 Fuel (Diesel) 34-340 Mechanical F 0.0267 $ (7,110.00) $ (970.00)
No. 2 Fuel (Diesel) 34-340 Mechanical GF 0.0267 $ (30,600.00) $ (4,170.00)
No. 4/5 Fuel <34 Dispersants 0.362 $ 13,000.00 $ 1,770.00
No. 4/5 Fuel <34 In-situ Burning 0.197 $ 7,040.00 $ 960.00
No. 4/5 Fuel <34 Mechanical 0.724 $ 25,900.00 $ 3,530.00
No. 4/5 Fuel <34 Mechanical F 0.832 $ 21,700.00 $ 2,960.00
No. 4/5 Fuel <34 Mechanical GF 0.832 $ (1,760.00) $ (240.00)
No. 4/5 Fuel > 34,000 Dispersants 0.00181 $ 64.80 $ 8.83
No. 4/5 Fuel > 34,000 In-situ Burning 0.000984 $ 35.20 $ 4.80
No. 4/5 Fuel > 34,000 Mechanical 0.00362 $ 130.00 $ 17.70
No. 4/5 Fuel > 34,000 Mechanical F 0.00416 $ (7,920.00) $ (1,080.00)
No. 4/5 Fuel > 34,000 Mechanical GF 0.00416 $ (31,400.00) $ (4,280.00)
No. 4/5 Fuel 1,700-3,400 Dispersants 0.0272 $ 972.00 $ 133.00
No. 4/5 Fuel 1,700-3,400 In-situ Burning 0.0148 $ 528.00 $ 72.00
No. 4/5 Fuel 1,700-3,400 Mechanical 0.0543 $ 1,940.00 $ 265.00
No. 4/5 Fuel 1,700-3,400 Mechanical F 0.0624 $ (5,840.00) $ (796.00)
No. 4/5 Fuel 1,700-3,400 Mechanical GF 0.0624 $ (29,300.00) $ (4,000.00)
No. 4/5 Fuel 3,400-34,000 Dispersants 0.00905 $ 324.00 $ 44.20
No. 4/5 Fuel 3,400-34,000 In-situ Burning 0.00492 $ 176.00 $ 24.00
No. 4/5 Fuel 3,400-34,000 Mechanical 0.0181 $ 648.00 $ 88.30
No. 4/5 Fuel 3,400-34,000 Mechanical F 0.0208 $ (7,330.00) $ (999.00)
No. 4/5 Fuel 3,400-34,000 Mechanical GF 0.0208 $ (30,800.00) $ (4,200.00)
No. 4/5 Fuel 340-1,700 Dispersants 0.0489 $ 1,750.00 $ 239.00
No. 4/5 Fuel 340-1,700 In-situ Burning 0.0266 $ 951.00 $ 130.00
No. 4/5 Fuel 340-1,700 Mechanical 0.0977 $ 3,500.00 $ 477.00
No. 4/5 Fuel 340-1,700 Mechanical F 0.112 $ (4,050.00) $ (552.00)
No. 4/5 Fuel 340-1,700 Mechanical GF 0.112 $ (27,500.00) $ (3,750.00)
No. 4/5 Fuel 34-340 Dispersants 0.118 $ 4,210.00 $ 574.00
No. 4/5 Fuel 34-340 In-situ Burning 0.0639 $ 2,290.00 $ 312.00
No. 4/5 Fuel 34-340 Mechanical 0.235 $ 8,420.00 $ 1,150.00
No. 4/5 Fuel 34-340 Mechanical F 0.270 $ 1,610.00 $ 219.00
No. 4/5 Fuel 34-340 Mechanical GF 0.270 $ (21,900.00) $ (2,980.00)
No. 6 Fuel <34 Dispersants 0.141 $ 5,050.00 $ 689.00
M
A
N
U
S
C
R
I
P
T

A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
No. 6 Fuel <34 In-situ Burning 0.0768 $ 2,750.00 $ 375.00
No. 6 Fuel <34 Mechanical 0.282 $ 10,100.00 $ 1,380.00
No. 6 Fuel <34 Mechanical F 0.325 $ 3,550.00 $ 484.00
No. 6 Fuel <34 Mechanical GF 0.325 $ (19,900.00) $ (2,720.00)
No. 6 Fuel > 34,000 Dispersants 0.000706 $ 25.30 $ 3.45
No. 6 Fuel > 34,000 In-situ Burning 0.000384 $ 13.70 $ 1.87
No. 6 Fuel > 34,000 Mechanical 0.00141 $ 50.50 $ 6.89
No. 6 Fuel > 34,000 Mechanical F 0.00162 $ (7,990.00) $ (1,090.00)
No. 6 Fuel > 34,000 Mechanical GF 0.00162 $ (31,500.00) $ (4,290.00)
No. 6 Fuel 1,700-3,400 Dispersants 0.0106 $ 379.00 $ 51.70
No. 6 Fuel 1,700-3,400 In-situ Burning 0.00576 $ 206.00 $ 28.10
No. 6 Fuel 1,700-3,400 Mechanical 0.0212 $ 758.00 $ 103.00
No. 6 Fuel 1,700-3,400 Mechanical F 0.0243 $ (7,190.00) $ (981.00)
No. 6 Fuel 1,700-3,400 Mechanical GF 0.0243 $ (30,700.00) $ (4,180.00)
No. 6 Fuel 3,400-34,000 Dispersants 0.00353 $ 126.00 $ 17.20
No. 6 Fuel 3,400-34,000 In-situ Burning 0.00192 $ 68.70 $ 9.36
No. 6 Fuel 3,400-34,000 Mechanical 0.00706 $ 253.00 $ 34.50
No. 6 Fuel 3,400-34,000 Mechanical F 0.00811 $ (7,770.00) $ (1,060.00)
No. 6 Fuel 3,400-34,000 Mechanical GF 0.00811 $ (31,200.00) $ (4,260.00)
No. 6 Fuel 340-1,700 Dispersants 0.0191 $ 682.00 $ 93.00
No. 6 Fuel 340-1,700 In-situ Burning 0.0104 $ 371.00 $ 50.60
No. 6 Fuel 340-1,700 Mechanical 0.0381 $ 1,360.00 $ 186.00
No. 6 Fuel 340-1,700 Mechanical F 0.0438 $ (6,500.00) $ (886.00)
No. 6 Fuel 340-1,700 Mechanical GF 0.0438 $ (30,000.00) $ (4,090.00)
No. 6 Fuel 34-340 Dispersants 0.0459 $ 1,640.00 $ 224.00
No. 6 Fuel 34-340 In-situ Burning 0.0249 $ 893.00 $ 122.00
No. 6 Fuel 34-340 Mechanical 0.0918 $ 3,290.00 $ 448.00
No. 6 Fuel 34-340 Mechanical F 0.105 $ (4,290.00) $ (585.00)
No. 6 Fuel 34-340 Mechanical GF 0.105 $ (27,800.00) $ (3,790.00)