EPA 600/R-04/120
September 2004
Characteristics of Spilled Oils, Fuels, and Petroleum
Products: 3a. Simulation of Oil Spills and Dispersants
Under Conditions of Uncertainty
by
James W. Weaver
Ecosystems Research Division
National Exposure Research Laboratory
Athens, Georgia 30605
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Notice
The U.S. Environmental Protection Agency through its Office of Research and Development
funded and managed the research described here. It has been subjected to the Agency's peer and
administrative review and has been approved for publication as an EPA document. Mention of
trade names or commercial products does not constitute endorsement or recommendation for use.
11
-------
Foreword
The National Exposure Research Laboratory's Ecosystems Research Division (ERD) in Athens,
Georgia, conducts research on organic and inorganic chemicals, greenhouse gas biogeochemical
cycles, and land use perturbations that create direct and indirect, chemical and non-chemical
stresses, exposures, and potential risks to humans and ecosystems. ERD develops, tests, applies
and provides technical support for exposure and ecosystem response models used for assessing
and managing risks to humans and ecosystems, within a watershed / regional context.
The Regulatory Support Branch (RSB) conducts problem-driven and applied research, develops
technology tools, and provides technical support to customer Program and Regional Offices,
States, Municipalities, and Tribes. Models are distributed and supported via the EPA Center for
Exposure Assessment Modeling (CEAM) and through access to Internet tools
(www.epa.gov/athens/onsite).
At the request of the US EPA Oil Program Center, ERD is developing an oil spill model that
focuses on fate and transport of oil components under various response scenarios. This model
includes various simulation options, including the use of chemical dispersing agents on oil slicks.
The dispersant simulation is backed by empirical data on the effectiveness of dispersants and oil
composition and properties. The model is offered as a tool for oil spill response and planning.
Rosemarie C. Russo, Ph.D.
Director
Ecosystems Research Division
Athens, Georgia
in
-------
Oil Spill Report Series
A series of research reports is planned to present data and models for oil spill planning and
response. To date, these include:
1. Oil Composition
Zhendi Wang, B.P. Hollebone, M. Fingas, B. Fieldhouse, L. Sigouin, M. Landriault, P. Smith, J.
Noonan, and G. Thouin, 2003, Characteristics of Spilled Oils, Fuels, and Petroleum
Products: 1. Composition and Properties of Selected Oils, United States Environmental
Protection Agency, National Exposure Research Laboratory, EPA/600/R-03/072.
2. Dispersants
George Serial, Subhashini Chandrasekar, James W. Weaver, 2004, Characteristics of Spilled
Oils, Fuels, and Petroleum Products: 2a. Dispersant Effectiveness Data for a Suite of
Environmental Conditions - The Effects of Temperature, Volatilization, and Energy,
United States Environmental Protection Agency, National Exposure Research Laboratory,
EPA/600/R-04/119.
3. Simulation Models
James W. Weaver, 2004, Characteristics of Spilled Oils, Fuels, and Petroleum Products: 3a.
Simulation of Oil Spills and Dispersants Under Conditions of Uncertainty, United States
Environmental Protection Agency, National Exposure Research Laboratory, EPA/600/R-
04/120.
As more reports are added to the series, they may be found on EPA's web site at:
http://www.epa.gov/athens/publications.
IV
-------
Table of Contents
Notice ii
Foreword iii
Oil Spill Report Series iv
List of Figures vii
List of Tables ix
1. Introduction 1
2. Implementation of the Model 2
The Four Applications Contained Within the ERO3S Frame 2
Oil Spreading 5
Floating Oil 7
Drift due to Winds and Currents 14
Mass Conservation 14
Prototype Equations for Non-Weathering Oils 16
Prototype Equations for Weathering Oils 17
Dispersal of Oil 18
Simulation of Oil Slicks 19
Uncertainty Analysis 20
Example 22
3. User's Guide 24
Applet versus Application 24
Software Requirements 24
Basic Interface Options 24
4. Conclusions 31
References 32
Appendices 34
Acronyms 34
Appendix: Latitude-Longitude Coordinates in ERO3S 35
Appendix: Model Development Within the MDP Model Development Platform 38
-------
Procedural Outline for Creating MDP Applications 38
Directory and Package Structure 38
Appendix: Serial et al. (2004) Dispersant Data 42
Appendix: Wang et al. (2003) Physical Properties and Chemical Composition of Alaska North
Slope Crude Oil 55
Appendix: Wang et al, (2003) Physical Properties and Chemical Composition of South
Louisiana 67
Appendix: Wang et al. (2003) Physical Properties and Chemical Composition of Fuel Oil No.
2/Diesel 79
VI
-------
List of Figures
Figure 1 ERO3S model selection screen 2
Figure 2 Model implementation schematic showing user inputs and data, simulations, and the
numerical solver. Dashed lines indicate inputs and solid lines represent results. The
multiple circles or ellipses indicate separate options within the model 3
Figure 3 Computed radius (top) and height of an oil slick caused by the release of 11,000,000
gallons of Alaska North Slope crude oil. The formula used assumes an instantaneous
release of the oil and is based on order of magnitude estimates 10
Figure 4 Spacial concept used in the proposed model 15
Figure 5 Relationship of uncertainty to model data availability 21
Figure 6 Required options for use of ERO3S in Microsoft Internet Explorer 25
Figure 7 Introductory ERO3S screen 27
Figure 8 ERO3S identification screen 28
Figure 9 ERO3S model selection screen 28
Figure 10 ERO3S model screen example 29
Figure 11 Illustration of latitude-longitude input for ship location and locations of shore line
points 35
Figure 12 Shorelines entered from lat-long data. The top shoreline's coordinates are shown in
Figure . The ship itself is barely visible between the two shorelines, but can be seen in
the close up at right 35
Figure 13 Latitude-Longitude calculation in ERO3S 36
Figure 14 Directory structure of the ERO3S MDP application. The two first level directories,
ERO3S and models]avabase, are the locations that contain the code that is unique to
ERO3S (ERO3S) and the MDP framework (modelsjavabase) 39
Figure 15 Directory structure of the Java MDP showing the required subdirectories: bak, classes,
doc, package cache, and src. An src subdirectory is created by the MDP user for each
new model and must contain additional subdirectories: gov, epa, first level name (here:
modelsjavabase), and second level names (here: awtextend, doc, fw2, etc) 39
Figure 16 Estimated vs Measured % Dispersal of PBC with No Dispersant 44
Figure 17 Estimated vs Measured % Dispersal of 2FO with No Dispersant 44
Figure 18 Estimated vs Measured % Dispersal of SLC with No Dispersant 44
Figure 19 Estimated vs Measured % Dispersal of PBC with Dispersant "A" 44
Figure 20 Estimated vs Measured % Dispersal of 2FO with Dispersant "A" 44
Figure 21 Estimated vs Measured % Dispersal of SLC with Dispersant "A" 44
Figure 22 Estimated vs Measured % Dispersal of BPC with Dispersant "B" 44
Figure 23 Estimated vs Measured % Dispersal of 2FO with Dispersant "B" 44
Figure 24 Estimated vs Measured % Dispersal of SLC with Dispersant "B" 44
Figure 25 Comparison of regression equations (curves) against measured Prudhoe Bay Crude/no
dispersantefficiency 46
Figure 26 Comparison of regression equations (curves) against measured Prudhoe Bay
Crude/dispersant "A" efficiency 46
vn
-------
Figure 27 Comparison of regression equations (curves) against measured Prudhoe Bay
Crude/dispersant "B" efficiency 47
Figure 28 Comparison of regression equations (curves) against measured South Louisiana
Crude/No Dispersant efficiency 48
Figure 29 Comparison of regression equations (curves) against measured South Louisiana
Crude/Dispersant "A" efficiency 49
Figure 30 Comparison of regression equations (curves) against measured South Louisiana
Crude/Dispersant "B" efficiency 50
Figure 31 Comparison of regression equations (curves) against measured No. 2 Fuel Oil/No
Dispersant efficiency 51
Figure 32 Comparison of regression equations (curves) against measured No. 2 Fuel
Oil/Dispersant "A" efficiency 52
Figure 33 Comparison of regression equations (curves) against measured No. 2 Fuel
Oil/Dispersant "B" efficiency 53
Vlll
-------
List of Tables
Table 1 Example values of surface tension, interfacial tension and spreading forces for a suite of
oils (Environment Canada, 1999) 6
Table 2 Density and viscosity of selected oils (Environment Canada, 1999, (a)2004) 12
Table 3 Variable parameters for uncertainty analysis example 22
Table 4 Example uncertainty analysis results 23
Table 5 Coefficients of Regression Equations with Terms Determined by Step-Wise Linear
Regression 43
IX
-------
1. Introduction
The EPA Research Object-Oriented Oil Spill (ERO3S) model has been developed as a
Federal-employee lead effort to develop a model that
1. Is in the public domain,
2. Is developed and understood by EPA employees,
3. Where EPA's commitment transcends limited-term contracting/grant/cooperative-
agreement/Interagency Agreement (IGA) vehicles,
4. can be modified to meet needs identified by EPA Regional Offices, and
5. that includes state-of-the-art concepts in modeling note especially the uncertainty
analysis capability and implementation in an object-oriented language.
The model was designed using concepts of object-oriented programming. Although this may
seem to be a programming detail, object-oriented software design is viewed as enhancing
maintainability and re-use of computer code. The approach taken for ERO3S naturally has these
features, but two others are important: 1) by implementing the model in Java, a version of the
code can run as a client-side applet from a web page and 2) the oil spill problem naturally aligns
with object-oriented programming.
The last point bears amplification. Oil spills are commonly observed to be "patchy." Oil
does not spread as a uniform pancake across the surface of the water. It is broken, rather, into
separate bodies of oil. These, however, share common characteristics: they are all composed of
oil, they all spread according to the same basic physics, they weather according to the same
physio-chemical processes. For example, the differences between patches depend more on
factors of time-in-water, current variations or shorelines encountered, and direct application of a
dispersant to part of the whole spill. From these attributes, common behavior can be
programmed for a single generic oil slick and individual differences among slicks essentially
become differences in their data. Thus transport of a patchy oil slick aligns closely with an
object-oriented approach, more so than many other problems.
-------
2. Implementation of the Model
The EPA Research Object-Oriented Oil Spill (ERO3S) was designed using object-
oriented programming concepts. Hence the inclusion of the phrase "Object-Oriented" in the title.
Use of common components (classes) is the software implementation feature that allows the
parts of each model to be recombined and reused in the more complex applications. These same
ideas are used in the ERO3S frame to provide components of the graphical user interface (GUI).
These features are mostly invisible to the model users, but the repetition of model input or output
screens is evidence of component reuse.
ERO3S
Run hteBBTicatKm
Select a Model
Select one of the oil spill models/data sets
Select Model/Test Problem;
Empirical Dispsrsant Data
Laboratoiy Flask Simulation
Patchy Oil Slick Model
Uncertain Patchy Oil Slick
,,
Select a model before pushing 'run'
Pause
Resume
Stop
Figure 1 ERO3S model selection screen.
The Four Applications Contained Within the ERO3S Frame
Figure 1 indicates four choices of models within the ERO3S frame. These present various
oil spill research results and increase in complexity following the sequence:
-------
1. Empirical Dispersant Data. Data from laboratory experiments on dispersion of oil in
the US EPA baffled flask test (BFT) are presented in a graphical form.1
2. Laboratory Flask Simulation. A simulation of the baffled flask test data is performed
as if dispersant is added at specified times during the simulation. Data on oil composition
and weathering are also used in the simulation.
3. Patchy Oil Slick Model. Because most observers note that oil slicks are patchy or
broken up into small bodies of oil, an oil slick model is implemented that simulates
spreading of small patches of oil that compose the entire oil slick. These sub-slicks are
independently spread, weathered, transported and dispersed with dispersant.
Components from each of the prior applications and a model of spreading and transport
of a single slick are used in the patchy oil slick model.
4. Uncertain Patchy Oil Slick Model. All model input parameters have associated levels
of uncertainty; some are not measured or even measurable. For example, the release rate
or volume is seldom known, the oil composition varies, the wind and current magnitude
and directions change or are only approximately known. Consequently, a modeling
analysis should include evaluation of uncertainties. In ERO3S, uncertainty is included in
the calculation by a simple method: ranges of several input parameters are selected, then
the model is run for all combinations of these inputs. Extreme values of a set of model
outputs are generated and given as outputs. This approach gives the model user an
indication of the uncertainty in model results, given the specified ranges of model inputs.
JEPA currently requires results from a swirling flask test for placement of products on the
Subpart J products list. The baffled flask test has been proposed as a better alternative to the
swirling flask test.
-------
Oil
Composition
Data
DIspersant lnPut
DitaStt Parameter
Values
Graphics
Laboratory Routlrws
Flask HO Spec
Simulation §. Results
Single Oil Slick
3ohf.rT.it R,P1W1 Differential
_ .. _.._,. , RKF1(2) Equation
Patchy Oil Slick RKFip) Solver
Figure 2 Model implementation schematic showing user inputs and data, simulations, and the
numerical solver. Dashed lines indicate inputs and solid lines represent results. The multiple
circles or ellipses indicate separate options within the model.
An outline of the model structure is contained in an Appendix that describes the Model
Development Platform (MDP) that was created for this work. MDP contains functions for all the
required elements of the model: the user interface, graphics, numerical input/output, numerical
solvers, and utility routines to execute the model. The model itself is created as a specialized
subset of the MDP, which is combined with the generic capability to become the model (i.e.,
ERO3S).
Figure 2 illustrates the internal structure of the MDP that allows the four options
described above. Multiple circles or ellipses on the figure indicate separate choices or options
that are available under certain circumstances. The dashed lines indicate internal passage of data.
User input is collected on input screens (Input Parameter Values) and combined with built-in
data on oil composition and dispersant effectiveness. These are passed to one of several models:
Single Oil Slick, Patchy Oil Slick or Laboratory Flask Simulation. Of these, the first two use an
ordinary differential equation solver to generate the solution. The Laboratory Flask Simulation,
-------
being a set of simple calculations requires no numerical solver. The solver uses an explicit
Runge-Kutta method. The specific method is selectable from a number of choices that include
embedded time stepping control, implemented through Runge-Kutta Felhberg (RKF) algorithms
(Fehlberg, 1969). The coefficients are represented by the ellipse indicating RKF 1(2) and
RKF2(3) methods. These methods allow for automated time-step controlling through the
embedding of a more accurate higher-order method within a lower-order method. RKF 1(2), for
example, uses a first order method to generate its solution and a second order method to estimate
error and thus modify time steps, if necessary (Fehlberg, 1969). Methods of up to seventh order
are available (RFK7(8)) in the framework, but rarely necessary (Hairer et al, 1993). These
procedures allow the model to increase or decrease the time step as needed to maintain accuracy
of the results.
The structure also allows for time step changes that are required by the characteristics of
the problem. For example, the model should have a time step that begins at the time(s) that
dispersant is applied. Such a happening in the model is called a "solution" event and the model
time steps are adjusted so that these occur at the beginning/ending of a time step.
Oil Spreading
In any of the oil spreading applications in ERO3S, the same basic physics apply; either if
the slick is composed of a single pancake or a set of individual slicks. The following section
describes the spreading algorithms.
The tendency for oil to spread is given by (e.g., Canevari, 1969)
F = °W - °o - ° (1)
where F is the spreading force, ow is the surface tension of water, o0 is the surface tension of oil
and oow is the oil/water interfacial tension. Table 1 lists surface and interfacial tensions for a
number of crude oils and petroleum products. The estimated spreading force for salt and fresh
water is given in the last two columns (assuming that the surface tension of water is 65 dyne/cm).
Of all the oils listed, only Jet A-l has a negative spreading force. The surface and interfacial
tension data suggest that all other of these oils tend to spread over the surface of water, at least as
long as the surface and interfacial tensions are unchanged by weathering.
Observation of oil spills provide insight on the minimum oil thickness that may occur.
For example, the US Coast Guard Fact Sheet on Small Diesel Spills (500-5000 gallons) states
that heavy diesel sheens contain about 1000 gallons per square nautical mile of surface and that
silver sheens contain about 75 gallons per square nautical mile. These observations imply
thicknesses of 1.18 x 10"2 mm down to 8.91 x 10"4mm. Reports on slick thicknesses suggest
that a bounding thickness is reached that may be due to surface tension changes (Fay, 1969).
-------
The model needs to account for the thickness of floating oil bodies, because at a
minimum the volume and thickness are prime determinants of the area of contact between the
slick and the water, and between the slick and the air. Both of these areas of contact are
important for determining the mass transfer rate of chemicals into these media. The observations
on lens thickness and surface tensions suggest three interrelated components of spreading:
transient, short-lived, thick oil layers
surface tension driven spreading of thin layers
weathering induced changes in surface and interfacial tension
The following section provides a simplified methodology for including the first two of these
components. Because of the complexity of interactions during weathering, the effect of
weathering on surface and interfacial tension is included empirically through data collected on oil
composition and properties (Wang et al., 2002).
-------
Fuel or Grade Oil
Santa Clara
Prudhoe Bay
Arabian Medium
Alaska North Slope
Kuwait
Iranian Heavy
Alberta Sweet Blend
Louisiana
West Texas Intermediate
Barrow Island
Diesel Fuel
Aviation Gas 100
JetB
JP-4
JetA-1
Leaded Gasoline
Surface
Tension
dyne/cm
28.7
28.3
27
28.1
27.8
26.1
25.6
25.9
26.6
26.2
26.5
20
23
22.8
26
19.8
Oil/Salt Water
Interfacial
Tension
dyne/cm
23.3
9.7
20.8
27.4
22.9
22.5
8.4
19.6
18.9
15.9
28
42 2
10.8
17
38.4
18.6
Oil/Fresh
Water
Interfacial
Tension
dyne/cm
25.7
16.9
21.7
29.4
28.6
22.5
21.5
21.1
19.1
18.1
29.4
42 2
12.4
36
40.4
18
Spreading Force
dyne/cm
Salt
Water
13.1
27.0
17.2
9.5
14.3
16.4
31.0
19.5
19.5
22.9
10.5
2.8
31.2
25.2
0.6
26.6
Fresh
Water
10.6
19.8
16.3
7.5
8.6
16.4
17.9
18.0
19.3
20.7
9.1
2.8
29.5
6.2
-1.4
27.2
Table 1 Example values of surface tension, interfacial tension and spreading forces for a suite of
oils (Environment Canada, 1999).
Floating Oil
Mass and momentum conservation for the oil as a separate phase determines how oil
floats on the water surface. Three stages have been identified in the spreading of an oil slick
(Fay, 1969, Hoult, 1972, Buckmaster, 1973, U.S. Coast Guard, 1994). In the first stage gravity
and inertial forces control the spreading of oil across the surface. In the second stage, the inertial
forces become negligible in comparison with viscous drag across the surface. In the third stage
interfacial forces become dominant and provide the driving force to propel spreading. Thus, at
equilibrium the floating oil could spread across the surface or it could form a lens. The
occurrence of these two types of behaviors depends on the relative magnitudes of the surface and
interfacial tension forces. Data presented in the table above show that for almost all of the oils in
-------
fresh and salt waters, the final stage would be characterized by spreading across the water,
because the values of the spreading force, F, are positive. This is, of course, true if the values in
the table well represent these oils and if weathering in stages 1 and 2 does not affect the surface
and interfacial tension values.
The most commonly used simple formulations for oil spreading were developed by Fay
(1969). The relations derived by Fay were intended to "estimate the order of magnitude of the
rate of spread of an oil slick on the surface of still water, i.e., water which is free of motions
induced by wind, wave, tidal currents." The analysis can clearly be seen to be only based on
order of magnitude estimates by Fay's omission of TI and other constants in formulas for volume
and area (equation 1 and 3 of Fay, 1969). Fay's discussion indicates his belief that gravity,
inertia, viscous and surface tension forces are dominant in causing spreading, even when subject
to winds, waves, and currents. Hoult (1972) describes a commonly held assumption that the
spread of an oil slick is composed of two parts, the first driven by winds and currents, and the
second due to the innate tendency of the oil to spread on even a calm surface. These two
physical phenomena are built into the ERO3S model. The necessity of this approach is fairly
obvious, as oil spilled into a calm environment would spread simply because of its innate
tendencies, while oil spilled into rapidly flowing water would spread both to its spreading
tendency and because of the action of wind and waves. The point being that the motion of the oil
may be dominated by one or the other of these two phenomena depending upon the situation.
Following Fay's analysis for spreading on calm seas, the gravity force is proportional to
the volume of the oil, V,
V = d2h~ d2h (2)
4
where d is the oil slick diameter and h is the slick height. Both d and h are assumed to vary with
time in this analysis. The gravity force per unit volume of oil is proportional to
(3)
where Ap is the difference in density between water and oil, and g is the acceleration of gravity.
The net effect of surface tension is to spread the oils listed in Table 1. The magnitude of the
surface tension force is given by
ad
(4)
V
where o is the net surface tension. Comparing the gravity and surface tension force expressions
shows that for a fixed volume, V, the magnitude of the gravity force will decrease with time as
-------
the diameter, d, increases; and that the magnitude of the surface tension force will increase with
diameter (time). Equating the two forces gives a critical diameter beyond which spreading is
dominated by surface tension. The critical diameter, dc, is proportional to
a
Inertial and viscous forces act to retard the spreading of the oil. The inertial force is the
product of density and acceleration
d
Viscous forces are proportional to
The ratio of viscous to inertia forces is proportional to
d2 tl/2 (8)
showing that inertial forces are dominant at early times (because at early times t, and thus t1/2 is
small, meaning that inertial forces are much greater than viscous forces).
Extending Fay's analysis by equating all four forces so that gravity and surface tension
are balanced by viscous and inertial resistance to flow gives an equation for the slick radius at all
times:
-41-^ = 0 «
t J yt/2
d
The surface tension term may be positive or negative, depending on the magnitude of the
spreading force, F. Thus the equation allows for surface tension to act against spreading. This
equation must be solved iteratively because it is nonlinear, but it combines all the mechanisms
used in Fay's analysis. By doing so, the original assumption that the three stages are driven by
mutually exclusive force pairs (gravity-inertia, gravity- viscous, and surface tension-viscous) is
relaxed. In Fay's analysis transition times would be calculated and an equation is used to
corresponds to a the appropriate force-pair is used. In the ERO3S approach the extended Fay
-------
equation applies at all times and the relative magnitudes of each term determine which process is
dominant.
Figure 1 shows the computed results for a spill similar to the Exxon Valdez. The oil was
assumed to have the same properties as the Alaska North Slope Crude described in Tables 1 and
2; and the volume was taken as 11,000,000 gallons. Data from spills, however, is needed to
estimate proportionality constants that have been ignored in the Fay analysis. The radius
estimate also shows the effects of assuming an instantaneous release-all eleven million gallons
were released instantaneously. Figure 3 shows that the initial increase in radius is very rapid,
followed by a continuing increase over time. The slick height decreases very rapidly to a
thickness of 10 cm and decreases to less than 0.1 mm before the end of one year. Despite its
limitations, this model shows that the lens increases in radius and decreases in thickness rapidly,
in agreement with generalized observations.
10
-------
400001
30000-
« 20000 H
w
10000
1E-005
1 2
Time (years)
Time(yeara)
Figure 3 Computed radius (top) and height of an oil slick caused by the
release of 11,000,000 gallons of Alaska North Slope crude oil. The
formula used assumes an instantaneous release of the oil and is based on
order of magnitude estimates.
11
-------
Fuel or Crude Oil
Santa Clara
Prudhoe Bay
Arabian Medium
Alaska North Slope
Kuwait
Iranian Heavy
Alberta Sweet Blend
Louisiana
West Texas Intermediate
Barrow Island
Diesel Fuel
Aviation Gas 100
JetB
JP-4
JetA-1
Leaded Gasoline
Density
g/cm3
0°C
0.93
0.92
0.89
0.90
0.88
0.89
0.85
0.86
0.85
0.85
0.84
0.73
0.77
0.77
0.82
0.75
15 °C
0.92
0.91
0.88
0.89
0.87
0.88
0.84
0.85
0.83
0.84
0.84
0.72
0.76
0.83
0.80
0.74
Dynamic
Viscosity
cP (or mPa s)
0°C
1278
577
59
23
90
43
47
15
15
4
4
1
1
1
2
0.8
15 °C
304
68
29
(a)12
22
20
9
8
7
2
2
1
1
1
1
0.6
Table 2 Density and viscosity of selected oils (Environment
Canada, 1999, (a)2004).
Fay's analysis also includes the release of oil at a fixed flow rate, Vf, into a flowing
current of velocity, u. The formula analogous to that given above, combining gravity, inertial,
viscous, and surface tension forces is
d2u
u pu
d\ - j"
X
= 0
(10)
where x is the distance down stream from the source.
12
-------
The coefficients which make the preceding relationships precise can be determined either
by experiment or by theoretical analysis. As an example of the latter, the Buckmaster (1973)
analysis begins with the governing equations for the stage 2 problem-gravity/viscous flow. The
governing equations for the oil are the mass and momentum conservation equations, given by
V du , x oh
and
dh d(qh)
+ -r^ = 0 (12)
dt dx
where h(x,t) is the thickness of the slick, q(x,t) the oil velocity, g is the acceleration due to
gravity, Ap is the fractional density difference between the oil and water, v is the kinematic
viscosity of the oil.
For the water
du du du d2u
dt dx dy dy2
and
du dv
+ = 0 (14)
ox oy
Buckmaster outlines a solution approach for computing the size and shape of the lens as a
function of time. Ultimately Buckmaster derived from this theory an expression for the radius of
slick, R, as
R(t) = 1.76^A/?0-25K0-3333V-0-125/0-375 (15)
where V is the volume of the slick. The leading coefficient (1.76) differed from a previously
observed empirical value of 1.5 (Hoult, 1972) by 15%.
13
-------
Drift due to Winds and Currents
In Hoult's conceptualization, the effects of winds and currents can be separated from the
spreading discussed above. In essence, the two sets of phenomena are superimposed upon each
other as the processes which cause spreading in calm seas are viewed as inexorable processes
also occurring when currents drive flow in a certain direction. Hoult (1972) estimated drift due
to the wind by equating shear stresses in the water and air and derived the result that
Uw = 0.03Ua (16)
where Uw is the wind-induced drift velocity and Ua is the wind speed. The coefficient 0.03
results from the differing densities of air and water. Further, the movement of the center of
mass, X, of the oil slick is given by
dx
= Uc + 0.03Ua (17)
where Uc is the velocity of the current itself. The importance of the last equation is that it
provides a simplified means of accounting movement of the slick caused by wind and currents.
Mass Conservation
The implementation of ERO3S is based upon the equations given in the following
summary. The two fundamental equations of fluid mechanics for incompressible flows are
conservation of mass and conservation of momentum. Conservation of mass is invoked in the
proposed model by
Conservation of momentum is included indirectly by using the extended Fay equation given
above
-0
t J Yf/2
The relationship between the volume, V, and characteristic planar dimension, d, is assumed to be
governed by a circle in the absence of winds and currents (Figure 4). Since winds and currents
will almost always be important for oil spills, this spatial relationship will be changed to an
14
-------
ellipse as the slick evolves (i.e., the circle is only an initial condition for the ellipse, which is
anticipated to be used in virtually every case). An ellipse has the form
where the coefficients a, and b describe the properties of the ellipse (in the same way that the
single coefficient, the radius, completely describes the properties of a circle). An circle is
centered in space at a single point, the center; while an ellipse is characterized by two focii. For
the oil spill problem, the release of oil will be assumed to occur at one focus of the ellipse. As
long as oil is being released the location of the release focus will remained fixed. Once the
release ends, the entire slick will be allowed to drift with wind and currents. Elongation and
drift will be calculated from the relationship
/= J (Uc + 0.03 £71* (21)
t=ov '
where 1 represents the length of the ellipse. Note that the initial length is equal to zero. As the
length of the ellipse increases with time, the volume of the slick is conserved and the shape is
governed by the extended Fay equation. The result will be that the slick increases in length and
the distance between the focii of the ellipse increases. Thus the shape of the lens is determined
by the effects of wind and current according to the relationship for spreading that relates gravity,
inertial, viscous and surface tension forces. Hoult's conception that the spreading occurs
independently of drift is incorporated into this proposed approach.
Figure 4 Spacial concept used in the proposed model. At
time of tj there has been no impact of wind nor currents
on the soil slick and its shape is circular. At t2 the slick
has elongated because of wind and/or waves. At time t3
the release has ended and the slick is drifting away from
the source. For most cases of interest the circular shape of
tj is expected only to be an initial condition of the model
and elongation is expected from the start.
15
-------
After the release ends the slick may drift away from the source. Thus the position of the
center of mass of the slick is given by
= x + U + 0.
.03Ua)dt (22)
where x is the location of the center of mass of the slick and xe is the center of mass position at
the end of the release.
Prototype Equations for Non-Weathering Oils
The prototype equations for the mathematical system are
D
= R <23>
where V is the oil volume and R is the rate of release of the oil. By treating the release as an
ordinary differential equation, the restrictions of Fay's analysis are relaxed: the release does not
have to be instantaneous nor continuous (equation 10). Since the ordinary differential equation
solver was designed to use automated time step control based upon Fehlberg's method
embedding approach (Hairer, 1993), the solver can be used also to step to specified times in the
solution. These are called solution events, and are used extensively in controlling the execution
of the ERO3S model. The oil spill ending time is the first of the important solution events
included in the model
The extended Fay equation (9) has the general form:
F
-------
_ _
dd dt dV dt ,,~
* = w (26)
dd
The two differential equations can be solve numerically for V(t) and d(t). The size of the lens
then determined from
dl da ( \
l=2a => -=2=(Uc + 0.03E/J (27)
and
(28)
The lens height, h, is determined from the ellipse volume V = h TI a b.
Prototype Equations for Weathering Oils
The prototype equations are extended to include the effects of weathering. The prototype
equations for the mathematical system become
D
= R <29>
where V is the oil volume and R is the rate of release of the oil.
The extended Fay equation (9) has the general form, if weathering effects on the oil
physical properties are included:
F(V(t\d(t\t±p(t\p(t\(j(t\v(t\t-g} = 0 (30)
where the volume, dimensions and physical properties (density, viscosity, surface tension) are
now all potentially functions of time. Therefore
DF df df dV df dd df d^p df dp df da df dju
- = + -- + -- + -- + + -- + = 0 (31)
Dt dt dV dt dd dt d^p dt dp dt d
-------
dd dt dV dt dt±p dt dp dt d
-------
%E = (A + B)]nt (36)
or
%E = (A + B)i(f) (37)
where %E is the weight percent evaporated, T is the sea surface temperature (°C), t is the time in
minutes, and A and B are coefficients that are determined experimentally, is used to approximate
volatilization and consequent property changes of the oils. As appear in the appendices, the
density, viscosity, surface and interfacial tensions vary with volatilization and data on these are
used in ERO3S to track changes in these properties. The data sets also include compositional and
emulsification data which can also be driven by volatilization of the oil. Volatilization then
provides a "master variable" that is used to access a variety of other empirical data on these oils.
Once the amount of weathering has been determined, the data developed by Sorial et al
(2004) are used with the temperature, energy level and dispersant type to determine the amount
of dispersal of the oil that is possible ( E(t,w,e) in equation 33). The Sorial data set contain data
for various volatilization weatherings (up to the maximum possible for the oil), temperatures
from 5 °C to 35 °C, speeds of rotation of the test apparatus that represent varying energy levels,
and three dispersant treatments. These treatments are two dispersants code-named "A" and "B"
and a set of data collected with no dispersant. These served both as controls for experiments
with "A" and "B," as well as providing data on natural dispersion of these oils.
After the oil is dispersed the slick dimension and total volume of floating and dispersed
oil are recalculated based upon the new volume of the slick or each slick if the patchy oil slick
model is used.
Simulation of Oil Slicks
Spreading and dispersal of oil slicks are two basic functions of the code. As noted by
Hoult (1972), transport with wind and waves are of equal importance. These are implemented in
the most flexible fashion in the patchy oil slick model and the subsequent discussion will focus
on this model.
The overall mass balance is tracked by solving equation 23, which has the effect of
simply adding up the mass of oil released2. The mass is divided in the model, however, into
2Even though this calculation could be performed simply as a summation of the oil
released, including it as an ordinary differential equation to solve forces the mass balance
19
-------
separate oil slicks which represent patches of the complete slick. This procedure provides the
flexibility to allow:
volatilization to change the composition of parts of the slick depending upon their time-
in-water,
different portions of the slick to move in response to variation in current or wind, and
the dispersal efficiency to vary both due to oil properties and the amount of dispersant
available.
The procedure that is followed for each individual patch is as follows:
the beginning and ending times of its release are determined,
the release rate is set,
the initial composition is set, and
a direction of transport and release point are assigned.
One thousand of these individual slicks are used to compose the entire oil slick. The spreading
of each slick is determined from one run of the single slick model. A single slick model is
executed for the oil, climatic conditions and release rate of the entire slick. This model is run
until a minimum thickness is achieved and the results are saved for use in the individual slicks of
the patchy oil slick model. The ERO3S model then tracks each of these slicks, determines their
changes in composition and oil properties, and when the time comes for dispersant application,
disperses a fraction of their oil in accordance with the properties of the oil and dispersant subject
to the limitation of dispersant availability.
Uncertainty Analysis
Models, though, are commonly viewed as useful tools for understanding contaminant
transport (Oreskes et al., 1994) and determining future risk (ASTM, 1995). The degree of
predictive capability of most transport models has, in fact, not been established. This follows
from the models' reliance on unmeasured input parameters and calibration to specific incidents.
Given that the values of all the parameters and the forcing function were known, and that the
assumptions behind the model were exactly met, the model equations could be solved for the
required outputs. In the real world, however, the values are not exactly known and because
response is the priority at oil spills, data collection for modeling or other purposes is generally
not undertaken. Time is also critical, as decisions must be made quickly to minimize
environmental impacts. This constraint alone imposes severe limitations on data collection.
Simple examples of limited knowledge are the release rate or volume, composition and
properties of the oil and climatic conditions controlling transport.
calculation to always be in synch with all other equations solved.
20
-------
Thus, models are more likely to provide a framework for understanding transport than for
predicting future exposure and risk. At an oil spill rapid response is required and sufficient data
is not likely to be collected for calibrating a model. How then should models be used in
situations where they can not or will not be calibrated? What are the plausible ranges of output
given uncertainty in inputs? Can worst case parameter sets be selected that always provide a
bound on plausible outcomes?
Figure 5 shows a conceptual relationship between uncertainty and data availability. With
small amounts of either measured input data or calibration data, the resulting model uncertainty
is high. Models may still be useful in these cases, but their uncertainty should be quantified so
that their results are not taken falsely as inerrant.
I vi-yi; Estimates and a few site-specific parameter values
Uncertainty i
I Model calibrated to heads/concentration data
I
; Calibrated model tested
I against, prediction
Data: Input Parameters or Calibration Data
Figure 5 Relationship of uncertainty to model data availability.
Approach
Several approaches to uncertainty analysis have been developed. Generally these require
knowledge of parameter values and their statistical distributions including correlations between
individual parameters. For the purpose of the ERO3S code it is presumed that data collection is
not sufficiently detailed to determine values for some of the parameters, let alone their statistical
distributions and correlations.
A widely-used alternative is to assume knowledge of the statistical properties by using
scientific literature values as substitutes. These approaches allow assignment of probabilities to
the various outcomes, but suffers from obvious lack of incident-specificity. Where results depend
strongly on assumed distributions, it is not possible to determine how much error is introduced
into the results from the distributions. Alternatively, if it is assumed only that plausible ranges of
input parameters are known, similar outcomes can be determined, but probabilities cannot be
assigned. Because of presumed lack of knowledge of certain parameters and their underlying
probability distributions, a method based on ranges of inputs was developed.
21
-------
In the ERO3S code, bounding values of selected parameters are input. The minimum and
maximum values of these determine the range of possibility for each of the inputs. Since the
patchy oil slick model contains the most realism, it is used to run all possible combinations of the
input parameters. Significant model outputs are defined and the minimum and maximum of
these are collected from the multiple runs of the patchy slick model.
In effect this approach presumes that the statistical distribution of each parameter is
uniform. This distribution is said to be useful "..when an expert is only willing/able to estimate
an upper and lower bound for a quantity..." and "...is used frequently in exposure assessment."
(Cullen and Frey, 1999, page 71).
As the initial approach to uncertainty analysis of the ERO3S model, six parameters were
assumed to be variable: source location, spill rate, spill duration, current speed, wind speed and
temperature. With this selection of inputs there are two values each for six parameters: the
minimum value and the maximum value. This leads to a total of 26 or 64 unique combinations of
parameters. This calculation highlights an assumption of this method: That each parameter value
is equally likely and can occur in combination with each other parameter value. In other words
that each parameter is uniformly distributed and uncorrelated. The outcomes of interest were
picked for this initial approach were the volume of floating oil, areal extent of the oil slick,
volume of potentially beached oil and volume of dispersed oil. Because differing parameter sets
produce the best and worst case results for each of these outputs, there generally is no generic
worst case parameter set: the worst case parameter set depends, rather, on the output of interest.
Example
An example uncertainty analysis is presented in the on-line users guide (see Section 3
User's Guide). In this example dispersant is applied at three times (4, 10 and 13 hours) after the
an oil spill begins. For each application, 1000 gallons of dispersant is available and is applied to
all floating oil. Table 3 shows the values of five variable parameters used in the analysis. The
spill location also varied by 2 minutes and 30 seconds of longitude.
22
-------
Parameter
Leak Rate (gal/day)
Duration (days)
Wind (knot)
Current Speed (m/s)
Temperature (°C)
Minimum Value
1000
1.0
1
0.001
1
Maximum Value
2000
1.5
5
0.005
5
Table 3 Variable parameters for uncertainty analysis example.
In this case the patchy oil slick model required roughly 90 seconds to execute for each
simulation and a total of 64 simulations were performed. Thus the total execution time is
approximately 1.5 hours for this example. From these three outputs were selected: volume of
floating oil, oil extent and volume of dispersed oil. For each of these the minimum and
maximum values were recorded by the model. Table 4 shows that the resulting range of these
outputs, for this case, can range by as much as an order of magnitude.
Outcome
Floating Oil
Extent
Dispersed Oil
Minimum
658 gal
3.78x 109ft2
3 17 gal
Maximum
2368 gal
4.08 x 109ft2
687 gal
Table 4 Example uncertainty analysis results.
23
-------
3. User's Guide
The user's guide for the ERO3S model is available on-line by following the links at the
EPA web site: http://www.epa.gov/athens/research/projects/eros.3 A short introduction is given
here.
Applet versus Application
The model is supplied in two forms: 1) a simplified version that runs as an "applet" from
the EPA web site, 2) the full version that runs as an "application" on the user's computer. The
applet version is limited in that it does not access databases, nor allow for saving of inputs or
outputs. Conversely, the application version will provide these functions and will be down
loaded from the EPA web site at http://www.epa.gov/CEAM.
Software Requirements
The applet version of ERO3S runs within browser windows. Either Microsoft's Internet
Explorer or Netscape Navigator are suitable. An additional requirement, however, is that the
Java Run-Time Environment is equal to that used to create ERO3S. The required version is Sun
Java 2 vl.4.2_05 or higher, which is available from http://java.com/en/index.jsp.
If Microsoft Internet Explorer is used, then the SUN Java plug-in must be selected as
shown in Figure 6.
Basic Interface Options
The interface contains three introductory screens that are shown in Figures 7, 8, and 9.
The identification screen (Figure 8) can be used to record general information concerning the
spill and simulation.
The first step toward running the model is to select the code (Figure 9). As noted above,
there are four choices in the ERO3S framework:
1. Empirical Dispersant Data Explore the character of the empirical dispersant data.
3 Temporarily and for the purposes of peer review of this document, access is granted
only to selected reviewers at
http://intranet.epa.gov/nerlintr/athens/research/projects/oilspills/index.html. Permanent access
will be provide at the address given above in the text.
24
-------
Laboratory Flask Simulation Perform a simulation of the baffled flask experiment with
dispersant application at varying times.
Patchy Oil Slick Model Simulate the movement of an oil slick composed of individual
patches and disperse at specified times
Uncertain Patchy Oil Slick Model Evaluate the uncertainty associated with some of the
input parameters of the patchy oil slick model.
The applications build in complexity and use components from lower-numbered models. A
realistic oil slick should be simulated with the forth or fifth options: (the patchy oil slick models)
as the prior models are too limited for most situations.
When the selection of one of the four has been made, more screens are added to the
interface. (See Figure 10, for an example.) These will include all necessary input and output
screens for the selected model. Changing the model selection causes the interface to reconfigure
itself for the newly-selected option. All previous information is lost when a new selection is
made. This effect, in fact, can be useful to assure that all old results have been cleared from the
model.
Since each model contains an example problem, the "Run" button can be pushed anytime
after the model selection has been made. (The button can also be pushed before a selection is
made, but there won't be any calculations made.) Changes to the model inputs are made on each
input screen, prior to running the model. There are three things to know about the current version
of the model:
1. This "applet" version of the model does not allow storage of data. This is a requirement
of running from the Intranet (or Internet). Later a PC version of the model will be
supplied that will allow storage and retrieval of data.
2. Most inputs in this version are selected by choosing from "drop-down" lists. The PC
version of the model will allow direct input of numerical values.
3. The source code for the model is about 500 pages long when printed and requires a
download of approximately 364 kbytes.
Specific instructions for each of the four ERO3S options are given in the on-line user's
guide.
25
-------
General Security Privacy j Content Connections j Programs Advanced
Settings:
Use inline AutoComplete «
Use Passive FTP (for firewall and DSL modem compatibility)
Use smooth scrolling
HIT TP 1.1 settings
UseHTTPU
Use HTTP 1.1 through prowy connections
va (Sun)
Use Java 2 v1.4.2_05 for
-------
ERO3S
Run
a
EPA's Oil
Laboratory
Office of and
Juiy,
Run
Payee
Resume
Stop
a "run1
Figure 7 Introductory ERO3S screen.
27
-------
Run Identification
Select a
Spill Name
Spill Location
Spill Date
User Name
Simulation
Comment
Run |
Select a model before pushing 'run1
Pause
Resume
Stop
Figure 8 ERO3S identification screen.
28
-------
ERO3S
Run Identification
Select a Model
Select one of the oil spill models/data
Select Model/Test Problem!
Run
Selects model before pushing 'run'
Empirical Dispersatrt Data
Laboratory Flask Simulation
Patchy Oil Slick
Uncertain Patchy Gii Siick Motlei;
Pause
Resume
Stop
Figure 9 ERO3S model selection screen.
29
-------
Run
Select a Model
Event
and
OH ft
one oil
on
Run
Stop
a "run1
Figure 10 ERO3S model screen example.
30
-------
4. Conclusions
EPA's Object-Oriented Oil Spill (ERO3S) model was developed to create a public-
domain model for the purpose of oil spill response planning. The model contains a suite of four
applications that range from an exploration of laboratory scale dispersant effectiveness data
through a patchy oil slick model to an uncertainty calculation that addresses fundamental and
irreducible limitations to oil spill model applications.
The primary focus of the work presented herein was to establish a framework for
developing and evaluating various formulations of oil spill models. Work is continuing in
several areas:
1. Inclusion of alternate weathering models and databases (including Stiver and MacKay,
1984 and Stiver et al. 1988),
2. Expansion of physio-chemical phenomena included in the model,
3. Inclusion of alternate spreading formulations (including those of Lehr et al., 1989 and
Buckmaster (1972): equation 13), and
4. Enhancement of the graphical user interface
Because of the established structure of the model, this work represents incremental enhancement.
Each feature that is to be included can be encapsulated as a separate object and used to modify
one or more of the existing four applications contained in ERO3S or used to create new
applications that add to the four.
Future work will address linking of the model to existing water quality and hydrodynamic
models, testing against spill data and inclusion of new approaches to estimating the impacts of
dispersants on oil slicks.
31
-------
References
Buckmaster, J., 1973, Viscous-gravity spreading of an oil slick, J. Fluid Mech, 59(3), 481-491.
Canevari, G.P., 1969, The role of chemical dispersands in oil cleanup, in Oil on the Sea, David.
P. Hoult, ed., Plenum Press, New York, 29-51.
Cullen, A. C., and H.C. Frey, 1999, Probabilistic Techniques in Exposure Assessment, Plenum
Press, New York, 335pp.
Environment Canada, 1999, A Catalogue of Crude Oil and Oil Product Properties,
Environmental Technology Center, Emergencies Science Division, Environment Canada,
www. etcentre.org.
Fay, J.A., 1969, The spread of oil slicks on a calm sea, in Oil on the Sea, David. P. Hoult, ed.,
Plenum Press, New York, 53-63.
Felhberg, E., 1969, Low-Order Classical Runge-Kutta Formulas with Stepsize Control and Their
Application to Some Heat Transfer Problems, NASA, Technical Report R-315.
Fingas, M., 2001, The Basics of Oil Spill Cleanup, Lewis Publishers, Boca Raton, 233 pp.
Hairer, E., S.P. Norsett, and G. Wanner, 1993, Solving Ordinary Differential Equations I,
Nonstiff Problems, Second Revised Edition, Springer Verlag, New York, 528 pp.
Hoult, 1972, Oil Spreading on the Sea, Annual Reviews of Fluid Mechanics, 4, 341-368.
Li, W.H. and S. H. Lam, 1976, Principles of Fluid Mechanics, Addison-Wesley, Reading,
Massachusetts, 374pp.
Meyer, B., 1997, Object Oriented Software Construction, 2ed, Prentice Hall, Upper Saddle
River, New Jersey, 1254 pp.
Oreskes, N, et al., 1994, Verification, Validation, and Confirmation of Numerical Models in the
Earth Sciences, Science, 263, 641.
Sorial, G., C., Chandrasekar, and J. W. Weaver, 2004, Characteristics of Spilled Oils, Fuels, and
Petroleum Products: 2a. Dispersant Effectiveness Data for a Suite of Environmental
Conditions - The Effects of Temperature, Volatilization, and Energy, United States
Environmental Protection Agency, National Exposure Research Laboratory, EPA/600/R-
04/xxx.
32
-------
Stiver, W., and D. Mackay, 1984, Evaporation rate of spills of hydrocarbons and petroleum
mixtures, Environmental Science and Technology, 18(11), 834-840.
Stiver, W., W.Y. Shiu, and D. Mackay, 1989, Evaporation times and rates of specific
hydrocabons in oil spills, Environmental Science and Technology, 23(1), 101-105.
U.S. Coast Guard, 1994, Adios Automated Data Inquiry for Oil Spills, Version 1.1, National
Oceanic and Atmospheric Administration, Hazardous Materials Response and
Assessment Division, Seattle, Washington, 98115
U.S. Environmental Protection Agency, 1993, Understanding Oil Spills and Oil Spill Response,
EPA540-K-93-003.
U.S. Environmental Protection Agency, 2003, Draft Guidance on the Development, Evaluation
and Application of Regulatory Environmental Models, Council on Regulatory
Environmental Modeling, Office of Science Policy, USEPA,
Wang, Z., B.P. Hollebone, M. Fingas, B. Fieldhouse, L. Sigouin, M. Landriault, P. Smith, J.
Noonan, and G. Thouin, 2003, Characteristics of Spilled Oils, Fuels, and Petroleum
Products: 1. Composition and Properties of Selected Oils, United States Environmental
Protection Agency, National Exposure Research Laboratory, EPA/600/R-03/072.
33
-------
Appendices
Acronyms
2FO
BFT
ERO3S
EPA
GUI
GPS
MDP
ODE
PBC
RKF
SLC
No. 2 Fuel Oil
Baffled Flask Test
EPA's Research Object-Oriented Oil Spill Model
Environmental Protection Agency
graphical user interface
global positioning system
Model Development Platform
ordinary differential equation
Prudhoe Bay Crude
Runge-Kutta-Felhberg
South Louisiana Crude
34
-------
Appendix: Latitude-Longitude Coordinates in ERO3S
Geographic locations are indicated in ERO3S with latitude-longitude4 coordinates.
Latitudes are measured from the equator and are assumed to increase positively to the North.
Longitudes are measured from the prime meridian and assumed to increase positively to the
West. So that the longitudes used in ERO3S are longitudes "west". Figure 11 shows the data-
entry screen for the locations of the stern of the ship (or other oil container) and 10 shoreline
points. These locations are presumed to be available from a global positioning system (GPS)
receiver and/or a nautical chart. The resulting shoreline is shown in Figure 12. These data show
the Great South Bay of Long Island New York between Oakdale and Moriches.
Degrees
Stern of Ship
Shore
Shore
Shore
Shore
Shore
Shore
Shore
Shore
Shore
Shore
Line Point
Line Point
Line Point
Line Point
Line Point
Line Point
Line Point
Line
Line Point
Line Point
-0
-1
-2
-3
-4
-5
-i
-7
-8
-9
No
No
J4Q
J40
No
No
No
No
J40
J40
J40
0
0
0
0
0
0
0
0
0
0
0
Minutes
JJ42.Q
JJ42.0
JI43.0
JI43.0
|l43.0
|l45.0
|l44.0
|l44.0
|l45.0
|l45.0
|l45.0
Seconds
ijo.o
iio.o
IMS
iho
IJ50
ijo.o
!J45
IJ40
l|30
l|20
l|20
0
0
0
0
0
0
0
0
Degrees
JJ73.0
JJ73.0
JJ73.0
JJ73.0
|l73.0
|l73.0
|l72.0
|l72.0
|l72.0
|l72.0
|l72.0
Minutes
J6.0
J9.0
js.o
|e.o
J2.0
J1.0
J58.0
J56.0
J55.0
|54.0
|54.0
Seconils
|30.0 |
|o.o |
lo.o I
lo.o I
lo.o |
lo.o |
Iso.o |
Iso.o |
Iio.o 1
lo.o 1
lo.o 1
Figure 11 Illustration of latitude-longitude input for ship location and locations of shore line
points.
Abbreviated as lat-long.
35
-------
Figure 12 Shorelines entered from lat-long data. The top shoreline's coordinates are
shown in Figure 11. The ship itself is barely visible between the two shorelines, but can
be seen in the close up at right.
\
Figure 13 Latitude-Longitude calculation in ERO3S.
Figure 13 illustrates the points involved in calculating the distance between two lat-long points.
Because the distance between any two lines of longitude depends upon the latitude, the
36
-------
difference in longitude is calculated by assuming that points are at the latitude of a) the first point
(x0) and b) the second point (xj). These two results are averaged to obtain the difference in East-
West direction. The North-South distance is calculated simply from the difference in longitude.
37
-------
Appendix: Model Development Within the MDP Model
Development Platform
The ERO3S model was developed within the MPD model development platform. This
platform facilitates the creation of object-oriented models in the Java programming language.
MDP includes components for
automated generation of Graphical User Interfaces (GUIs)
selection from among a number of related models under a single interface
access to a series of numerical solvers including,
a family of Runge-Kutta-Fehlberg (RKF) Ordinary Differental Equation (ODE)
solvers
a linear equation solver
an uncertanty range solver
graphical display routines for both animated drawings and charted output
The correct usage of the MDP platform gives a flexible approach to creating and solving model
equations.
Procedural Outline for Creating MDP Applications
Each MDP application contains several required files/classes:
Main routine,
Applet frame,
Application frame,
One or more interfaces,
One or more models,
Title Screens, and
Background Screens.
Directory and Package Structure
Java requires that file names and directory structures match the internal package structure
of an application. The structure that has been created for MDP applications and applets5 is
5In Java applications are codes that run on PCs in much the same fashion as any other
windows software. Applets run as embedded objects in web pages. Applets are subject to a
series of constraints, both security-related and practical, that limit their comprehensiveness. For
brevity both applications and applets will be indicated by use of the term application. Where
necessary to distinguish between the two, the context will clearly indicate which is indicated.
38
-------
shown in Figurel4. The general components of MDP are located in the directory/package
models]avabase and are separate from any application. MDP contains basic utilities for creating
and running applications in Java. MDP contains:
Graphical User Interface (GUI) components in the fw2 subdirectory,
charting software in the MSBChart and GraphControl subdirectories,
general purpose drawing software in the graphics subdirectory,
numerical methods in the models and Numerics subdirectories, and
various utilities in the utils subdirectory.
ERO3S components are contained in the directory/package ERO3S, likewise separate from MDP
and any other application. ERO3S consists of three parts:
EmpiricalDispersantData,
EROS, and
OilSlick.
EmpiricalDispersantData contains results from the dispersant studies of Sorial et al. (2004).
These are used as the basis for dispersing oil slicks in ERO3S. EROS contains the main routines
that run the model. OilSlick contains the single and multiple oil slick implementations of the
model.
The structure shown in Figure 14 organizes the source code (subdirectory src), Java class files
(classes), documentation (doc), backup (bak) into standardized locations for any MDP
application. The subdirectories gov and epa under src and classes follow Java naming
convention and trace the codes back to development within the [Federal] EPA. FigurelS shows
the structure within "classes" and how it is mimics the structure of src. Unlike "src", "classes" is
automatically generated upon compilation when using Borland's JBuilder for creating
applications. Thus the major chore in setting up an application is creating the bak, doc, classes
and src directories, then placing source codes in subdirectorys:
src/'gov/epa/'first level name/second level name
where the first level name is chosen by convention to match the application name (i.e., ERO3S)
and as many second level names as needed are added to match sub parts of the application (i.e.,
EmpiricalDispersantData, OilSlick, etc.)
39
-------
id -_l ModelsJavaZOXX
E
± -'Hi bak
I \ ± _J classes
| j E Hi src
| | . 3 LJ gov
III - Hi epa
III --i _J eroSs
1 I j _J EmpiricalDispersantData
1 1 i -J EROS
| | j B _J OilSlick
i ! fe _j Fay
! ; _ll test
! E 1 modelsjavabase
! j _j bale.
! E - _| classes
\ E Hi doc
| 1 package cache
; ^
I - | src
' i r*?-,t t
; I gov
I 3 Hi epa
| 3 i modelsjavabase
! '"1 awtextend
| ;±j Hi doc
i Hi fw2
! El -Hi GraphControl
! 1 graphics
! 1 InterfacePanels
i Hi model
i * _J MSBChart
I :+l 1 Numerics
\ Hi package cache
! 3 -_J Transport
I 'I UtilS
- 1 modelsjavabase
HE HJ bak
F 1 classes
j | qov
^ . j epa
i E 1 modelsjavabase
1 -1
1 -^
i ;T Hj
1 -Hi
1 -_i
i . -J
i F Hi
1 E Hj
- is
i i :± -Hi
.Jj
awtextend
fv.,'2
GraphControl
graphics
InterfacePanels
model
MSBChart
Numerics
Transport
utils
1 package cache
F Hi doc
1 package cache
t 3 src
^
i 1 gov
s
ri ! I epa
E 1 modelsjavabase
Hi
:± -Hi
Hi
E -Hi
Hi
Hi
; -H
E -_J
E -Hi
; Hi
;T Hj
-Hi
awfcextend
doc
fw2
GraphControl
graphics
InterfacePanels
model
MSBChart
Numerics
package cache
Transport
utils
Figure 14 Directory structure of the ERO3S
MDP application. The two first level
directories, EROSS and modelsjavabase, are the
locations that contain the code that is unique to
ERO3S (EROSS) and the MDP framework
(modelsjavabase).
Figure 15 Directory structure of the Java MDP
showing the required subdirectories: bak,
classes, doc, package cache, and src. An src
subdirectory is created by the MDP user for each
new model and must contain additional
subdirectories: gov, epa, first level name (here:
modelsjavabase), and second level names (here:
awtextend, doc, fw2, etc).
40
-------
41
-------
Appendix: Serial et al. (2004) Dispersant Data
A linear regression empirical model was fit to the experimental data of Serial et al.,
(2004) for each of the oil/dispersant combinations. The model takes the following form and is
used directly in each ERO3S application:
% Efficiency ^^ = (J0 + Pww + p,f + (J^
+ PH*W* + Pw* + M* + P*tsts
+ 2 + 2 + (38)
where w represents weathering in %, t is the temperature (water) in °C, and s is the speed in
RPM. The terms were chosen to include linear and parabolic effects of each variable and possible
two- and three-factor interactions. If all were statistically significant, the model would include 15
terms. Because for each oil/dispersant combination there are no more than 27 data points, no
additional interaction or non-linear terms were included in the model. Data from the replicate
study were used to enhance the regressions: each non-replicated point for the speed of 200 rpm
and dispersants "A" and "B" was replaced by the average result from the replicate study. As seen
in the results, only a few terms were significant for a given oil/dispersant combination as
determined by step-wise multiple regression with an acceptance/rejectance level of 0.05.
Between 4 and 9 terms represented the data for these experiments. Notably the step-wise
regression showed that adding more of the 15 possible terms did not improve the fits.
The various parameters of Equation 38 for the various oil - dispersant combinations are
given in Table 5 together with R2 values, which indicates the linearity of the model. Generally,
values above 90% indicate good linear fits. With the exception of 2FO with no dispersant
(86.9%), all the R2 values were above 90%. Regression equation terms that include weathering
as a varaible are highlighted in Table 5 with gray shading. Note from the table that none of the
regressions include weathering alone as a term. This indicates the secondary nature of
weathering as a variable as described previously for each oil. Figures 16 to 24 show comparisons
of estimated and measured values of dispersal efficiency. Each of the plots show that the data
cluster along the 1 : 1 line, indicating, obviously, a close match. Prudhoe Bay Crude with either
dispersant (Figures 19 and 22) and the South Louisiana Crude with dispersant B (Figure 24)
show particularly tight clustering along this line.
42
-------
Table 5 Coefficients of Regression Equations with Terms Determined by Step-Wise Linear Regression
Factor0'
constant
w
t
s
wt
ws
ts
wts
w2
t2
s2
W2t2
w2s2
t2S2
w2 12 s2
R2
Prudhoe Bay Crude
No Dispersant Dispersant A Dispersant B
-5.9325 -264.6 -15.16
1.2090 4.222 3.506
2.609
-8.386e-3
-4.120e-3 -8.386e-3
-4.845e-5
-1.038e-2
-1.979e-2 -9.697e-2 -2.817e-2
1.468e-4 -5.409e-3 1.433e-3
2.6e-7
91.1% 97.5% 98.2%
No. 2 Fuel Oil
No Dispersant Dispersant A Dispersant B
1.490 -112.0 -17.65
10.67 3.032
0.6617
-2.452e-3
-1.4089e-3 -2.435e-2
6.996e-3 -0.2000 -6.313e-2
9.871e-5 1.256e-3
1.39e-6
9e-8 1.30e-6
86.9% 96.7% 94.8%
South Louisiana Crude
No Dispersant Dispersant A Dispersant B
-17.25 41.39 -69.24
-0.1381 -8.873 -9.149
0.1680 0.1762 1.322
-6.391e-3
-1.631e-3
7.656e-4 4.092e-2 4.132e-2
4.382e-3 0.1516 0.1178
-3.3750e-4 -2.970e-3
9.99e-6
5e-8
-2.87e-6 -2.26e-6
98.2% 90.8% 98.6%
(1) , =
w = weathering, t = temperature, s = speed
43
-------
15 i
10
I
LU
15 i
10
5
15 i
10
5
^ I ' I ' I
0 5 10 15 0 5 10 15 0 5 10 15
Measured Measured Measured
Figure 16 Estimated vs Measured %Figure 17 Estimated vs Measured % Figure 18 Estimated vs Measured %
Dispersal of PB C with No Dispersal of 2FO with No Dispersal of SLC with No
Dispersant. Dispersant. Dispersant.
100 i
100 i
100 |
80 100
I ' I ' I
0 20 40 60 80 100 0 20 40 60
Measured Measured
Figure 19 Estimated vs Measured % Figure 20 Estimated vs Measured %
Dispersal of PBC with Dispersant
I ' I ' I ' I
20 40 60 80 100
Measured
Figure 21 Estimated vs Measured %
Dispersal of 2FO with Dispersant "A" Dispersal of SLC with Dispersant
100 i
100 i
100 i
20
40 60
Measured
80
100
20
40 60
Measured
80 100
Figure 22 Estimated vs Measured % Figure 23 Estimated vs Measured % Figure 24 Estimated vs Measured %
Dispersal of BPC with Dispersant Dispersal of 2FO with Dispersant "B". Dispersal of SLC with Dispersant
"B". "B".
Figure 25 shows a comparison of the regression equations and measured values plotted for
the Prudhoe Bay Crude with no dispersant. The squares, for example, should cluster about the
200 rpm dashed line. The measured values, however, span almost the entire range of dispersal for
speeds of 150 rpm to 250 rpm. This result indicates that the measured variation in dispersal at
200 is as great as the fitting error in the regression equations. The coefficients for these
regressions contain no terms that involve weathering. Thus the amount of volatilization
weathering that occurs does not affect the dispersal efficiency. So the three curves for the
different speeds represent all possibilities for dispersal of the oil.
44
-------
Figure 26 shows a comparison of the regression equations and measured values plotted for
the Prudhoe Bay Crude with dispersant A. The regression equations for this pair contain no terms
involving weathering (Table 5), so that the regression equations only need to be plotted for speed
and temperature. The graph shows the inverted parabolic shape of the curves (i.e., highest
dispersal at the mid-temperature), and the experimental data for each speed and percent
weathering. That the latter quantity is unimportant for this oil and dispersant is shown by the data
points falling generally near each other regardless of the amount of weathering.
45
-------
100 ,
90
80 -
03
i_
CD
Q.
W)
30
60
40
20
10
0
Prudhoe Bay No Dispersant
o
o
A
A
150 RPM, 0% Weathering
150RPM, 10% Weathering
150 RPM, 20% Weathering
200 RPM, 0% Weathering
200 RPM, 10% Weathering
200 RPM, 20% Weathering
250 RPM, 0% Weathering
250 RPM, 10% Weathering
250 RPM, 20% Weathering
Estimate for 150 RPM
Estimate for 200 RPM
Estimate for 250 RPM
0
40
10 20 30
Temperature (C)
Figure 25 Comparison of regression equations (curves) against measured Prudhoe Bay Crude/no dispersantefficiency.
46
-------
100
90
80
70
60
50
40
30
20
10
0
Prudhoe Bay Crude Dispersant "A"
O
A
A
150RPM,0%w
150 RPM, 10% w
150RPM,20% w
200RPM,0%w
200 RPM, 10% w
200 RPM, 20% w
250 RPM, 0% w
250 RPM, 10% w
250 RPM, 20% w
Estimate 150 RPM 0% w
Estimate 150 RPM 10% w
Estimate 150 RPM 20% w
Estimate 200 RPM 0% w
Estimate 200 RPM 10% w
Estimate 200 RPM 20% w
Estimate 250 RPM 0% w
Estimate 250 RPM 10% w
Estimate 250 RPM 20% w
0
30
40
10 20
Temperature (C)
Figure 26 Comparison of regression equations (curves) against measured Prudhoe Bay Crude/dispersant "A" efficiency.
47
-------
100
90
80
70
60
50 -
40 -
30
20
10
0
Prudhoe Bay Crude Dispersant "B"
O
A
A
150RPM,0%w
150 RPM, 10% w
150RPM,20% w
200 RPM, 0% w
200 RPM, 10% w
200 RPM, 20% w
250 RPM, 0% w
250 RPM, 10% w
250 RPM, 20% w
Estimate 150 RPM 0% w
Estimate 150 RPM 10% w
Estimate 150 RPM 20% w
Estimate 200 RPM 0% w
Estimate 200 RPM 10% w
Estimate 200 RPM 20% w
Estimate 250 RPM 0% w
Estimate 250 RPM 10% w
Estimate 250 RPM 20% w
0
30
40
10 20
Temperature (C)
Figure 27 Comparison of regression equations (curves) against measured Prudhoe Bay Crude/dispersant "B" efficiency.
48
-------
CO
2
CD
W
Q
100
90
80
70
60
50
40
30
20
10
0
South Louisiana Crude No Dispersant
A
A
A
150RPM, 0%w
150 RPM, 10% w
150 RPM, 20% w
200 RPM, 0% w
200 RPM, 10% w
200 RPM, 20% w
250 RPM, 0% w
250 RPM, 10% w
250 RPM, 20% w
Estimate 150 RPM 0%w
- - Estimate 150 RPM 10% w
- - Estimate 150 RPM 20% w
Estimate 200 RPM 0% w
Estimate 200 RPM 10%w
Estimate 200 RPM 20% w
Estimate 250 RPM 0% w
- - Estimate 250 RPM 10% w
- Estimate 250 RPM 20% w
0
40
10 20 30
Temperature (C)
Figure 28 Comparison of regression equations (curves) against measured South Louisiana Crude/No Dispersant efficiency.
49
-------
CO
CD
Q.
100
90
80
70
60
50
40
30
20
10
0
South Louisiana Crude Dispersanf'A"
O 150RPM, 0%w
^ 150RPM, 10%w
^ 150RPM, 20% w
Q 200 RPM, 0% w
Q 200 RPM, 10% w
| 200 RPM, 20% w
A 250 RPM, 0% w
A 250 RPM, 10% w
A 250 RPM, 20% w
----- Estimate 150 RPM 0%w
--------- Estimate 150 RPM 10% w
---------- Estimate 150 RPM 20% w
-- Estimate 200 RPM 0% w
--- Estimate 200 RPM 10% w
---- Estimate 200 RPM 20% w
- - Estimate 250 RPM 0% w
Estimate 250 RPM 10% w
Estimate 250 RPM 20% w
0
10
20
30
40
Temperature (C)
Figure 29 Comparison of regression equations (curves) against measured South Louisiana Crude/Dispersant "A" efficiency.
50
-------
CO
CD
Q.
W
Q
sP
100 -
90 H
80 -
70
60
50 -
40 -
30 -
20
10
0
South Louisiana Crude DispersanFB"
O
A
A
A
150 RPM, 0% Weathering
150 RPM, 10% Weathering
150 RPM, 20% Weathering
200 RPM, 0% Weathering
200 RPM, 10% Weathering
200 RPM, 20% Weathering
250 RPM, 0% Weathering
250 RPM, 10% Weathering
250 RPM, 20% Weathering
Estimate for 150 RPM
Estimate for 200 RPM
Estimate for 250 RPM
0 10 20 30 40
Temperature (C)
Figure 30 Comparison of regression equations (curves) against measured South Louisiana Crude/Dispersant "B" efficiency.
51
-------
cc
W
0)
40
30
0
100 -
90 |
80 -
70
60
50 |
No. 2 Fuel Oil No Dispersant
0
0
150RPM,0%w
150RPM, 10%w
150RPM, 20% w
200RPM,0%w
200RPM, 10%w
200 RPM, 20% w
250RPM,0%w
250 RPM, 10% w
250 RPM, 20% w
Estimate 150 RPM 0%w
Estimate 150 RPM 10% w
Estimate 150 RPM 20% w
Estimate 200 RPM 0% w
Estimate 200 RPM 10% w
Estimate 200 RPM 20% w
Estimate 250 RPM 0% w
Estimate 250 RPM 10% w
Estimate 250 RPM 20% w
A
A
A
o
40
10 20 30
Temperature (C)
Figure 31 Comparison of regression equations (curves) against measured No. 2 Fuel Oil/No Dispersant efficiency.
52
-------
cc
100
90
80 -
70 -
60
50
40
30
20
10 -
0
No. 2 Fuel Oil Dispersant "A"
O
o
A
A
A
150 RPM,0% Weathering
150RPM, 10% Weathering
150 RPM, 20% Weathering
200 RPM, 0% Weathering
200 RPM, 10% Weathering
200 RPM, 20% Weathering
250 RPM, 0% Weathering
250 RPM, 10% Weathering
250 RPM, 20% Weathering
Estimate for 150 RPM
Estimate for 200 RPM
Estimate for 250 RPM
0
40
10 20 30
Temperature (C)
Figure 32 Comparison of regression equations (curves) against measured No. 2 Fuel Oil/Dispersant "A" efficiency.
53
-------
03
£
CD
100
90
80
70
60
50
40
30
20
10
0
0
T
/
No. 2 Fuel Oil Dispersant "B"
o
o
A
A
150 RPM,0% Weathering
150RPM, 10% Weathering
150 RPM, 20% Weathering
200 RPM, 0% Weathering
200 RPM, 10% Weathering
200 RPM, 20% Weathering
250 RPM, 0% Weathering
250 RPM, 10% Weathering
250 RPM, 20% Weathering
Estimate for 150 RPM
Estimate for 200 RPM
Estimate for 250 RPM
10
30
40
20
Temperature (C)
Figure 33 Comparison of regression equations (curves) against measured No. 2 Fuel Oil/Dispersant "B" efficiency.
54
-------
Appendix: Wang et al. (2003) Physical Properties and Chemical
Composition of Alaska North Slope Crude Oil
1 Origin: Alaska, U.S.A (the oil was drawn as a line sample off the TAPS pipeline
where it spurs off to the Petrostar Refinery in Valdez on March 19, 2002)
Synonyms: ANS
Appearance: Brown-black, light, little odour, fine black particulates dispersed through-out
liquid.
Values are reported for the fresh oil and for artificially weathered fractions of 10.0%,
22.5% and 30.5% loss by weight. The notations "(n=2)" "(n=2)," "(n=4),n etc indicate
the number of replicates.
2 API Gravity
30.89 (calc)
3 Equation for Predicting Evaporation
%Ev = (2.86 + 0.045 7) In t
Where: %Ev = weight percent evaporated; T = surface temperature (°C); / = time (minutes)
4 Sulphur Content
Weathering
(weight %)
0
10.0
22.5
30.5
Sulphur
(weight %)
1.11
1.20
1.38
1.50
(n=3)
(n=3)
(n=3)
(n=3)
Water Content
Weathering
(weight %)
0
10.0
22.5
30.5
Water
(volume %)
<0.1
<0.1
<0.1
<0.1
(n=3)
(n=3)
(n=3)
(n=3)
55
-------
Flash Point
Weathering
(weight %)
0
10.0
22.5
30.5
Flash Point
<-8
19
75
115
(n=3)
(n=3)
(n=3)
(n=3)
Density
Weathering Temperature
(weight %) (°C)
0 0
15
10.0 0
15
22.5 0
15
30.5 0
15
Density
(g/mL)
0.8777
0.8663
0.9054
0.8940
0.9303
0.9189
0.9457
0.9340
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
Pour Point
Weathering
(weight %)
0
10.0
22.5
30.5
Pour Point
-32
-20
-9
-6
(n=2)
(n=2)
(n=2)
(n=2)
56
-------
Dynamic Viscosity
Weathering
(weight %)
0
10.0
22.5
30.5
Temperature Viscosity
(°Q (cP)
0 23.2
15 11.5
0 76.7
15 31.8
0 614
15 152
0 4230
15 624.7
(n=3)
(n=3)
(n=3)
(n=3)
(n=2)
(n=3)
(n=2)
(n=2)
10 Chemical Dispersibility
Weathering
(weight %)
0
10.0
22.5
30.5
Chemical Dispersibility
using Corexit 9500 ( %)
47
45
34
15
(n=6)
(n=6)
(n=6)
(n=6)
11 Adhesion
Weathering
(weight %)
0
10.0
22.5
30.5
Adhesion
(g/m2)
20
35
38
40
(n=4)
(n=4)
(n=4)
(n=4)
57
-------
12 Surface and Interfacial Tensions
12.1 Surface Tension (Oil/Air Interfacial Tension)
Weathering Temperature
(weight %) (°C)
0 0
15
10.0 0
15
22.5 0
15
30.5 0
15
Surface Tension
(mN/m)
27.3
26.4
29.8
28.4
31.2
30.4
33.1
31.8
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
72.2 Oil/Brine (33%o) Interfacial Tension
Weathering Temperature
(weight %) (°C)
0 0
15
10.0 0
15
22.5 0
15
30.5 0
15
Surface Tension
(mN/m)
22.5
20.2
25.3
23.1
26.8
24.2
30.1
25.6
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
58
-------
12.3 Oil/Fresh Water Interfacial Tension
13
Weathering Temperature
(weight %) (°C)
0 0
15
10.0 0
15
22.5 0
15
30.5 0
15
Surface Tension
(mN/m)
26.7
23.6
28.1
25.5
30.8
27.7
33.2
30.2
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
Emulsion Formation
Weathering Visual Stability
(weight %)
0 Unstable
10.0 Unstable
22.5 Unstable
30.5 Mesostable
Complex Modulus Emulsion
(Pa) Water Content (%)
155
72.9
59
-------
14 Boiling Point Distribution
Boiling Point
(°Q
40
60
80
100
120
140
160
180
200
250
300
350
400
450
500
550
600
650
Cumulative Weight Fraction (%)
0% 10.0% 22.5% 30.5%
weathered weathered weathered weathered
2.5
3.9
6.5
10.0
13.4
16.6
19.8
22.6
25.2
32.6
40.7
49.5
57.7
66.0
72.8
79.0
84.1
88.4
0.1
0.5
1.4
3.6
6.6
9.8
13.1
16.3
19.2
27.4
36.4
46.1
55.3
64.5
72.1
79.0
84.7
89.5
0.1
0.6
2.0
4.4
7.3
16.6
27.0
38.2
48.7
59.3
68.2
76.0
82.6
88.0
0.5
7.5
18.7
31.1
42.8
54.5
64.2
72.8
79.9
85.8
60
-------
15 Hydrocarbon Groups
16
Component
Saturates
Aromatics
Resins
Asphaltenes
Waxes
0%
weathered
75.0
15.0
6.1
4.0
2.6
Concentration
(weight %)
10.0% 22.5%
weathered weathered
72.1
16.0
7.4
4.4
2.9
69.2
16.5
8.9
5.4
o o
J.J
30.5%
weathered
64.8
18.5
10.3
6.4
3.6
Volatile Organic Compounds
Component
Benzene
Toluene
Ethylbenzene
Xylenesf
C3-Benzenes{
Total BTEX
Total BTEX and C3-
Benzenes}
Concentration
(ug/g oil)
0% 30.5%
weathered weathered
2866
5928
1319
6187
5620
16300
21920
0
0
0
0
30
0
30
f'Xylenes" include o-, m-, and^-xylene isomers.
|"C3-Benzenes" include eight isomers.
61
-------
17
tt-Alkane Distribution
n-Alkane Component
w-C8
w-C9
w-CIO
w-Cll
w-C12
w-C13
w-C14
w-C15
w-C16
w-C17
Pristane
w-C18
Phytane
w-C19
w-C20
w-C21
w-C22
w-C23
w-C24
w-C25
w-C26
w-C27
w-C28
w-C29
w-C30
w-C31
w-C32
w-C33
w-C34
w-C35
w-C36
w-C37
w-C38
w-C39
w-C40
w-C41
TOTAL
C17/PRISTANE
C18/PHYTANE
PRISTANE/PHYTANE
CPI
Concentration
0%
weathered
5.55
4.29
4.13
3.98
3.71
3.57
3.42
3.28
3.15
3.06
1.89
2.68
1.41
2.32
2.11
1.96
1.90
1.79
1.65
1.47
1.27
0.97
0.78
0.70
0.56
0.44
0.31
0.27
0.24
0.22
0.11
0.09
0.07
0.05
0.03
0.02
63.4
1.62
1.9
1.35
0.9
(mg/g oil)
30.5%
weathered
0.73
2.51
3.80
4.58
4.34
4.05
4.00
2.41
3.46
1.80
2.93
2.71
2.50
2.45
2.34
2.16
1.94
1.73
1.28
1.03
0.98
0.69
0.60
0.43
0.33
0.31
0.25
0.14
0.13
0.10
0.07
0.06
0.04
56.9
1.66
1.92
1.34
1.0
62
-------
I'
OQ ^
= 1
OJ
tt-
>-|
o
O
3-
O
c
n-C8
n-cio;
n-C12;
n C14 ~
n C1fi "
Phytane ~_
_
n-C20
n-C22 ;
n-C24 ;
n P9R
n-C28 ;
n-C30 i
n-C32 "
n-C34 _
n-C36 "
n-C38 _
n-C40 ;
Cone, (mg/g oil)
3 -> K3 CO -fc. Ol
[ ]
p " ''
I ,.,.,.,.,''','' >
1 ' z
1 (fl
, O
~___M ov
^ ' |'1
E3
a
3
a
i
i
i
c
n-C8
n-C10
n-C12
n-C14
n-C16
Pristane
Phytane
n-C22
n-C24
n-C26
n-C28
n-C30
n-C32
n-C34
n-C36
n-C38
n-C40
Cone, (mg/g oil)
D ^ IO
-------
18
PAH Distribution
Concentration (ug/g oil)
Alkylated PAH
Naphthalene
CO-N
Cl-N
C2-N
C3-N
C4-N
Sum
Phenanthrene
CO-P
Cl-P
C2-P
C3-P
C4-P
Sum
Dibenzothiophene
CO-D
Cl-D
C2-D
C3-D
Sum
Fluorene
CO-F
Cl-F
C2-F
C3-F
Sum
Chrysene
co-c
Cl-C
C2-C
C3-C
Sum
TOTAL
2-m-N/l-m-N
(3+2-m/phen)/(4-/9-+lm-phen)
4-m:2/3m:l-m-DBT
Other PAHs
Biphenyl
Acenaphthylene
Acenaphthene
Anthracene
Fluoranthene
Pyrene
Benz(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(e)pyrene
Benzo(a)pyrene
Perylene
Indeno(l,2,3 cd)pyrene
Dibenz(a,h)anthracene
Benzo(ghi)perylene
TOTAL
0%
weathered
261
1015
1800
1702
815
5594
209
666
710
486
296
2368
122
225
318
265
931
142
328
447
379
1295
48
74
99
84
306
10493
1.49
0.76
1 :0.65 :0.34
134.71
12.03
13.03
2.88
2.88
8.40
4.64
5.14
0.50
10.28
2.26
3.01
0.13
0.63
3.13
204
30.5%
weathered
167
1288
2716
2575
1174
7919
295
932
988
707
432
3354
174
319
456
362
1312
197
449
647
525
1819
68
107
141
115
430
14834
1.41
0.76
1 :0.65 :0.34
176.9
18.43
20.02
4.55
3.81
11.92
8.11
7.49
0.70
14.74
3.69
4.42
0.25
1.02
4.91
281
64
-------
ANS Fresh 15o1
i n _
,-, 5000 n "
'o /inno
fj sooo
a 2000
1 10 .nlllln -nnn.
o u
1 5 8 1 5 8
1
i
Dibenz
Other EPA Priority PAHs
03 S' CD CD
CD
Q Q Li. Li. ^ O O
A CO O CNI .C -A CO
O O O O O O O
ANS 30.5%
^ 5000 n
o 4000
fj 3000 -
a 2000
g 1000
W
.nlllln .
^ n Other EPA Priority PAHs
15§:ll
5-
-------
19
Biomarker Concentrations
Concentration (ug/g oil)
Biomarker
C23
C24
C29
C30
C31(S)
C31(R)
C32(S)
C32(R)
C33(S)
C33(R)
C34(S)
C34(R)
Ts
Tm
C27app steranes
C29app steranes
TOTAL
0%
weathered
65.6
40.8
77.6
116.6
46.9
33.0
34.2
21.6
21.6
13.3
15.1
8.6
20.9
28.5
73.6
84.2
702
30.5%
weathered
91.8
57.8
104.6
161.1
64.1
46.1
46.5
30.9
31.0
19.3
21.0
12.4
31.8
43.0
103.2
113.8
978
Diagnostic Ratios
C23/C24
C23/C30
C24/C30
C29/C30
C31(S)/C31(R)
C32(S)/C32(R)
C33(S)/C33(R)
C34(S)/C34(R)
Ts/Tm
C27aPP/C29aPP
1.61
0.56
0.35
0.67
1.42
1.58
1.62
1.76
0.73
0.87
1.59
0.57
0.36
0.65
1.39
1.51
1.60
1.70
0.74
0.91
66
-------
Appendix: Wang et al, (2003) Physical Properties and Chemical
Composition of South Louisiana
1 Origin: Baton Rouge, Louisiana, U.S.A. (Exxon-Mobil)
Synonyms: Louisiana
Values are reported for the fresh oil and for artificially weathered fractions of 10.9%,
19.7% and 27.7% loss by weight. The notations "(n=2)," "(n=2)," "(n=4),n etc indicate
the number of replicates.
2 API Gravity
32.72 (calc)
3 Equation for Predicting Evaporation
%Ev =(2.74 + 0.045 7) In t
Where: %Ev = weight percent evaporated; T = surface temperature (°C); / = time (minutes)
4 Sulphur Content
Weathering
(weight %)
0
10.9
19.7
27.7
Sulphur
(weight %)
0.49 (n=3)
0.71 (n=3)
0.79 (n=3)
0.88 (n=3)
Water Content
Weathering
(weight %)
0
10.9
19.7
27.7
Water
(volume %)
<0.1 (n=3)
<0.1 (n=3)
<0.1 (n=3)
<0.1 f»=.3j
67
-------
Flash Point
Weathering
(weight %)
0
10.9
19.7
27.7
Flash Point
<-10
42.3
80.7
>110
(n=2)
(n=3)
(n=3)
(n=2)
Density
Weathering Temperature
(weight %) (°C)
0 0
15
10.9 0
15
19.7 0
15
27.7 0
15
Density
(g/mL)
0.8668
0.8562
0.8888
0.8770
0.9025
0.8906
0.9135
0.9018
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
Pour Point
Weathering
(weight %)
0
10.9
19.7
27.7
Pour Point
-41
-19
-14
-11
(n=2)
(n=2)
(n=l)
(n=2)
68
-------
Dynamic Viscosity
Weathering
(weight %)
0
10.9
19.7
27.7
Temperature Viscosity
(°Q (cP)
0 18.5
15 10.1
0 54.8
15 23.7
0 217.3
15 48.9
0 515.9
15 141.0
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=2)
(n=3)
(n=3)
10 Chemical Dispersibility
Weathering
(weight %)
0
10.9
19.7
27.7
Chemical Dispersibility
using Corexit 9500 ( %)
26.5
23.5
15.8
10.3
(n=6)
(n=6)
(n=6)
(n=6)
11 Adhesion
Weathering
(weight %)
0
10.9
19.7
27.7
Adhesion
(g/m2)
24
34
50
28
(n=4)
(n=4)
(n=5)
(n=4)
69
-------
12 Surface and Interfacial Tensions
12.1 Surface Tension (Oil/Air Interfacial Tension)
Weathering Temperature
(weight %) (°C)
0 0
15
10.9 0
15
19.7 0
15
27.7 0
15
Surface Tension
(mN/m)
28.3
26.1
29.3
28.1
30.4
29.4
31.1
29.8
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
72.2 Oil/Brine (33%o) Interfacial Tension
Weathering Temperature
(weight %) (°C)
0 0
15
10.9 0
15
19.7 0
15
27.7 0
15
Surface Tension
(mN/m)
20.9
16.8
22.0
19.4
22.0
22.2
20.6
18.4
(n=2)
(n=3)
(n=3)
(n=2)
(n=3)
(n=2)
(n=4)
(n=3)
70
-------
12.3 Oil/Fresh Water Interfacial Tension
13
Weathering Temperature
(weight %) (°C)
0 0
15
10.9 0
15
19.7 0
15
27.7 0
15
Surface Tension
(mN/m)
20
15
25
15
25
22
24
21
.8 (n=3)
.5 (n=2)
.2 (n=3)
.8 (n=3)
.3 (n=3)
.3 (n=3)
.7 (n=3)
.9 (n=3)
Emulsion Formation
Weathering Visual Stability
(weight %)
Complex Modulus
(Pa)
Emulsion
Water Content (%)
0 Unstable
10.9 Unstable
19.7 Unstable
27.7 Unstable
71
-------
14 Boiling Point Distribution
Boiling Point
(°Q
40
60
80
100
120
140
160
180
200
250
300
350
400
450
500
550
600
650
Cumulative Weight Fraction (%)
0% 10.9% 19.7% 27.7%
weathered weathered weathered weathered
1.2
1.6
2.1
5.6
8.2
11.1
14.1
17.5
20.6
29.8
39.9
49.7
58.1
65.8
72.0
77.1
80.9
83.8
0.9
2.4
4.8
7.8
11.4
14.9
25.2
36.6
47.7
57.0
65.7
72.7
78.5
82.8
86.0
0.1
0.4
1.6
4.0
7.2
18.1
30.6
42.8
53.1
62.7
70.4
76.7
81.5
85.0
0.1
0.3
1.4
10.6
24.1
37.5
49.0
59.6
68.2
75.2
80.5
84.5
72
-------
15 Hydrocarbon Groups
16
Component
Saturates
Aromatics
Resins
Asphaltenes
Waxes
0%
weathered
80.8
12.6
5.9
0.8
1.7
Concentration
10.9% 19.7%
weathered weathered
80.4
12.3
6.4
0.9
1.8
78.4
12.5
8.0
1.1
2.0
27.7%
weathered
77.3
13.3
8.0
1.5
2.2
Volatile Organic Compounds
Component
Benzene
Toluene
Ethylbenzene
Xylenesf
C3-Benzenes{
Total BTEX
Total BTEX and C3-
Benzenes}
Concentration
(ug/g oil)
0% 27.7%
weathered weathered
1598
3552
891
6164
6680
12210
18890
0
10
0
2
190
12
202
f'Xylenes" include o-, m-, and^-xylene isomers.
|"C3-Benzenes" include eight isomers.
73
-------
17
tt-Alkane Distribution
n-Alkane Component
w-C8
w-C9
w-CIO
w-Cll
w-C12
w-C13
w-C14
w-C15
w-C16
w-C17
Pristane
w-C18
Phytane
w-C19
w-C20
w-C21
«-C22
w-C23
w-C24
w-C25
w-C26
w-C27
w-C28
w-C29
w-C30
w-C31
w-C32
w-C33
w-C34
w-C35
w-C36
w-C37
w-C38
w-C39
w-C40
w-C41
TOTAL
C17/PRISTANE
C18/PHYTANE
PRISTANE/PHYTANE
CPI
Concentration
0%
weathered
4.33
4.12
4.12
4.56
4.25
4.14
3.81
3.88
3.48
3.05
2.10
2.24
1.35
2.00
1.70
1.55
1.33
1.13
1.03
0.92
0.72
0.54
0.49
0.42
0.38
0.31
0.23
0.18
0.16
0.15
0.08
0.07
0.05
0.04
0.03
0.02
59.0
1.45
1.65
1.55
0.95
(mg/g oil)
27.7%
weathered
0.21
1.81
3.81
4.94
5.19
5.29
4.75
4.13
2.76
3.11
1.84
2.61
2.27
2.11
1.81
1.58
1.44
1.28
1.08
0.78
0.70
0.62
0.54
0.46
0.34
0.27
0.24
0.20
0.12
0.10
0.08
0.07
0.05
0.04
56.7
1.50
1.68
1.49
1.02
74
-------
t
ft
O
OJ
o
o
O
a-
CD
O
"3s
c^
OQ
o -» NJ oo -t^ cn co
i i i i i i
n-C8
n-C10 ;
n C"\ 9
n C"\ A
n C"\ R
Phytane "
n-C20 "
n-C22
n-C24 "
n-C26
n-C28 "
n-C30 "
n-C32
n-C34
n-C36 "
n-C38 "
n-C40 "
i
1 O
! 3
1 NJ
ssi r-j
13 g
a
Zl
i
a
i
I
I
C
n-C8
n PI n
-
n-C12
-
n-C16
Phytane
n-C20 ;
n-C22 ;
n-C24 ^
n-C26
n-C28 ^
n-C30 ;
n-C32
n-C34 .
n-C36
n-C38 ;
n-C40 ^
D -> K) CO -N Ol
'
:--
^3_i
"""""""""'."' &
_' 5"
3'
zz, g
1:3 3-
=1
Zl
i
1
1
1
-------
18
PAH Distribution
Concentration (ug/g oil)
Alkylated PAH
Naphthalene
CO-N
Cl-N
C2-N
C3-N
C4-N
Sum
Phenanthrene
CO-P
Cl-P
C2-P
C3-P
C4-P
Sum
Dibenzothiophene
CO-D
Cl-D
C2-D
C3-D
Sum
Fluorene
CO-F
Cl-F
C2-F
C3-F
Sum
Chrysene
co-c
Cl-C
C2-C
C3-C
Sum
TOTAL
2-m-N/l-m-N
(3+2-m/phen)/(4-/9-+lm-phen)
4-m:2/3m:l-m-DBT
Other PAHs
Biphenyl
Acenaphthylene
Acenaphthene
Anthracene
Fluoranthene
Pyrene
Benz(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(e)pyrene
Benzo(a)pyrene
Perylene
Indeno(l,2,3 cd)pyrene
Dibenz(a,h)anthracene
Benzo(ghi)perylene
TOTAL
0%
weathered
248.6
952.7
1500.1
1765.7
886.3
5353
134.4
569.8
654.6
427.4
251.8
2038
40.0
125.7
237.4
205.5
609
67.3
181.7
291.4
246.0
804
23.0
58.8
81.6
69.1
233
9037
1.63
1.00
1:0.62:0.31
94.32
8.15
17.90
2.47
3.70
8.64
5.19
2.10
0.37
4.07
0.49
30.37
0.50
0.86
1.23
180
27.7%
weathered
164.1
1058.9
1965.6
2403.6
1222.3
6815
188.3
777.8
887.1
574.6
349.6
2777
55.4
172.4
323.1
272.6
823
94.8
253.2
396.4
354.1
1098
30.4
80.1
108.4
90.7
310
11823
1.59
1.01
1:0.61:0.31
120.60
10.70
24.27
3.61
5.10
11.33
6.35
3.73
1.24
5.97
0.62
38.95
1.12
1.12
1.99
237
76
-------
South Louisiana
2500 n
2000
1500
1000
500
0
Fresh
150 1 Other ERA Priority PAH
75 n
o 1 1 _ « _
^OO
ZOOIXOO-QOOOOOOO
Q
Figure 2 PAH Distribution for South Louisiana crude oil (n-g/g oil)
77
-------
19
Biomarker Concentrations
Concentration (ug/g oil)
Biomarker
C23
C24
C29
C30
C31(S)
C31(R)
C32(S)
C32(R)
C33(S)
C33(R)
C34(S)
C34(R)
Ts
Tm
C27app steranes
C29app steranes
TOTAL
0%
weathered
16.9
11.2
59.9
81.5
31.0
27.5
20.1
13.6
12.2
8.8
6.1
4.4
19.0
23.1
65.0
72.8
473
27.7%
weathered
22.7
14.7
75.9
105.6
40.2
35.7
25.1
17.4
15.4
10.5
7.3
5.2
24.3
30.3
85.8
94.3
610
Diagnostic Ratios
C23/C24
C23/C30
C24/C30
C29/C30
C31(S)/C31(R)
C32(S)/C32(R)
C33(S)/C33(R)
C34(S)/C34(R)
Ts/Tm
C27aPP/C29aPP
1.50
0.21
0.14
0.73
1.13
1.48
1.39
1.37
0.82
0.89
1.54
0.21
0.14
0.72
1.13
1.44
1.46
1.41
0.80
0.91
78
-------
Appendix: Wang et al. (2003) Physical Properties and Chemical
Composition of Fuel Oil No. 2/Diesel
1 Origin: Local Retailer, Ontario, Canada (Stinsons' Gas)
Synonyms: "Summer" Diesel, Fuel Oil No. 2
Appearance: Golden-coloured, light, characteristic "fuel" odour.
Values are reported for the fresh oil and for artificially weathered fractions of 7.2%,
14.2% and 22.0% loss by weight. The notations "(n=2)" "(n=2)," "(n=4),n etc indicate
the number of replicates.
2 API Gravity
37.52 (calc)
3 Equation for Predicting Evaporation
%Ev =( 0.02 + 0.013 7) sqrt(f)
Where: %Ev = weight percent evaporated; T = surface temperature (°C); / = time (minutes)
4 Sulphur Content
Weathering
(weight %)
0
7.2
14.2
22.0
Sulphur
(weight %)
0.09 (n=3)
0.10 (n=3)
0.10 (n=3)
0.10 (n=3)
Water Content
Weathering
(weight %)
0
7.2
14.2
22.0
Water
(volume %)
<0.1 (n=3)
<0.1 (n=3)
<0.1 (n=3)
<0.1 f»=.3j
79
-------
Flash Point
Weathering
(weight %)
0
7.2
14.2
22.0
Flash Point
54
65
76
85
(n=2)
(n=2)
(n=2)
(n=2)
Density
Weathering Temperature
(weight %) (°C)
0 0
15
7.2 0
15
14.2 0
15
22.0 0
15
Density
(g/mL)
0.8423
0.8310
0.8468
0.8350
0.8493
0.8383
0.8524
0.8416
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
Pour Point
Weathering
(weight %)
0
7.2
14.2
22.0
Pour Point
-50
-49
-43
-41
(n=2)
(n=2)
(n=2)
(n=2)
80
-------
Dynamic Viscosity
Weathering Temperature Viscosity
(weight %) (°C) (cP)
0 0 4.08
15 2.76
7.2 0 4.55
15 3.27
14.2 0 5.16
15 3.42
22.0 0 5.59
15 4.18
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=2)
(n=3)
(n=2)
10 Chemical Dispersibility
Weathering Chemical Dispersibility
(weight %) using Corexit 9500 ( %)
0 72
7.2 71
14.2 64
22.0 66
(n=6)
(n=6)
(n=6)
(n=6)
11 Adhesion
Weathering Adhesion
(weight %) (g/m2)
0 2
7.2 12
14.2 13
22.0 8
(n=4)
(n=4)
(n=3)
(n=4)
81
-------
12 Surface and Interfacial Tensions
12.1 Surface Tension (Oil/Air Interfacial Tension)
Weathering Temperature
(weight %) (°C)
0
7.2
14.2
22.0
72.2 Oil/Brine (33%o) Interfacial
0
15
0
15
0
15
0
15
Tension
Weathering Temperature
(weight %) (°C)
0
7.2
14.2
22.0
0
15
0
15
0
15
0
15
Surface Tension
(mN/m)
28.7
27.5
28.8
27.7
28.6
28.1
29.3
28.3
Surface Tension
(mN/m)
21.5
18.1
24.8
19.5
26.6
20.7
28.5
21.9
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=8)
(n=3)
(n=3)
(n=2)
(n=3)
(n=3)
(n=3)
82
-------
12.3 Oil/Fresh Water Interfacial Tension
13
Weathering Temperature
(weight %) (°C)
0 0
15
7.2 0
15
14.2 0
15
22.0 0
15
Surface Tension
(mN/m)
25.0
21.6
28.1
23.9
28.5
24.3
29.1
25.7
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=3)
(n=2)
Emulsion Formation
Weathering Visual Stability
(weight %)
Complex Modulus Emulsion
(Pa) Water Content (%)
0 Unstable
7.2 Unstable
14.2 Unstable
22.0 Unstable
83
-------
14 Boiling Point Distribution
Boiling Point
(°Q
40
60
80
100
120
140
160
180
200
250
300
350
400
450
500
550
600
650
Cumulative Weight Fraction (%)
0% 7.2% 14.2% 22.0%
weathered weathered weathered weathered
0.2
0.5
1.2
2.8
7.8
16.4
26.8
57.4
84.1
96.4
97.9
98.1
98.2
98.3
98.4
98.6
0.1
0.1
0.1
0.7
4.0
11.8
22.4
55.4
84.5
98.1
99.7
99.9
0.1
1.4 0.3
7.1 3.2
17.0 11.2
51.7 46.7
83.3 81.4
98.1 97.8
99.8 99.7
84
-------
15 Hydrocarbon Groups
Concentration
(weight %)
Component
Saturates
Aromatics
Resins
Asphaltenes
Waxes
0%
weathered
88.2
10.2
1.7
0.0
1.7
7.2%
weathered
86.1
11.9
2.0
0.0
1.8
14.2%
weathered
86.1
11.7
2.2
0.0
2.0
22.0%
weathered
85.6
11.4
3.0
0.0
1.8
16 Volatile Organic Compounds
Component
Benzene
Toluene
Ethylbenzene
Xylenesf
C3-Benzenes{f
Total BTEX
Total BTEX and C3-
Benzenes}
Concentration
(ug/g oil)
0% 22.0%
weathered weathered
136 0
1024 0
619 0
3774 7
13780 2260
5550 7
19330 2267
fNote that the "Xylenes" include o-, m-, and^-xylene isomers.
f f Note that the "C3-Benzenes" include eight isomers.
85
-------
17
tt-Alkane Distribution
n-Alkane Component
w-C8
w-C9
w-CIO
w-Cll
w-C12
w-C13
w-C14
w-C15
w-C16
w-C17
Pristane
w-C18
Phytane
w-C19
w-C20
w-C21
w-C22
w-C23
w-C24
w-C25
w-C26
w-C27
w-C28
w-C29
w-C30
w-C31
w-C32
w-C33
w-C34
w-C35
w-C36
w-C37
w-C38
w-C39
w-C40
w-C41
TOTAL
C17/PRISTANE
C18/PHYTANE
PRISTANE/PHYTANE
CPI
Concentration
0%
weathered
1.15
4.24
10.93
13.43
13.23
13.02
12.33
11.98
10.96
9.22
3.81
6.72
2.52
4.72
3.01
1.70
0.85
0.41
0.19
0.09
0.04
0.02
0.02
0.01
0.01
0.01
124.6
1.58
1.61
1.25
0.99
(mg/g oil)
22.0%
weathered
3.96
11.79
15.25
16.51
15.77
15.58
13.70
11.37
4.82
8.20
3.10
5.88
3.74
2.11
1.06
0.52
0.24
0.11
0.05
0.03
0.02
0.01
0.01
0.01
133.8
1.52
1.58
1.23
1.03
86
-------
I'
OJ
-
I
oo
p
to
B
v>
OQ
O
c
n-C8
n-C10"
n-C12
n-C14
n-C16
Pristane
Phytane
n-C20 ;
n-C22 "
n-C24 ;
n-C26
n-C28 ;
n-C30 "
n-C32 ;
n-C34 ;
n-C36 ;
n-C38 ;
n-C40 ;
Cone, (mg/g oil)
3 CO O3 CO K> Ol OO
'
1 1 (D
n lo
I °"~
5
C
n-C8
n-C10"
n-C12
n-C14
n-C16
Pristane
Phytane
n-C20 "
n-C22 "
n-C24 ;
n-C26 ;
n-C28 "
n-C30 "
n-C32 ;
n-C34 ;
n-C36 ;
n-C38 ;
n-C40 ;
Cone, (mg/g oil)
D GO CD CD NJ Ol 00
ii',',i]i',',i]i',',i]i',',i]i',',i]i',',i]i
.,:.!: ;.:!
1=3 55'
r--n (D
ca --«
-n
I (D
w
3-
-------
18
PAH Distribution
Concentration (ug/g oil)
Alkylated PAH
Naphthalene
CO-N
Cl-N
C2-N
C3-N
C4-N
Sum
Phenanthrene
CO-P
Cl-P
C2-P
C3-P
C4-P
Sum
Dibenzothiophene
CO-D
Cl-D
C2-D
C3-D
Sum
Fluorene
CO-F
Cl-F
C2-F
C3-F
Sum
Chrysene
co-c
Cl-C
C2-C
C3-C
Sum
TOTAL
2-m-N/l-m-N
(3+2-m/phen)/(4-/9-+lm-phen)
4-m:2/3m:l-m-DBT
0%
weathered
820
3664
6927
6636
2805
20852
437
1000
617
185
53
2293
65
110
99
38
312
567
799
756
360
2481
0.02
0.03
0.04
0.00
0.09
25938
1.56
1.50
1 :0.35 :0.16
22.0%
weathered
677
3968
8101
8163
3427
24337
557
1262
769
237
65
2890
82
137
123
50
392
713
1025
961
458
3157
0.03
0.04
0.04
0.00
0.12
30776
1.53
1.52
1 : 0.36: 0.17
Other PAHs
Biphenyl
Acenaphthylene
Acenaphthene
Anthracene
Fluoranthene
Pyrene
Benz(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(e)pyrene
Benzo(a)pyrene
Perylene
Indeno(l,2,3 cd)pyrene
Dibenz(a,h)anthracene
Benzo(ghi)perylene
TOTAL
839.73
34.87
153.55
13.08
6.60
30.88
0.25
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1080
1072.40
42.29
187.34
14.09
8.48
38.84
0.28
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1364
88
-------
Diesel Fresh
9000-1
=- 7500 -
o
5> 6000 -
O)
3 4500 -
o
o 3000-
o
1500-
n
ii
1000 -, Other EPA Priority PAHs
800 - 1-1
600 -
400 -
20° 1 1 n
Q.0r£<2~D-D-D-
< CO £ CO CO
n D n n D n n
Q-i i0i i c i i i '. E* i i
ZOOCLOO-QOOOOOQQ
Q
Diesel
22.0% w
9000
= 7500
§ 6000
^ 4500
c
0 3000
1500
r~
n
]
1500 , other EPA Priority PAHs
1000 - l-l
500
0 1 1 n
m < co S" co co1
CO
nDn^ ___ nDDn
^. ~Z. ~Z. CCLCL NQQLLLL >»OO
Q_ i i(i)i i C i i i ' ^ i i
OJ <- CO _c T- CO
-------
19
Biomarker Concentrations
Biomarker
Concentration (ug/g oil)
0% 22.0%
weathered weathered
C23
C24
C29
C30
C31(S)
C31(R)
C32(S)
C32(R)
C33(S)
C33(R)
C34(S)
C34(R)
Ts
Tm
C27app steranes
C29app steranes
4.0
1.4
5.3
1.8
TOTAL
Diagnostic Ratios
C23/C24
C23/C30
C24/C30
C29/C30
C31(S)/C31(R)
C32(S)/C32(R)
C33(S)/C33(R)
C34(S)/C34(R)
Ts/Tm
C27aPP/C29aPP
3.0
2.9
Note: except for the C23 and C24 terpanes,
no other biomarkers were detected.
90
------- |