Вы находитесь на странице: 1из 173

THE OIL SURFACE FLOW AND

SCREENING MODEL (OILSFSM)


BACKGROUND DOCUMENTATION
AND
USERS GUIDE
DRAFT
EPA Contract No. 68-C6-0020
Work Assignment No. 2-07: Task 3
June 10, 1999
THE OIL SURFACE FLOW AND SCREENING
MODEL (OILSFSM)
BACKGROUND DOCUMENTATION
AND
USERS GUIDE
DRAFT
Prepared by
HydroGeoLogic, Inc.
1155 Herndon Parkway, Suite 900
Herndon, VA 20170
EPA Contract No. 68-C6-0020
Work Assignment No. 2-07: Task 3
June 10, 1999
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 i
TABLE OF CONTENTS
Page
1.0 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1.1 ORGANIZATION OF THE DOCUMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1.2 KEY COMPONENTS OF OILSFSM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
1.3 OILSFSM CAPABILITIES, LIMITATIONS AND KEY ASSUMPTIONS . . . 1-2
1.4 MAJOR FEATURES OF OILSFSM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
1.5 CODE COMPILATION AND LINKING REQUIREMENTS . . . . . . . . . . . . . . 1-6
2.0 LITERATURE REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.2 LITERATURE SEARCH AND PERSONAL COMMUNICATIONS . . . . . . . . 2-2
2.3 SURFACE FLOW OF OIL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4
2.3.1 Viscous and Non-viscous Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4
2.3.2 Surface Flow of Non-viscous Fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
2.3.3 Surface Flow of Viscous Fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
2.3.3.1 Flow Over Horizontal Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . 2-6
2.3.3.2 Flow Over Inclined Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8
2.3.3.2.1 Two-dimensional Gravity Currents . . . . . . . . . . . . 2-8
2.3.3.3 Stability And Fingering On Inclined Surfaces . . . . . . . . . . . . . . 2-11
2.4 INFILTRATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11
2.4.1 Philip Infiltration Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12
2.4.2 Green-Ampt Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-13
2.5 EVAPORATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15
2.5.1 Pure Liquid Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2.5.2 Mixture Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
2.6 COUPLED SURFACE FLOW AND INFILTRATION PROCESSES . . . . . . . 2-18
3.0 PROPOSED METHODOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.1 SURFACE TENSION AND ROUGHNESS . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.2 THE ROLE OF A BERM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
3.3 FLOW OF NON-VISCOUS OIL OVER A PERMEABLE AND
RUGGED SURFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
3.3.1 Flow Over a Flat Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
3.3.2 Flow Over an Inclined Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3.3.3 Numerical Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3.3.3.1 Evaluation Of Volumetric Infiltration Flux . . . . . . . . . . . . . . . . . . 3-6
3.3.3.2 Evaluation Of Volumetric Volatilization Flux . . . . . . . . . . . . . . . 3-7
3.4 FLOW OF VISCOUS OIL OVER A PERMEABLE RUGGED
SURFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
TABLE OF CONTENTS
Page
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 i i
3.4.1 Horizontal Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
3.4.2 Inclined Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
3.4.2.1 Two-dimensional Flow From A Line Source . . . . . . . . . . . . . . 3-12
3.4.2.2 Axisymmetric Flow From A Point Source . . . . . . . . . . . . . . . . 3-13
3.4.2.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
3.4.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
3.4.3.1 Proportionality Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
3.4.3.2 Effective Oil Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
3.4.3.3 Treatment of Singularities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16
3.4.3.4 Evaluation of Maximum Up-Slope Extent x
T
. . . . . . . . . . . . . . 3-16
3.4.3.5 Treatment of Finite Duration Spill Sources . . . . . . . . . . . . . . . . 3-17
4.0 OILSFSM CODE DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.2 SIMULATION OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.3 SIMULATION ALGORITHMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.3.1 Algorithm For Surface Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.3.2 Algorithm For The Green-ampt Infiltration Model . . . . . . . . . . . . . . . . . . . 4-2
4.4 CODE CAPABILITIES AND LIMITATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
4.4.1 Source Spill Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
4.4.2 Types of Oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
4.4.3 Surface Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
4.4.4 Infiltration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
4.4.5 Volatilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
4.5 MAJOR CODE FEATURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
4.5.1 Input Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
4.5.2 Time Marching Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
4.5.3 Stopping Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
4.5.4 Run-time Message Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
4.5.5 Output Control For Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-7
4.6 MODULE DESCRIPTION AND FLOW CHARTS . . . . . . . . . . . . . . . . . . . . . . 4-8
5.0 INPUT GUIDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5.2 TYPES OF INPUT FILES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5.3 EXECUTION MODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5.4 OUTPUT FILES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.5 BATCH FILE FORMAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.6 COMPLETE INPUT INSTRUCTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
TABLE OF CONTENTS
Page
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 i i i
5.7 EXAMPLE MAIN INPUT FILE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-19
5.8 SOIL DATABASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-22
5.9 OIL DATABASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-23
6.0 OUTPUT INTERPRETATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.1 RUN-TIME LOG FILE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.2 LISTING FILE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.3 RESULT SUMMARY FILE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.3.1 Characteristics of Surface Spreading of Oil . . . . . . . . . . . . . . . . . . . . . . . . 6-2
6.3.1.1 Non-viscous Flow From A Line Source Spill . . . . . . . . . . . . . . . 6-2
6.3.1.2 Non-viscous Flow From A Point Source Spill . . . . . . . . . . . . . . 6-2
6.3.1.3 Viscous Flow From A Line Source Spill . . . . . . . . . . . . . . . . . . 6-2
6.3.1.4 Viscous Flow From A Point Source Spill . . . . . . . . . . . . . . . . . . 6-3
6.3.2 Mass Balance Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4
6.3.3 Observations of Infiltration Rates and Fronts . . . . . . . . . . . . . . . . . . . . . . . 6-4
7.0 CODE VERIFICATION AND EXAMPLE PROBLEMS . . . . . . . . . . . . . . . . . . 7-1
7.1 PURE SURFACE FLOW OF OIL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7.1.1 Non-viscous Oil Flow over Horizontal/inclined Surfaces . . . . . . . . . . . . . . 7-1
7.1.2 Viscous Oil Flow over Horizontal/inclined Surfaces . . . . . . . . . . . . . . . . . . 7-4
7.2 PURE INFILTRATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8
7.2.1 Green-ampt Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8
7.2.2 Philip Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-13
7.3 PURE VOLATILIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-13
7.3.1 Pure Liquid Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-13
7.3.2 Mixture Model I: the Mean Effective Molar Volume Approach . . . . . . . 7-16
7.3.3 Mixture Model Ii: the Activity Coefficient Approach . . . . . . . . . . . . . . . . 7-20
7.4 SURFACE FLOW OF OIL COUPLED WITH INFILTRATION
AND VOLATILIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-23
7.4.1 Flow of Viscous Oil over a Horizontal Surface . . . . . . . . . . . . . . . . . . . . 7-23
7.4.2 Flow of Non-viscous Oil over a Horizontal Surface . . . . . . . . . . . . . . . . 7-26
8.0 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
APPENDIX ALITERATURE SEARCH RESULTS
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 i v
LIST OF FIGURES
Page
Figure 1.1 A Schematic Illustration of Oil Spill on a Land Surface . . . . . . . . . . . . . . . . . . . . . 1-3
Figure 3.1 Schematic Diagram of Surface Tension and Surface Roughness . . . . . . . . . . . . . . . 3-2
Figure 3.2 Schematic Illustrations of the Shape and Extent of Non-Viscous
Flow of Oil Resulting from Point/Line Sources on Flat/Inclined Surfaces . . . . . . . . 3-5
Figure 3.3 Schematic Illustrations of the Shape and Extent of Viscous Flow of Oil
Resulting from Point and Line Sources on Horizontal and Inclined
Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
Figure 4.1 Program MAIN Flow Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-12
Figure 4.2 Subroutine MAINSUB Flow Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13
Figure 4.3 Subroutine SOLVER Flow Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-14
Figure 4.4 Subroutine RHSNVISC (Generic Name DERIV) Flow Chart . . . . . . . . . . . . . . . 4-15
Figure 4.5 Subroutine RHSVISC (Generic Name DERIV) Flow Chart . . . . . . . . . . . . . . . . 4-16
Figure 7.1 Extent of Non-Viscous Oil Spill on Flat/Inclined Surfaces Subject to
Pulse/Finite Duration Spills at a Point or Along a Line. . . . . . . . . . . . . . . . . . . . . . . 7-3
Figure 7.2 Extent of Viscous Oil Spill on a Flat Surface Subject to a Line Source
(a) and a Point Source (b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6
Figure 7.3 Extent of Viscous Oil Spill on an Inclined Surface Subject to a Line
Source (a) and a Point Source (b,c). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-7
Figure 7.4 Green-Ampt Model: Comparison of OILSFSM Simulated Results
with Those of Simulation D of Weaver et al. [1994] for a Pure Ponded
Infiltration Problem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12
Figure 7.5 Philip Model: Comparison of Simulated and Exact Results for a
Pure Ponded Infiltration Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-14
Figure 7.6 Pure Liquid Volatilization Model: Comparison of Simulated and Exact
Results. (a) Volumetric Volatilization Flux; (b) Volume Lost to
Volatilization per Unit Pan Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-15
Figure 7.7 Mixture Volatilization Model I for Iranian Heavy Oil: Simulated and
Exact Total Volumetric Volatilization Fluxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-18
Figure 7.8 Mixture Model I for Iranian Heavy Oil: Comparison of Simulated and
Exact Mass Losses to Volatilization for Each Volatile Compound. . . . . . . . . . . . . 7-19
Figure 7.9 Mixture Volatilization Model II for Iranian Heavy Oil: Simulated and
Exact Total Volumetric Volatilization Fluxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-21
Figure 7.10 Mixture Model II for Iranian Heavy Oil: Comparison of Simulated and Exact Mass Loses
to Volatilization for Each Volatile Compound. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-22
Figure 7.11 Surface Flow of Viscous Iranian Heavy Oil: Radius of Spreading and
Averaged Oil Lens Thickness and the Effect of Infiltration and
Volatilization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-24
Figure 7.12 Surface Flow of Viscous Iranian Heavy Oil: (a) Cumulative
Mass Loss to Volatilization and (b) Cumulative Mass Loss to Infiltration . . . . . . . 7-25
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 v
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 vi
Figure 7.13 Surface Flow of Non-Viscous Iranian Heavy Oil: Radius of
Spreading and Averaged Oil Lens Thickness and the Effect of Infiltration and
Volatilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-27
Figure 7.14 Surface Flow of Non-Viscous Iranian Heavy Oil: (a) Cumulative
Mass Loss to Volatilization and (b) Cumulative Mass Loss to Infiltration. . . . . . . . 7-28
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 vi i
LIST OF TABLES
Page
Table 5.1 Group Code, Contents, and Number of Potential Data Records . . . . . . . . . . . . . . 5-4
Table 5.2 Complete Input Instructions for Main Input Data File . . . . . . . . . . . . . . . . . . . . . . 5-5
Table 5.3 An Example Main Input Data File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-20
Table 5.4 Soil Names Available in the Soil Database file DBASE4SOILS.DAT . . . . . . . . . 5-22
Table 5.5 Oil Names Available in the Oil Database file DBASE4OILS.DAT . . . . . . . . . . . 5-23
Table 7.1 Description of Test Problems Associated with Pure Surface Flow of Oil . . . . . . . . 7-2
Table 7.2 Transition Time t
*
and the Maximum Up-slope Spill Extent x
T
for Viscous
Flow of Oil Over an Inclined Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5
Table 7.3 Parameters Used by OILSFSM to Simulate 1-D Infiltration Using the
Green-Ampt Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9
Table 7.4 OILSFSM Input Data File for Reproducing Infiltration Portion of
Weaver et al. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-10
Table 7.5 Parameters for Various Compounds in Iranian Heavy Crude Oil . . . . . . . . . . . . . 7-17
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 1- 1
THE OIL SURFACE FLOW AND SCREENING MODEL (OILSFSM)
BACKGROUND DOCUMENTATION
AND
USERS GUIDE
1.0 INTRODUCTION
Crude oil spills can occur on land, water, or from underground storage tanks. Many sites require screening
through the use of simplified models because of lack of time and characterization resources. One such
screening would be to determine the relative amount of spilled oil that infiltrates the ground surface versus
the amount of oil which may run-off the immediate area of the spill. Because of the need for relatively rapid
assessment of the spill conditions, an analytical or semi-analytical approach is often needed. The OIL
Surface Flow and Screening Model (OILSFSM) has been developed to provide for this need. It utilizes
a number of analytical solutions to simulate surface spreading, infiltration, and evaporation of an
aboveground, terrestrial oil spill.
1.1 ORGANIZATION OF THE DOCUMENT
This report describes the background theory, development, and method of application of the OILSFSM
screening model for assessing aboveground oil spills. It contains both the Background Documentation and
the Users Guide, and provides the user with the following information pertaining to OILSFSM:
A description of the conceptual model of an aboveground, terrestrial oil spill and the subsequent
infiltration, volatilization and runoff of oil as represented in the model
A detailed description of the code structure and major features
A complete set of instructions on how to prepare input files
A description and interpretation of output resulting from execution
A set of example problems that illustrate and verify the key algorithms
The balance of this introductory section presents an outline of the model components, capabilities,
limitations and major assumptions (Sections 1.2 and 1.3). Section 1.4 presents some major features of the
OILSFSM model. Section 1.5 describes code compilation and linking requirements. The remaining
sections of this report are divided into the following four parts:
Part 1: Model Formulation
Part 2: OILSFSM FORTRAN Code Development
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 1- 2
Part 3: OILSFSM Input Guide and Output Interpretation
Part 4: Code Verification
The first part is a background document. It includes a review of literature on modeling surface flow of oils,
infiltration and volatilization, and the conceptual formulation of the OILSFSM model. Part 2 addresses the
code development. Detailed code structure, capabilities and limitations, major features and flow charts of
key subroutines are presented. Part 3 contains a comprehensive input guide and addresses output
interpretations. The input guide describes the mode of execution, type of input files required, and output
files generated during a session. In particular, it gives step by step, detailed instructions for preparing input
data files required for proper execution of the code. It also addresses access to soil and oil databases
included with the code. Issues regarding output interpretation and analysis are also discussed. Part 4
addresses issues related to code verification, testing, and prototype simulations. Included in this part are
comparisons of results from OILSFSM with existing analytical solutions and peer reviewed publications,
as well as demonstration examples of modeling an aboveground oil spill with infiltration and evaporation.

1.2 KEY COMPONENTS OF OILSFSM
Figure 1.1 illustrates schematically a surface oil spill and its subsequent spreading, infiltration, and
volatilization as represented in OILSFSM. (Note: For the balance of this report, the term surface oil spill
should be understood to mean an aboveground spill on land unless otherwise indicated.) The oil is spilled
from a source (such as a storage tank). As it runs off the permeable surface, it infiltrates into the vadose
zone due to gravity and capillary forces and evaporates into the atmosphere due to its volatility.
OILSFSM is based on a simplified conceptualization of the oil spill scenario shown in Figure 1.1. It consists
of the following four key components:

Source of oil spill
Surface spreading of oil
Infiltration into the vadose zone
Volatilization into the atmosphere
The processes of spreading, infiltration and volatilization are coupled. As oil is spilled onto a permeable,
rough ground surface, it fills the surface voids and depressions (depression storage), infiltrates into the soil,
and volatilizes from the surface. The excess oil remaining on the ground surface runs off into the immediate
surrounding area depending on surface topographic conditions.
1.3 OILSFSM CAPABILITIES, LIMITATIONS AND KEY ASSUMPTIONS
OILSFSM Capabilities
OILSFSM can simulate one or more of the following processes simultaneously: surface flow, infiltration
and volatilization. Specifically, it is capable of simulating the following conditions:
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 1- 3
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 1- 4
Land Surface
Vadose Zone
x
N
(t)
h(x,t)
Source of Spill
z
f
(x,t)
Surface
Spreading
Volatilization
x
Extent of spreading
Air
Infiltration
Figure 1.1 A Schematic Illustration of Oil Spill on a Land Surface.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 1- 5
Surface flow of two types of oils on permeable surfaces:
< Viscous oil whose surface flow is governed by a balance between gravitational and viscous
forces
< Non-viscous oil whose surface flow is governed by a balance between gravitational and
inertial forces
Surface spreading on flat or inclined surfaces
Source spill patterns in time:
< A pulse spill in which the duration of the spill is neglected
< A finite duration spill with turn-off time less than the maximum simulation time
< A continuous spill that lasts as long as the maximum simulation time
< Constant spill rate
< Time-variable spill rate.
Source spill patterns in space:
< A point source
< A line source
Surface that is either smooth or rugged
Oils that consist of single or multiple volatile and non-volatile compounds
Presence of a berm
Compositional properties of oil on the surface as they vary with time due to selective volatilization
Output
Results from an OILSFSM simulation consist of three parts. The first part contains key variable values that
characterize the surface spreading of oil. It includes:
Information about the shape of spreading of oil
Thickness and areal extent of surface spreading
Volumetric infiltration and volatilization fluxes
The second portion provides mass balance information. This includes:
Cumulative oil mass spilled
Oil mass remaining on the ground surface
Mass fractions, mass losses to infiltration and volatilization for each compound

C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 1- 6
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 1- 7
The third portion contains estimates at user-specified locations of infiltration front and infiltration rate as they
vary with time.
Major Limitations and Assumptions
The major limitations and underlying assumptions in OILSFSM are:
Only one source location is allowed
Rugged surface is represented by a single parameter called roughness
For non-viscous oil (or gasoline), surface spreading is assumed to occur instantaneously
Variations of surface oil thickness in space are ignored. A spatially averaged thickness is used to
characterize the vertical dimension of the oil lens on the ground surface
Mixing and diffusion within the il phase is much faster than spreading or volatilization. As a result,
oil composition is assumed to be homogeneous in space at any given time.
Soil properties are homogeneous
Infiltration is assumed to be in the vertical direction only. It is modeled using a kinematic
approximation of the governing equation for fluid infiltration into an unsaturated porous medium
the Green-Ampt equation or the simplified Philip infiltration model. The former ignores the effect
of gravity, and is not recommended for simulating redistribution of oil after the thickness of oil drops
to zero (since the sharp infiltration front stops moving as soon as the thickness drops to zero).
In the presence of a berm, it is assumed that the oil front reaches the edges of the berm at the same
time at all points along the oil front, regardless of the shape of the oil lens on the surface.
1.4 MAJOR FEATURES OF OILSFSM
In developing the OILSFSM model, every effort was made to ensure that the code is easy to use, requires
as little as possible from the user, is efficient in the usage of computer memory and CPU time, is user
friendly, and facilitates future extensions and modifications. The following is a list of the attractive features
of OILSFSM:
Input preparation is made easier by the following options:
< In line comments in the input data file
< Classification of model input parameters into distinctive groups
< Default settings of some control parameters and constants, and the option of overwriting any
one or more of them
< Access to soil and oil database files, and the option of overwriting any one or more of the
values read from database files.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 1- 8
Dynamic allocation of memory that allows problems of arbitrary size and makes efficient use of
computer memory
Time-stepping control options:
< Time-stepping with constant time-step size
< Time stepping following pre-specified time values
< Adaptive time stepping
Run-time error information and output control:
< Run-time log file provides options of recording warnings and fatal errors that occur during a
session
< Automatic correction and adjustment of time step size inside solver if minor failures occur or
if smaller time step size is warranted
< User has control over amount and level of run-time message output to both the screen and log
file
< User has control over reporting frequency and level of detail of simulation results
Stopping rule control the user can specify one or more of the following basic conditions as the
simulation stopping rule:
< Spill stopped
< Edges of berm reached
< Frontal spreading stopped
< Oil lens thickness reached minimum allowed
< Maximum simulation time reached
Modular structure:
< Subroutines are highly independent of each other
< Future modifications and extensions are made easier
1.5 CODE COMPILATION AND LINKING REQUIREMENTS
The source code for OILSFSM was developed using ANSI Standard FORTRAN 90. The user should
have access to a compiler that supports the FORTRAN 90 features if the code is to be modified. There
is no specific requirement for minimum computer memory because of the memory allocation adopted in
OILSFSM. Actual memory used will depend on primarily, in decreasing order of importance, the number
of time steps taken, and the number of observation locations and recording frequencies of such
observations.
PART 1:
MODEL FORMULATION
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 1
2.0 LITERATURE REVIEW

2.1 INTRODUCTION
Overflow and leakage from aboveground storage tanks and pipelines occur frequently. A number of spill
incidents are documented at the National Response Center [EPA 1998]. Spills of crude oil and petroleum
products can result in contamination of the surrounding soil and subsequent contamination of ground water.
Mackay and Mohtadi [1975] have indicated that spills from the Mackenzie Valley pipeline in Canada were
expected to occur at a frequency of two spills per year. They also examined oil spill incidents from pipelines
in Alberta Canada, and found that spill volumes are frequently in the range of 200 barrels (31.8 m
3
), with
larger spills greater than 2,000 barrels (318 m
3
) also occurring.
Non-Aqueous Phase Liquids (NAPLs) spilled on the ground surface flow laterally away from the source.
As the spill size increases, parts of the spilled liquid infiltrate the soil surface to the vadose zone while other
parts evaporate to the atmosphere. Flow of spilled liquid on the ground surface is influenced by a number
of factors including: source characteristics (shape, rate, duration, and location with respect to the ground
surface), topography, surface conditions (surface inclination, surface cover, and type of vegetation), liquid
properties (density, viscosity, and surface tension), environmental conditions (temperature, wind, and
precipitation), and soil properties (permeability, porosity, and water saturation). Modeling the movement
of spilled liquids on the ground surface is challenging. Difficulty in quantifying small-scale variations in
topography and surface conditions as well as source characteristics plays an important role in the surface
flow of oil. However, with some simplifying assumptions, screening models can be designed to evaluate the
impacts of a potential release and the associated risks. In the absence of detailed data about the controlling
factors mentioned above, screening models provide a means for rapid assessment and design of remedial
measures.
Crude oil and refined oil products are the most widely used liquid chemicals. EPA [1998] defines oil
according to the Clean Water Act as "crude oil and refined petroleum products, edible animal and
vegetable oil, other oils of animal or vegetable origin, and other non-petroleum oils." Such a definition
incorporates a wide range of liquid chemicals with different properties. The term oil used here generally
refers to crude petroleum and petroleum products that consist of a mixture of different compounds, some
of which may be volatile.
After an oil spill occurs on a soil surface, infiltration will divert some of the fluid mass to the subsurface. If
the water table is well below the ground surface, the oil infiltrates through the vadose zone as a NAPL.
Depending on the composition of the oil, volatile compounds may evaporate from the spill surface. Both
infiltration and volatilization depends on surface flow characteristics such as area, extent, thickness, etc. The
three processes--surface flow, infiltration and volatilization, are expected to be coupled.
The literature review to be presented below covers previous research in modeling the above mentioned
processes. It begins with a literature search conducted on the subject of surface flow of oils (Section 2.2).
Review of earlier analytical/semi-analytical solutions of each one of the three processes surface flow of
oil, infiltration and volatilization, will be presented in Sections 2.3, 2.4, and 2.5. Section 2.6 summarizes
previous work on the coupled processes of surface flow of oil and infiltration.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 2
2.2 LITERATURE SEARCH AND PERSONAL COMMUNICATIONS
An extensive literature search was conducted to evaluate existing models and the approaches used for
simulating surface flow of petroleum and similar liquid chemicals. No existing model was found that could
simulate all processes involved in the spread of oil from a surface spill. As a result, the literature search was
broadened to include studies that addressed one or more of the main processes involved. These processes
include flow of fluids on rigid surfaces, infiltration of non-aqueous liquids in porous media, and evaporation
of liquid mixtures. The search was conducted by performing three main tasks: (1) electronic search of
scientific databases; (2) physical search of selected libraries; and (3) contacting researchers who have
published studies related surface oil spills.
The electronic search covered the following scientific databases:
GEOREF
AGRICOLA
Applied Science and Technology Abstracts
Dissertation Abstracts
General Science Abstracts
GEOBASE
Papers First
Engineering Index - Compendex
Mechanical Engineering Abstracts
Environmental Science and Pollution
Water Resources Abstracts
National Technical Information Services
The resulting reference list was compiled into a database and each reference was examined to determine
its relevance to the problem of the surface flow of oil. Irrelevant references were removed from the
database. For example a large body of literature was determined to be associated with oil spills in marine
environments. Marine spills involve significantly different processes from those associated with
aboveground terrestrial spills.
To expand the search to older references and engineering reports not identified through scientific literature
databases and to acquire appropriate references from the electronic search, a physical search of selected
libraries was conducted. The library search was conducted at the University of Maryland Engineering and
Applied Science Library, George Mason University Library, and the United States Geologic Survey
Library in Reston, Virginia. References found during the library search were added to the database.
Scientific research usually takes a considerable amount of time to reach publication. To ensure a thorough
literature search, a selected number of researchers who have conducted recent studies relevant to the
problem at hand were contacted. These researchers include:
Herbert E. Huppert
nstitute of Theoretical Geophysics
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 3
Department of Applied Mathematics and Theoretical Physics
University of Cambridge
Edward A. Sudicky
Department of Earth Sciences,
University of Waterloo
Ali El-Kadi
Department of Geology and Geophysics/Water Resources Research Center
University of Hawaii
M. Yavuz Corapcioglu
Environmental and Water Resources Engineering Division
Civil Engineering Department
Texas A&M University
Martinus T. van Genuchten
United State Salinity Laboratory
United State Department of Agriculture
Don Mackay
Department of Chemical Engineering & Applied Chemistry
University of Toronto
A meeting was held with Professor Huppert during his visit to Florida State University. In that meeting the
problem of viscous flow over porous media was discussed. Other researchers were contacted via
electronic mails and telephones.
None of the above mentioned researchers have directly addressed the problem of oil flow on the ground
surface coupled with infiltration and evaporation. Each has published research papers, however, related
to one or more of the processes involved. Huppert's pioneering research on viscous gravity currents is
directly applicable to the spreading of oil on rigid surfaces. Sudicky has published a considerable amount
of work on NAPL flow in both porous and fractured media. El-Kadi published two papers on the analytical
solution of flow of NAPL in porous media. Similar to Sudicky's work, Corapcioglu developed a number
of compositional models for subsurface NAPL flow. Over the years, the Salinity Laboratory has conducted
a number of studies on infiltration into different types of soil. Van Genuchten was approached to solicit his
expertise on analytical models for infiltration. Finally, Mackay published a number of research papers on
both spreading of oil on the ground surface and evaporation from oil spills. These researchers guided us
to additional published work, which was reviewed and added to the reference database. Appendix A
provides a list of published studies related to surface oil spills.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 4
2.3 SURFACE FLOW OF OIL
Surface flow of oil can be considered as a special case of gravity currents. Gravity currents occur when
a fluid flows into another fluid of lesser density [Huppert, 1982a]. Gravity is the main driving force for
surface flow while inertia and viscosity are retarding forces. Flow due to gravity currents has been a subject
of intensive research [Benjamin, 1968; Fay, 1969; Hoult, 1972; Smith, 1973; Buckmaster, 1973; Huppert
and Simpson, 1980; Didden and Maxworthy, 1982; Huppert, 1982a,b; Huppert, 1986; Lister and Kerr,
1989; Lister, 1992; and Fukushima, 1998]. The problem has a wide range of applications in geophysics,
industrial engineering and geology. Huppert [1986] presented applications of the propagation of gravity
currents in the flow of volcanic lava.
2.3.1 Viscous and Non-viscous Flow
Didden and Maxworthy [1982] studied the rate of propagation of the front of two-dimensional and
axisymmetric gravity current flow on a horizontal, rigid surface resulting from a constant rate of release.
Two-dimensional gravity current flow arises from a release of fluid from a line source, whereas the
axisymmetric gravity current flow results from a release of fluid from a point source. From order-of-
magnitude analysis of the buoyancy forces, inertial forces and viscous forces, Didden and Maxworthy
[1982] showed that at initial stages of release, inertia forces dominate over viscous forces and the spreading
due to gravity currents is governed by a balance between gravity and inertia forces. At later stages of
release, viscous forces become more dominant and the spreading is governed by a balance between gravity
and viscous forces. Surface spreading characteristics of fluid in these two kinds of flow regions are quite
different.
Huppert [1982a] generalized the analysis of Didden and Maxworthy [1982], and considered the more
general case where the source volume V(t) increases with time according to the power law
V(t)=qt
"
for a line source [L
3
]
V(t)=Qt
"
for a point source [L
3
/L]

where the value of " denotes the spill pattern in time, i.e.,
" = 0 for a pulse release
" = 1 for a constant volumetric flux release
and any other value of " represents time-variable rate of release. Thus, for " = 0 (pulse release), Q is the
volume released instantaneously from a point source [L
3
], and q is the volume released per unit width of
source from a line source [L
3
/L]. For " = 1 (constant flux release), Q is the constant volumetric flux from
a point source [L
3
/T], and q is the constant volumetric flux per unit width of source from a line source
[L
3
/L/T]. The transition time, t
v
, at which inertia and viscous forces are of the same order of magnitudes
was found to be [ Huppert, 1982a]

C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 5
(2.1b)
(2.1a)
where : is the oil dynamic viscosity [Force T/L
2
], Dis the oil density [M/L
3
], and g is the gravitational
acceleration [L/T
2
]. As can be seen from Eq. (2.1), there exists a critical value, "
c
, for which the exponent
in (2.1a) or (2.1b) is infinity. This critical value is "
c
= 3 for a point source, and "
c
= 7/4 for line source. The
order-of-magnitude analysis of Huppert [1982a] can be summarized as follows:
For " < "
c
, the flow is initially dominated by inertial forces. Viscous force increases monotonically
with time until it exceeds the inertial force at time t = t
v
, after which it dominates over that of inertial
one. The case considered by Didden and Maxworthy [1982] is in this region.
For " > "
c,
the flow is initially dominated by viscous forces. As time goes on, viscous force
decreases monotonically with time so that at t
v
it equals the inertial force. Afterwards, the inertial
force dominates over the viscous one.
For " ="
c
, the inertial force and viscous force increase with time at the same rate, and the gravity
current propagates under the same balance of forces for all time and has no transitional time.
2.3.2 Surface Flow of Non-viscous Fluid
For the case of fluids whose viscosity is negligible such as gasoline, a balance between inertia and gravity
forces are expected to dominate the surface flow. Flow under an inertia-gravity regime is described by the
well-known shallow-water equations [Hoult 1972] consisting of a mass balance equation for the thickness
and a momentum balance equation for the velocity. Grundy and Rottman [1986] presented similarity
solutions to these equations. Their solution forms are too complicated for use in a screening model, and are
not discussed here. Raisbeck and Mohtadi [1974] adopted a much simpler model for the surface flow of
non-viscous oil. They assumed that the surface spreading occurs rapidly, such that it reaches a constant
thickness instantaneously. The constant thickness is controlled by the surface tension.
2.3.3 Surface Flow of Viscous Fluid
For the case of surface flow of viscous fluid, the driving force is gravity, and the retarding force is simply
the fluids own viscosity. Huppert [1982a,b; 1986] and Lister [1992] conducted theoretical and
experimental studies of the flow of viscous fluid on horizontal and inclined surfaces. The following
paragraphs summarize their work.
2.3.3.1 Flow Over Horizontal Surfaces
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 6
(2.2)
(2.3)
(2.4a)
(2.4b)
Huppert [1982a] investigated the viscous fluid flow over a horizontal surface due to gravity currents for two
types of source release a point source and a line source. A point source release produces an
axisymmetric spreading, whereas a line source release leads to a two-dimensional spreading that is primarily
in the direction perpendicular to the line. The governing equation for the surface flow of viscous fluid under
these two types of source release can be generalized by the following depth averaged continuity equation
subject to the global continuity or mass balance equation
where R =Dg/3: [1/LT], m=0, 1 denotes line and point sources, respectively, h is the fluid thickness or
depth [L], x is the coordinate (radius for point source case), t is time, x
N
(t) is the extent of the spreading
at time t or the distance the front has traveled away from the source, and V(t) is the cumulative volume of
fluid spilled.
While numerical solutions to (2.2)-(2.3) for any initial condition could be obtained, it is desirable to obtain
some simpler analytical solutions. A similarity solution is a special kind of analytical solution. It uses
transformation of variables to reduce the original complex partial differential equations to a simpler form
amiable to the derivation of an analytical solution. With the supposition that the cumulative volume V(t)
increases with time according to the power law, V(t)=qt
"
[L
3
/L] for a line source, and V(t)=Qt
"
[L
3
] for
a point source, Huppert [1982a] was able to obtain similarity solutions to (2.2)-(2.3) that describe the
shape and rate of the propagation of the spreading. For similarity variables defined as
and
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 7
(2.5)
(2.6)
(2.7)
(2.8)
(2.9a)
(2.9b)
the thickness and front extent of spreading are given by
where the function N(0), 0 = >/>
N
, satisfies the ordinary differential equation
Note that in deriving (2.7), the power laws V(t)=qt
"
for a line source, and V(t)=Qt
"
for a point source was
used.
A useful result is that >
N
is independent of x and t, and it is related to N(0) through
For a pulse spill (" = 0), >
N
can be explicitly evaluated to be 1.41124477 and 0.893913848 for a line
source and a point source, respectively. For all other values of ", including the constant rate scenario of
"=1, the value of >
N
has to be evaluated numerically using (2.8), which requires the solution of (2.7). To
solve (2.7), Huppert [1982a] suggested the use of the condition N(1)=0, and the following approximate
solution to N for 0 near 1
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 8
(2.10)
(2.11)
Eqs.(2.9a,b) are sufficient to specify an unique solution of (2.7). For any value of ", the solution
is obtained by numerically integrating (2.7) inward from 0=1 and using (2.9) as a starting condition.
Huppert [1982a] conducted a number of experiments to compare with his similarity solutions. For
axisymmetric spreading (resulting from a point source), he found that results from experiments with "=0
(or a pulse) and "=1 (or a constant flux) compared very well with his similarity solutions. He observed that
the mode of initiation or initial conditions makes no difference after a few seconds from the start of release.
This is supported by Eqs.(2.5)-(2.6) since the thickness and extent depend more on the volume of fluid
released up to time t than the actual release history. The release history or the value of " only has an impact
on the proportionality constant >
N
as is evident in (2.8) and Figure 3 of Huppert [1982a]--the maximum
value of >
N
occurs at " =0, which decreases gradually as " increases, and stays fairly constant for " >1.0.
2.3.3.2 Flow Over Inclined Surfaces
Lister [1992] expanded Hupperts [1982a,b] solution and studied the propagation of gravity currents from
line and point sources over an inclined surface. The presentation below follows that of Lister [1992].
Because of the complexity involved, gravity currents resulting from these two types of source release are
discussed separately.

2.3.3.2.1 Two-dimensional Gravity Currents
The depth averaged continuity equation for viscous gravity currents from a line source over an inclined and
impermeable surface is given by
subject to the global mass balance equation
where x is the coordinate (positive in the down-slope direction, with the origin of the coordinate system
being chosen at the source), 2 is the angle of inclination of the surface with respect to the horizontal
direction, x
N
(t) is the extent of the spreading in the down-slope direction, and x
T
(t) is the extent in the up-
slope direction.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 9
(2.12)
(2.13b)
(2.13a)
(2.14)
(2.15)
With the volume V(t) [L
3
/L] increasing according to the power law q t
"
, Lister [1992] noted that a spill
from a line source initially propagates in both the up- and down-slope directions until time t
*
when the
down gradient component of gravity dominates the flow. The transition time t
*
was estimated to be
The similarity solutions for the extent of the spreading at any time t are
where >
N
and >
N
*
are the proportionality constants for small time (t < t
*
) and large time (t > t
*
),
respectively. As suggested by Lister [1992], at small time, the solution mimics that of Huppert [1982a] for
gravity currents resulting from a line source, so >
N
can be approximated using exactly the same procedure
outlined in Section 2.3.3.1. The long time solution for >
N
*
is [Lister, 1992, Eq.3.3a]
It follows that >
N
*
=3/2
2/3
for "=0 ( a pulse release), and for >
N
*
=1 for " =1 (a constant flux release)..
2.3.3.2.2 Viscous gravity currents from a point source
The propagation of gravity currents from a point source over an inclined surface is governed by
subject to the global mass balance equation
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 10
(2.16)
(2.17)
(2.18c)
(2.18b)
(2.18a)
where the perimeter of the flow is given by |y|=y
p
(x,t) in -x
T
(t)< x <x
N
(t), x
N
(t) is the extent of the
spreading in the down-slope direction, x
T
(t) is the extent in the up-slope direction, and y
p
(x, t) is the lateral
extent of the spreading at a particular location x and time t.
Lister [1992] showed that release from a point source initially forms a circular shape until time, t
*
, when
the down gradient component of gravity dominates the flow. With the supposition that the volume V(t)
released increases according to the power law Qt
"
, the transitional time t
*
was estimated to be
Defining y
M
(t) to be the maximum value of y
p
(x,t) at time t, Listers similarity solutions for the extent of
spreading at any time are
It follows from (2.18) that at initial stages of viscous spreading from a point source, the shape of the current
is circular with the cross gradient extent equal to the down gradient extent. After some time, the current
spreads down gradient at a higher rate. Similar to the two-dimensional gravity currents over an inclined
surface, the short-time solution to >
N
can be approximated using exactly the same procedure proposed by
Huppert [1982a] for gravity currents induced by a point source over a horizontal surface, as summarized
in Section 2.3.3.1. The long-time solution for the proportionality constants >
N
*
and 0
M
*
were derived by
Lister [1992] for two regions in ". One is
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 11
(2.19)
" =0; the other is " >0. The values of >
N
*
are 1.0, 0.8, 0.64, 0.60 for " =0.2, 1.0, 2.0, and 3.0,
respectively [Lister, 1992, Fig. 7]. The values of 0
M
*
are 1.2,1.075, 1.0 for " =0.2, 0.4, 0.6, respectively.
For " > 0.6, 0
M
*
stays fairly constant at 1.0. For " = 0, the values of >
N
*
and 0
M
*
are given by [Lister,
1992, Eq.B7b,c]
where s = ln (t/t
*
) (s >0 since t > t
*
), and t
*
is as defined in (2.17).
2.3.3.3 Stability And Fingering On Inclined Surfaces
Huppert [1982b] presented results from a number of laboratory experiments in an attempt to compare with
his similarity solutions for viscous currents over an inclined surface induced by a line source. Those
experiments showed that the viscous spreading initially follows his similarity solution. However, after some
time, flow from the line source formed a series of small amplitude waves (fingers). The amplitude of the
waves (length of the fingers) increases with time. He attributed the formation of these fingers to surface
tension. Hocking [1990] studied the instability that occurs at the front of a gravity current that leads to the
development of fingers. A number of researchers have considered the effect of conditions at the front of
the gravity current where a current head is formed [Greenspan, 1978; Britter and Simpson, 1978;
Fukushima, 1998). However, Huppert's experimental results indicated that during viscous spreading, the
conditions at the front do not affect the shape of the current or its size except in the formation of fingers on
inclined slopes.
2.4 INFILTRATION
As spilled oil flows over a porous, permeable ground surface, part of the oil infiltrates to the vadose zone.
Most crude oil and oil products are immiscible with soil water, and can be treated as NAPLs. NAPL flow
in the vadose zone is influenced by the interactions between the three immiscible fluids, namely, air, water
and NAPL. Flow of NAPL in porous media has been the focus of a large body of research literature.
Mercer and Cohen [1992] reviewed NAPL behavior in the subsurface. Parker [1989] and Kueper and
Frind [1992] described the principles of modeling NAPLs in the subsurface and reviewed a few of the
available models. HydroGeoLogic [1992a] presented a review of a selected number of numerical, semi-
analytical, and analytical models for NAPL flow and transport. Numerous numerical models have been
developed and discussed in Abriola and Pinder [1985a, 1985b], Hochmuth and Sunada [1985],
Corapcioglu and Baehr [1987], Kaluarachchi and Parker [1989], Kueper and Frind [1991], Letniowski
and Forsyth [1991], and HydroGeoLogic [1995].
The governing equation for oil infiltration to the vadose zone is nonlinear and coupled with equations
describing flow of the water and gas phases. With some simplifying assumptions, the governing equation
for oil infiltration can be reduced to Richards equation for water flow in the vadose zone. These
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 12
(2.20)
assumptions include:
The air phase is immobile
The water phase is at or near residual saturation
Water saturation is uniform
During infiltration, oil displaces air only
The soil is homogeneous with respect to physical and mechanical properties
Flow of oil occurs vertically downward
Raisbeck and Mohtadi [1974], Darco [1978], Reible et al. [1990] and El-Kadi [1992] have formulated
analytical solutions for NAPL infiltration. In these models the air phase is assumed to be inactive. Flow of
NAPLs was decoupled from the flow of water by assuming that the NAPL displaces air only and that
water saturation remains constant. As mentioned above, Raisbeck and Mohtadi [1974] neglected both the
pressure from ponding depth and flow due to gravity and considered capillary pressure only. Their
infiltration model was based on the capillary bundle model of unsaturated flow [Jury et al., 1991]. Darco
[1978] calculated the time for infiltration to reach some depth from point, line, and area sources. El-Kadi
[1992] presented a similar approach for sudden and continuous releases in one, two, and three dimensions.
Reible et al. [1990] formulated an analytical model for NAPL flow. Their model was one-dimensional and
was based on the assumption of sharp infiltration and drainage fronts. The model parameters, including
effective permeability and effective capillary suction at the infiltrating and drainage fronts, were determined
experimentally. Weaver et al. [1994] extended the Green-Ampt model for water infiltration [Jury et al.,
1988; Chaw et al., 1988] to simulate NAPL infiltration. In their formulation, a sharp front was assumed
during infiltration. The redistribution stage, which starts after infiltration at the surface ceases, was modeled
using the kinematic wave approximation. Model parameters such as relative permeability and capillary
suction are calculated from physical properties of the oil and the porous medium. Illangasekare et al.
[1994] tested both Reible et al. [1990] and Weaver et al. [1991] models against experimental results. Test
results showed that during infiltration the assumption of a sharp front was reasonable. However, no sharp
drainage fronts were observed during redistribution, as assumed by Reible et al., 1990.
Since simple analytical/semi-analytical solutions to the oil infiltration model are of interest in this report, two
relatively simple solutions to the oil infiltration problem (the Philip model and the Green-Ampt model), are
presented below.
2.4.1 Philip Infiltration Model
Philip [1957] developed an analytical solution for Richards equation using a power series. He used the first
two terms to approximate the solution for the infiltration rate q
i
, as follows
where S is the sorptivity of the soil, and B is a parameter related to the saturated hydraulic conductivity of
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 13
(2.21)
(2.22)
(2.23)
the soil. An implicit assumption in (2.20) is that infiltration starts at time t=0. The above model can be
extended to simulate oil infiltration. Chong and Green [1983] showed that the sorptivity can be related to
the saturated hydraulic conductivity and the wetting front pressure head. However, there is some ambiguity
in the definition of B. For long term infiltration, B should be evaluated as the saturated hydraulic
conductivity, K
s
, while for short-term infiltration, the value of B ranges from K
s
/3 to K
s
/2 [Knight, 1983].
2.4.2 Green-Ampt Model
The Green-Ampt model is a physically based model for infiltration of water through the vadose zone [Chow
et al., 1988; Jury et al., 1991]. In addition to the assumptions listed at the beginning of Section 2.4, the
Green-Ampt model assumes that infiltration is one-dimensional and occurs vertically downward and that
infiltrating fluid forms a sharp front that advances in the vadose zone in a piston-like flow. Weaver et al.
[1994] and Charbeneau [1995] extended the model to simulate infiltration of oil and other hydrocarbons
in the presence of uniform water saturation. The following presentation follows their formulation closely.
Let h be the oil thickness on the ground surface [L], z
f
the depth of the sharp infiltration front [L], S
1
the
oil saturation behind the advancing front, K
1
(S
1
) (which depends on the oil saturation S
1,
) be the oil
conductivity of the soil behind the infiltration front [L/T], and H
f
the pressure head at the wetting front.
According to Darcys law, the volumetric flux of oil per unit surface area q
i
is
Assuming a sharp infiltration front, the depth of the oil front, z
f
(t), is related to the volumetric infiltration flux
as
where n = porosity. Combining (2.21)-(2.22) yields a differential equation in the unknown z
f
(t). Given initial
condition z
f
(t
0
) at t
0
, an analytical solution for z
f
(t) ( t >t
0
)is
As indicated by Weaver et al. [1994], Eq.(2.23) is preferred over (2.21)-(2.22) because the high initial
pressure gradient causes problems in the numerical solution of Eqs.(2.21)-(2.22).
The oil conductivity of the soil K
1
is evaluated from the following relation
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 14
(2.24)
(2.25)
(2.26)
(2.27)
(2.28)
where k is the intrinsic permeability of the soil, k
ro
is the relative permeability of oil, and Dand : are the
density and viscosity of the oil, respectively. The relative permeability k
ro
as a function of any oil saturation
S
o
can be calculated using Brook and Corey's model as
where g = (2+38)/8, 8 is the Brook and Corey's pore size distribution index, S
or
is the residual oil
saturation, S
w
is the antecedent water saturation, and S
wr
is the residual or irreducible water saturation.
Water is assumed to fill a fixed volume of the pore space. S
w
is a site-specific parameter and should be
estimated based on the antecedent moisture condition in the field prior to the spill. It is expected that S
w
will remain constant since the overlying oil layer will prevent recharge to the vadose zone. Behind the
assumed sharp infiltration front, the void space is filled with water at a saturation level corresponding to S
w
and oil at a saturation level corresponding to S
o
. Moreover, trapped air fills part of the space, hence
reducing the available space for oil. Weaver et al. [1994] used the following expression for estimating the
residual air phase saturation S
ar
,
where K
e
is the effective oil conductivity and K
s
is the saturated oil conductivity. S
ar
is estimated by
assuming that the maximum effective hydraulic conductivity, K
e
, is 50% of the saturated hydraulic
conductivity, K
s
, due to the presence of air [Weaver et al., 1994]. The oil saturation above the interface
will be S
1
= 1-S
w
-S
ar
, which is constant and remains so during the infiltration of oil. The effective pressure
head H
f
is estimated as
where h
ceao
is the air/oil entry head [Weaver et al., 1994; Charbeneau et al. 1995]. Weaver et al. [1994]
related the air/oil entry head h
ceao
to the air/water entry head h
ceaw
as
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 15
(2.29)
where the subscripts aw and ao are used to indicate air/water and air/oil interfacial tension, respectively.
If the antecedent soil water saturation is equal to the residual water saturation (S
w
= S
wr
), the effective
pressure head at the infiltration front is evaluated by
which can be used to verify the numerical integration of (2.27).
2.5 EVAPORATION
The term "oil" used in this report encompasses a wide range of liquids that consist of a mixture of different
chemical compounds, some of which may be volatile. For example, crude oil contains volatile organic
compounds (VOCs) such as benzene, toluene, ethylbenzene, xylene, and C3-benzenes. Estimation of
evaporation rates is important to quantifying the mass lost to the atmosphere. The VOCs released during
evaporation from oil spills can pose a threat through inhalation and are considered a fire hazard. Also, the
release of these compounds affects properties of the spilled oil such as viscosity and density, thereby
affecting the spreading and infiltration rates.
Evaporation depends on a number of factors. Resistance to evaporation occurs due to liquid phase
resistance and gas phase resistance [Mackay and Matsugu, 1973; Thomann and Mueller, 1987]. Liquid
phase resistance influences transfer of the VOC from the bulk of the fluid to the surface. Mackay and
Matsugu [1973] indicated that, for spills in arctic regions, the liquid phase resistance is high due to the high
viscosity of the oil. Under such conditions, a thin film is formed on the surface that essentially stops
evaporation. Mixing caused by molecular diffusion and oil movement can reduce the liquid phase resistance
considerably. Gas phase resistance is usually the limiting factor in evaporation from oil spills. Gas phase
resistance controls transfer of contaminant in the gas phase (which is in equilibrium with the liquid phase)
from the oil spill surface to the atmosphere.
The simplest evaporation models are those presented by Jokuty et al. [1998]. The weight percent
evaporated, E
v
, is related to surface temperature, T, (degree Celsius) and time, t (minutes), through an
empirical equation of the form
where A and B are constants, to be determined from pan evaporation.
Stiver and Mackay [1984] developed an analytical algorithm to predict the volume fraction evaporated.
Parameters needed for their analytical algorithm were estimated from distillation data. Stiver and Mackay's
algorithm was implemented in the Spill Impact Mapping model (SIMAP) for oil spill in marine environments
[Applied Science Associates, 1998]. In SIMAP, the change in viscosity due to evaporation is considered
by increasing the viscosity by a factor that depends on the volume fraction evaporated (i.e. assuming that
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 16
(2.30)
(2.31)
(2.32)
the volatile compounds are less viscous). The algorithm does not consider change in oil composition due
to evaporation of the VOCs.

More useful and accurate evaporation models have explicitly considered the effect of exposure conditions
and compound properties [Mackay and Matsugu, 1973; Thomann and Mueller, 1987; Stiver et al.,
1989]. Major exposure conditions include exposure area, time, wind velocity, etc. Major compound
properties that control evaporation rates are the compounds vapor pressure, mass fraction of each
component, and temperature. In addition, the evaporation rate depends on the interactions between the
compounds. In the following, a review of three such models is presented. Since all of these models are
empirically based, explicit forms of dimensions for all the parameters involved are given.
2.5.1 Pure Liquid Model
For the case of a pure liquid, Stiver et al.[1989] suggested the following model for calculating q
e
, the
volumetric flux per unit exposure area [cm/s]
where k
g
is the gas phase mass transfer coefficient [cm/s], v is the liquid molar volume [cm
3
/mol], P is the
vapor pressure [kPa], R is the universal gas law constant [kPa-cm
3
/mol-K], and T is the temperature [K].
To evaluate the gas-phase mass transfer coefficient for evaporation from oil, Mackay and Matsugu [1973]
proposed the following empirical relation for k
g
(modified for units)
where D
g
is the diffusivity of the compound in air [cm
2
/sec], <
g
is the kinematic viscosity of air [cm
2
/s], U
is the wind speed [m/hr], and D
p
is the diameter of the evaporation pan [cm]. The diffusivity D
g
for any
compound is related to the molecular weight w of the compound through
where D
Mw
is the diffusion of water vapor in air [cm
2
/s][Thomann and Mueller, 1987].
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 17
(2.33)
(2.34)
(2.35)
2.5.2 Mixture Models
Most crude oils consist of many compounds. Evaporation from such oils are best approached using
compositional models. Stiver et al. [1989] developed a compositional approach to simulate evaporation
from oil spills. The approach was based on calculating the change in the mole fraction of each volatile
component in the oil because of evaporation. For each compound j in the oil, Stiver et al. [1989] proposed
the following formula for calculating the volumetric evaporation flux q
ej
of component j per unit exposure
area
where <
c
is the mean effective oil molar volume [cm
3
/mol], V
o
is the total oil volume [cm
3
], n
j
is the number
of moles of component j, <
j
is the molar volume of compound j [cm
3
/mol], and P
j
is the vapor pressure of
component j [kPa]. Other parameters are the same as defined before.
By definition the mean effective molar volume is related to the initial molar volume <
c
o
, activity coefficient
(
j
and fraction moles evaporated F
x
through
The mean effective oil molar volume <
c
varies with oil composition, and because the oil composition
changes as evaporation proceeds, so also does <
c
. The analysis of Stiver et al. [1989] showed that <
c
is relatively constant compared with variations in the term P
j
/RT in (2.33). Therefore, they suggested that
in the absence of direct measurements, a value of 400 cm
3
/mol is a reasonable estimate for the mean
effective molar volume of most crude oils.
An alternative method of estimating the evaporation rate of a compound is the activity coefficient approach
[Stiver et al.,1989] If the initial mass fraction of each compound is known and the activity coefficient can
be estimated, the following relation can be used to determine the volumetric evaporation rate of j-th volatile
compound from the mixture
where m
j
is the mole fraction of compound j. The rest of the parameters are constant, assuming that a
representative value of the activity coefficient can be estimated.
Whichever method is used to estimate q
ej
, the overall volumetric volatilization rate per unit area is given by
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 2- 18
(2.36)
(2.37)
where NVOC is the number of volatile compounds that comprise the oil.
2.6 COUPLED SURFACE FLOW AND INFILTRATION PROCESSES
Coupling between surface flow and infiltration was addressed by Raisbeck and Mohtadi [1974]. They
formulated an analytical approach for simulating the spread of oil on a soil surface. Their formulation is
based on the assumption that the thickness of the oil spill remains constant in time. Such an assumption
implies that surface spreading occurs at a fast rate such that the oil spill reaches a constant thickness
instantaneously. This thickness is the minimum thickness above which surface runoff starts. The assumption
of fast surface flow can be justified in the case of less viscous oil products such as gasoline. However, this
assumption is less likely to hold for viscous crude oil. Also, the infiltration component of their model neglects
gravity forces and the pressure head of the oil thickness on the surface. Evaporation from the oil surface
is not considered in their model. The simplified model of Raisbeck and Mohtadi [1974] was intended to
simulate oil spills in arctic regions, where the soil evaporation rates are low and infiltration may be hindered
by the existence of snow cover and/or permafrost. Other research efforts in modeling the three coupled
processes are limited. Huppert [1998, personal communication] suggested that surface spreading of initial
volume V, coupled with infiltration will generate a spill of a size x
N
according to the proportional relation
where $ is some function of the hydraulic properties of the porous media. Huppert suggested that the
proportionality constant should be close to 1 and that the parameter $ should be determined experimentally.
Such an approach would be very efficient to implement but will require a great deal of experimental effort
to quantify $ for different soils.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 1
(3.1)
(3.2)
3.0 PROPOSED METHODOLOGY
As oil spreads on the surface of a rugged and permeable soil, surface flow, infiltration and evaporation
remove mass from spill to the vadose zone and to the atmosphere, respectively. Infiltration and evaporation
rates are dependent on the spill size and the thickness of the spilled oil. Evaporation is component-specific
so that the properties of the unevaporated oil change as evaporation proceeds. As a result, coupling
between the three processes is expected.
This section presents combined analytical/semi-analytical approaches to the simulation of surface flow of
oil over flat or inclined surfaces subject to infiltration and volatilization. The section is organized as follows.
Section 3.1 introduces the concept of surface tension and surface roughness. Section 3.2 examines the role
of a berm. The proposed methodology for non-viscous flow on flat and inclined surfaces and its
implementation are given in section 3.3. The proposed methodology for viscous flow on flat or inclined
surfaces and its implementation are given in section 3.4.
3.1 SURFACE TENSION AND ROUGHNESS
In most cases, land surfaces are rugged. As is illustrated in Figure 3.1, when an oil spill incident occurs the
initial spilled volume tends to fill in the depression storage in the immediate area of the spill location.
Excessive oil on the land surface flows or spreads laterally. Surface flow continues until the oil thickness
reaches a minimum thickness defined by
where h
d
is the surface roughness or depression storage per unit surface area [L], and h
s
is the thickness
due to surface tension given by [Raisbeck and Mohtadi, 1974]
where F is the surface tension of the oil [Force/L], and n is the contact angle between oil and soil solids.
Depression storage is created by oil accumulation in hollows over the ground surface below which ground
roughness and interfacial tension impede lateral flow.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 2
Figure 3.1 Schematic Diagram of Surface Tension and Surface Roughness
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 3
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 4
(3.3)
(3.4a)
3.2 THE ROLE OF A BERM
When an actual oil spill occurs on the land surface, it is often preferable to set up a berm to contain the
further spreading of the oil away from the source. Spreading stops when the oil front reaches the edge of
the berm. Build-up in the oil lens thickness may begin, depending on the source release scenario and
surface conditions. In the formulation that follows, the effect of such a bounding condition on the surface
spreading of oil is considered. For simplicity, it is assumed that a berm is characterized by its area, A
b
, and
that its shape is such that all oil fronts will reach the edges of the berm at the same time.

3.3 FLOW OF NON-VISCOUS OIL OVER A PERMEABLE AND RUGGED
SURFACE
When viscous drag and surface tension forces at the front are neglected, the governing equations for the
surface flow of liquids under an inertia-gravity regime reduce to the shallow water equations [Hoult, 1972
and Grundy and Rottman, 1986]. Details of the equations and the imposed boundary conditions can be
found in Grundy and Rottman [1986] who developed a similarity solution for the case of plane and
axisymmetric flow. Because of its complexity, their solution is not considered in this report. Instead, simpler
solutions appropriate for screening purposes are developed.
3.3.1 Flow Over a Flat Surface
Since surface flow under the inertia-gravity regime is expected to be fast, the spill will quickly reach a
constant thickness. Hence, surface flow of non-viscous oil can be reasonably assumed to occur
instantaneously. Any oil added at the source will cause the spill size to increase while maintaining the same
thickness. Under these conditions, the thickness of oil will be controlled by the surface tension of the oil and
the ruggedness of the soil surface, as is defined by Eq.(3.1). The size of the spill will depend on the source
release scenario. In the presence of a berm, spilled oil moves laterally before reaching the berm.
Afterwards, the spreading area stays at a constant, whereas the thickness increases. The general mass
balance for surface flow of non-viscous oil over a flat surface is given by:
where V
t
is the volume of oil on the land surface [L
3
], Q
s
is the volumetric source flux [L
3
/T], Q
i
is the
volumetric infiltration flux [L
3
/T] and Q
e
is the volumetric volatilization flux [L
3
/T]. The area A and actual
thickness h of spreading at any time are given by
and
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 5
(3.4b)
where A
b
is the berm area. It follows from Eqs. (3.4a,b) that the area of spreading is V
t
/ h
min
and the
thickess is h
min
before the oil front reaches the edge of the berm. Afterwards, the area of spreading stays
at A=A
b
and the thickness of the oil lens is h = V
t
/A
b
.
The extent of the spill depends on the source release scenario and the surface condition. Figures 3.2a,b
show the shape and extent of non-viscous oil spill from a line source and a point source, respectively, on
a horizontal surface. Oil spilled from a line source of width w
d
is assumed to be rectangular with a constant
width w
d
(assuming w
d
is large enough) and length . Thus, the extent of spill in each direction of
propagation is given by . Oil spilled from a point source on horizontal surfaces is assumed
to form a circular shape. The radius of the circular area of spreading is calculated as .
3.3.2 Flow Over an Inclined Surface
For flow of non-viscous oil over an inclined surface, the governing equation is the same as that for the flat
surface case. To simplify the problem, the flow is assumed to form a stream of constant width w
d
. This is
equivalent to assuming the presence of a line source with a pre-specified width. The shape of spreading is
rectangular with a constant width w
d
, and a length x
N
(t) that changes with time according to .
Figure 3.2c presents a schematic view of the shape and extent of non-viscous oil spill from a line source
on an inclined ground surface. The thickness of the oil lens is controlled by surface roughness and surface
tension before the oil reaches the edge of the berm. Afterwards, the thickness of the oil lens will start to
build up first near the edge of the berm. It then backs up, depending on the angle of inclination, berm area,
spilled volume and amount of infiltration and volatilization. In this study, simulation will be stopped as soon
as the oil reaches the edges of the berm; back-up of oil in the up-slope direction will not be simulated.
3.3.3 Numerical Implementation
Non-viscous flow of oil over either a horizontal or an inclined surface is governed by an ordinary differential
Eq. (3.3) subject to the constraints in Eq. (3.4a,b). It will be solved numerically using Runge-Kutta
integration with adaptive time step control [Press et al., 1992]. The approach utilizes Fehlbergs method
for estimating the error using Cash-Karp parameters. For a detailed description of the Runge-Kutta
algorithm, the user is referred to the routine ODEINT in Press et al. [1992].
In the numerical implementation time is discretized into discrete time levels t
j
(j =0, 1, 2, ...). These are
referred to as the outer time steps, and they can be specified by the user. Let t
o

C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 6
Line Spill
w
d
x
N
x
N
Plane view
Cross-section view
Point Spill
x
N
Plane view
Cross-section view
o
x N
Line Spill
w
d
x
N
Plane view
Cross-section view
Line source

(a) flat surface, line source (b) flat surface, point source
(c) inclined surface, line source
Figure 3.2 Schematic Illustrations of the Shape and Extent of Non-Viscous Flow of Oil
Resulting from Point/Line Sources on Flat/Inclined Surfaces.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 7
(3.5)
(3.6)
represent the spill start time. The initial condition V
t
(t
o
) is in general zero except for an impulse spill, in which
case V
t
(t
o
) is set equal to the volume of oil spilled. Given the initial conditions at t
j
the solution toV
t
(t) at
time t
j+1
is solved by Runge-Kutta. An examination of Eq.(3.3) reveals that this requires the evaluation of
the volumetric infiltration and volatilization fluxes, respectively, at several time points between t
j
and t
j+1
.
The issue is discussed below.
3.3.3.1 Evaluation Of Volumetric Infiltration Flux
As oil spreads laterally, part of the volume is lost to the vadose zone through infiltration. The volumetric
infiltration flux Q
i
(t) in Eq.(3.3) is by definition the integral of the spatially varying infiltration rate q
i
at time
t over the whole spreading area A(t). Because of the assumption of homogeneity of soil properties and
uniform oil lens thickness and the assumption of reasonably small time step size, the evaluation of volumetric
infiltration flux can be greatly simplified.

During each time step, the oil front may move forward an incremental subarea, depending on the available
oil on the ground surface, the source release scenario, the infiltration and evaporation capabilities, and the
presence of a berm. Let *A
l
(l=1, 2, ..., k) be the k incremental subareas the oil front has visited during
the first j time steps, and *A
k+1
be an estimate of the incremental subarea the oil front will visit from time
t
j
to t (t
j
#t #t
j+1
). Because of the potential restriction of a berm, the condition k < j will hold. At any time
t (t
j
#t #t
j+1
), the k incremental subareas *A
j
(j=1, 2, ..., k) stay constant; however, the magnitude of the
(k+1)-th incremental subarea *A
k+1
will vary with time t until t
j+1
has been attained or the surrounding berm
has been reached. The cumulative area of spreading at time t can be written into two parts, as follows
Since A(t) is easily calculated using Eq.(3.4a) given the volume of oil on the ground, the oil thickness at any
time t and berm area, the incremental subarea *A
k+1
of the oil front scanned from time t
j
to t can be
computed using (3.5) given A(t) and *A
l
(l=1, 2, ..., k).
Under the assumption of homogeneity of soil properties and relatively small time-step size, the infiltration
rates at any point in each one of the k+1 incremental subareas at any time can be assumed equal. Let q
i
(l,
t) (l =1, 2, ..., k+1) be the infiltration rates at time t (t
j
#t #t
j+1
) in each one of the (k+1) incremental
subareas. The volumetric infiltration flux at time t (t
j
#t #t
j+1
) over the whole spreading area can then be
computed as
The infiltration rate q
i
(l, t) at time t (t
j
#t #t
j+1
) in any subarea l (l =1, 2, ..., k+1) depends on the arrival
time t
a,l
of the front in subarea l, the infiltration model to be used and soil parameters. For the Philip model
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 8
(3.7)
(3.8)
(3.9)
(3.10)
[cf. Eq. 2.20], computation of q
i
(l, t) is straight forward, i.e.,
for l=1, 2, ...k+1. The infiltration front z
f
(l, t) in subarea l is calculated using
where n is the soil porosity, and S
1
is the oil saturation behind the infiltration front.
To evaluate q
i
(l, t) for the Green-Ampt model [Eqs.(2.21)-(2.22)], the approach of Weaver et al. [1994]
is followed (see Eq.2.23). Assume that the oil lens thickness on the ground surface does not change very
much during a small time interval from t
j
to t. The infiltration front z
f
(l, t) at time t (t
j
< t #t
j+1
) in any one
of the k+1 subareas is calculated given the initial infiltration front z
f
(l, t
j
) at t
j
by solving a modified form
of Eq.(2.23) or
The Newton-Raphson method is used to solve Eq. (3.9) for the infiltration front. Once the infiltration front
is solved, Eq.(2.21) is used to obtain the infiltration rate q
i
(l,t).
The presence of a berm is accounted for by the incremental subarea counter k. When the front of the spill
reaches the edge of the berm at any time t (t
j
#t #t
j+1
) or A(t)>A
b
, A(t) is set equal to A
b
and *A
l
=0 for
all l > k+1.
3.3.3.2 Evaluation Of Volumetric Volatilization Flux
Unlike infiltration, volatilization rate does not vary in space or with incremental subareas. This is based on
the assumption that mixing of different compounds occurs so fast that an instantaneous equilibrium is
reached as soon as additional oil is spilled on the ground. Thus, the volumetric volatilization flux, Q
e
(t) at
any time t is simply calculated using
where q
e
(t), the volatilization rate over a unit exposure area. q
e
(t) can be calculated using (2.30) for a pure
liquid model or (2.33)/(2.35) for a mixture model.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 9
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 10
(3.11)
(3.12)
Note that the gas phase transfer coefficient as calculated by (2.30) or (2.33) was based on an evaporation
pan that is circular with a diameter of D
p
. When the shape of the surface is not circular, an effective
diameter is defined such that
where A(t) is the area of spreading at time t.
It is further assumed that the mole fraction and volume fraction, density and viscosity of each compound
in the source are known. At each time step, one has to keep track of and update the number of moles of
each compound, the mass fraction and volume fraction of each compound, the overall oil density and
viscosity for oil on the land surface. The update of mole, mass and volume fractions of each individual
compound follows mass conservation laws. The overall viscosity is computed using
where NCMP is the total number of compounds that make up the oil, n
j
= n
j
(t) is the mole fraction of
compound j at time t, and :
j
= :
j
(t) is the compounds viscosity. The overall oil density is calculated by
a weighted average of the individual compounds density with the weight being the volume fraction of the
compound.

3.4 FLOW OF VISCOUS OIL OVER A PERMEABLE RUGGED SURFACE
Section 2.3.3 provides the similarity solutions of Huppert [1982a,b] and Lister [1992] for the shape and
extent of spreading of gravity currents over a rigid surface. Their solutions were based on the following
assumptions:
Flow is driven by gravitational forces
Surface is smooth, impermeable
Viscous drag forces are much greater than inertial forces
Surface tension at the current head does not affect the spreading of oil
The area of the spill is much larger than its thickness.
Flow of viscous oil over permeable soil surface is expected to occur at a slower rate than that over a rigid
and impermeable surface. Infiltration and evaporation operate on the oil volume during surface flow, hence
reducing the available volume for spreading. In the discussion that follows, a combined analytical/semi-
analytical approach is proposed to simulate the surface flow of viscous oils on rugged and permeable soil
surfaces. This approach results from extending the similarity solutions of Huppert [1982a,b] and Lister
[1992].
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 11
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 12
(3.13)
(3.14)
(3.15)
3.4.1 Horizontal Surfaces
For flow of viscous oil over a horizontal permeable surface subject to infiltration and evaporation, two sink
terms representing the loss due to infiltration and evaporation need to be added to the right hand side of
Eq.(2.2). The resulting depth averaged continuity equation is
where R =Dg/3: , q
i
(x,t) is the infiltration rate [L/T], q
e
(t) is the evaporation rate [L/T] (time variable but
spatially homogeneous), and m=0, 1 denote a two-dimensional plane flow resulting from a line source and
an axisymmetric flow resulting from a point source, respectively. Other terms are the same as those defined
earlier. The global continuity Eq. (2.3) also needs to be modified to reflect the loss to depression storage,
infiltration and volatilization. The modified global continuity equation reads
where V(t) is the cumulative volume of oil spilled, as defined in Eq.(2.3). The integral on the right-hand-side
of (3.7) accounts for cumulative loss to infiltration and evaporation, and the last term reflects loss to the
depression storage.
The system of Eq. (3.13)-(3.14)) are not amenable to similarity solutions using the Huppert [1982a]
approach because of the presence of the sink terms in Eq. (3.13) and the complex nature of the loss terms
in Eq. (3.14). With some simplifying assumptions, however, the similarity solutions of Huppert [1982a]
presented in Section 2.3.3.1 can be extended to Eq. (3.13)-(3.14). To proceed, it is assumed that the rates
of infiltration and evaporation are slow in comparison to the changes in h(x,t) in both time and space as
reflected by the terms on the LHS of Eq. (3.13). The contribution of both infiltration and volatilization
processes to the governing equation can then be neglected. The contribution to the global continuity
equation (3.14)from the three loss terms , however, has to be included to maintain a correct mass balance.
To proceed, an effective oil volume V
eff
(t) is defined as
which represents the total volume of oil on the ground at time t minus the loss to depression storage. This
effective volume is the part of the oil on the ground surface that is the driving force for lateral spreading. In
the absence of losses to infiltration, volatilization and depression storage, V
eff
(t) is simply V(t). In principle,
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 13
(3.16b)
(3.16a)
(3.17)
the extent of spreading can thus be approximated by substituting (3.15) into (2.6). The problem is that
V
eff
(t) by definition depends on the characteristics of the surface spreading such as extent and thickness and
it may not vary with time t according to power laws for which Huppert [1982a,b] similarity solutions were
derived (cf. Eq.2.17). As a result, similarity solutions to (3.13) subject to (3.14)-(3.15) can only obtained
in an approximate sense. For this purpose, the effective volume is assumed to vary with time according to
the power law
where q
eff
and Q
eff
are some effective source parameters. The effective source parameters (q
eff
and Q
eff
) defined in such a manner are fictitious source parameters that would produce a cumulative spill volume
of V
eff
(t) at time t in the absence of any volume loss. The determination of q
eff
and Q
eff
can be made given
V
eff
(t), ", and time t from (3.16). Substituting (3.16) into (2.6), taking the derivative with respect to time,
and assuming that q
eff
or Q
eff
are slowly varying in time yields the following rate of change in the extent of
spreading
Eq.(3.17) can be solved using Runge-Kutta integration scheme for ordinary differential equations. Once
the extent is solved, the spatially averaged oil lens thickness can be easily calculated given the volume of
oil on ground and the area of spreading. Detailed implementation for (3.17) is discussed in Section 3.4.3.
Note that for a point source (m =1), the shape of spreading is circular with radius equal to x
N
(see Figure
3.2), and for a line source (m =0), the shape of spreading is rectangular with width equal to the source
width, and the total length being 2x
N
(x
N
is the extent of spreading in one direction). Figures 3.3a and 3.3b
provide a schematic illustration of the shape of a viscous oil spill from a line source and a point source,
respectively, on a flat surface.

C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 14
Line Spill
w
d
x
N
x
N
Plane view
Cross-section view
Point Spill
x
N
Plane view
Cross-section view
o
x N
Line Spill
w
d
x
N
x
T
Plane view
Cross-section view
Point Spill
x
N
Plane view
Cross-section view
x
N
Point source
x
T
2y
N

(a) horizontal surface, line source (b) horizontal surface, point source


(c) inclined surface, line source (d) inclined surface, point source
Figure 3.3 Schematic Illustrations of the Shape and Extent of Viscous Flow of Oil Resulting
from Point and Line Sources on Horizontal and Inclined Surfaces
.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 15
(3.18)
(3.19)
(3.20)
(3.21a)
3.4.2 Inclined Surfaces
The viscous flow of oil over a permeable rugged and inclined surface subject to infiltration and volatilization
is governed by Eq.(2.10) with the addition of two sink terms on the right-hand-side, i.e.,
where 2 is the angle of inclination of the surface (measured from the horizontal). This scenario is also
subject to the appropriate global continuity equation. In the discussion that follows, Listers [1992] similarity
solutions to gravity currents over impermeable inclined surfaces are extended to include infiltration and
volatilization as presented below.
3.4.2.1 Two-dimensional Flow From A Line Source
In the presence of surface roughness, infiltration and volatilization, the global continuity (2.11) for gravity
currents over a smooth and rigid surface is modified to give
where the effective oil volume is given by
Following the assumptions regarding the effective oil volume and the procedures proposed for the horizontal
surface case, the similarity solution of Lister [1992] for gravity currents over an inclined surface, Eq.(2.13),
can be extended to include infiltration and volatilization. Thus, the rate of change of the extent of spreading
is given by
and
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 16
(3.21b)
(3.22)
(3.23)
(3.24b)
(3.24a)
where the parameters are the same as defined in (2.13). As in the case of viscous gravity currents over a
smooth and rigid surface, it is assumed that the transition time t
*
can be evaluated using (2.12). As is shown
in Figure 3.3c, the spreading is in both down- and up-slope directions during the time period up to t
*
.
Afterwards, spreading is strictly in the down slope direction. A detailed description of the solution
procedures for solving (3.21) is provided in Section 3.4.3.
3.4.2.2 Axisymmetric Flow From A Point Source
In the case of viscous flow on an inclined surface resulting from a point source release (see Figure 3.2), the
global continuity equation is modified based on (2.16) to give
where the effective volume is defined by
Following the same procedures as in the horizontal surface case, the similarity solutions (Eq. 2.28), of Lister
[1992] for axisymmetric gravity currents over a smooth and rigid inclined surface are extended to include
infiltration and volatilization. The resulting rate of change in the extent of spreading is given by
and
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 17
(3.24c)
where the parameters are the same as defined in (2.18). As in the case of viscous gravity currents over a
smooth and rigid surface, it is assumed that the transition time t
*
can be evaluated using (2.17). The shape
of the spill is assumed to be circular for the initial stages of the release ( t < t
*
). For t >t
*
the down-slope
dimension of the spill increases at a faster rate than the cross-slope dimension. The shape of the spill after
time t
*
is approximated by an ellipse, as is shown in Figure 3.3d. The major axis of the ellipse is equal to
(x
N
(t) +x
T
), and the minor axis is equal to 2y
M
(t) (see Figure 3.2). Implementation of (3.24) is described
in Section 3.4.3.
3.4.2.3 Discussion
It should be noted that the approach presented above for flow on inclined surfaces provides an
approximate description of the shape and dimensions of the oil spill. For a point source, the spill shape is
expected to deviate from the circular shape (for t <t
*
) and elliptical shape (for t>t
*
) assumed above
because of the formation of fingers at the front of the spill. The fingers channel the flow down-slope, thus
inhibiting further cross-slope spreading. The formation of fingers on an irregular ground surface is very
complicated and cannot be predicted with the current approach due to lack of quantitative representation
of the irregularities and lack of knowledge about the processes involved.
3.4.3 Implementation
The proposed methodology for quantifying the surface flow of viscous oils on a permeable and rugged
horizontal or inclined soil surface produces one or more ordinary differential equations that describe the rate
of change in the extent of surface spreading. These ordinary differential equations are best solved using a
Runge-Kutta integration scheme with adaptive time-step control [Press et al., 1992], as for the surface flow
of non-viscous oils. Essentially, the same implementation procedures proposed for non-viscous flow (see
Section 3.3.3) can be adopted for the viscous case. The difference is that in the non-viscous case, there
is only one unknown variable to solve volume of oil on the land surface; whereas in the viscous case,
there are potentially two unknowns the extent of spreading in both x and y directions. In all cases, spatial
variations in oil thickness on the ground surface are ignored.
As before, let t
j
(j =0, 1, 2, ...) be the outer time levels and t
o
represent the spill start time. The initial
conditions for the extent of spreading are, in general, zero. At any time step j, the problem reduces to the
solution of the extent of spreading at time t
j+1
given conditions at t
j
. This is accomplished using Runge-Kutta
integration [Press et al., 1992]. The Runge-Kutta scheme requires the evaluation of the right-hand-sides
of the rates of change in the extent of spreading at several time points between t
j
and t
j+1
. An examination
of the rate of change in the extent of spreading as represented in Eq. (3.17), (3.21), and (3.24) reveals that
the evaluation of the
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 18
(3.25)
(3.26)
derivatives requires knowledge of the appropriate proportionality constants (>
N
, 0
N
,...), the effective volume
V
eff
(t), the time variable oil density and viscosity (if mixture models are used for calculating volatilization
rates), and the transitional time t
*
(for the case of an inclined surface).
3.4.3.1 Proportionality Constants
In the proposed approach, the proportionality constants >
N
, 0
N
... are calculated based on the temporal and
spatial patterns of the source spill (pulse or constant-flux or time-variable; point source or line source),
surface condition (horizontal or inclined) and, in the case of flow on an inclined surface, on t
*
, as discussed
in Section 2.3. The value for t
*
can be evaluated from (2.12) or (2.17), without accounting for the effect
of infiltration and volatilization. In the particular case of viscous flow from a point source over an inclined
surface, the long-time solutions to the proportionality constants >
N
*
and 0
M
*
in Eq. (2.19) will be invoked
only when t > 1.05 t
*
to avoid the singularity associated with the term s=ln(t/t*) in the denominator of
Eq.(2.19).
3.4.3.2 Effective Oil Volume
In order to find V
eff
(t), the area of spreading A(t) is calculated first given the shape and an estimate of the
extent of spreading at time t. For flow over a horizontal surface, the shape is rectangular or circular for a
line source or a point source, respectively. For flow over an inclined surface, the shape is rectangular for
a line source with width equal to the source width and length equal to x
N
(t)+ x
T
(t). For a point source,
the shape is circular with radius x
N
(t) for t < t
*
, and elliptical with major and minor axises
a=(x
N
(t)+x
T
(t))/2 and b=y
M
(t), respectively for t >t
*
. Given A(t) and the incremental subareas before
time t
*
,

( t
j
#t #t
j+1
), the incremental subarea the oil front scanned from time t
*

to t is given by
Evaluation of the effective oil volume on the ground surface at any time t ( t
j
#t #t
j+1
) is performed
according to the following generalized discrete version of (3.15), (3.20), and (3.23)
where V
t
(t
j
) is the estimated total oil volume remaining on the ground at t
j
(assuming V
t
(t
0
)=0), *V(t) is
the volume of oil spilled from t
j
to t, and other terms are the same as defined earlier. The minimum operator
on the last term in (3.26) is to ensure that the loss to depression storage is no greater than V
t
(t
j
)+*V(t),
the maximum total volume available on the surface at time t. Calculation of the infiltration rates q
i
(l, t) at
time t requires knowledge of average oil thickness on the ground at time t, which is estimated as
(V
t
(t
j
)+*V(t))/A(t). Procedures for the evaluation of infiltration and volatilization rates are then the same
as outlined in Section 3.3.3. After accounting for loss to infiltration and volatilization, the total oil volume
remaining on the ground at time t is
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 19
(3.27)
(3.28)
(3.29)
(3.30)
from which the average oil lens thickness at any time is calculated as
3.4.3.3 Treatment of Singularities
The evaluation of the right-hand-sides of the ODEs represented by Eq. (3.17), (3.21) and (3.24) at any
time is made using the effective oil volume calculated earlier. Because of the singularities of these ODEs at
t=0, special considerations need to be given to the evaluation of the right-hand-sides of these ODEs.
Consider, for example, the flow over a horizontal surface. The rate of change in the extent of spreading is
given by Eq.(3.17), which is singular at t=0. Let t
,
be some pre-specified small time. The value of x
N
at
t = t
,
can be approximated by (cf. Eq.2.6)
where V
eff
is the effective oil volume on the ground at time t= t
,
. To circumvent the singularity of the
derivative dx
N
/dt at t=0, an effective value of dx
N
/dt for t < t
,
is defined as
In actual computer implementations, the first time-step size can be set to a small value. The value of t
,
is
then chosen to be equal to the first time-step size, but no less than 0.1 sec. The above argument applies to
viscous flow over inclined surfaces as well.
3.4.3.4 Evaluation of Maximum Up-Slope Extent x
T
For viscous flow of oil over an inclined surface, it is important to estimate the maximum extent x
T
of spill
in the up-slope direction, as defined in Eqs.(3.21a) and (3.24a). This value is used in estimating the shape
and the area of surface spreading of oil over an inclined surface. At first, this appears trivial, since it is
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 3- 20
simply the integration of the rate of change of x
T
from spill start time to t
*
, the transition time between the
short-time solution and the long-time solution. As indicated earlier, when adopting the Runge-Kutta scheme
with adaptive time step control to solve the ODEs (3.21) or (3.24) for the extent of spill at time t
j+1
given
conditions at t
j
, evaluation of the RHS of these ODEs is required at several intermediate time values
between t
j
and t
j+1
. It is when evaluating the RHS of these ODEs that the estimation of the maximum up-
slope extent, x
T
is made. The maximum value of the up-slope extent x
T
is found by setting it equal to the
short-time solution of x
N
(t), where t is some time between t
j
and t
j+1,
when t is the nearest value to t
*
.

3.4.3.5 Treatment of Finite Duration Spill Sources
When a source stops spilling beyond a turnoff time, T
turnoff
, the rate of change in the extent of the spill will
not be as large as in the case where the source spills continuously. To account for this, the value of " in the
derivatives (3.17), (3.21) and (3.24) will be set to zero for t > T
turnoff
. Such a treatment is in keeping with
the observations of Huppert [1982a] that the mode of release of fluid (i.e, whether it is a pulse or a constant
flux spill) made no difference after a few seconds from the start of release. The proportionality constants,
however, will be kept the same as before the source turnoff time, T
turnoff.

PART 2:
OILSFSM FORTRAN CODE DESCRIPTION
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 1
4.0 OILSFSM CODE DESCRIPTION
4.1 INTRODUCTION
A computer simulation program, OIL Surface Flow and Screening Model (OILSFSM), has been
developed to simulate the coupled surface flow, infiltration, and volatilization of oil using the methodology
proposed in Section 3.0. In this section, a detailed description of the code is presented.
4.2 SIMULATION OBJECTIVES
The purpose of the OILSFSM model is to provide a screening tool for characterizing crude oil and
petroleum hydrocarbon fuel spills on the land surface, its subsequent spreading, infiltration and volatilization.
In particular, it is designed to simulate the shape, extent, thickness and area of oil spreading on the land
surface, and the volumetric infiltration and volatilization fluxes as they vary with time.
4.3 SIMULATION ALGORITHMS
For both non-viscous and viscous flow over a horizontal or inclined surface, the primary dependent
variables (PDVs) of solution can be reduced to one or two ordinary differential equations (ODEs) (see
Sections 3.3 and 3.4). These ODEs express the rates of change of the PDVs as a function of source spill
scenario, infiltration, and volatilization. A powerful and accurate method for solving such ODEs is Runge-
Kutta integration with adaptive step control [Press et al., 1992]. In OILSFSM, the Runge-Kutta routine
ODEINT from the Numerical Recipes [Press et al., 1992] is adopted to solve the system of ODEs. The
Runge-Kutta routine ODEINT requires the beginning time, ending time, values of PDVs at the beginning
time, error tolerance, and an external subroutine with the generic name of DERIVS. The purpose of
DERIVS is to evaluate the right-hand-side of the ODEs.
4.3.1 Algorithm For Surface Flow
For non-viscous flow, there is only one PDV the volume of oil on the ground surface at any time. The
right-hand-side of the ODE is as given in Eq. (3.3). Procedures for calculating all the necessary terms are
given in Section 3.3.3. Special consideration is given to the case of an impulse spill. Such a source spill
scenario is handled by setting the initial volume on the ground surface equal to the volume spilled. For all
other spill scenarios, the flux from the source is added to the right-hand-side of the ODE. Since oil lens
thickness on the surface is assumed to be controlled by surface roughness and surface tension, it is known
a priori. Once the volume is found, the area is calculated via Eq.(3.4) bounded by the area of a berm.
For viscous flow, there is either one or two PDVs. For a line source, a single PDV represents the extent
of spreading in the direction perpendicular to the source. For a point source, a single PDV represents the
radius of spreading for a spill on a horizontal surface. For a point source spill on an inclined surface, there
are two PDVs that represent the extent of spreading in the down-slope and cross-slope directions.
Procedures for calculating the right-hand-sides of these ODEs are as given in Section 3.4.3. Special
consideration is given to the evaluation of the right-hand-sides of the ODEs at small time because of the
singularities at t =0 (cf. Eqs. 3.17, 3.21, 3.24). For very small time, say t <1 sec, the rate of change in
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 2
the PDVs is assumed to be constant so that the integration of the rate of change in the PDVs from t =0
to t =1 sec is equal to the corresponding values of these PDVs as given by (2.6), (2.13) and (2.18)
evaluated at t =1 sec, assuming oil density, viscosity, and effective oil volume are constant.

4.3.2 Algorithm For The Green-ampt Infiltration Model
If the Green-Ampt model is used for simulating infiltration, a Newton-Raphson scheme is employed to solve
Eq.(3.9) for the infiltration front. The solution is then used along with Eq. (2.21) to obtain the infiltration
rate. The maximum number of nonlinear iterations allowed, ITERMX, and the tolerance error EPS (cm)
are input by the user. When the difference between the solutions to the infiltration front at two successive
iterations is less than EPS, a satisfactory solution is found and the iterations stop. If, however, the tolerance
level EPS is not met after ITERMX iterations but the solution is converging, the solution at the final iteration
is considered as the true solution. A warning message regarding this can be made available through the
run-time message output control switch KRTMSG2, as will be discussed later.

4.4 CODE CAPABILITIES AND LIMITATIONS
To use OILSFSM correctly, the user must be aware of the capabilities, limitations and key assumptions
of the model. These are described in the following paragraphs.
4.4.1 Source Spill Scenarios
A source spill scenarios in OILSFSM is described by its spatial and temporal patterns of release. In space,
a spill can occur at a point or along a line that is assumed to be infinitely long. In the former case, the
subsequent spreading can be in any direction away from the source. In the later case, surface spreading
is assumed to be in the directions perpendicular to the line. Note that the code cannot simulate the surface
flow of non-viscous oil over an inclined surface that is induced by a spill from a point source.
The temporal pattern of a spill can be one of the following:
A pulse (instantaneous) spill with a fixed volume V
A finite duration spill (that lasts until a known turnoff time) with a constant flow rate Q
A continuous spill (that lasts until the maximum simulation time) with a constant flow rate Q
A finite duration spill (that lasts until a known turnoff time) with a cumulative spill volume increasing
according to the power law: [L
3
] (where " is a non-negative real parameter, and Q is
a constant whose dimension is such that V must be in [L
3
]
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 3
A continuous spill (that lasts until the maximum simulation time) with a cumulative spill volume
increasing according to the power law:

Note that for finite duration spill scenarios, the simulation of surface flow, infiltration and volatilization may
continue beyond the source turnoff time.
4.4.2 Types of Oil
The types of oil spilled can be one of two kindsnon-viscous and viscous types. (See Section 2.3.1 for
a discussion of the difference in model conceptualization for viscous and non-viscous fluid flow.) Flow of
non-viscous oil is fast and is characterized by the instantaneous establishment of an constant oil lens
thickness. Gasoline can be considered as an example of non-viscous oil. Flow of viscous oil, however, is
driven by gravity, and retarded by viscous forces; the solution to the corresponding flow equations is based
on extension of the similarity solutions of Huppert [1982a,b] and Lister [1992] developed for gravity
currents over smooth and rigid surfaces. If the user is uncertain of the oil type, the code will decide the
appropriate type according to Eq. (2.1) given the oil density, viscosity and source spill scenario. Although
Eq.(2.1) was developed for gravity currents over a horizontal surface, it is assumed that it applies to an
inclined surface as well. The choice of viscous or non-viscous flow is made before the simulation begins,
and the flow type is not changed during the process of the simulation regardless of the potential changes
in the oil density and viscosity.
4.4.3 Surface Conditions
In OILSFSM, surface conditions are characterized by the surface inclination, roughness, the presence of
a berm and wind velocity.
The land surface can be horizontal or inclined. If it is inclined, the user must provide the angle of inclination.
The angle of inclination must not be too large since the similarity solutions developed by Lister [1992] for
gravity currents over inclined surfaces assume that the angle is small. Surface roughness, also called
depression storage per unit surface area, or simply depression storage in this report, reduces the effective
volume of oil on the surface available for flow, thus impeding surface spreading.
The presence of a berm prevents the oil on the ground from further lateral spreading. In OILSFSM, the
edges of the berm are considered to be reached once the area of spreading at any time is greater than or
equal to the berm area. An implicit assumption in this treatment is that the shape of the berm is similar to
that of the oil lens on the surface. This assumption is adopted because it greatly simplifies the numerical
implementation.
Wind velocity has an effect only on the volatilization. The wind direction is not taken into account in the
OILSFSM model. A zero wind velocity results in no volatilization.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 4
4.4.4 Infiltration
In OILSFSM, the following assumptions are made regarding infiltration
Infiltration is one-dimensional, vertically downward
Soil properties are homogeneous everywhere
Air phase is immobile
Air and water saturations are uniform
During infiltration, oil displaces air only
There exists a sharp infiltration front
No volatilization is simulated once oil has infiltrated into the soil
Two infiltration models are available in the code the Philip model and the Green-Ampt model. For a
detailed discussion on these two models, the user is referred to Sections 2.4.1 and 2.4.2. The limitation
of both models is that they cannot simulate the redistribution of the oil in the soil once the oil lens thickness
on the ground surface goes to zero. Further, the oil front stops advancing once the oil disappears from the
land surface.
4.4.5 Volatilization
Volatilization or evaporation of oil from the surface depends on the exposure conditions and compound
properties (see Section 2.5). Two kinds of volatilization models are available in OILSFSM a pure liquid
model and mixture models. In the pure liquid model, composition of the oil is ignored; whereas in the
mixture models, the oil consists of NVOC number of volatile compounds and NNVOC number of non-
volatile compounds. The mixture models are further subdivided into two sub models: the mean effective
molar volume model (See Eq.2.33) and the activity coefficient model (Eq.2.35). The former assumes that
the mean effective molar volume of the oil is known or can be estimated. The latter assumes that the
activity coefficient of each volatile compound in the oil is known. The key assumptions made in these
volatilization models are:
Thermal mixing is fast compared to surface flow and volatilization such that temperature reaches
equilibrium over the whole spreading area instantaneously.
In the mixture model, oil composition reaches equilibrium over the whole spreading area
instantaneously.

4.5 MAJOR CODE FEATURES
In developing the OILSFSM model, every effort was made to ensure that the code is easy to use, requires
as little as possible from the user, is efficient in the usage of computer memory and CPU time, is user
friendly, and facilitates future extensions and modifications. The following presents the major features of the
code OILSFSM.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 5
4.5.1 Input Features
In OILSFSM, input preparation is made easier by the following options:
In-line comments in the input data file
Classification of model input parameters into distinctive groups
Default settings of some control parameters and constants, and the option of overwriting any one
or more of them
Access to soil and oil database files, and the options of overwriting any one or more of the values
read from database files.
4.5.2 Time Marching Control
Time-step control for the outer time levels in OILSFSM is accomplished by the time marching scheme
switch KTMSCH selected by the user. Three time marching schemes are available:
Time marching with constant time-step size (KTMSCH=1)
Time marching with variable time-step sizes (KTMSCH=1)
Adaptive time-stepping (KTMSCH=1)
In the first marching scheme, the user specifies the beginning time t
0
, ending time t
max
, and the total number
of time steps, NTS. The constant time-step size is then calculated as ) t=(t
max
-t
0
)/NTS. Time-stepping will
then proceed with time levels t
0
, t
1
(=t
0
+) t), t
2
(=t
0
+2) t), ..., until t
max
is reached or some stopping rule
is met (see Section 4.5.3). In the second marching scheme, the user specifies the total number of time-steps
NTS, and the time sequence, t
0
, t
1
, t
2
, ..., t
nts
(t
max
=t
nts
). The difference t
j+1
-t
j
is simply the j-th time-step
size (j=1, 2, ..., NTS). For adaptive time-stepping, the user provides the initial time t
0
, the maximum
simulation time t
max
, the initial time-step size ) t, the maximum allowable time-step size ) t
max
, and the time-
step multiplier DTMUL (DTMUL >1.0). Subsequent time-step sizes are decided adaptively as the time-
stepping proceeds, taking into consideration the maximum allowable time-step size, the time levels at which
intermediate results are desired, and the number of Newton-Raphson iterations required to obtain solutions
to the nonlinear Green-Ampt infiltration equation in the previous time-step.
4.5.3 Stopping Rule

The execution of OILSFSM continues until some stopping rule has been satisfied. In OILSFSM, the
stopping rule is represented by KSRULE, a string 5 characters long that consists of combinations of the
binary values of 0 or 1. The most basic stopping rules are:
KSRULE=00001 Source stopped spilling
KSRULE=00010 Front stopped advancing
KSRULE=00100 Edges of berm reached
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 6
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 7
KSRULE=01000 Oil thickness on surface reached zero
KSRULE=10000 Maximum simulation time attained
The actual stopping rule can be any combination of the basic rules specified above. For example, the
stopping rule represented by KSRULE=100101 implies that the execution of OILSFSM continues until
one or more of the three conditions: (1) source stopped spilling; (2) edges of berm reached; or (3)
maximum simulation time attained, are satisfied.
The default stopping rule for both viscous and non-viscous flow on a flat surface is the attainment of the
maximum simulation time t
max
, or the oil lens thickness becoming zero, which is represented by
KSRULE=11000. The default stopping rule for both viscous and non-viscous flow on an inclined surface
is the attainment of the maximum simulation time t
max
, or the oil lens thickness becoming zero or the oil front
reaching the edge of the berm, which is represented by KSRULE=11100.
4.5.4 Run-time Message Output

Run-time messages include error messages and execution status. There are potentially two types of errors
during the execution of OILSFSM: warning errors and fatal errors. Examples of warning errors are (1)
the maximum Newton-Raphson iterations are exceeded without convergence; or (2) the time-step size is
too large for the Runge-Kutta routine ODEINT. An example of a fatal error is that the time-step size is
reduced to below the minimum level specified without obtaining a satisfactory solution. In this case, a fatal
error occurs after all corrective measures have failed.
During each session, a run-time log file will be created. Run-time messages can be directed to two output
devicesthe monitor screen and the run-time log file. The amount and level of run-time message output
to these two output devices is controlled by the integer control switches KRTMSG1 and KRTMSG2,
respectively. The values of KRTMSG1 and KRTMSG2 and their corresponding amount and level of run-
time message output are as follows:
Run-time message output to the monitor
KRTMSG1=1 (Default) fatal error message only
KRTMSG2=2 Fatal + warning error messages
KRTMSG3=3 Fatal + warning error messages plus time-stepping information
Run-time message output to log file
KRTMSG2=1 (Default) fatal error message only
KRTMSG2=2 Fatal + warning error messages
KRTMSG2=3 Fatal + warning error messages plus values of key solution variables

The first time the code is run, it is desirable to set KRTMSG1 and KRTMSG2 to 3 such that the maximum
amount of information concerning run-time errors is directed to either the screen or the run-time log file.
This will be helpful for trouble-shooting, should it prove necessary. If it is expected that execution will
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 8
proceed smoothly, setting KRTMSG1=1 and KRTMSG2=1 would greatly reduce the I/O time, and
enhance the overall efficiency.
4.5.5 Output Control For Simulation Results
Execution of OILSFSM generates three types of results at user desired time levels. The first type consists
of extent of spreading (radius/frontal distances from source), shape and area of spreading, spatially
averaged oil lens thickness, volumetric infiltration flux and volatilization flux over the spreading area. The
second type consists of mass balance information. The third type consists of time history of infiltration rates
and fronts at user selected locations. The frequency and level of detail of these three types of results are
controlled by the integer control switches KOUT1, KOUT2, and KOUT3 respectively. The values of
KOUT1, KOUT2 and KOUT3 and their implications are discussed below.
Control switch KOUT1
KOUT1=0 Report results only at final time or when the stopping rule is met
KOUT1=n1 (n1 >1) Report results every n1 time-steps plus the final time
KOUT1=! n1 (n1>1) Report results every n1 levels from the user specified vector TOUTLEV(j),
j=1, 2, ..., NOUTLEV.

In the case of KOUT1=! n1, the user is required to provide the number of time levels in NOUTLEV and
output time vector TOUTLEV(j). Should this option be chosen, the time-step size adjusts appropriately
in order to obtain solutions at targeted time levels in TOUTLEV. Such an adjustment is not considered as
a time step; thus, it has no bearing on the time-step counter and time-step size output onto the screen. This
option is especially useful when the user is interested in the results at a fixed number of time values.
Control switch KOUT2
The reported mass balance information includes, for any given time level, the cumulative oil mass spilled,
total mass remaining on the ground, mass fraction of each compound, and cumulative mass losses to
infiltration and volatilization for each compound. The reporting frequency and level of detail of the mass
balance results are controlled by the integer control switch KOUT2, as noted below:
KOUT2=0 Report results only at final time or when the stopping rule is met
KOUT2=n2 (n2 >1) Report results every n2 time-steps plus the final time
KOUT2=! n2 (n2>1) Report results every n2 time levels in the user specified vector
TOUTLEV(j), j=1, 2, ..., NOUTLEV.
Control switch KOUT3
The reporting frequency and level of detail of the infiltration rates and fronts at any number of selected
locations are controlled by KOUT3, as noted below:
KOUT3=0 Report results only at final time or when the stopping rule is met
KOUT3=n3 (n3 >1) Report results every n3 time-steps plus the final time
KOUT3=! n3 (n3>1) Report results every n3 time-levels in the user specified vector
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 9
TOUTLEV(j), j=1, 2, ..., NOUTLEV.
When time histories of infiltration rates and fronts are desired, it is necessary specify the number of such
observation (NOBS), and the desired spatial locations, {XOBS(j) YOBS(j)} (j=1, 2, ..., NOBS). Because
infiltration rates and fronts are assumed to be the same within each incremental subarea (cf. Sections 3.3.3
and 3.4.3), it is desirable to spread observation points along the frontal propagation direction in order to
obtain a better description of the infiltration profile in space.
4.6 MODULE DESCRIPTION AND FLOW CHARTS
The OILSFSM code consists of a main program and several relatively independent modules or
subprograms. Flow charts for the most important modules are shown in Figures 4.1-4.5. The following is
a description of the modules/subroutines that comprise the code, and the major functions of each of the
routines.
Module Name Description of Module Functions
MAIN (Figure 4.1) To provide the user with an interactive channel to submit
either a single job or batch (multiple) jobs to the model during each
session. A log file is created for each sessionit records all I/O, run-time
warning and fatal messages, and execution status. For each problem set,
the user is expected to provide the main input file name, listing file name,
and output file name. The listing file is generated by a subroutine called
LISTDATA, which summarizes all input parameters. The output file
contains all solution results. If the listing file and output file share the same
name, input parameter listings and results will be directed to the same file.
After processing I/O file names, MAIN calls MAINSUB, the main
subroutine, to solve an individual problem set. The completion status of
each problem set will be echoed to the screen after each call to
MAINSUB. If fatal errors occur when solving a particular problem,
processing of that problem set is aborted, and the code processes the next
problem set, if any.
MAINSUB (Figure 4.2) To process and solve a single problem set. It calls subroutine
GETDATA to read in all model parameters, subroutine CHKDATA to
check the validity of those parameter values, subroutine LISTDATA to
print listings of input parameters along with their definitions, and subroutine
INITCOND to perform initialization of variables and arrays. It then
performs time-stepping according to the user selected time marching
scheme, taking into consideration the time levels at which results are to be
directed to output devices. The time-stepping in MAINSUB is called the
outer time loop. It is not restricted by the kind of solution scheme. During
each time step (from, say TIME1 to TIME2), MAINSUB calls the solver
subroutine SOLVER, which uses the generic external subroutine
DERIVS. The SOLVER returns solutions of oil lens area, spreading
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 10
extent, average thickness, volumetric infiltration and volatilization fluxes
(cm
3
/s) at time=TIME2 given conditions at time=TIME1. If desired,
MAINSUB also records the infiltration rate and infiltration depth below
the surface at any selected spatial location and at any user desired time
level . The outer time loop continues until some stopping rule (see Section
4.5.3) is met.
GETDATA To read all model input parameters from a main input data file,
supplemented by access to soil and oil databases (a Users Guide
provides guide lines for preparation of an input data file). Default control
parameters and some physical parameters are provided for the users
convenience.
DERIVDATA To derive model parameters that cannot be read from the input data file.
These parameters include overall oil properties, effective oil conductivity
and pressure head, proportionality constants, etc.
CHKDATA To check for possible errors in the model parameters. If fatal errors are
found, execution of the problem is aborted, and the execution status is
returned to MAIN.
LISTDATA To print a summary listings of all input parameters.

SOLVER (Figure 4.3) To solve for the spreading characteristics of the oil spill, such
as lens area, thickness, extent, volume remaining on ground surface,
infiltration flux and volatilization flux at time =TIME2 (=TIME1+DELT)
given conditions at time=TIME1, where DELT is the outer time-step size.
For the very first attempt, T1 is set equal to TIME1, and T2 to
TIME1+DTMP, where DTMP is the sub time-step size taken inside
SOLVER. By default DTMP is set equal to DELT. For viscous flow of
oil, however, DTMP is set to a value small enough to avoid singularities
associated with the rate of change in the extent of spreading (see Section
3.4.3.3). SOLVER then calls ODEINT, a subroutine taken from
Numerical Recipes (Press et al., 1992) to perform the integration of the
appropriate ODE (ordinary differential equation) from T1 to T2 using
adaptive time-step control. This requires evaluation of the right-hand-side
derivative of the ODE. The generic subroutine DERIVS (whose specific
name is RHSNVISC for non-viscous flow and RHSVISC for viscous
flow) is passed to ODEINT for this purpose. Subroutine ODEINT makes
repeated calls to DERIVS to evaluate the right-hand-side function of the
generic ODE dy/dt = f(t,y), where y is the primary dependent variable
(vector) and t is time (for non-viscous flow of oil, y is total volume of oil
on ground surface, and for viscous flow of oil, y is the extent). If ODEINT
fails to find solutions at T2, then a middle point, T2= (T1+T2)/2, is
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 11
chosen, and another call to ODEINT is carried out to find the solution at
T2 given conditions at T1. If solution at T2 is found, and T2 < TIME2,
T1 is set equal to T2, and T2 to T1+DTMP with T2 being constrained by
T2 < TIME2. This process is repeated until a solution to y at
time=TIME2 has been obtained.
ODEINT To perform the integration of an ODE of the form: dy/dt = f(t, y) from T1
to T2 using the Runge-Kutta scheme with adaptive step size control.
ODEINT makes repeated calls to a generic external subroutine called
DERIVS (whose specific name is RHSNVISC for non-viscous flow of
oil, and RHSVISC for viscous flow of oil) to evaluate right-hand-side
function f(t, y) at intermediate time values between T1 and T2. ODEINT
returns the solution to y at time=T2 to the calling program SOLVER. For
a detailed description of this routine, the user is referred to Press et al.
[1992].
RHSNVISC (Figure 4.4) RHSNVISC is the specific name of the generic subroutine
DERIVS for evaluating the right-hand-side of the ODE, Eq.(3.3). For a
complete description of the implementation procedure, see Section 3.3.3.
RHSNVISC (Figure 4.5) RHSVISC is the specific name of the generic subroutine
DERIVS for evaluating the right-hand-side of the rate of change in the
extent of spreading, Eqs.(3.17), (3.21), or (3.24), depending on source
release scenarios and surface conditions. For a complete description of
the implementation procedure, see Section 3.4.3.
GREEN_AMPT To evaluate the infiltration front, infiltration rate and overall infiltration flux
at any time t given the conditions at t
0
for the Green-Ampt model. For a
complete description of implementation procedures, see Eq.(3.9) and
related text.
JRPHILLIP To evaluate the infiltration front, infiltration rate and overall infiltration flux
for Philips model. For further description, see Eq.(3.8) and related text.
VLTSUB1 To evaluate volatilization flux for the pure liquid model.
VLTSUB2 To evaluate volatilization flux for the mixture model using the effective
molar volume approach.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 12
VLTSUB3 To evaluate volatilization flux for the mixture model using the activity
coefficient approach .
GLQZW To find the points and weights for Gauss integration scheme.
ETAXIN To solve the ordinary differential Eq. (2.9a,b) for the similarity solution N
using Runge-Kutta integration inward from 0=1, and solve Eq. (2.8) for
the proportionality constant for arbitrary " using Gauss integration.
In addition to the above mentioned subroutines, the code also contains two important modules. One is
called MODELPARMS, which stores all the model parameters that the code reads from the input data file,
and the other is SOLNVARS, which stores all key solution variables. The two modules are accessed by
any subroutine through the USE MODULE statements provided by FORTRAN 90. Dynamic allocation
of arrays is used to optimize computer memory requirements as well as for ease of use.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 13
START
Echo program and copyright information
Read name of log file that records run-
time error messages and execution status
Read in job mode from screen
no
yes
Read name of batch file
from screen
Read main input,
listing and output file
names from screen
Read main input ,
listing and output file
names from batch file
Call MAINSUB to solve
a single problem set. Returns
ierr, error indicator
yes
no
no
yes
END
See MAINSUB
flow chart for details
Batch job?
Batch job?
End
of batch file?
Error condition handling and message output:
If ierr=0, execution successful
If ierr=1, at most warning errors occurred
If ierr=2, fatal error occurred
Figure 4.1 Program MAIN Flow Chart
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 14
Call GETDATA (to return ierri), CHKDATA (to return ierrc)
DERIVDATA, and LISTDATA to read, check, derive and
print model parameters. Set ierr(1)=ierri, ierr(2)=ierrc
Initialization of variables and
time stepping parameters
Print arrays containing
observation results
yes
no
Called From MAIN
Start time stepping
Select time step size t: Considering
a) user desired time marching scheme;
b) t
min
<t < t
max
;
c) times at which results are expected
Update time level
t
i+1
= t
i
+ t
Call SOLVER (with dummy RHSNVISC for non-viscous flow of oil or
RHSVISC for viscous flow of oil ) to find solution at time t
i+1
given
conditions at time t
i
. Returns ierrs. Set ierr(3)=max(ierr(3),ierrs)
Print results and/or record observations at
time level t
i+1
into arrays if necessary
no
yes
End
Return to MAIN with ierr
See flow chart
for SOLVER
for details
Fatal errors?
Stopping rule met or
kstop >1?
ierr(j) =0 (j=1,2,3 correspond to
GETDATA, CHKDATA and SOLVER, resp.)
Fatal error?
no
yes
Fi gur e 4.2 Subr outi ne MAINSUB Fl ow Char t
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 15
t
1
=t
i
, t =initial timstep size
If surface spreading continues:
KAREA2=KAREA1+1 (incremental areas counter)
Reallocate dynamic arrays associated with sub areas
Called from MAINSUB with dummy DERIVS
DERIVS is a generic name of the external subroutine
RHSNVISC (for non-viscous flow of oil), or RHSVISC (for viscous flow of oil) that
evaluates RHS of the general ODE of the form: dy/dt = f(t, y).
Objective: Find soln at t
i+1
given conditions at t
i
Call to SOLVER failed
See flow chart
for DERIVS
for details
Call ODEINT (Num. Recipes, Press et al., 1992),
passing to it DERIVS, to solve dy/dt = f(t, y)
for y (y=volume of oil for non-viscous flow of oil
or y=extent of spreading for viscous flow of oil)
at t
2
given conditions at t
1
.
t
2
= t
1
+ t
kstop=0 (stopping rule not satisfied by default), ierr=0 (no error by default)
y = y
0
=volume on ground for non-viscous oil, or extent of spreading for viscous oil at t
i
Call DERIVS
to evaluate
f(t,y) at inter-
mediate times:
t
1
< t < t
2
yes
no
Reduce time step
t = t *2/3
y = y
0
yes
no
yes
no
Call DERIVS to find y and other variables at
time t
2
. Find state of spreading : KSTATE
Update solution variables
KAREA1=KAREA2, ...
Call to
ODEINT error free?
t<t
min
? Is
stopping rule met ?
ierrs = 2
kstop= # stop cond. satisfied
End
yes
no
t
2
= t
i+1
?
t
1
= t
2
Set next t
=min(2t,
t
i+1
- t
1
)
Return to MAINSUB
with kstop and ierrs
Fi gur e 4.3 Subr outi ne SOLVER Fl ow Char t
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 16
Find spreading area AREA(t)= V/h
where h= thickness due to surface tension roughness
Called from ODEINT
RHSNVISC is the specific name of the
generic external subroutine DERIVS. Its objective is to evaluate the
Right Hand Side of dV/dt = Qs-Qi-Qv for Non-VISCous flow on flat and inclined surfaces.
Input Argument: t (time), V(volume of oil on ground)
Other input: KAREA1,KAREA2, Spreading area, infiltration rates and fronts in all subareas
at end of last time step, and model input parameters
CALL GREEN_AMPT or JRPHILLIP to find
volumetric infil. flux Qi over area = AREA(t). Store
infil. rates and fronts in each KAREA2 subareas
dV/dt = Qs-Qi-Qv
End
AREA(t) > Berm Area
?
no
yes
AREA(t)=Berm Area
h = V/Berm Area
Find volumetric source flux
Qs at time
CALL volatilization subs to find volumetric volat. flux
Qv over the area = AREA(t). Store volatilization fluxes
of individual volatile compounds
Return to Calling Subroutine
Fi gur e 4.4 Subr outi ne RHSNVISC (Gener i c Name
DERIVS) Fl ow Char t
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 17
Find spreading area AREA(t) based on
extent Xn and shape of spreading
Called from ODEINT
RHSVISC is the specific name of the generic external subroutine DERIVS.
Its objective is to evaluate the Right Hand Side of dXn/dt for VISCous flow on flat and inclined surfaces.
Input Argument: t (time),Xn (spreading extent in x {and y} directions)
Other input: KAREA1,KAREA2, Spreading area, infiltration rates and fronts in
all subareas at end of last time step, and model input parameters
CALL GREEN_AMPT or JRPHILLIP to find
volumetric infil. flux Qi over area = AREA(t). Store
infil. rates and fronts in each KAREA2 subareas
Find effective oil volume V
eff
at time t using procedure
outline in Eq.(3.26) of report
End
AREA(t) > Berm Area
?
no
yes
AREA(t)=Berm Area
Adjust Xn accordingly
CALL one of volat. subs to find volumetric volat. flux
Qv over the area = AREA(t). Store volat. fluxes of
individual volatile compounds
Return to Calling Subroutine
Find volume spilled from t
1
to t and rate of spill at t.
Estimate volume and average thickness of oil on
ground (not incl. loss to infil and volatilization)
Find dXn/dt as a function of time t, V
eff
and oil properties
Fi gur e 4.5 Subr outi ne RHSVISC (Gener i c Name
DERIVS) Fl ow Char t
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 4- 18
PART 3:
OILSFSM INPUT GUIDE AND OUTPUT INTERPRETATION
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 1
5.0 INPUT GUIDE
5.1 INTRODUCTION
Using OILSFSM to simulate an aboveground surface oil spill and the subsequent infiltration and
volatilization involves the following three steps:
Create input data file(s)
Execute OILSFSM
Perform post-processing and output analysis
This input guide provides detailed instructions on how to perform each of these three steps. The
input guide is organized as follows: First, the types of input files expected from the user for the proper
execution of the OILSFSM is discussed. Second, execution mode and output files generated by the code
during a session are described. Third, instructions on preparing a batch run are provided. Next, complete
instructions for the preparation of the main input data file are given. This is followed by descriptions of the
soil and oil databases included with the code.
5.2 TYPES OF INPUT FILES
Four kinds of input files, some of which are optional, are needed for proper execution of the OILSFSM
model. They are:
A batch file (optional)
Main input data files(s) (mandatory)
An soil database file (optional)
An oil database file (optional)

The above mentioned input files are ASCII text files. A batch file is required when the user wishes to
conduct multiple runs during a single session (see description below). The main input data file(s) are
mandatory. The soil and/or oil database files are existing data files included with the code. Access to one
or both of the two database files is optional, depending on users preference. The user should use care in
modifying the soil and oil database files.
5.3 EXECUTION MODE
During each session, OILSFSM can be executed in one of two modessingle job mode or multiple
(batch) job mode. In the single job mode, the code will prompt the user for the name of the main input data
file. It will then read, check and process the data, and perform the required simulations. In the multiple
jobs or batch mode, the user must provide a batch file that contains the names of at least one or more
independent input data files; the code will read, check, and process input data, and perform the required
simulation for each input data file sequentially until all the input data files have been processed.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 2
5.4 OUTPUT FILES
Three kinds of files are generated during each computational session. They are:

A run-time log file
One or more listing files
One or more result summary files
The run-time log file records all the important information regarding execution of the code during a session.
The information includes starting and ending date and time of the session, important I/O processes, run-time
fatal and warning error messages and execution status.
A listing file provides a descriptive summary of all model input parameters associated with an input data
file. The default name for a listing file associated with an input data file is the same as that of the input data
file (without the extension, if any) plus the extension {.LST}. For example, if the input data file is called
TEST.DAT, then the default listing file name is TEST.LST. The default name will be overwritten if the user
specifies a name for the listing file.
A result summary or output file contains the spill extent (radius, frontal spreading distances from spill
location), average lens thickness, area, infiltration and volatilization fluxes at user desired time levels, time
history of infiltration rate and front at user specified spatial locations, and time history of mass remaining and
mass fractions on ground, mass losses to infiltration and volatilization. The default name for a result summary
file associated with an input data file is the same as that of the input data file (without the extension, if any)
plus the extension {.OUT}. For example, if the input data file is called TEST.DAT, then the default result
summary file name is TEST.OUT. The default name will be overwritten if the user specifies a name for the
result summary file.
The listing file and result summary file can share the same name, in which case, the result summary will be
appended to the model parameter listings.

5.5 BATCH FILE FORMAT
The batch file is optional; it is required only when the user wishes to run OILSFSM in multiple jobs mode.
The batch file may contain an arbitrary number of records. Each record should look like the following:
Input_file_name, {listing_file_name, result_summary_file_name}
Free format is used. File names may consist of any meaningful ASCII characters such as A-Z, 0-9, etc.,
but no commas, space, or semi-colons are allowed for file names. File names can be separated by a
comma, a space, or a semi-colon. The listing file and result summary file names are optional. In case they
are not provided, default names will be created by the code. Note that omission of the listing file name is
not permitted if a result summary file name is provided; the result summary file, however, can be omitted
if the listing file name is provided.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 3
5.6 COMPLETE INPUT INSTRUCTIONS
An input data file consists of a collection of recordseach record corresponds to one line of information
in the data file. There are two kinds of records data records and comment records. Data records
contain, as the name implies, data values to be used for proper execution of the code. Comment records
are either blank lines or any records that begin with the character string ** (two asterisks). Comment
records are not used in any way by OILSFSM; their purpose is to elucidate the meaning of the data
records and improve readability of the data file for the user. Comment records may be interspersed freely
with data records.
There are seven different groups of data records. Each group is identified by a two-character code in the
first two columns of the record. Table 5.1 provides a description of the group-codes, their description and
the number of potential data records (excluding comment records) in each group.
The order in which the seven groups appear in the data file can be arbitrary, i.e., the actual order of groups
may not follow the one given in the table above. Free format is used for all data records. The restrictions
on records in each group are:
Number of characters in any record may not exceed 132.
The first field (two characters long) should always be the group code
The second field in a data group other than the title group must contain the record number for that
group
A data record can contain at most 10 real (floating point) values after the record number. A
continuation record is needed if more than 10 real values are needed for that record. The first two
fields in a continuation record must begin with a group code and a record number. The record
number in a continuation record should be the same as the one from which it is continued
Table 5.2 provides the input instructions for all groups of data required by a simulation. Table 5.3 presents
an example input data file prepared based on the instructions given in Table 5.2.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 4
Table 5.1
Group Code, Contents, and Number of Potential Data Records
Group
Code Group Contents
No.
Potential
Records Remarks
TI Problem title records Arbitrary All records optional
SP Soil parameters records 10 Some records optional
OP Oil parameters records 13 Some records optional
SS Source and surface parameters 8 Most mandatory
TO Time marching & output control parameters 21 All records optional
SV Solver control parameters 4 All records optional
CP Constant parameters 5 All records optional
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 5
Table 5.2
Complete Input Instructions for Main Input Data File
**Problem Title Group: Optional
Group
Code
Variable
Name
Data
Type Descriptions
TI Titles(j) String (Optional) J-th line of title. Repeat for as many lines
as desired.
**Soil Parameters Group: The first record is mandatory, and the rest of the records are optional.
If a soil database file is accessed, and the user wishes to overwrite some or all of the parameters
read from the database, this can be done by providing the corresponding records that contain the
desired parameters.
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
SP 1 KINFIL Integer (Mandatory) Infiltration model switch
=0: No infiltration is simulated (rest of this group
can be omitted).
=1: Green-Ampt model (records 13 & 14 can be
omitted)
=2: J.R. Philip model (records 3, 4, 6-12 can be
omitted)
SP 2 Sw Real (Mandatory if KINFIL=1; optional otherwise)
Water saturation
SP 3 Soi Real (Mandatory if KINFIL=1; not used otherwise) Initial
oil saturation
SP 4 KSoilDBase
Soilname
Integer
String
(Mandatory if KINFIL>1; optional otherwise)
Integer control switch denoting whether soil
database file called DBASE4SOIL. DAT is to be
accessed
=0: No
=1: Yes
(Mandatory if KSoilDBase=1) Name of soil (at most
15 characters long, and in upper case).
If KSoilDBase=1, the code will search the soil
database for the soil name string given by Soilname.
If a match is found, the soil parameter values
associated with the name are read. If any parameter
values read from database are to be overwritten, the
user may provide the corresponding records below.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 6
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
SP 5 theta Real (Mandatory if KINFIL>1 and KsoilDBase=0;
optional otherwise ) Porosity
SP 6 kperm Real (Mandatory if KINFIL=1 and KsoilDBase=0;
optional otherwise) Intrinsic permeability (cm
2
)
SP 7 sro Real (Mandatory if KINFIL=1 and KsoilDBase=0;
optional otherwise ) Residual oil saturation
SP 8 srw Real (Mandatory if KINFIL=1 and KsoilDBase=0;
optional otherwise) Residual water saturation
SP 9 kbcvg
bcvg1
bcvg2
Integer
Real
Real
(Mandatory if KINFIL=1 and KsoilDBase=0;
optional otherwise) Integer switch denoting the
constitutive relation to be used
=1: Brooks-Corey model
=2: van Genuchten model
First parameter in the permeability-saturation
relationship
=Air-entry head (cm) in Brooks-Corey if kbcvg=1
=a (1/cm) in van Genuchten if kbcvg=2
Second parameter in the permeability-saturation
relationship
= Pore size index l in Brooks-Corey if kbcvg=1
= n in van Genuchten if kbcvg=2
SP 10 jrpa
jrps
So
Real
Real
Real
(Mandatory if KINFIL=2; optional otherwise ) J.R.
Philip model parameter A (cm/sec) and S
(cm/sec
1/2
). Note that these two parameters are
currently not available in the soil database file.
Maximum infiltration behind infiltration front
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 7
**Oil Parameters Group: Some records are mandatory and some optional. When the oil database
file is accessed, and the user wishes to overwrite one or more parameters read from the database,
the user must provide the record or records that contain the desired parameters.
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
OP 1 KVOLAT Integer (Mandatory) Volatilization model switch
=0: No volatilization is simulated
=1: Pure liquid model
=2: Mixture model I. The mean effective molar
volume approach
=3: Mixture model II. The activity coefficient
approach.
OP 2 temp Real (Mandatory if KVOLAT>0) Oil temperature (K)
OP 3 cang Real (Mandatory) Contact angle (radians)
OP 4 KOilDBase
Oilname
Integer
String
(Mandatory) Integer control switch denoting whether
oil database file called DBASE4OIL. DAT is to be
accessed
=0: No
=1: Yes
(Mandatory if KOilDBase=1) Name of oil (at most
15 characters long, and in upper case)
If KOilDBase=1, the code will search the oil
database for the oil name string given by Oilname. If
a match is found, the oil parameter values associated
with the name are read. If any parameter values read
from database are to be overwritten, the user must
provide the corresponding records detailed
below.
OP 5 sigma Real (Mandatory if KoilDBase=0, optional otherwise) Oil
surface tension (dyne/cm)
OP 6 effmvol Real (Mandatory if KOilDBase=0 and KVOLAT=2,
optional otherwise) Mean effective molar volume of
oil (cm
3
/mol). Suggested value = 400.0.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 8
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
OP 7 NVOC
NONVOC
Integer
Integer
(Mandatory if KoilDBase=0, optional otherwise)
Number of volatile and non-volatile compounds that
comprise the oil under consideration, respectively.
Note: Total number of compounds NCOMP =
NVOC+NONVOC. It is important to remember
that the volatile compounds are listed first followed
by the non-volatile compounds in all the property
arrays to be provided below.
OP 8 mui(j) Real (Mandatory if KoilDBase=0, optional otherwise)
Dynamic viscosity of compound j=1 (dyne-sec/
cm
2
). Repeat for j=2, 3, ..., NCOMP on same
record. A total of MOD(NCMP-1,10) continuation
records are needed, if required. If less than
NCOMP values are entered in the input data file, the
viscosity values for the remaining compounds are
automatically set equal to the very last value entered.
OP 9 rhoi(j) Real (Mandatory if KoilDBase=0, optional otherwise)
Density of compound j=1 (g/cm
3
). Repeat for j=2,
3, ..., NCOMP on same record. A total of
MOD(NCMP-1,10) continuation records are
needed, if required. If less than NCOMP values are
entered in the input data file, the density values for
the remaining compounds are automatically set equal
to the very last value entered.
OP 10 molwi(j) Real (Mandatory if KoilDBase=0 and KVOAT>0,
optional otherwise) Molecular weight of compound
j=1 (g/mol). Repeat for j=2, 3, ..., NCOMP on
same record. A total of MOD(NCMP-1,10)
continuation records are needed, if required. If less
than NCOMP values are entered in the input data
file, the molecular weight values for the remaining
compounds are automatically set equal to the very
last value entered.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 9
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
OP 11 mfraci(j) Real (Mandatory if KoilDBase=0 and KVOAT>1,
optional otherwise) Initial mole fraction (or mass
fraction if negative) of compound j=1. Repeat for
j=2, 3, ..., NCOMP on same record. A total of
MOD(NCMP-1,10) continuation records are
needed, if required. If less than NCOMP values are
entered, the mole (or mass) fraction values for the
remaining compounds are automatically set equal to
the very last value entered.
Note: Sum of mfraci(j) must equal 1.0.
OP 12 vpri(j) Real (Mandatory if KoilDBase=0, optional otherwise)
Vapor pressure of the first volatile compounds
(kPa).Repeat for j=2, 3, ..., NVOC on same record.
A total of MOD(NVOC-1,10) continuation records
are needed, if required. If less than NVOC values
are provided in the input data file, the vapor pressure
values for the remaining compounds are
automatically set equal to the very last value entered.
OP 13 acoefi(j) Real (Mandatory if KoilDBase=0 and KVOAT=3,
optional otherwise) Activity coefficients of volatile
compound j=1. Repeat for j=2, 3, ..., NVOC on
same record. A total of MOD(NVOC-1,10)
continuation records are needed, if required. If less
than NVOC values are provided in the input data
file, the activity coefficient values for the remaining
compounds are automatically set equal to the very
last value entered.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 10
**Spill Source and Surface Parameters Group: Most records mandatory
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
SS 1 KTSRS
T
turnoff
Integer
Real
(Mandatory) Oil spill pattern in time
=0: Instantaneous (pulse) spill with a volume V
0

=1: Finite duration spill at a constant rate q
0
that
lasts until time = T
turnoff
=2: Finite duration spill at a time-variable rate
according to q =q
0
a t
a-1
that lasts until
time = T
turnoff
.
=-1: Continuous spill at a constant rate q
0

=-2: Continuous spill at a time variable rate
according to q = q
0
a t
a-1
.
The two parameters defining the spill volume or flux
profile are to be provided in record 3 of this group.
(Mandatory if KTSRS=1 or 2; not used otherwise)
Source turnoff time (sec).
SS 2 KXSRS
DIMSRS
Integer
Real
(Mandatory) Oil spill pattern in space
=0: Point source at origin.
=1: Line source with width DIMSRS (see next
value of same record). For non-viscous flow
on an inclined surface, the source has to be a
line source.
(Mandatory if line source; optional otherwise) Width
of line source (cm).
SS 3 V
0
or q
0
a
Real
Real
(Mandatory) First source parameter defining the
source profile
= Total volume V
0
spilled (m
3
) if KTSRS=0
= Parameter q
0
if KTSRS=-2, -1, 1 or 2. Note: q
0

has units (m
3
/sec) if |KTSRS|=1; or for a time
variable spill, flux q=q
0
a t
a-1
is in (m
3
/s) if
|KTSRS|=2
(Mandatory if |KTSRS|=2, not used otherwise)
Second parameter defining the spill volume/flux
profile.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 11
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
SS 4 KFLOTYP Integer (Mandatory) Flow type index
=0: Non-viscous flow
=1: Viscous flow
=-1:Uncertain (Code will decide flow type given
the source scenario and oil parameters).
SS 5 KSLOPE
ANGLE
Integer
Real
(Mandatory) Type of surface
=0: Horizontal
=1: Inclined
(Mandatory if KSLOPE=1, not needed otherwise)
Angle of inclination (in radians)
SS 6 HSRS Real (Mandatory) Surface roughness (cm)
SS 7 ABERM Real (Optional) Area of berm (m
2
) (Default=1.0e30 or no
berm at all). If the user provides a value of 0.0,
ABERM will be reset to 1.0e30.
SS 8 VWIND Real (Mandatory if KVOLAT>0) Wind speed (m/hrs)
used for calculating volatilization flux (A zero value
for VWIND is equivalent to setting KVOLAT=0..
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 12
**Time Marching and Output Control Group: All records optional. Default values used unless
specified otherwise.
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
TO 1 KTMSCH Integer (Optional) Time marching scheme index
=1: Time marching with equal time-step sizes of
DELT=(TMAX-TIME0)/NTS. Values
TMAX, TIME0, and NTS are to be
provided below.
=2: Time marching using the time levels given
by the vector TMVEC(j) (j=0,1,2,...,NTS) to
be provided in Record No. 4 below.
=3: Adaptive time-stepping given initial time-
step size, maximum time-step size, and
a time-step multiplier (Code
Default).
TO 2 NTS Integer (Mandatory if KTMSCH=1 or 2, optional
otherwise) Number of time-steps to be taken.
TO 3 ITMXMUL Integer (Mandatory if KTMSCH=3 and KINFIL=1)
Trigger for increasing time-step size DELT in the
adaptive time marching scheme. If the number of
nonlinear Newton-Raphson iterations performed
when solving Green-Ampt infiltration equation is less
than ITMXMUL, the previous time-step size will be
multiplied by a factor of DTMUL (see record 10 of
this group).
ITMXMUL should be between 5 and 10.
(Default=5)
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 13
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
TO 4 TMVEC(j)
(j=1, 1, 2,
...,NTS+1)
Real (Mandatory if KTMSCH=2, optional otherwise)
NTS+1 time levels (sec) correspond to the second
time marching scheme. TMVEC(1) has to be the
simulation starting time, and TMVEC(NTS+1) is the
end time. A total of MOD(NTS, 10) continuation
records are needed if the user wishes to provide all
the NTS+1 terms. Automatic generation of
remaining values is made if the number of values
entered is less than NTS+1. Automatic generation of
remaining values is based on two scenarios: (a) If
only two values, TMVEC(1) and TMVEC(2) are
entered, automatic generation is based on a constant
increment of DT = TMVEC(2)-TMVEC(1), i.e.,
TMVEC(j)= TMVEC(j-1)+DT for j=3,4,...,
NTS+1. (b) If at least three values of TMVEC are
entered, automatic generation is based on the most
recent three values entered, assuming a constant
growth rate in the increment (a growth rate of 1 for
the increment is equivalent to a constant increment).
That is, TMVEC(j) = TMVEC(j-1)+DT(j), where
DT(j)= r DT(j-1), and r is the constant rate of
growth in the increments given by r=DT(j-1)/DT(j-
2).
TO 5 TIME0 Real (Optional) Starting time of simulation (seconds)
(Default =0.0) TIME0 will be overwritten by
TMVEC(1) if KTMSCH=2).
TO 6 TMAX Real (Optional) Maximum simulation time (seconds)
(Default=1.0e30). TMAX will be overwritten by
TMVEC(NTS+1) if KTMSCH=2.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 14
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
TO 7 DTMIN Real (Optional) Minimum time-step size (seconds), to be
used in ODE solver. DTMIN should be less than
(TMAX-TIME0)/NTS if KTMSCH=1, or
TMVEC(j+1)-TMVEC(j) if KTMSCH=2, or
DELT (Record 8 below) if KTMSCH=3
(Default=0.01).
TO 8 DELT Real (Optional) Initial time-step size to be taken by the
adaptive time-stepping scheme (seconds) (Default=1
seconds).
TO 9 DTMAX Real (Optional) Maximum time-step size (seconds)
(Default=60 seconds).
TO 10 DTMUL Real (Optional=3) Time-step size multiplier, see record 3,
ITMXMUL, of this group for meaning
Recommended value is 1.0 to 5.0 (Default = 1.05).
TO 11 KRTMSG1 Integer (Optional) Integer control switch for run-time
information output onto screen (each higher value
indicates more screen output)
=1: Fatal error messages only.
=2: 1 + warning error messages.
=3: 2 + time-stepping information such as time
step no., time-step size, time level (Default)
TO 12 KRTMSG2 Integer (Optional) Integer control switch for run-time
information output onto log file (each higher value
indicates more information output)
=1: Fatal error messages only.
=2: 1 + warning error messages.
=3: 2 + values of key variables when fatal/warning
errors occur (Default)
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 15
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
TO 13 KOUT1 Integer (Optional) Integer control switch for output
frequency of the extent of spreading, average
thickness, volumetric infiltration and volatilization
fluxes to result summary file
=0: Only at end of simulation
=n
1
: (n
1
>0) Every n time steps + final time
(Default = 10).
=-n
1
: (n
1
>0) Every n
1
time levels in the vector
TOUT(j) (provided in Record 17)+final
time.
TO 14 KOUT2 Integer (Optional) Integer control switch for output of mass
balance, mass fractions and mass losses to infiltration
and volatilization
=0: Only at end of simulation
=n
2
: (n
2
>0) Every n time steps + final time
(Default =10).
=-n
2
: (n
2
>0) Every n
2
time levels in the vector
TOUT(j) (provided in Record 17)+final
time.
TO 15 KOUT3 Integer (Optional) Integer control switch for output
frequency of infiltration rate and infiltration front at
selected observation locations (the recorded values
will be directed to the result summary file)
=0: Only at end of simulation.
=n
3
: (n
3
>0) Every n time steps + final time
(Default, n=10).
=-n
3
: (n
3
>0) Every n
3
time levels in the vector
TOUT(j) (provided below in Record 17) +
final time.
TO 16 OUTLEV Integer (Mandatory if KOUT1 or KOUT2 or if both are
negative) Number of time levels in the vector
TOUT(j) to be provided on the next record.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 16
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
TO 17 TOUT(j)
(j=1, 2, ...,
NOUTLEV)
Real (Mandatory if KOUT1 and/or KOUT2 and/or
KOUT3 are negative) Potential time levels (in
seconds) at which output of either general simulation
results and/or observations of infiltration rate and
front and/or mass balance/mass fractions/mass losses
to infiltration and volatilization is desired. A total of
MOD(NOUTLEV-1, 10) continuation records are
needed if the user wishes to provide all the
NOUTLEV terms. Automatic generation of
remaining values is made if the number of values
entered is less than NOUTLEV. Automatic
generation of remaining values is based on two
scenarios: (a) If only two values, TOUT(1) and
TOUT(2) are entered, automatic generation is based
on a constant increment of DT = TOUT(2)-
TOUT(1), i.e., TOUT(j)= TOUT(j-1)+DT for
j=3,4,..., NOUTLEV. (b) If at least three values of
TOUT are entered, automatic generation is based on
the most recent three values entered, assuming a
constant growth rate in the increment (a growth rate
of 1 for the increment is equivalent to a constant
increment). That is, TOUT(j) = TOUT(j-1)+DT(j),
where DT(j)= r DT(j-1), and r is the constant rate
of growth in the increments given by r=DT(j-
1)/DT(j-2).
TO 18 NOBS Integer (Mandatory if KOUT2 0 and KINFIL>0)
Number of observation locations at which time
history of infiltration rate and infiltration front will be
recorded. If NOBS>1, the observation coordinates
{XOBS(j), YOBS(j)} (j=1,2,..., NOBS) must be
provided in records 19 and 20, respectively.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 17
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
TO 19 XOBS(j)
(j=1, 2, ...,
NOBS)
Real (Mandatory if KOUT2 0 and KINFIL>0, and
nobs >0) NOBS values of the X- coordinates of
observation location pairs {XOBS(j), YOBS(j)}
(cm). A total of MOD(NOBS-1, 10) continuation
records are needed if the user wishes to provide all
the NOBS terms. Automatic generation of remaining
values is made if the number of values entered is less
than NOBS. Automatic generation of remaining
values is based on two scenarios: (a) If only two
values, XOBS(1) and XOBS(2) are entered,
automatic generation is based on a constant
increment of DX = XOBS(2)-XOBS(1), i.e.,
XOBS(j)= XOBS(j-1)+DX for j=3,4,..., NOBS.
(b) If at least three values of XOBS are entered,
automatic generation is based on the most recent
three values entered, assuming a constant growth
rate in the increment (a growth rate of 1 for the
increment is equivalent to a constant increment). That
is, XOBS(j) = XOBS(j-1)+DX(j), where DX(j)= r
DX(j-1), and r is the constant rate of growth in the
increments given by r=DX(j-1)/DX(j-2).
TO 20 YOBS(j)
(j=1, 2, ...,
NOBS)
Real (Mandatory if KOUT2 0 and KINFIL>0, and
nobs >0) NOBS values of the Y- coordinates of
observation location pairs {XOBS(j), YOBS(j)
(cm). A total of MOD(NOBS-1, 10) continuation
records are needed if the user wishes to provide all
the NOBS terms. Automatic generation of remaining
values is made if the number of values entered is less
than NOBS. The procedure is the same as that for
XOBS.
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 18
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
TO 21 KSRULE Char*5 (Optional) A character string of 5 characters long
specifying the simulation stopping rule. The basic
rules are:
=00001: If spill stops.
=00010: If frontal spreading stops.
=00100: If berm is reached.
=01000: If thickness of oil lens reaches zero; or
02000: If thickness drops below hmin, oil
thickness due to surface tension.
=10000: If maximum time TMAX is reached.
One may use any combination of the five basic rules.
For example, if KSRULE=00110', simulation
would stop if either frontal spreading stops or the
berm is reached, whichever comes first.
Default: KSRULE =1100 for flow over a flat
surface, and KSRULE=11100 for flow over an
inclined surface.
**Solver Control Group: All records optional. Default values used unless otherwise specified
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
SV 1 KSOLVER Integer (Optional, currently inactive ) Analytical/semi-
analytical solution solver switch
SV 2 ITERMX Integer (Optional) Maximum number of Newton-Raphson
iterations to be taken when solving the Green-Ampt
infiltration model (Default=100)
SV 3 EPSRK Real (Optional) Error tolerance to be used in Runge-
Kutta integration (dimension-less). Should be a small
number (Default = 1.0e-4)
SV 4 EPSNR Real (Optional, required only if KINFIL=1) Error
tolerance to be used in solving the Green-Ampt
equation using Newton-Raphson method (cm).
Should be a small number (Default = 1.0e-4 cm)
Tabl e 5.2 (conti nued)
Compl ete Input Instr ucti ons for Mai n Input Data Fi l e
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 19
** Constant Parameters Group: All records optional. Default values used unless otherwise
specified.
Group
Code
Rec.
No.
Variable
Name(s)
Data
Type(s) Descriptions
CP 1 cnst(1) Real (Optional) Water viscosity (g/cm/sec) (Default
=0.01 at T=293K)
CP 2 cnst(2) Real (Optional) Water density (g/cm
3
) (Default =1.0)
CP 3 cnst(3) Real (Optional) Water vapor diffusion coefficient in air
(cm
2
/sec) (Default =0.239)
CP 4 cnst(4) Real (Optional) Kinematic viscosity of air (cm
2
/sec)
(Default=0.15)
CP 5 cnst(5) Real (Optional) Water surface tension (dyne/cm) (Default
=65)
5.7 EXAMPLE MAIN INPUT FILE
The following is an example main input data file that was created for reproducing a portion of the results
in Weaver et. al. (1994), simulation D, which was a pure infiltration problem. The section on code
verifications explains the problem in more detail.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 20
Table 5.3
An Example Main Input Data File
**----------------------------------------------------------------------
** Title group: Title group may consist of up to 20 records.
**----------------------------------------------------------------------
TI This is an example input data file for OILSFSM
TI It was used to replicate the results in Weaver et al., WRR 30(1),
TI pp.93-105,1994,Table 1, Simulation D.
TI Set up for Weaver et al. problem assumes non-viscous flow
TI on a horizontal surface subject to pulse spill and a berm. The ponding
TI depth of 6.5cm was obtained by setting a berm area of 1 sq meters
TI and a spilled volume of 0.065 cubic meters.
**-----------------------------
** Soil parameters group
**-----------------------------
SP 1 1 ! infil model index-- 0: no infil;1:Green Ampt;3: Philip
SP 2 0.0588 ! Water saturation
SP 3 0.0 ! Initial oil saturation
SP 4 0 ! kdbase,{soil name}
SP 5 0.37 ! porosity
SP 6 9.202628e-7 ! intrinsic permeability (cm^2)
SP 7 0.05 ! residual saturation of oil
SP 8 0.0588 ! residual saturation of water
SP 9 1 24.8 4.84 ! Brooks-Corey model, Air entry pres and pore
**-----------------------------
** Oil parameters group
**-----------------------------
OP 1 0 ! kvolat-- 0: no volat; 1: mixture;=2: pure liquid
OP 2 293. ! Oil temperature in K
OP 3 .00 ! Contact angle in radians
OP 4 0 ! flat: Read from DBASE4OIL.DAT ? 0: NO; =1: YES
OP 5 25. ! surface tension (dyne/cm)
OP 7 1 0 ! Nvoc, Nonvoc: number of volatile & nonvolatile
OP 8 0.0476 ! component's viscosity (dynes-sec/cm^2, or poise)
OP 9 0.79 ! component's density (g/cm^3)
OP 10 12. ! molar weights (g/mole)
OP 11 1.0 ! initial mole fraction
**---------------------------------------------------------------------
** Source and surface parameters: All records Mandatory
**---------------------------------------------------------------------
SS 1 0 !ktsrstyp(0,-1,1,-2,2),{turnoff(sec) if ktsrstyp>=1}
SS 2 0 10. !kxsrstyp(0: point;1:line;2: area),{dimsrs if kxsrstyp>=1}
SS 3 0.065 ! v0 (pulse volume) (m^3)
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 21
SS 4 0 ! kflotyp(0: non-viscous;1: viscous;=-1: uncertain)
SS 5 0 !kslope(0,1),{angle if kslope=1}
Table 5.3 (continued)
An Example Main Input Data File
SS 6 0.00 !surface roughness (cm)
SS 7 1. ! berm area (m^2)
**---------------------------------------------------------------------------------
** Time marchinig and output control parameters group: all optional
**---------------------------------------------------------------------------------
TO 1 3 ! TMSCH -- 1: DELT=const ; =2: DELT=tmvec(j)-tmvec(j-1); =3: adaptive
TO 3 6 ! itmxmul, used for ktmsch=3
TO 5 0.0 ! initial time (sec)
TO 6 6000 ! max time (sec)
TO 7 0.0001 ! smallest time-step size dtmiin (sec)
TO 8 .1 ! initial time-step size (sec); not needed if TMSCH=1 or 2
TO 9 10. ! max time-step size (sec)
TO 10 1.05 ! time-step multiplier; not needed if TMSCH=1 or 2
TO 11 1 ! krtmsg(1) for run-time message output to screen
TO 12 1 ! krtmsg(2) for run-time message output to log file
TO 13 -1 ! koutput(1):to print summary results every 1 intervals in TOUTLEV
TO 14 -1 ! koutput(2): koutput(1):to print observation results every 1 intervals in TOUTLEV
TO 16 30 ! noutlev : number of potential output time levels
TO 17 1 10 20 30 50 60 70 80 90 100 ! output time levels (sec)
TO 17 110 120 180 240 300 360 420 480 540 600
TO 17 800 1200 1800 2400 3000 3600 4200 4800 5400 6000
TO 18 1 ! nobs = no. of obs locations at which infiltration rate and fronts are to be recorded
TO 19 0. ! {xobs(i) ,i=1,...,nobs )
TO 20 0. ! {yobs(i), i=1,...,nobs)
TO 21 10000 ! stopping rule string (stop if TMAX is reached)
**----------------------------------------------------------------------------
** Solver control parameters group. Omitted, default values used.
**----------------------------------------------------------------------------
**-----------------------------------------------------------------------
** Constant parameters group. Omitted, default values used
**-----------------------------------------------------------------------
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 22
5.8 SOIL DATABASE
OILSFSM provides the user with the option of accessing a soil database for typical soil parameter values.
The database file created for this purpose is called DBASE4SOILS.DAT. To make use of this database,
the user needs to specify 1" for the integer control variable KsoilDBase (SP Group, record no. 4). The
user must also specify the name of the soil, which has to be in upper case and has to match one of the soil
names available in the soil database file. At present, parameter values available in the soil database are:
Porosity
Intrinsic permeability (cm
2
)
Residual saturation of oil
Residual saturation of water
Brooks-Corey or van Genuchten (but not both) model parameters, used only if KINFIL =1
(group SP, record 1)
Soil names for which the above mentioned soil parameters are available are provided in Table 5.4.
Table 5.4
Soil Names Available in the Soil Database file DBASE4SOILS.DAT
SAND LOAMY_
SAND
SAND_L
OAM
LOAM SILT_L
OAM
SAND_
CLAY_
LOAM
CLAY_
LOAM
SILTY_
CLAY_
LOAM
SILTY_
CLAY
CLAY
The user may include additional soil names and their associated parameters in the database file
DBASE4SOILS.DAT. There is no restriction on the number of soil names. The input formats for the soil
database are as follows:
All records should be at most 132 characters long
Any blank record or record beginning with the two asterisks ('**') are considered as a comment
record, and will be ignored by the code
Comment records can be interspersed anywhere inside the file
1st non-commented record: soil name, which can be up to 15 characters long
2nd record: porosity
3rd record: intrinsic permeability (cm
2
)
4th record: residual oil saturation
5th record: residual water saturation
6th record: kbcvg, bcvg1, bcvg2:
where
kbcvg=1: implies Brooks-Corey model is used
kbcvg=2: implies van Genuchten model is used
bcvg1=air entry pressure (cm) for Brooks-Corey relation or " (1/cm) for van Genuchten function;

bcvg2=pore distribution parameter 8 for Brooks-Corey relation or n for van Genuchten function
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 23
The above format can be repeated for each one of the soil names and the associated soil parameters.
5.9 OIL DATABASE
OILSFSM also provides the user with the option of accessing an oil database for typical oil parameter
values. The database file is called DBASE4OILS.DAT. To make use of this database, the user needs to
specify 1" for the integer control variable KOilDBase (OP Group, record no. 4). The user must also
specify the name of the oil, which has to be in upper case and has to match one of the oil names available
in the oil database file. At present, the parameter values available in the oil database are:
Number of volatile and non-volatile compounds that constitute the oil
Surface tension of the oil (dynes/cm)
Effective molar volume of the oil (cm
3
/mol) (used only if KVOLAT=2 on Group OP, record 1)
Component viscosities (dynes-sec/cm
2
)
Component densities (g/cm
3
)
Component molecular weights (g/mol)
Component mole fractions in the oil
Vapor pressure of volatile compounds (kPa)
Activity coefficients of volatile compounds (used only if KVOLAT=3 on Group OP, record 1)
Brooks-Corey or van Genuchten (but not both) model parameters
Oil names for which the above mentioned oil parameters are available are provided in Table 5.5.
Table 5.5
Oil Names Available in the Oil Database file DBASE4OILS.DAT
ARABIAN LIGHT
CRUDE
PRUDHOE BAY ARABIAN MEDIUM
CRUDE
IRANIAN HEAVY
The user may include additional oil names and their associated parameters in the database file
DBASE4OILS.DAT. There is no restriction on the number of oil names. The input formats for the oil
database are as follows:
All records should be at most 132 characters long.
Any blank record or record beginning with the two asterisks ('**') are considered as a comment
record, and will be ignored by the code
Comment records can be interspersed anywhere inside the file
1st record: oil name, which can be up to 15 characters long
2nd record: NVOC, and NONVOC, the number of volatile and non-volatile compounds that
make up the oil
3rd record: surface tension (dynes/cm)
4th record: effective molar volume (cm
3
/mol)
5th record: component dynamic viscosities ( NVOC+NONVOC values) (dynes-sec/cm
2
)
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 5- 24
6th record: component densities (NVOC+NONVOC values) (g/cm
3
)
7th record: component molecular weights (NVOC+NONVOC values) (g/mol)
8th record: component mole fractions (NVOC+NONVOC values) (g/mol)
9th record: component vapor pressures (NVOC values) (kPa)
10th record: component activity coefficients (NVOC values)
Records 1 through 10 can be repeated for each one of the oil names and the associated
parameters.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 6- 1
6.0 OUTPUT INTERPRETATION
Three types of files are generated during each simulation session. They are:
A run-time log file
One or more listing files
One or more result summary files
A detailed description of the output files along with their interpretation is provided below.
6.1 RUN-TIME LOG FILE
The run-time log file records all important information concerning the execution of the code during a
session. The information includes starting and ending date and time of the session, important I/O processes,
run-time fatal and warning error messages and execution status. The amount of run-time message output
on the run-time log file is controlled by the integer control switch KRTMSG2 (see Section 5.0, Input
Guide, Record 12 of Group OT). The input range for KRTMSG2 is 1 through 3. Each higher value renders
more information to be directed to the run-time log file when errors occur.
Error messages are classified into two types by the OILSFSM code warning errors and fatal errors.
When a warning error occurs, the execution continues. When a fatal error occurs, however, the execution
is aborted. For example, a warning error message is generated when the maximum number of allowed
Newton-Raphson iterations is exceeded in SUBROUTINE GREEN_AMPT when solving the Green-
Ampt equation for the infiltration front. When this occurs, the maximum absolute error in the iterative
solution of the infiltration front can be made available by specifying KRTMSG2>1. If the error is not too
large, it is not a concern. Specifying a larger maximum allowable number of iterations (see Group SV,
Record 2 of Input Guide) or a larger error tolerance level (see Group SV, Record 4 of Input Guide) may
solve the problem. A fatal error is reported if all correction measures fail. When a fatal error occurs, there
are usually serious errors in one or more of the input parameters.
6.2 LISTING FILE

A listing file provides a descriptive summary of all model input parameters associated with an input data
file. Note that a listing file may exclude some of the redundant input information provided in the input data
file. The listing file may be used to check the input parameters.
6.3 RESULT SUMMARY FILE
A result summary or output file consists of three portions. The main portion contains the extent of surface
spreading (radius or frontal spreading distances from spill location), average lens thickness, area, and
volumetric infiltration and volatilization fluxes at user desired time levels. The second portion provides mass
balance information, i.e., cumulative oil mass spilled, mass remaining on the ground surface, mass fractions
for oil on the ground, and cumulative mass losses to infiltration and volatilization at user desired time levels.
The final portion (if requested), presents a time history of infiltration rates and fronts at user specified spatial
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 6- 2
locations at user specified time observation levels.
6.3.1 Characteristics of Surface Spreading of Oil
The main portion of the result summary file contains values of key variables characterizing surface flow of
oil at user desired time levels. The time levels at which these key variables are printed are controlled by the
integer control switch KOUT1 (see Group OT, Record 13 of Input Guide). These key variables depend
on the following factors
Type of flow whether viscous or non-viscous flow
Type of surface whether horizontal or inclined
Type of source spill whether a point source or a line source
6.3.1.1 Non-viscous Flow From A Line Source Spill
Non-viscous flow resulting from a line source spill applies to horizontal and inclined surfaces. The shape
of surface spreading is a rectangle of width equal to the source width. The spreading occurs in two opposite
directions perpendicular to the line source for the case of horizontal surfaces (cf. Figure 3.2a), and in the
down-slope direction only for the case of a inclined surface (cf. Figure 3.2c). The result summary file
contains the following key variables:
Time levels (1
st
column)
Extent of spreading or distance of moving front from line source (2
nd
column)
Spatially averaged oil thickness (3
rd
column)
Area of spreading (4
th
column)
Volumetric infiltration flux over whole spreading area (5
th
column)
Volumetric volatilization flux over whole spreading area (6
th
column)

6.3.1.2 Non-viscous Flow From A Point Source Spill
Non-viscous flow resulting from a point source spill applies only to the horizontal surface case. The shape
of surface spreading is a circle centered around the source location (cf. Figure 3.2b). The result summary
file contains the following key solution variables:
Time levels (1
st
column)
Radius of oil lens (2
nd
column)
Spatially averaged oil thickness (3
rd
column)
Area of spreading (4
th
column)
Volumetric infiltration flux over whole spreading area (5
th
column)
Volumetric volatilization flux over whole spreading area (6
th
column)

6.3.1.3 Viscous Flow From A Line Source Spill
Viscous flow over a horizontal surface resulting from a line source spill is a rectangle of width equal to the
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 6- 3
source width. Spreading occurs in two opposite directions perpendicular to the line source (cf. Figure
3.3a). The result summary file contains the same key variables as those for non-viscous flow over a
horizontal surface resulting from a line source spill (see Section 6.2.1.2).
Viscous flow over an inclined surface resulting from a line source spill propagates equally in the up-slope
and down-slope directions at the initial stage when t < TSTAR, where TSTAR is the critical time after
which the oil stops propagating in the up-slope direction. The shape of spreading is a rectangle of width
equal to the source width, and length equal to twice XN for t < TSTAR, and XNSTAR+XN for t >
TSTAR, where XN is the maximum extent of spreading in the down-slope direction and XNSTAR is the
maximum extent of spreading in the up-slope direction, as defined in Section 3.3.4 (cf. Figure 3.3c). The
result summary file contains the following key variables:
Time levels (1
st
column)
The extent of down-slope spreading (2
nd
column). Note this value can also be interpreted as the
up-slope extent for time < TSTAR.
Spatially averaged oil thickness (3
rd
column)
Area of spreading (4
th
column)
Volumetric infiltration flux over whole spreading area (5
th
column)
Volumetric volatilization flux over whole spreading area (6
th
column)
In addition, the values of TSTAR and the maximum up-slope extent, XNSTAR (both time independent),
are also provided.
6.3.1.4 Viscous Flow From A Point Source Spill
Viscous flow over a horizontal surface resulting from a point source spill forms a circle centered around the
source location (cf. Figure 3.3b). The result summary file is the same as that for non-viscous flow over a
horizontal surface induced by a point source (see Section 6.2.1.1).
Viscous flow over an inclined surface resulting from a point source spill is a circle centered around the
source location at the initial stage (or when time < TSTAR), and an ellipse (cf. Figure 3.3d) with major axis
of (XN+XNSTAR) and minor axis of twice YN around a moving centroid at ((XN-XNSTAR)/2, 0),
where TSTAR = t* the critical time as defined in Section 3.3.4, XN is the distance of the down-slope oil
front from the source, XNSTAR is the maximum extent of propagation in the up-slope direction from the
source location, and YN is the maximum cross-slope extent or half the minor axis. The result summary
contains the following key variables
Time levels (1
st
column)
XN, the maximum down-slope extent with respect to source (2
nd
column)
YN, the maximum cross-slope half-extent with respect to source (3
rd
column)
Spatially averaged oil thickness (4
th
column)
Area of spreading (6
th
column)
Volumetric infiltration flux over whole spreading area (6
th
column)
Volumetric volatilization flux over whole spreading area (7
th
column)
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 6- 4
In addition, the values of TSTAR and XNSTAR (both time independent) are also provided.
6.3.2 Mass Balance Results
The second portion in the output file contains mass balance information at user desired time levels. The time
levels are controlled by the mass balance output control switch KOUT2 (Group OT, Record 14 of Input
Guide). The mass balance output is divided into two parts. The first summarizes overall mass balance. The
second part provides detailed descriptions of component mass balance. Contents in the overall mass
balance table are:
Time levels (1
st
column)
Cumulative oil mass spilled up to the time level in column 1 (2
nd
column)
Total mass remaining on the ground surface, including that in the depression storage (3
rd
column)
Cumulative oil mass loss to infiltration (4
th
column)
Cumulative oil mass loss to volatilization (5
th
column)
Overall density of oil on the ground surface (6
th
column)

The component mass balance results may consist of several tables, the number of which depends on the
total number of components. Each table lists the time level, mass fraction, cumulative mass loss to
infiltration, and cumulative mass loss to volatilization for at most two components.
6.3.3 Observations of Infiltration Rates and Fronts
The final portion of the result summary file consists of the observations of infiltration rates and infiltration
fronts (or depths) at user selected locations and desired time levels. The time levels are controlled by the
observation output control switch KOUT3 (Group OT, Record 15 of Input Guide). The locations are
specified through the input variables NOBS and {XOBS, YOBS} (see Group OT, Records 18 and 19
of Input Guide). On an inclined surface, the infiltration front is measured from the surface, not from any
reference datum. Depending on the number of observation points, the observation results may be broken
into multiple tables. Each table contains observed infiltration front (depth) and rate associated with one
location.
PART 4:
CODE VERIFICATION
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 1
7.0 CODE VERIFICATION AND EXAMPLE PROBLEMS
A number of benchmark problems were analyzed to test and verify the three main processes modeled by
OILSFSM, namely:

Surface flow
Infiltration
Volatilization

This section presents results from verification and example runs. Sections 7.1 to 7.3 are each devoted to
the verification of one of the main processes. Section 7.1 focuses on pure surface flow of non-viscous and
viscous oil over horizontal and inclined surfaces. Section 7.2 is devoted to pure infiltration problems and
their solutions. Section 7.3 presents solutions to pure volatilization problems. Whenever possible, solutions
obtained with the OILSFSM model were compared with existing analytical solutions and/or peer reviewed
publications. Finally, Section 7.4 illustrates the use of OILSFSM for simulating coupled oil spills, surface
flow, infiltration and volatilization processes through example problems.
7.1 PURE SURFACE FLOW OF OIL
The test problems presented in this subsection illustrate the capabilities of OILSFSM to correctly simulate
the surface flow of non-viscous and viscous oils over horizontal and inclined surfaces. Table 7.1 provides
a description of the test problems, and the oil and source parameters used. For each problem, four source
scenarios were considered (1) pulse spill from a line source; (2) finite duration spill from a line source;
(3) pulse spill from a point source; and (4) finite duration spill from a point source. The finite duration spill
was designed so that the cumulative oil volume spilled from beginning T=0 to T = T
turnoff
was the same as
that of a pulse spill. The total volume spilled in each case was 0.065 m
3
. For all cases, a gravitational
acceleration of g = 981 cm/sec
2
was used, no berms were included, and the surface roughness was set to
zero.
7.1.1 Non-viscous Oil Flow over Horizontal/inclined Surfaces
For non-viscous flow, the minimum oil lens thickness due to surface tension and roughness was calculated
to be h
min
= 7.14 10
- 2
cm according to Eqs. (3.1)-(3.2). It then follows from (3.4a) that the area of
spreading is:
A(t) = 91.08 m
2
for a pulse spill for all time
A(t) = 91.08 m
2
x (elapsed time t)/T
turnoff
for t <T
turnoff
and A(t)=91.08 m
2
for t >T
turnoff
for a
finite duration spill
The analytical solution to the extent of surface spreading, x
N
(t), is inferred from the shape of spreading to
be:
For flow from a line source with width w
d
over a horizontal surface, x
N
(t) =A(t)/2w
d

For flow from a line source with width w
d
over an inclined surface, x
N
(t) = A(t)/w
d
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 2
For flow from a point source over a horizontal surface, x
N
(t) =oA(t)/B.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 3
Figure 7.1 shows the simulation results obtained with OILSFSM and the corresponding analytical solutions.
Excellent agreement between OILSFSM and analytical solutions is observed. The oil lens thickness on the
ground surface obtained using OILSFSM is identical to the hand calculated value of h
min
= 7.14 10
- 2
cm
for all cases.
Table 7.1
Description of Test Problems Associated with Pure Surface Flow of Oil
Problem
ID
Oil
Type
1
Surface Type
2
(h
d
=0) Oil Parameters Source Scenarios
3
P711 NV H
D= 0.79 g/cm3
F = 25 dynes/cm
n = 0.4 (contact angle)
(a) Pulse in time, line in space
(b) Finite in time, line in space
(c) Pulse in time, point in space
(d) Finite in time, point in space
For pulse spill, V=0.065m
3
For finite duration spill, T
turnoff
=
600 s with V=0.065(t/T
turnoff
) for t
< T
turnoff
, and V=0 for t >T
turnoff
For line source, w
d
= 1.0m
P712 NV I (2=0.1745 or 10
o
)
P713 V H D= 0.79 g/cm3
F = 25 dynes/cm
n = 0.4 (contact angle)
P714 V I ((2=0.05 or 2.826
o
)
1. NV = non-viscous; V= viscous
2. H =horizontal; I = inclined
3. Scenarios (c) and (d) (point source) do not apply to P712--non-viscous flow over an inclined surface
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 4
(a) Horizontal Surface, Line Source
0
1000
2000
3000
4000
5000
6000
0 120 240 360 480 600 720 840 960 1080 1200
Time (sec)
x
N

(
c
m
)
pulse spill
finite duration spill
(b) Horizontal Surface, Point Source
0
200
400
600
800
1000
0 120 240 360 480 600 720 840 960 1080 1200
Time (sec)
x
N

(
c
m
)
pulse spill
finite duration spill
(c) Inclined Surface, Line Source
0
2000
4000
6000
8000
10000
12000
0 120 240 360 480 600 720 840 960 1080 1200
Time (sec)
x
N

(
c
m
)
pulse spill
finite duration spill
Figure 7.1 Extent of Non-Viscous Oil Spill on Flat/Inclined Surfaces Subject to Pulse/Finite
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 5
Duration Spills at a Point or Along a Line. Solid line: Solutions of OILSFSM Model;
Dotted line: Exact Solutions.
7.1.2 Viscous Oil Flow over Horizontal/inclined Surfaces
For viscous flow over a horizontal surface, the analytical solutions for the extent of oil spreading from point
and line sources are obtained using Eq.(2.6). Strictly speaking, Eq.(2.6) cannot be used to find the extent
of the spill resulting from a finite duration source when the time is greater than the source turnoff time T
turnoff
.
As an approximation, however, it is assumed that Eq.(2.6) holds for t > T
turnoff
; the volume V(t) in (2.6)
being a constant for t > T
turnoff
. This approximate analytical solution was used to verify the solution from
OILSFSM. The results are presented in Figures 7.2(a)-(b). Figure 7.2(a) compares the solutions for the
case of a pulse and a finite duration with line source, while Figure 7.2(b) compares the solutions for the
case of a pulse and a finite duration with point source. Both figures show excellent agreement between
OILSFSM simulation results and the exact solution.
For viscous flow over an inclined surface, the analytical solutions to the extent of spreading resulting from
a line source or a point source are calculated using Eqs.(2.13) and Eq.(2.18), respectively. The transition
time t
*
defined in these equations and the maximum up-slope extent x
T
for all cases considered are
presented in Table 7.2. The transition time given by the code agrees with that of the exact solution since
both were calculated using the same formulas, Eqs.(2.12) and (2.17). The difference in the maximum up-
slope extent x
T
between the two solutions is due to the fact that in OILSFSM, x
T
is calculated based on
Runge-Kutta integration of the corresponding rate of change in the up-slope spill extent from time 0 to time
t
*
. The difference is particularly large for the instantaneous line and point source cases where t
*
is smaller
than 0.1 second. This is in consequence of the fact that some effective rate of change in the extent of
spreading has been used to overcome singularity problems when t < 0.1 second (see Section 3.4.3.3).
Similar to the case of a horizontal surface, Eqs.(2.13) and (2.18) cannot be used to find the exact analytical
solution to the extent of spreading resulting from a finite duration spill when the time is greater than the
source turnoff time. A similar approximation was adopted, where the volume of the spill is assumed to be
a constant after T
turnoff
, and the equations are still assumed to be valid. The simulated spill extent and the
corresponding exact solutions are presented in Figures 7.3(a)-(c). Figure 7.3(a) compares the down-slope
extent obtained using OILSFSM and the corresponding exact solution for the case of a pulse and a finite
duration spill with line source. Figure 7.3(b) compares the solutions for a point source. Figure 7.3(c)
compares the cross-slope extent solutions for the case of a pulse and a finite duration spill with line source.
All figures demonstrate good agreement between the OILSFSM results and the exact solution.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 6
Table 7.2
Transition Time t
*
and the Maximum Up-slope Spill Extent x
T
for Viscous Flow of Oil Over an
Inclined Surface (exact solutions are enclosed in parentheses)
Source type
Instantaneous
Line Source
Finite Duration
Line Source
Instantaneous
Point Source
Finite Duration
Point Source
t
*
(seconds)
7.3110
-2
(7.3110
-
2
)
1.47
(1.47)
8.5710
-2
(8.5710
-
2
)
7.8410
-1

(7.8410
-1
)
Up-slope Extent
x
T
(cm)
250.0
(224.0)
8.0
(7.9)
52.0
(123.0)
10.1
(10.5)
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 7
(a) Horizontal Surface, Line Source
0
200
400
600
800
1000
1200
1400
1600
0 120 240 360 480 600 720 840 960 1080 1200
Time (sec)
x
N

(
c
m
)
pulse spill
finite duration spill
(b) Horizontal Surface, Point Source
0
50
100
150
200
250
300
350
400
450
0 120 240 360 480 600 720 840 960 1080 1200
Time (sec)
x
N

(
c
m
)
pulse spill
finite duration spill
Figure 7.2 Extent of Viscous Oil Spill on a Flat Surface Subject to a Line Source (a) and a
Point Source (b). Solid Line: Solutions of OILSFSM Model; Dash Line: Exact Solutions.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 8
(a) Inclined Surface, Line Source
0
2000
4000
6000
8000
10000
0 120 240 360 480 600 720 840 960 1080 1200
Time (sec)
x
N

(
c
m
)
pulse spill
finite duration spill
(b) Inclined Surface, Point Source
0
1000
2000
3000
4000
5000
6000
0 120 240 360 480 600 720 840 960 1080 1200
Time (sec)
x
N

(
c
m
)
pulse spill
finite duration spill
(c) Inclined Surface, Point Source
0
50
100
150
200
250
0 120 240 360 480 600 720 840 960 1080 1200
Time (sec)
y
N

(
c
m
)
pulse spill
finite duration spill
Figure 7.3 Extent of Viscous Oil Spill on an Inclined Surface Subject to a Line Source (a) and
a Point Source (b,c). (a) and (b) Show Down-Slope Extent, (c) Shows Cross-Slope
Extent. Solid Line: Solutions of OILSFSM Model; Dash Line: Exact Solutions.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 9
7.2 PURE INFILTRATION

In OILSFSM, infiltration is assumed to be one-dimensional in the vertical direction and to possess a sharp
infiltration front. Soil properties are considered homogeneous in space. Infiltration at any time and any
location depends only on the present time, the arrival time of propagating front at this location, the oil lens
thickness at this location above the ground surface as it varies with time, and the infiltration model and
model parameters. In what follows, testing of the two built-in infiltration models, the Green-Ampt model
and the Philip model, is presented.
7.2.1 Green-ampt Model
Weaver et al. [1994] developed KOPT (Kinematic Oily Pollutant Transport), a screening model for flow
of a NAPL and associated chemical transport in the vadose zone. During infiltration under ponded
conditions or when the NAPL flux exceeds the maximum effective conductivity of the soil, their model
reduces to the Green-Ampt model. During the redistribution stage, their model is based on a kinematic
wave theory. They conducted a simple laboratory column experiment in order to evaluate the NAPL flow
portion of the KOPT model. The experimental setup consists of a 1m-long, 0.05m-diameter glass column
with a coarse, porous glass frit at the bottom that was modified by adding seven air release vents. A
saturated sand pack was used as a permeameter for determining the hydraulic conductivity under steady-
state flow conditions. Manometers attached to the vents on one side of the column were employed to
determine the drop in head over each 10 cm long section of the glass column. The sand was chosen to be
coarse enough to have a low capillary fringe, and the water saturation above the fringe was assumed to be
near residual after hours of drainage. Infiltration was initiated by releasing 100 g of dyed soltrol 220 at the
surface and allowing it to enter the column. The initial ponding depth was at 6.5cm. The ponding depth and
location of the NAPL infiltration front as they varied with time were recorded. Four simulations were then
conducted using KOPT: Simulations A, B, C, and D. These simulations differed slightly in their choice of
the permeability, porosity, and water saturation. The soil and oil parameters used in their simulations were
determined independently of the transient flow experiment to avoid fitting or adjusting the parameters.
Results from all four simulations compared well with those from the experiment.
To verify the Green-Ampt module in OILSFSM, simulation D in Weaver et al. [1994] was used as a basis
for comparison. Table 7.3 provides the soil and oil parameters employed in simulation D. To simulate the
infiltration problem as described above using OILSFSM, the option of a non-viscous oil flow over a flat
surface bounded by a berm is adopted. An instantaneous or pulse spill of a volume V=0.065 m
3
of oil was
initiated. The berm area was set at A
b
=1 m
2
, giving an initial ponding depth of 6.5 cm. An observation
location at the origin was selected to record the time-variable infiltration front in the soil. The input data file
for OILSFSM is presented in Table 7.4. Results from the OILSFSM model are presented in Figures 7.4
and 7.5, along with the experimental and simulation D results of Weaver et al.[1994]. Figure 7.4a compares
the measured and simulated ponding depths at the surface. Figure 7.4b shows the oil front positions
measured at the left edge, center, and right edge of the column, along with simulated results. Both figures
show that OILSFSM reproduces the experimental and simulation results very well during the infiltration
phase of the oil spill process. The redistribution stage, which starts 6 min after the spill, however, is not
simulated by OILSFSM.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 10
Table 7.3
Parameters Used by OILSFSM to Simulate 1-D Infiltration Using the Green-Ampt Model
(Soil and Oil Parameters Taken from Weaver et al., 1994, Simulation D in Table 1)
Soil Parameters Oil Parameters Source Parameters
Permeability k (cm
2
) =9.202610
-7
Porosity n =0.37
Water Saturation S
w
= 0.0588
Initial Oil Saturation S
oi
=0
Residual Oil Saturation S
or
=0.05
Residual Water Saturation S
wr
=0.0588
Air Entry Head (cm) h
ceaw
=24.8
Pore Size Distribution 8 = 4.84
Oil Density D(g/cm3)= 0.79
Viscosity (dyne-sec/cm
2
) : =0.0476
Surface Tension (dynes/cm ) F= 25
Contact Angle n = 0.4
Pulse Spill of Non-Viscous Oil
with a volume of V=0.065m
3
on a flat surface with a berm
whose area is A
b
=1m
2
. This
would be equivalent to
specifying an initial ponding
depth of 6.5 cm.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 11
Table 7.4
OILSFSM Input Data File for Reproducing Infiltration Portion of
Weaver et al. [1994, Simulation D]
**---------------------------------------------------------------------
** Title group: Title group may consist of up to 20 records.
**---------------------------------------------------------------------
TI Problem 7.2.1: Pure infiltration problem
TI Purpose: To replicate the infiltration part of results in Weaver et al.,
TI WRR 30(1), pp.93-105,1994, Table 1, Simulation D.
TI Set up in Weaver et al. were realized by using the non-viscous flow
TI on a flat surface subject to pulse spill and a berm. The ponding
TI depth of 6.5cm was obtained by setting a berm area of 1 sq meters
TI and a spilled volume of 0.065 cubic meters.
TI Note the saturated conductivity in Table 3 of Weaver et al. is
TI the saturated conductivity for water. It has been converted to
TI the permeability of the soil.
**---------------------------------------------------------------------
** Soil parameters group
**---------------------------------------------------------------------
SP 1 1 ! infil model index--0: no infil;1:Green Ampt;3: Philip
SP 2 0.0588 ! Water saturation
SP 3 0.0 ! Initial oil saturation
SP 4 0 ! kdbase,{soil name}
SP 5 0.37 ! porosity
SP 6 9.202628e-7 ! intrinsic permeability (cm^2)
SP 7 0.05 ! residual saturation of oil
SP 8 0.0588 ! residual saturation of water
SP 9 1 24.8 4.84 ! Brooks-Corey model, Air entry pres and pore
** size dist
**---------------------------------------------------------------------
** Oil parameters group:
**---------------------------------------------------------------------
OP 1 0 ! kvolat-- 0: no volat; 1: mixture;=2: pure liquid
OP 2 293. ! Oil temperature in K
OP 3 .00 ! Contact angle in radians
OP 4 0 ! flat: Read from DBASE4OIL.DAT ? 0: NO; =1: YES
OP 5 25. ! surface tension (dyne/cm)
OP 7 1 0 ! Nvoc, Nonvoc: number of volatile & nonvolatile
OP 8 0.0476 ! component's viscosity (dynes-sec/cm^2, or poise)
OP 9 0.79 ! component's density (g/cm^3)
OP 10 12. ! molar weights (g/mole)
OP 11 1.0 ! initial mole fraction
**---------------------------------------------------------------------
** Source and surface parameters: All records Mandatory
**---------------------------------------------------------------------
SS 1 0 !ktsrstyp(0,-1,1,-2,2),{turnoff(sec) if ktsrstyp>=1}
SS 2 0 !kxsrstyp(0:point;1:line;2:area),{dimsrs if kxsrstyp>=1}
SS 3 0.065 ! pulsed volume released.
SS 4 0 ! kflotyp (0: non-viscous;1: viscous;=-uncertain)
SS 5 0 !kslope(0,1),{angle if kslope=1}
SS 6 0.00 !surface roughness (cm)
SS 7 1. ! area surrounded by berm (m^2)
**---------------------------------------------------------------------
Tabl e 7.4 (conti nued)
OILSFSM Input Data Fi l e for Repr oduci ng Infi l tr ati on
Por tion of
Weaver et al . [1994, Si mul ati on D]
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 12
** Time marchinig and output control parameters group: all optional
**---------------------------------------------------------------------
TO 1 3 ! TMSCH -- 1: DELT=const ; =2: DELT=tmvec(j)-tmvec(j-1); =3: adaptive
TO 3 6 ! itmxmul, used for ktmsch=3
TO 5 0.0 ! initial time (sec)
TO 6 6000 ! max time (sec)
TO 7 0.01 ! smallest time step size dtmiin (sec)
TO 8 .1 ! initial time step size (sec); not needed if TMSCH=1 or 2
TO 9 1. ! max time step size (sec)
TO 10 1.05 ! time step multiplier; not needed if TMSCH=1 or 2
TO 11 1 ! krtmsg(1) for run-time message output to screen
TO 12 1 ! krtmsg(2) for run-time message output to log file
TO 13 -1 ! koutput(1):intermediate results control (0,-n,n)
TO 14 -1 ! koutput(2):observation results control(0,-n,n)
TO 15 -1 ! koutput(3):mass balance results control(0,-n,n)
TO 16 110 ! noutlev : number of time levels at which to print results
TO 17 1 2 5 10 20 30 50 60 120 180 ! 10 values on reach line
TO 18 1 ! nobs = no. of obs locations
TO 19 0. ! xi,i=1,...,nobs )
TO 20 0. ! yi,i=1,...,nobs )
TO 21 10000 ! stopping rule string (stop if TMAX is reached)
**---------------------------------------------------------------------
** Solver control parameters group. Omitted, default values used.
**---------------------------------------------------------------------
**---------------------------------------------------------------------
** Constant parameters group. Omitted, default values used.
**---------------------------------------------------------------------
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 13
Figure 7.4 Green-Ampt Model: Comparison of OILSFSM Simulated Results with Those of
Simulation D of Weaver et al. [1994] for a Pure Ponded Infiltration Problem: (a)
Ponding Depth versus Time, (b) Infiltration Depth versus Time. Both measured and
simulation D results were taken from Weaver et al. [1994]. Results for simulations A, B,
and C were excluded.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 14
7.2.2 Philip Model
An alternative infiltration model available in OILSFSM is the Philip model. To verify the Philip module in
OILSFSM, ponded infiltration along a semi-infinite long soil column was simulated using OILSFSM. An
initial ponding depth of 6.5cm was assumed. The Philip model parameters are: B = 0.004 cm/second,
S=0.1 cm/second
1/2
, n(porosity) =0.37, and S
o
(oil saturation)=0.7682. Again, the infiltration problem was
simulated as non-viscous flow on a flat surface subject to infiltration and bounded by a berm. The initial
ponding depth of 6.5cm was realized by application of a pulse spill of volume V= 0.065 m
3
and a berm of
area 1 m
2
. The simulated ponding depth and infiltration front are presented in Figures 7.5a and Figure 7.5b,
respectively, along with the corresponding exact solutions. Excellent agreement between the OILSFSM
model results and exact solutions are observed. Note that there is no correlation between the choice of the
Philip model parameters and those of the Green-Ampt model.
7.3 PURE VOLATILIZATION
Three alternative approaches are available in OILSFSM for simulating the volatilization process: the pure
liquid model (Section 2.5.1), the mixture model that employs an effective molar volume, and the mixture
model that employs activity coefficients (Section 2.5.2). To verify the three volatilization procedures,
OILSFSM runs were conducted to simulate volatilization from an evaporation pan with area 1 m
2
. The
initial oil depth inside the pan is 0.5m. A wind speed of 1610 m/hr (=10 mi/hr) is assumed. Other
parameters used in the analysis are unchanged from the default ones, i.e., water viscosity = 0.01 g/cm/sec,
diffusion of water vapor in air = 0.239 cm
2
/sec, and kinematic viscosity of air = 0.15 cm
2
/s. To model
volatilization from an evaporation pan using OILSFSM, the non-viscous flow on a flat surface is selected.
Surface roughness is considered zero. The berm area is 1 m
2
, the volume spilled equals 0.5 m
3
, and the
surface tension related parameters were set such that thickness due to surface tension is much less than 0.5
m. This allows the spilled oil to reach the edge of the berm instantaneously, creating an evaporation pan.
Subsequent volatilization was simulated for a period of 24 hours. Comparison between simulated
volatilization flux and cumulative mass loss to volatilization with the exact solutions were made for all cases.
7.3.1 Pure Liquid Model
The oil considered in this case was assumed to consist of a single compound. Oil properties are as follows:
oil density = 0.79 g/cm
3
; dynamic viscosity = 0.0476 dynes-sec/cm
2
; molecular weight = 100 g; vapor
pressure = 2 kPa; and temperature T=20
o
C. The exact solution to volatilization flux was found using
Eq.(2.30). Figure 7.6 shows the simulated and exact volumetric volatilization fluxes per unit area, and
volume of oil lost via volatilization per unit area. As expected, the volatilization flux stays constant over time,
and the volume lost is a linear function of time. After one day, most of the oil has been lost to evaporation.
The simulated results agree with the exact solutions perfectly.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 15
(a) Ponding Depth vs. Time
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
0 2 4 6 8 10 12 14
Time (minutes)
D
e
p
t
h

(
c
m
)
(b) Infiltration Depth vs. Time
0
20
40
60
80
0 5 10 15 20
Time (minutes)
D
e
p
t
h

(
c
m
)
Figure 7.5 Philip Model: Comparison of Simulated and Exact Results for a Pure Ponded
Infiltration Problem: (a) Infiltration Front versus Time; (b) Ponding Depth versus Time.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 16
Solid Line: OILSFSM Solution; Dash Line: Exact Solution.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 17
(a) Volumetric Volatilization Flux
0.00
0.10
0.20
0.30
0.40
0.50
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
Q
v
(
c
m

3
/
s
)

(b) Volume Loss per Unit Area
0
1
2
3
4
5
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
V
o
l
u
m
e

L
o
s
s

(
c
m
)
Figure 7.6 Pure Liquid Volatilization Model: Comparison of Simulated and Exact Results.
(a) Volumetric Volatilization Flux; (b) Volume Lost to Volatilization per Unit Pan
Area. Solid Line: OILSFSM Solution; Dash Line: Exact Solution.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 18
(7.1)
(7.2)
(7.3)
7.3.2 Mixture Model I: the Mean Effective Molar Volume Approach
To simulate volatilization of a hydrocarbon mixture using the effective molar volume approach, a crude oil
called Iranian Heavy Crude Oil was considered. The Iranian Heavy consists of five volatile compounds
(benzene, toluene, ethylbenzene, xylene, and C3-benzenes) plus other less volatile compounds. For
simulation purposes, the less volatile compounds were grouped into a single non-volatile compound. The
overall density and viscosity of Iranian Heavy are 0.8657g/cm
3
and 0.2 at 15
o
C, respectively [Jokuty et
al., 1998]. A mean effective molar volume of v
c
=400 cm
3
/mole was used. Table 7.5 illustrates the various
parameters of the compounds that comprise the oil. A temperature of 15
o
C was used. Other parameters
such as those for the source and surface conditions were the same as in the example problem for the pure
liquid volatilization model. Figure 7.7 shows the simulated and exact total volumetric volatilization fluxes.
Figure 7.8 illustrates the simulated and exact cumulative mass loss for each of the volatile compounds.
Excellent agreement between simulated and exact results are observed. In both Figures 7.7 and 7.8, the
exact solution was obtained as follows [Stiver et al., 1989]. Let j denote one of the volatile compounds.
Since the mass fraction of the volatile compounds is very small, Eq.(2.33) can be rewritten as
where n
j
is number of moles of compound j remaining, and V
o
, the total volume of oil in the evaporation
pan, is considered a constant. Solution to (7.1) is obtained as
where n
j
0
is the initial number of moles of compound j. The cumulative mass loss of compound j is given
by the initial mass of that compound minus the mass remaining for the compound. The mass remaining is
the product of the number of moles remaining and the molecular weight. The evaporation or volatilization
flux q
e j
is calculated as
The total volumetric volatilization flux is the sum of q
e j
for individual volatile compounds. It follows from
(7.2)-(7.3) that the volatilization rate is proportional to the vapor pressure of the compound, and inversely
proportional to its molecular weight (since the mass transfer coefficient k
g
is inversely proportional to the
compounds molecular weight via Eqs.(2.31)-(2.32)). This is confirmed by Figures 7.8a-e, in which
benzene, with the highest vapor pressure and the lowest molecular weight, exhibits the greatest volatilization
rate. On the contrary, C3-benzenes have the smallest vapor pressure and the largest molecular weight
among the VOCs, and their rate of volatilization is the lowest.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 19
Table 7.5
Parameters for Various Compounds in Iranian Heavy Crude Oil
Benzene Toluene Ethylbenzene Xylene C3-Benzenes non-volatile
Density
(g/cm
3
)
0.8765 0.8669 0.876 0.8642 0.8652 0.8657
Viscosity
(dynes.s/cm
2
)
0.604 0.56 0.631 0.603 0.600 0.1934
Molecular
Weight (g)
78.11 92.14 106.17 106.17 120.14 130.0
MassFraction
(ppm)
830 2520 1310 5100 7570 982670
Vapor Pressure
(kPa)
12.7 3.79 1.28 1.13 0.33
Note: (a) Density, viscosity, molecular weight, and vapor pressure for each volatile compound were obtained from CRC
[1995-1996].
(b) Mass fractions were obtained from Jokuty et al. [1998].
(c) Viscosity for C3-Benzene was estimated since no value was available in literature. Values for non-volatile
compounds were back calculated from those for the volatile ones and the oil as a whole. Molecular weight for
the non-volatile compound was estimated.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 20
0.00
0.01
0.02
0.03
0.04
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
Q
v
(
c
m
3
/
s
)

Figure 7.7 Mixture Volatilization Model I for Iranian Heavy Oil: Simulated and Exact Total
Volumetric Volatilization Fluxes. Solid Line: OILSFSM Solution; Dash Line: Exact
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 21
(a) Benzene
0.00
10.00
20.00
30.00
40.00
50.00
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

(b) Toluene
0
50
100
150
200
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

(d) Xylene
0
50
100
150
200
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

(e) C3-Benzenes
0
20
40
60
80
100
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

(c) Ethylbenzene
0
10
20
30
40
50
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

Solution.
Figure 7.8 Mixture Model I for Iranian Heavy Oil: Comparison of Simulated and Exact Mass
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 22
Losses to Volatilization for Each Volatile Compound. Solid Line: OILSFSM
Solution; Dash Line: Exact Solution.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 23
(7.4)
(7.5)
7.3.3 Mixture Model Ii: the Activity Coefficient Approach
To verify the activity coefficient approach to volatilization from a hydrocarbon mixture, the same Iranian
Heavy crude oil is adopted. The oil composition and model parameters are the same as those used in
Section 7.3.2 for testing the mean effective molar volume approach. For simplicity, it was assumed that
activity coefficients are unity for all the volatile compounds. Results from the OILSFSM model runs are
presented in Figures 7.9 and 7.10 along with the corresponding exact solutions. Figure 7.9 shows the
simulated and exact total volumetric volatilization fluxes. Figure 7.10 illustrates the simulated and exact
cumulative mass losses for each of the volatile compounds. Again, excellent agreement between simulated
and exact results are observed.
In both Figures 7.9 and 7.10, the exact solution was obtained using similar procedures as those for the
mean effective molar approach. Let n
T
be the total number of moles in the evaporation pan. Eq.(2.35) can
then be rewritten
Because of the very low mole fractions of the volatile compounds, it is reasonable to assume that n
T
remains a constant throughout the volatilization process. The solution to (7.4) is then given by
from which the volumetric volatilization flux for volatile compound j can be derived according to Eq. (7.3).
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 24
0.00
0.01
0.01
0.02
0.02
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
Q
v
(
c
m

3
/
s
)

Figure 7.9 Mixture Volatilization Model II for Iranian Heavy Oil: Simulated and Exact Total
Volumetric Volatilization Fluxes. Solid Line: OILSFSM Solution; Dash Line: Exact
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 25
(a) Benzene
0.00
10.00
20.00
30.00
40.00
50.00
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

(b) Toluene
0
20
40
60
80
100
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

(e) C3-Benzenes
0
20
40
60
80
100
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

(d) Xylene
0
50
100
150
200
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

(c) Ethylbenzene
0
10
20
30
40
50
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Hours)
M
a
s
s

L
o
s
s

(
g
)

Solution.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 26
Figure 7.10 Mixture Model II for Iranian Heavy Oil: Comparison of Simulated and Exact
Mass Losses to Volatilization for Each Volatile Compound. Solid Line: OILSFSM
Solution; Dash Line: Exact Solution.
7.4 SURFACE FLOW OF OIL COUPLED WITH INFILTRATION AND
VOLATILIZATION
The individual components (the pure surface flow, infiltration, and volatilization) of the OILSFSM model
have been independently tested and verified in Sections 7.1-7.3. The availability of exact solutions and/or
peer reviewed publications for all cases tested make the verifications possible. In what follows, however,
the focus will be on simulations of fully coupled surface flow, infiltration and volatilization resulting from
hypothetical oil spills. The purpose of presenting these example problems is to illustrate how surface
spreading of oils would be affected by infiltration and volatilization processes.
7.4.1 Flow of Viscous Oil over a Horizontal Surface
Consider a hypothetical spill of an Iranian Heavy Crude Oil on a flat soil surface. The oil parameters are
(at 288K): density= 0.8657g/cm
3
; dynamic viscosity= 0.2 dynes-s/cm
2
; and surface tension =27 dynes/cm
[Jokuty et al., 1998]. A contact angle of 0.5 radian was assumed. Compositional properties of the oil are
given in Table 7.5. Volatilization, if simulated, is based on the effective mean molar volume approach. The
mean effective molar volume of the oil was set at 400 cm
3
/mol. The soil was assumed to be a silt with
parameters [EPA, 1995]: soil permeability = 7.078910
-10
cm
2
; porosity=0.46; residual water saturation
= 0.034; residual oil saturation =0.05; and the van Genuchten parameters n=1.37, "=0.016(1/cm). Initial
water saturation was set equal to the residual water saturation. Surface roughness was set at zero. A finite
duration point source with a volumetric flux of 3.33 10
-4
m
3
/s and duration of 5 minutes was assumed
(total volume of oil spilled is 1 m
3
after 5 minutes). Surface spreading was not bounded or restricted by a
berm.
Figure 7.11 illustrates the OILSFSM simulated radius of the spill and oil lens thickness as they vary with
time for three cases: (1) pure surface flow without infiltration or volatilization; (2) coupled surface flow and
volatilization only, and (3) coupled surface flow, volatilization and infiltration. For verification purposes, the
exact solutions to the radius and thickness of the oil lens are also presented for the case of pure surface flow
with no volatilization and infiltration. Figure 7.12 shows the cumulative mass loss to volatilization and
infiltration. A careful examination of Figures 7.11 and 7.12 leads to the following conclusions:
For pure surface flow, both the radius and the thickness of the oil lens simulated using OILSFSM
agree with those of exact solutions.
As the oil lens expands, the spatially averaged thickness decreases.
Volatilization has little impact on the radius and thickness of the spill because of the low mass
fraction of volatile compounds.
For the case of surface flow coupled with volatilization only, cumulative mass loss to volatilization
stabilizes after about 24 hours, indicating that most of the volatile compounds have been lost to
volatilization.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 27
Infiltration has a significant impact on surface spreading of the oil as a whole. After about 5 hours
from the start of spill, much of the oil originally spilled has been lost to the soil via infiltration.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 28
(a) Radius of Spreading versus Time
0
200
400
600
800
1000
1200
1400
0 4 8 12 16 20 24
Time (hours)
R
a
d
i
u
s

(
c
m
)
Surface Flow Only (Exact)
Surface Flow Only (OILSFSM)
Surface Flow + Volatilization (OILSFSM)
Surface Flow+Infiltration+Volatilization (OILSFSM)
(b) Oil Lens Thickness versus Time
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 4 8 12 16 20 24
Time (hours)
h

(
c
m
)
Surface Flow Only (Exact)
Surface Flow Only (OILSFSM)
Surface Flow + Volatilization (OILSFSM)
Surface Flow+Infiltration+Volatilization (OILSFSM)
Figure 7.11 Surface Flow of Viscous Iranian Heavy Oil: Radius of Spreading and Averaged
Oil Lens Thickness and the Effect of Infiltration and Volatilization. (a) Radius of
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 29
(a) Cumulative Mass Loss to Volatilization
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
0 4 8 12 16 20 24
Time (hours)
M
a
s
s

(
g
)
Surface Flow+Volatilization Only
Surface+Volatilization+Infiltration
(b) Cumulative Mass Loss to Infiltration
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
900,000
0 1 2 3 4 5
Time (hours)
M
a
s
s

(
g
)
Surface Flow+Volatilization+Infiltration
spreading versus time; (b) Oil lens thickness versus time.

C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 30
Figure 7.12 Surface Flow of Viscous Iranian Heavy Oil: (a) Cumulative Mass Loss to
Volatilization and (b) Cumulative Mass Loss to Infiltration.
7.4.2 Flow of Non-viscous Oil over a Horizontal Surface
In this case, it is assumed that the flow resulting from the spill of an Iranian Heavy Crude Oil is of non-
viscous type, instead of the viscous type as was assumed in 7.4.1. The oil and soil parameters are the same
as those used in Section 7.4.1. The source spill patterns also remain the same. Unlike in Section 7.4.1, it
is assumed the surface spreading of oil is bounded by a berm whose area is 100 m
2
, which corresponds
to a radius of 5.64m.
Figure 7.13 illustrates the OILSFSM simulated radius of spill and oil lens thickness as they vary with time.
Again the results correspond to three cases : (1) pure surface flow without infiltration or volatilization; (2)
coupled surface flow and volatilization only, and (3) coupled surface flow, volatilization and infiltration. For
verification purpose, the exact solutions to the radius and thickness of the oil lens are also presented for the
case of pure surface flow. Figure 7.14 shows the cumulative mass loss to volatilization and infiltration. The
following observations are made by examining Figures 7.13 and 7.14:
For pure surface flow, both the radius and thickness of oil lens simulated using OILSFSM agree
with those of the exact solutions.
The oil reached the edge of the berm after about 30 seconds. Beyond this time, the oil builds up
inside the area surrounded by the berm. The oil thickness increases from 0.0882 cm (the minimum
thickness due to surface tension) to about 1 cm.
Neither volatilization nor infiltration seem to have a great impact on the radius of spreading. In the
case of surface flow coupled with both volatilization and infiltration, the oil lens thickness drops
gradually with time because of the volume loss to infiltration.
Mass lost to infiltration is significantly higher than the mass loss to volatilization.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 31
(a) Radius of Spreading versus Time
0
100
200
300
400
500
600
0 1 2 3 4 5
Time (Min)
R
a
d
i
u
s

(
c
m
)
Surface Flow Only (Exact)
Surface Flow Only (OILSFSM)
Surface Flow + Volatilization (OILSFSM)
Surface Flow+Infiltration+Volatilization (OILSFSM)
(b) Oil Lens Thickness versus Time
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 30 60 90 120
Time (Min)
h

(
c
m
)
Surface Flow Only (Exact)
Surface Flow Only (OILSFSM)
Surface Flow + Volatilization (OILSFSM)
Surface Flow+Infiltration+Volatilization (OILSFSM)
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 7- 32
(a) Cumulative Mass Loss to Volatilization
0
1,000
2,000
3,000
4,000
5,000
6,000
0 30 60 90 120
Time (min)
M
a
s
s

(
g
)
Surface Flow+Volatilization Only
Surface Flow+Infiltration+Volatilization
(b) Cumulative Mass Loss to Infiltration
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
900,000
0 1 2 3 4 5
Time (hours)
M
a
s
s

(
g
)
Surface Flow+Volatilization+Infiltration
Figure 7.13 Surfa
ce Flow of Non-Viscous
I r a n i a n Heavy Oil:
Radius of Spreadi ng
a n d Av e r a g e d
Oil Lens Thi c kne s s
and t he E f f e c t o f
Infiltration a n d
Volatilizati on. (a) Radius
of spreading versus time;
(b) Oil lens t h i c k n e s s
versus time.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 8- 1
Figure 7.14 Surface Flow of Non-Viscous Iranian Heavy Oil: (a) Cumulative Mass Loss to
Volatilization and (b) Cumulative Mass Loss to Infiltration.
8.0 REFERENCES
Abriola, L. M. and G. F. Pinder, A multiphase approach for modeling of porous media contamination by
organic compounds, 1. Equation development, Water Resources Research, 21(1), 11-18, 1985a.
Abriola, L. M. and G. F. Pinder, A multiphase approach for modeling of porous media contamination by
organic compounds, 2. Numerical simulation, Water Resources Research, 21(1), 19-26, 1985b.
Applied Science Associates, Spill Impact Mapping (SIMAP); Technical Manual, Version 2.5, Applied
Science and Technology, Narragansett, RI, 1998.
Benjamin, T. B., Gravity currents and related phenomena, Journal of Fluid Mechanics, 31, 209-248,
1968.
Britter, R. E. and J. E. Simpson, Experiments on the dynamics of gravity current head, 88, 223-240, 1978.
Buckmaster, J., Viscous-gravity spreading of an oil slick, Journal of Fluid Mechanics, 59, 481-491,
1973.
Carsel, R.F., and R.S. Parrish, Developing joint probability distributions of soil water retention
characteristics, Water Resour. Res., 24(5), 755-769, 1985.
Charbeneau, R. J., J. W. Weaver, and B. K. Lien, The Hydrocarbon Spill Screening Model (HSSM)
Volume 2: Theoretical Background and Source Codes, US EPA, EPA/600/R-94/039b, 1994.
Coats, K.H., An Equation of state compositional model, Society of Petroleum Engineering Journal,
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 8- 2
363-376, 1980.
Chong, S. K. and R. E. Green, Sorptivity measurement and its applications, in Advances in Infiltration,
proceeding of the national conference on advances in infiltration, American Society of Agriculture
Engineers, St. Joseph, MI, 82-91,1983.
Chow, V. T., D. R. Maidment, and L. W. Mays, Applied hydrology, McGraw-Hill Book Company,
publisher, New York, NY, 1988.
Corapcioglu, M. Y. and A. Baehr, A compositional multiphase model for groundwater contamination by
petroleum products, I, theoretical considerations, Water Resources Research, 23, 201-243, 1987.
CRC, Handbook of Chemistry and Physics, D. R. Lide, Editor-in-Chief, 76
th
Ed., CRC Press, 1995-
1996.
Darcos, T., Theoretical considerations and practical implications on the infiltration of hydrocarbons in
aquifers, paper presented at International Symposium on Groundwater Pollution by Oil Hydrocarbon,
Int. Assoc. of Hydrol., Prague, June 5-9, 1978.
Didden N. and T. Maxworthy, The viscous spreading of plane and axisymmetric gravity currents, Journal
of Fluid Mechanics, 121, 27-42, 1982.
El-Kadi, A. I., Applicability of sharp-interface models for NAPL transport: 1, Infiltration, Groundwater,
30, 849-856, 1992.
Environmental Protection Agency, Oil spill notification, response, and recovery, the National Response
Center, 1998.
Environmental Protection Agency, Review of mathematical modeling for evaluating soil vapor extraction
systems, EPA/540/R-95/513, Office and Development, Washington DC, July, 1995.
Fay, J. A., The spread of oil slicks on a calm sea, in "oil on the sea," D. P. Hoult (ed.), 53-63, 1969.
Forsyth, P. A., Simulation of nonaqueous phase groundwater contamination, Advances in Water
Resources, 11, 74-83, 1988.
Fukushima Y., Numerical simulation of gravity current front, Journal of Hydraulic Engineering, ASCE,
124(6), 1998.
Greenspan, H. P., On the motion of a small viscous droplet that wets a surface, 84, 125-143, 1978.
Grundy R. E. and J. W. Rottman, Self similar solutions of shallow-water equations representing gravity
currents with variable inflow, Journal of Fluid Mechanics, 169, 337-351, 1986.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 8- 3
Hochmuth D. P. and D. K. Sunada, Ground-water model of two-phase immiscible flow in course material,
Ground Water, 23(5), 1985.
Hocking, L. M., Spreading and instability of viscous fluid sheet, Journal of Fluid Mechanics, 173, 557-
594, 1986.
Hoult, D. P., Oil spreading on the sea, Annual Reviews in Fluid Mechanics, 4, 341-368, 1972.
Huppert H. E. and J. E. Simpson, The slumping of gravity currents, Journal of Fluid Mechanics, 99, 785-
799, 1980.
Huppert, H. E., The propagation of two-dimensional and axisymmetric viscous gravity currents over a rigid
surface, Journal of Fluid Mechanics, 121, 43-58, 1982a.
Huppert, H. E., Flow and instability of a viscous current down a slope. Nature 300, 427-429, 1982b.
Huppert, H. E., The intrusion of fluid mechanics into geology, Journal of Fluid Mechanics, 173, 557-594,
1986.
HydroGeoLogic, Criteria for the use of Aqueous/Non-Aqueous Phase Liquid (NAPL) models for
assessing subsurface migration of oily waste, Prepared for US EPA, contract no. 68-WO-0029, 1992.
HydroGeoLogic, EPA's Composite Model for Oily Waste, EPACMOW, Background document and
user's guide, US EPA, Office of Solid Wate, Washington DC, 1995.
Illangasekare, T., D. Znidarcic, G. Walser, and J. Weaver, An Experimental Evaluation of Two Sharp
Front Models for Vadose Zone Non-Aqueous Phase Liquid Transport, US EPA, EPA/600/R-94/197,
1994.
Jokuty, P., S. Whiticar, Z. Wang, M. Fingas, P. Lambert, B. Fieldhouse, and J. Mullin, A catalogue of
crude oil and oil product properties, Manuscript Report EE-157, Environment Canada Ottawa,
Ontario, on the web at http://www.etcentre.org, 1998.
Jury, W. A., Gardner W. R., and Gardner, W. H., Soil physics, John Wiely & Sons, publishers, New
York, NY, 1991.
Kaluarachchi and Parker, An efficient finite element method for modeling multiphase flow, Water
Resources Research, 25(1), 43-54, 1989.
Knight, J. H., Infiltration functions from exact and approximate solutions of Richards' equations, in
Advances in Infiltration, proceeding of the national conference on advances in infiltration, American
Society of Agriculture Engineers, St. Joseph, MI, 82-91,1983.
Kueper, B. H. and E. O. Frind, Two-phase flow in heterogeneous porous media: 1 Model development,
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 8- 4
Water Resource Research, 27(6), 1049-1057, 1991.
Kupper, B. H. and E. O. Frind, Numerical modeling of multiphase/multicomponent flow and transport, in
"Subsurface Contamination by Immiscible Fluids," Weyer (ed), 1992.
Letniowski F. W. and P. A. Forsyth, A control volume finite element method for three-dimensional NAPL
groundwater contamination, International Journal for Numerical Methods in Fluids, 13, 955-970,
1991.
Lister J. R., Viscous flows down an inclined plane from point and line sources, Journal of Fluid
Mechanics, 242, 631-653, 1992.
Logan J. D., Applied mathematics, a contemporary approach, John Wiley & Sons, 1987.
Mackay, D. and R. S. Matsugu, Evaporation rates of liquid hydrocarbon spills on land and water, the
Canadian Journal of Chemical Engineering, 51, 434-439, 1973.
Mackay D. and M. Mohtadi, The area affected by oil spills on Land, Canadian Journal of Chemical
Engineering, 53(4), 140-143, 1975.
Mercer, J. W. and R. M. Cohen, A review of immiscible fluids in the subsurface, properties, models,
characterization and remediation, Journal of Contaminant Hydrology, 6, 107-163, 1990.
Neuman, S. P., Wetting front pressure head in the infiltration model of Green and Ampt, Water Resource
Research, 12, 564-566, 1976.
Parker, J. C., Multiphase flow and transport in porous media, Reviews of Geophysics, 27(3), 311-328,
1989.
Philip, J. R., The infiltration equation and its solution. Soil Sciences, 83, 345-357, 1957.
Press, W. H., S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery, Numerical Recipes in Fortran: The
Art of Scientific Computing, Cambridge University Press, 1992.
Raisbeck, J. M. and M. F. Mohtadi, The environmental impacts of oil spills on land in the arctic
regions, Water Air & Soil Pollution. 3, 195-508, 1974.
Reible, D. D., T. H. Illangasekare, D. V. Doshi, and M. E. Malhiet, Infiltration of immiscible contaminants
in the unsaturated zone, Groundwater, 28(5), 685-692, 1990.
Smith, P. C., A similarity solution for slow viscous flow down an inclined plane, Journal of Fluid
Mechanics, 58, 275-288, 1973.
Stiver, W. and D. Mackay, Evaporation rate of spills of hydrocarbons and petroleum mixture,
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 8- 5
Environmental Science and Technology, 18, 384-840, 1984.
Stiver, W., W. Y. Shiuu, and D. Mackay, Evaporation times and rates of specific hydrocarbons in oil spills,
Environmental Sciences and Technology, 23, 101-105, 1989.
Thomann R. V. and J. A. Mueller, Principles of surface water quality modeling and control, Happer
& Row, Publishers, New York, 1987.

Weaver, J. W., R. J. Charbeneau, and B. K. Lien, A screening model for nonaqueous phase liquid
transport in the vadose zone using Green-Ampt and kinetic wave theory, Water Resource Research,
30(1), 93-105, 1994.
APPENDIX A

LITERATURE SEARCH RESULTS
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 1
APPENDIX A
LITERATURE SEARCH RESULTS
Abdin, A. E., J. J. Kaluarachchi, Stochastic analysis of three-phase flow in heterogeneous porous media:
1. Spectral/perturbation approach. Water Resources Research, 33(7), 549-1558, 1997.
Abdin, A. E., J. J. Kaluarachchi, Stochastic analysis of three-phase flow in heterogeneous porous media:
2. Numerical simulations. Water Resources Research, 33(7), 1559-1566, 1997.
Abdul, A. S., Migration of Petroleum Products Through Sandy Hydrogeologic Systems. Ground Water
Monitoring Review, 8(4), 73-81, 1988.
Abin, A., J. J. Kalurachchi, M. W. Kemblowski, and C. M. Chang, Stochastic analysis of multiphase flow
in porous media; II, Numerical simulations. Stochastic Hydrology and Hydraulics, 10(3), 231-251,
1996.
Abriola, L. M., Mathematical modeling of the multiphase migration of organic compounds in a porous
medium. Department of Civil Engineering. Princeton, NJ, Princeton University, 236 pp, 1983.
Abriola, L. M. and G. F. Pinder, A multiphase approach to modeling of porous media contamination by
organic compounds, 1, Equation development. Water Resources Research, 21(1), 11-18, 1985a.
Abriola, L. M. and G. F. Pinder, A multiphase approach to modeling of porous media contamination by
organic compounds, 2, Numerical simulation. Water Resources Research, 21(1), 19-26, 1985b.
Abriola, L. M., T. J. Dekker, and K. D. Pennell, Surfactant-enhanced solubilization of residual dodecane
in soil columns. 2. Mathematical modeling. Environmental Science & Technology, 27, 2341-51,
1993.
Ahmadi, A., M. Quintard, J. J. Gomez-Hernandez, Large-scale properties for two-phase flow in random
porous media, Effective parameter estimation for flow and transport in the subsurface. Journal of
Hydrology, 183(1-2), 69-99, 1996.
Alberto, S. C., Three-phase flow analysis of oil spills in partially water-saturated soils (groundwater,
contamination, immiscible fluids, petroleum, infiltration). Civil Engineering, Colorado State University,
47-02b: 148 pages, 1985.
Alexander, A. G., D. D. Zettwoch, M. D. Unthank, and R. B. Burns, Contamination of Soil, Soil Gas, and
Ground Water by Hydrocarbon Compounds Near Greear, Morgan County, Kentucky. Denver, CO,
US Geological Survey, 37 pp, 1992.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 2
Allen, F. R., and D. A. Puckett, Theoretical and experimental studies of rate-dependent two-phase
immiscible flow. SPE Production Engineering, 1, 62-74, 1986.
Annable, M. D., R. Hatfield, and G. Wood, Use of Partitioning tracers for measuring residual NAPL
distribution in a contaminated aquifer: preliminary results from a field-scale test. 2nd Tracer Workshop,
Austin, TX, 1994.
Ashley, R. P., D. N. Lerner, J. W. Lloyd, H. S. Wheater, P. A. C. Raats, and A. C. Armstrong,
Distribution and degradation of diesel oil in the unsaturated zone following an oil spill on a Chalk aquifer
Field, laboratory and Modelling studies of flow and transport processes. Journal of Hydrology,
159(1-4), 43-59, 1994.
Baehr, A. L. and M. Y. Corapcioglu, A compositional multiphase model for groundwater contamination
by petroleum products, 2, Numerical solution. Water Resources Research, 23(1), 201-213, 1987.
Bear, J., V. Ryzhik, C. Braester, and V. Entov, On the movement of an LNAPL lens on the water table.
Transport in Porous Media, 25(3), 283-311, 1996.
Beckett, G. D. and D. Huntley, Soil properties and design factors influencing free-phase hydrocarbon
cleanup. Environmental Science & Technology, 32, 287-93, 1998.
Benjamin, T. B., Gravity currents and related phenomena. J. Fluid Mechanics, 31 (part 2), 209-248,
1968.
Betelu, S., J. Diez, L. Thomas, R. Gratton, and B. Marino, Boundary element method for viscous gravity
currents. International Journal for Numerical Methods in Fluids, 25(1), 1-19, 1997.
Brannaka, L. K., Aqueous transport of trichloroethylene in an unsaturated uniform fine sand under transient
rainfall conditions, The Pennsylvania State University, 54-07B, 476pp, 1993.
Britter, S., Experiments on the dynamics of a gravity current head. J. Fluid Mechanics, 88 (part 2), 223-
240, 1978.
Bruehl, D. H., N. K. Chung, and W. F. Diesl, Geological Studies of Industrially-Related Contamination:
Soil and Ground Water Investigations. EPA National Conference on Management of Uncontrolled
Hazardous Wastes, Washington, DC, 1980.
Buckmaster, J., Viscous-gravity Spreading of an oil slick. Journal of Fluid Mechanics, 59 (Part 3), 481-
491, 1973.
Buyevich, Y. A., and V. A. Ustinov, Hydrodynamic conditions of transfer processes through a radial jet
spreading over a flat surface. International Journal of Heat & Mass Transfer, 37(1), 165-173,
1994.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 3
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 4
Cary, J. W., J. F. McBride, and C. S. Simmons, Observations of water and oil infiltration into soil; some
simulation challenges. Water Resources Research, 25(1), 73-80, 1989.
Cary, J. W., C. S. Simmons, and J. F. McBride, Predicting oil infiltration and redistribution in unsaturated
soils. Soil Science Society of America Journal, 53(2), 335-342, 1989.
Cary, J. W., C. S. Simmons, and J. F. McBride, Infiltration and redistribution of organic liquids in layered
porous media. Soil Science Society of America Journal, 58(3), 704-711, 1994.
Catalan, L., Gravity drainage of waterflood residual oil assisted by inert gas injection, University of
Waterloo (Canada), 54-08b, 249pp, 1992.
Chang, C. C., and P. B. Bedient, Multiphase unsaturated zone flow and transport model for ground water
contamination by hydrocarbon. Proceedings of the Petroleum hydrocarbons and organic
chemicals in ground water; prevention, detection, and restoration; a conference and exposition,
Water Well Journal Pub. Co., Dublin, OH, United States, 1991.
Chang, C. C. W., R. W. Klopp, R. A. Bedient, and B. Philip, In-situ source release mechanism study at
an aviation gasoline spill site; Traverse City, Michigan. Proceedings of the Fourth national outdoor
conference on Aquifer restoration, ground water monitoring and geophysical methods, National
Water Well Association, Dublin, OH, United States, 1990.
Charbeneau, R., Kinematic Models for soil moisture and solute transport. Water Resources Research,
20(6), 699-706, 1984.
Charbeneau, R. and R. Asgian, Simulation of the transient soil water content profile for a homogeneous
bare soil. Water Resources Research, 27(6), 1271-1279, 1991.
Charbeneau, R. J., J. W. Weaver, and V. J. Smith, Kinematic Modeling of Multiphase Solute Transport
in the Vadose Zone, US EPA, 158 pp, 1989.
Charbneeau, R. J., J. W. Weaver, and B. K. Lien, The Hydrocarbon Spill Screening Model (HSSM)
Vol.2: Theoretical Background and Source Codes. Ada, OK, EPA, Kerr Enviro. Research Lab,
Office of R&D, EPA/600/039b, 1994.
Chen, M. R., Multiphase flow during contaminant transport and remediation in subsurface, University of
Houston, 57-02B, 175 pp, 1995.
Chen, Z., and R. E. Ewing, From single-phase to compositional flow; applicability of mixed finite elements.
Transport in Porous Media, 27(2), 225-242, 1997.
Chevalier, L. R., Evaluation of hydraulic controls to promote surfactant dissolution of trapped residual
NAPL. Michigan State University, 175 pp, 1994.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 5
Chiayang, L., and E. A. Voudrias, Migration and sorption of jet fuel aliphatic vapors in unsaturated soil.
Water Research, 28(12), 2447-2456, 1994.
Corapcioglu, M. Y., and A. L. Baer, A compositional multiphase model for groundwater contamination
by petroleum products 1: theoretical considerations. Water Resources Research 23(1), 191-200,
1987.
Corapcioglu, M. Y. and M. A. Hossain, Ground-water contamination by high-density immiscible
hydrocarbon slugs in gravity-driven gravel aquifers. Ground Water, 28(3), 403-412, 1990.
Corapcioglu, M. Y., L. Rajasekhar, and V. Haisler, Analytical Prediction of Gasoline thickness on the
water table. Water Resources Sessions at Water Forum, Baltimore, MD, American Society of Civil
Engineers, 1996.
Corapcioglu, M. Y., S. Chowdhury, S. Roosevelt, K. Bullock, R. Kukreja, and D. Hodko, An
investigation of residual hydrocarbon dissolution by micromodels. Annual Meeting Abstracts -
American Association of Petroleum Geologists and Society of Economic Paleontologists and
Mineralogists, 1996.
Corapcioglu, M. Y., K. Tuncay, and B. K. Ceylan, Oil mound spreading and migration with ambient
groundwater flow in coarse porous media. Water Resources Research, 32(5), 1299-1308, 1996.
Cruz-Sanjulian, J., and M. Olias, Characterization of hydrocarbon spreading in an alluvial aquifer by cross
correlation study of precipitation and contaminant content data. Water, Air and Soil Pollution, 81(3-
4), 337-347, 1995.
Curtis, G. P., I. M. Cozzarelli, and M. J. Baedecker, Simulation of solute transport influenced by coupled
microbial and inorganic geochemical reactions at a crude-oil spill site. Eos, Transactions, American
Geophysical Union, 78(17) (Suppl.), 167, 1997.
Dahmani, M., A. Forster, K. Ashfeld, D. Hoag, G. E. Carley, R. J. Chen, J. Kostecki, P. T. Calabrese,
and J. Edward, Ed., Measurement of motor fuel contaminated soil leachate. Hydrocarbon
contaminated soils and groundwater, 1993.
Dalen, V., W. G. Gray, G. F. Pinder, and C. A. Brebbia, Immiscible flow by finite elements. Finite
elements in water resources; proceedings of the first international conference, Pentech Press, London,
United Kingdom, 1977.
Dawson, H. E. and P. V. Roberts, Influence of viscous, gravitational, and capillary forces on DNAPL
saturation. Ground Water, 35(2), 261-269, 1997.
Dayal, P., NAPL remediation by vacuum enhanced recovery; laboratory and model evaluations.
Environmental Sciences, The University of Arizona, 57-04B, 112 pp, 1996.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 6
Dekker, T. J., An Assessment Of The Effects Of Field-Scale Formation Heterogeneity On Surfactant-
Enhanced Aquifer Remediation, 57-07b, 235 pp, 1996.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 7
Dengen, Z. and M. Blunt, Effect of spreading coefficient on the distribution of light non- aqueous phase
liquid in the subsurface, 1997.
Di Luise, G., Unsaturated medium hydrocarbons pollution evaluation. First Int Conf Health Saf Environ
Oil Gas Explor Prod., Richardson, TX, Soc of Petroleum Engineers of AIME, 1990.
Didden, N. and T. Maxworthy, The viscous spreading of plane and axisymmetric gravity currents. Journal
of Fluid Mechanics, 121: 27-42, 1982.
Dillard, L. A., H. I. Essaid, and W. N. Herkelrath, Multiphase flow modeling of a crude-oil spill site using
geostatistical simulation of soil hydraulic properties. Eos, Transactions, American Geophysical
Union, AGU fall meeting, 73(43) Suppl, 185, 1992.
Dillard, L. A., Essaid, I. Hedeff, and W. N. Herkelrath, Multiphase flow modeling of a crude-oil spill site
with a bimodal permeability distribution. Water Resources Research, 33, 1617-32, 1997.
Dorgarten, H. W., H. E. Kobus, and W. Kinzelbach, Finite element simulation of immiscible and slightly
soluble pollutants in soil and groundwater. Proceedings of the international symposium on
Contaminant transport in ground water, 1989.
Doshi, D. V. and D. D. Reible, Application of the Boundary Element Method to Moving Boundary
Problems Arising During Non-Aqueous Phase Liquid (NAPL) Migration in Soils, 1992.
Douglas, J., B. L. Darlow, and M. F. Wheeler, Self-adaptive finite element and finite difference methods
for one-dimensional two-phase immiscible flow. Computer Methods in Applied Mechanics and
Engineering, 47, 119-30, 1984.
Dracos, T., Theoretical consideration and practical implications of the infiltration of hydrocarbons in
aquifers. International symposium on groundwater pollution by oil hydrocarbons, Prague, 1987.
Dracos, T., Ed., Immiscible Transport of Hydrocarbons Infiltrating in Unconfined Aquifers. Oil in
Freshwater: Chemistry, Biology, Countermeasure Technology, 1987.
Drakos, T., and R. Muzikar, Theoretical considerations and practical implications on the infiltration of
hydrocarbons in aquifer. International symposium on ground water pollution by oil hydrocarbons, 1978.
Drolet, C., O. Banton, P. Lafrance, and J. P. Villeneuve, Assessing the fate of polynuclear aromatic
hydrocarbons from oily waste land spreading by modelling. Canadian Journal of Civil Engineering,
23(1), 211-217, 1996.
Duffy, J. J., E. Peake, and M. F. Mohtadi, Subsurface persistence of crude oil spilled on land and its
transport in groundwater. Oil spill conference (prevention, behavior, control, cleanup), API Publication,
1977.
Eckberg, D. K. and D. K. Sunada, D. K., Nonsteady three-phase immiscible fluid distribution in porous
media. Water Resources Research, 20, 1891-1897, 1984.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 8
El-Kadi, A. I., Applicability of sharp-interface models for NAPL transport: infiltration. Ground Water,
30(6), 849-56, 1992.
EPA, Estimating Potential for Occurrence of DNAPL at Superfund Sites. Ada, OK, EPA, office of
emergency and remedial response, 9, 1992.
EPA, Ground Water Sampling -- A Workshop Summary. Dallas, TX, EPA, R. Kerr Environmental
Research Lab, 97, 1993.
Essaid, H. I. and K. M. Hess, Monte Carlo Simulations of Multiphase Flow Incorporating Spatial
Variability of Hydraulic Properties. Ground Water, 31(1), 123-134, 1993.
Essaid, H. I., W. N. Herkelrath, and K. M. Hess, Simulation of fluid distributions observed at a crude oil
spill site incorporating hysteresis, oil entrapment, and spatial variability of hydraulic properties. Water
Resources Research, 29(6), 1753-1770, 1993.
Essaid, H. I., B. A. Bekins, E. M. Godsy, E. Warren, M. J. Baedecker, I. M. Cozzarelli,, G. M.
Hornberger, and M. P. Anderson, Simulation of aerobic and anaerobic biodegradation processes at
a crude oil spill site. Water Resources Research, 31(12), 3309-3327, 1995.
Essaid, H. I., M. J. Baedecker, I. M. Cozzarelli, D. W. Morganwalp, and D. A. Aronson, Use of
simulation to study field-scale solute transport and biodegradation at the Bemidji, Minnesota, crude-oil
spill site. proceedings of the technical meeting Water-Resources Investigations - U. S. Geological
Survey, U.S. Geological Survey Toxic Substances Hydrology Program; U. S. Geological Survey,
[Reston, VA], United States, 1996.
Fabbri, P., A. Ferronato, and P. Zangheri, A case of groundwater contamination by organo-chlorine
compounds. Hydrogeologie, 3, 207-215, 1993.
Fay, J., Ed., Spread of oil slicks on a calm sea. Oil on the Sea. Cambridge, MA, MIT, 1969.
Feldman, P. R., Hydrogeology of a contaminated industrial site on filled land. College of Engineering &
Mines, Department of Hydrology & Water Resources. Tucson, University of Arizona, 1988.
Fenwick, D. H., M. J. Blunt, A. S. Mayer, and R. J. Lenhard, Three-dimensional modeling of three phase
imbibition and drainage Multiphase flow and chemical transport. Advances in Water Resources,
21(2), 121-143, 1998.
Ferrara, S. M. M. S., Control and mitigation of inland gasoline releases from petroleum pipelines to surface
water. Environmental Sciences, Fullerton, California State University. 31-02, 133 pp, 1992.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 9
Finsterle, S. and K. Pruess, Optimizing multiphase aquifer remediation using ITOUGH2. Berkeley, CA,
United States, Lawrence Berkeley Laboratory, University of California, 29pp, 1994.
Fiori, C. M., Climatic Effects On Near-Surface Contaminant Transport Routes In Fine-Grained Soils,
Drexel University. 58-06B, 200 pp, 1997.
Forsyth, P., Simulation of NAP groundwater contamination. Adv. Water Resources, 11, 74-83, 1988.
Forsyth, P., A control volume finite element approach to NAPL groundwater contamination. Siam
Journal of Scientific Stat. Comput. 12(5), 1029-1057, 1991.
Forsyth, P. A. and B. Y. Shao, Numerical simulation of gas venting for NAPL site remediation. Adv.
Water Resources, 14(6), 354-367, 1991.
Forsyth, P. A., A positivity preserving method for simulation of steam injection of NAPL site remediation.
Advances in Water Resources, 16, 351-370, 1993.
Forsyth, P. A., Three-dimensional modeling of steam flush for DNAPL site remediation. International
Journal for Numerical Methods in Fluids, 19, 1055-1081, 1994.
Forsyth, P. A. and E. A. Sudicky, Discrete wellbore simulations of pump-and-treat strategies for
remediation of LNAPL-contaminated aquifers. Journal of Contaminant Hydrology 31(1-2), 57-81,
1998.
Fortin, J., Experimental and theoretical investigation of rate-limited processes in soil under water-saturated
conditions (pesticides), University Of California, Riverside, 201 pp, 1995.
Fukushima, Y., Numerical simulation of gravity current front. Journal of Hydraulic Engineering, 124(6),
572-8, 1998.
Galin, T., C. McDowell, and B. Yaron, The effect of volatilization on the mass flow of a non-aqueous
pollutant liquid mixture in an inert porous medium: experiments with kerosene. The Journal of Soil
Science, 41(4), 631-641, 1990.
Ginzbourg, I. and P. M. Adler, Surface tension models with different viscosities Multiphase flow in porous
media. Transport in Porous Media, 20(1-2), 37-76, 1995.
Glass, R. J., M. J. Nicholl, T. S. Steenhuis, C. J. Ritsema, and L. W. Dekker, Physics of gravity fingering
of immiscible fluids within porous media; an overview of current understanding and selected
complicating factors. Fingered flow in unsaturated soil; from nature to model, Geoderma, Elsevier,
Amsterdam, Netherlands, 1996.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 10
Gomez-Lahoz, C., R. A. Garcia-Delgado, and D. J. Wilson, A model with mass transport limitations for
pump and treat remediation of soils polluted with NAPL. Environmental Monitoring and
Assessment, 32(2), 161-186, 1994.
Grady, D. E., Remediation of LNAPL in a naturally fractured clay till: physical and numerical modeling
(hydrocarbons, soil vapor extraction). Dai, 58-09b, 277 pp, 1997.
Gratton, J. and F. Minotti, Self-similar viscous gravity currents. Phase-plane formalism. Journal of Fluid
Mechanics, 210, 155-182, 1990.
Gravelle, C. F., M. A. Knight, and R. J. Mitchell, Measurement of two-fluid saturations in a fine to medium
sand. Canadian Geotechnical Journal, 33(6), 1014-1017, 1996.
Greenspan, H. P., On the motion of a small viscous droplet the wets a surface. J. Fluid Mechanics, 84
(part 1), 125-143, 1978.
Grubb, D. and N. Sitar, Evaluation of Technologies for In-situ Cleanup of DNAPL Contaminated sites.
Ada, OK, EPA, Kerr Enviro. Research Lab., 171, 1994.
Guigard, S., W. H. Stiver and R. G. Zytner, The infiltration and movement of immiscible chemicals in
unsaturated soil. Environmental Technology, 17(10), 1123-1130, 1996.
Guiguer, N., Numerical modelling of the fate of residual immiscible fluids in saturated porous media.
Pollution, protection and control of groundwater. Proceedings of the IAWPRC International Seminar,
held in Porto Alegre, Brazil, Porto Alegre, Brazil, Oxford [England]; New York: Pergamon Press,
c1992.
Hadad, A., J. Bensabat, and H. Rubin, Simulation of immiscible multiphase flow in porous media; a focus
on the capillary fringe of oil-contaminated aquifers. Transport in Porous Media, 22(3), 245-269,
1996.
Hartz, M. J. and T. J. Overcamp, The geometry of nonwetting liquids in the unsaturated zone. Ground
Water, 34(5), 837-841, 1996.
Hatfield, K. and T. B. Stauffer, Analytical models for one, two, and three-dimensional non-equilibrium
solute transport through heterogeneous porous media contaminated with NAPL. Eos, Transactions,
AGU Fall meeting, 1992.
Hatfield, K. and T. B. Stauffer, Transport in porous media containing residual hydrocarbon. I: model.
Journal of Environmental Engineering - ASCE 119(3), 540- 558, 1993.
Hatfield, K., J. Ziegler, and D. R. Burris, Transport in porous media containing residual hydrocarbon. II:
experiments. Journal of Environmental Engineering - ASCE 119(3), 559- 575, 1993.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 11
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 12
Hathaway, D. L. and C. B. Andrews, Fate and transport modeling of organic compounds from a gasoline
spill. Ground Water Management, Proceedings of Petroleum hydrocarbons and organic chemicals
in ground water; prevention, detection, and restoration; a conference and exposition, Water Well
Journal Pub. Co., Dublin, OH, United States, 1990.
Hayden, N. J., R. B. Wallace, M. D. Annable, and T. C. Voice, Change in gasoline constituent mass
transfer during soil venting, 1994.
Hess, K. M., Simulation of fluid distributions observed at a crude oil spill site incorporation hysteresis, oil
entrapment, and spation variability of hydraulic properties. Water Resources Research, 29(6), 1753-
1770, 1993.
Heyes, E. P., Mass transfer between organic and aqueous phases: investigations using a continuously stirred
flow cell (groundwater, soil organic matter), University of Florida. Dai, 56-11B, 285 pp, 1994.
Higuera, F. J., Steady creeping flow down a slope. Physics of Fluids, 7(11), 2918-2920, 1995.
Hinkley, R. E., Numerical simulation of subsurface, multiphase contaminant infiltration and remediation
(groundwater), University of Houston; 0087. 57-10B, 264 pp, 1996.
Hinkley, R. K., T. F. J. Russell, R. E. Ewing, C. A. Brebbia, W. G. Gray, and G. F. Pinder, Simulation
of multi-phase contaminant transport in the vadose zone; model case studies and an exploration of air-
phase dynamics. Computational Methods in Water Resources IX, 1992.
Hochmuth, D. P. A. S., and K. Daniel, Ground-water model of two-phase immiscible flow in coarse
material. Ground Water, 23, 617-26, 1985.
Hocking, L. M., Sliding and Spreading of thin two-dimensional drops. Journal of Mechanics and
Applied Mathematics XXXIV, pt 1, 1981.
Holzer, T., Application of ground-water flow theory to a subsurface oil spill. Ground Water, 14(3), 138-
145, 1976.
Hossain, M. A. and M. Y. Corapcioglu, Modifying the USGS solute transport computer model to predict
high- density hydrocarbon migration. Ground Water 26(6), 717-723, 1988.
Host-Madsen, J. A. J. and K. Hogh, Laboratory and numerical investigations of immiscible multiphase flow
in soil. Journal of Hydrology 135(1-4), 13-52, 1992.
Hughes, D. S. A. M. P., Use of a Monte Carlo method to simulate unstable miscible and immiscible flow
through porous media. SPE Reservoir Engineering, 3, 1129-36, 1988.
Huling, S. and J. Weaver, Dense non-aqueous phase liquids. Ada, OK, EPA, 21, 1991.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 13
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 14
Huppert, H., Flow and instability of a viscous current down a slope. Nature, 300, 427-429, 1982a.
Huppert, H., The propagation of two-dimensional and axi-symmetric viscous gravity currents over a rigid
horizontal surface. J. Fluid Mech., 121, 43-58, 1982b.
Huppert, H., The intrusion of fluid mechanics into geology. J. Fluid Mech., 173, 557-594, 1986.
Huppert, H. A. S., The slumping of gravity currents. J. Fluid Mech., 99 (part 4), 785-799, 1980.
Huppert, H. A. W., Gravity-driven flows in porous layers. Journal of Fluid Mechanics, 292, 55-69,
1995.
Huyakorn, P. S., A. P., G. F., A new finite element technique for the solution of two-phase flow through
porous media. Advances in Water Resources, 1(5), 285-298, 1978.
Huyakorn, P. S., Y. S. Wu, S. Kuo, and Y. Chin. A numerical model for multiphase flow and transport
assessment of migration of petroleum contaminants in the subsurface. Engineering Hydrology, 964-
969, 1993.
Huyakorn, P. S., Y. S. Wu; S. Panday, N. S. Park, J. B. Kool. SAMFT2D; single-phase and multiphase
flow and transport in 2 dimensions; Version 2, Documentation and user's guide, 317, 1991.
Huyakorn, P. S. , Y. S. Wu, N.S. Park, An improved sharp-interface model for assessing NAPL
contamination and remediation of groundwater systems. Journal of Contaminant Hydrology, 16(3),
203-234, 1994.
Huyakorn, P.S. and S. Panday, MAGNAS3 Multiphase Analysis of Groundwater, NAPL and Soluble
component. Herndon, HydroGeoLogic, Inc., 1994.
HydroGeoLogic, Technical Report: Criteria for the use of Aqueous/Non-Aqueous Phase Liquid (NAPL)
Models for Assessing Subsurface Migration of Oily Waste. Herndon, EPA Office of solid waste, 69,
1992.
HydroGeoLogic, EMACMOW: EPA's composite model for oily wastes version 2.0. Herndon, EPA,
1995.
Illangasekare, T., D. Znidarcic, G. Walser, and J. Weaver, Experimental evaluation of two sharp front
models for vadose zone non-aqueous phase liquid transport, Environmental Protection Agency, 197,
1994.
Illangasekare, T., D. Znidarcic, and G. Walser, An Experimental Evaluation of Two Sharp Front Models
for Vadose Zone NAPL Transport. Ada, OK, EPA, 1994.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 15
Illangasekare, T. H. R., L. James, Jr., Jensen, H. Karsten, and M. B. Butts, Experimental study of
movement and distribution of dense organic contaminants in heterogeneous aquifers. Journal of
Contaminant Hydrology, 20(1-2), 1-25, 1995.
Jawitz, J. W., M. D. Annable, and P. S. C. Rao, Miscible fluid displacement stability in unconfined porous
media; two dimensional flow experiments and simulations. Journal of Contaminant Hydrology,
31(3-4), 211-230, 1998.
Jensen, O. E. and J. B. Grotberg, Insoluble surfactant spreading on a thin viscous film: shock evolution and
film rupture. Journal of Fluid Mechanics, 240 (July '92), 259-88, 1992.
Jinquan, W., Modeling water flow and chemical transport in soil and groundwater systems (rainfall-
recharge, root density, hydraulic conductivity), University Of Wyoming, 196 pp, 1996.
Joseph, A. T., Development of a computer model to simulate oil migration through the unsaturated soil and
modification of the vadose zone interactive processes (VIP) model, Utah State University, 288 pp,
1993.
Joseph, A. T., W. J. Grenney, and D. K. Stevens, Oil migration through unsaturated soils and its effect on
the vadose zone interactive processes (VIP) model output. Water Science & Technology, 30(3), 39-
51, 1994.
Kaluarachchi, J. and J. C. Parker, An Efficient finite element method for modeling multiphase flow. Water
Resources Research, 25(1), 43-54, 1989.
Kaluarachchi, J. J., and J. C. Parker, Multiphase flow with a simplified model for oil entrapment.
Transport in Porous Media, 7(1), 1-14, 1992.
Kao, C. S. S., Analysis and verification of a wetting front model for infiltration into soils. University of
California, Berkeley, 171 pp, 1996.
Katyal, A. K., J. J. Kaluarachchi, and J. C. Parker, MOFAT; a two-dimensional finite element program
for multiphase flow and multicomponent transport, Environmental Protection Agency, 1991.
Keller, A. A., Single and multiphase flow and transport in fractured porous media (NAPLs, granite,
sandstone), Stanford University, 235 pp, 1996.
Kershaw, G. P., Movement of crude oil in an experimental spill on the SEEDS simulated pipeline right-of-
way, Fort Norman, NWT. Arctic 43(2), 176-183, 1990.
Kessler, A. and R. Hillel, Relationships between water infiltration and oil spill migration in sandy soils.
Journal of Hydrology, 91(3-4), 187-204, 1987.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 16
Kessler, A. R., H. and W. R. Waldrop, Development of a practical model simulating oil spill migration
(OSPIM) in soils. Hydraulics and hydrology in the small computer age, 1480-1485, 1985.
Kim, J, Modeling dissolution and volatilization of LNAPL sources migrating on the groundwater table
(groundwater contamination, NAPL migration), Texas A&M University. 59-04B, 123pp, 1985.
Kiriakidis, D. G., G. H. Neale, and E. Mitsoulis, Slow immiscible displacement flow in disordered porous
media. The Canadian Journal of Chemical Engineering, 70, 1220-3, 1992.
Kissling, W., M. McGuinness, G. W. Weir, and S. R. Young, Vertical two-phase flow in porous media.
Transport in Porous Media, 8(2), 99-131, 1992.
Knight, M. A., Physical and numerical modelling of light non-aqueous phase liquid releases into unsaturated
fine sand. Civil Engineering, Queen's University at Kingston (Canada); 57-07B, 227pp, 1996.
Kool, J. B. and J. C. Parker, Analysis of the inverse problem for transient unsaturated flow. Water
Resources Research, 24(6), 817-830, 1988.
Kueper, B. H. and E. O. Frind, The behavior of DNAPL in heterogeneous porous media, 1: Model
development, verification, and validation. submitted to Water Resources Research, 1989.
Kuhlthau, R. H., Collocation finite element model of three-phase flow in porous media. Civil Engineering,
Colorado State University. 56-04B, 170pp,1994.
Lane-Serff, G. F., L. M. Beal, and T. D. Hadfield, Gravity current flow over obstacles. Journal of Fluid
Mechanics, 292, 39-53, 1995.
Lee, B. Y. C., A three-dimensional study of transport and remediation of organic compounds in variably
saturated porous media (Pentane), University of Toronto (Canada). 35-01, 169 Pages, 1996.
Lee, K. Y., C. V. Chrysikopoulos, and H. J. Morel-Seytoux, Numerical modeling of contaminant migration
from dissolution of stagnant multicomponent NAPL pools in saturated three-dimensional porous media.
Proceedings of Fourteenth annual American Geophysical Union: Hydrology Days, American
Geophysical Union, 1994.
Lee, K. Y., and C. V. Chrysikopoulos, Numerical modeling of three-dimensional contaminant migration
from dissolution of multicomponent NAPL pools in saturated porous media. Environmental Geology,
26, 157-65, 1995.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 17
Lee, S. C., L. Eastcott, W. Y. Shiu, and D. Mackay, Ed, Petroleum Contaminated Soil: Chemistry and
Modeling. Principles and Practices for Petroleum Contaminated Soils. Boca Raton, Florida., Lewis
Publishers, 1993.
Lehotsky, J. and R. Muzikar, Dynamics of motion and alteration in quality of petroleum hydrocarbons in
soil and water environment. International symposium on ground water pollution by oil hydrocarbons,
Prague, Czechoslovakia, Stavebni Geol., 1978.
Lenhard, R. and J. Parker, A model for hysterectic constitutive relation governing Multiphase flow 2.
Permeability-saturation relations. Water Resources Research, 23(12), 2197-2206, 1987.
Lenormand, R., Transport equations for fluid displacements in stratified porous media: the MHD model.
The 1995 SPE Annual Technical Conference and Exhibition, Richardson, TX, (USA), Society of
Petroleum Engineers, Spe., 1995.
Letniowsky, F. and P. Forsyth, A control volume finite element method for three-dimensional NAPL
groundwater contamination. International Journal for Numerical Methods in Fluids, 13, 955-970,
1991.
Lister, J. R. and R. C. Kerr, Propagation of two-dimensional and axisymmetric viscous gravity currents
at a fluid interface. Journal of Fluid Mechanics, 203, 215-249, 1989.
Lujan Sierraalta, C. A. (1985). Three-phase flow analysis of oil spills in partially water-saturated soils.
Thesis or dissertation Record Level: Monographic: 132 p, 1985.
Mallard, G. E., M. J. Baedecker, and H. I. Essaid, Hydrocarbon transport and degradation in ground
water. Geotimes, 38, 18-20, 1993.
Mariner, P. E., M. Jin, and R. E. Jackson, An algorithm for the estimation of NAPL saturation and
composition from typical soil chemical analyses. Ground Water Monitoring & Remediation, 17, 122-
9, 1997.
Martinez, M. J. and D. F. McTigue, The steady distribution of moisture beneath a two-dimensional surface
source. Water Resources Research, 27(6),1193-1206, 1991.
Mayer, A. S. and C. T. Miller, Simulation of NAPL distributions in groundwater systems, using a
laboratory pore-scale model. Eos, Transactions, American Geophysical Union, 70(15) (Third
scientific assembly of the International Association of Hydrological Sciences): 337, 1989.
Mayer, A. S., An investigation of residual nonaqueous phase liquid dissolution in saturated groundwater
systems (water quality). Chapel Hill, University of North Carolina at Chapel Hill. 53-07B, 228 Pages,
1992.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 18
Mayer, A. S., C. T. Miller, T. F. Russell, R. E. Ewing, C. A. Brebbia, W. G. Gray, and G. F. Pinder,
Simulating nonaqueous phase liquid dissolution in heterogeneous porous media. the International
Conference on Computational Methods in Water Resources IX; Mathematical modeling in water
resources, Boston, MA, 1992.
Mayer, A. S. and C. T. Miller, Experimental investigation of pore-scale distributions of nonaqueous phase
liquids at residual saturation. Transport in Porous Media, 10(1), 57-80, 1993.
McBride, J. F., C. S. Simmons, and J. W. Cary, Interfacial spreading effects on one-dimensional organic
liquid imbibition in water-wetted porous media. Journal of Contaminant Hydrology, 11(1-2), 1-25,
1992.
McCaulou, D., D. Jewett, and S. Huling, NAPL compatibility with materials used in well construction,
sampling, and remediation. Ada, OK, EPA, 11, 1995.
McHale, G., M. I. Newton, and S. M. Rowan, The spreading of small viscous stripes of oil. Journal of
Physics. D, Applied Physics, 28, 1925-9, 1995.
McWhorter, D. and D. Sunada, Exact Integral Solution of two-phase flow. Water Resources Research,
26(23), 399-413, 1990.
Mendoza, C. and E. Frind, Advective-dispersive transport for dense organic vapors in the unsaturated
zone 1: model development. Water Resources Research, 26(3), 379-387, 1990.
Mercer, J. and R. Cohen, A Preview of immiscible fluids in the subsurface: properties models,
characterization and remediation. J. Contaminant Hydrology, 6, 107-163, 1990.
Metcalfe, D. E. and G. Zukovs, Rapid Assessment Model for Spills on Soil of Oily Fluids that are
Immiscible with Water. NWWA/API Conference on Petroleum Hydrocarbons and Organic Chemicals
in Ground Water: Prevention, Detection and Restoration., Houston Texas, National Water Well
Association, Dublin, OH, 1986.
Mott, H. V., A model for determination of the phase distribution of petroleum hydrocarbons at release
sites. Ground Water Monitoring & Remediation, 15(3), 157-167, 1995.
Mualem, Y., A new model for predicting the hydraulic conductivity of unsaturated porous media. Water
Resources Research, 12(3), 513-522, 1976.
Murarka, I., MOSES 2.0 released for beta testing. (Mineral Oil Spill Evaluation System). SOURCE:
EPRI Journal, 20, 40-2, 1995.
Odeh, A. S., A proposed technique for simulation of viscous fingering in one-dimensional immiscible flow.
SPE Reservoir Engineering, 4, 304-8, 1989.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 19
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 20
O'Dell, C. H., A laboratory investigation of the relationship between groundwater velocity and mass
transfer from a simulated aromatic hydrocarbon spill at the water table interface. Thesis or
dissertation Record Level: onographic United States English: 233 pp, 1988.
Oostrom, M., C Hofstee, and J. H. Dane (1997). Light nonaqueous-phase liquid movement in a variably
saturated sand. Soil Science Society of America Journal, 61(6), 1997.
Oostrom, M., R. J. Lenhard, and A. S. Mayer, Comparison of relative permeability-saturation-pressure
parametric models for infiltration and redistribution of a light nonaqueous-phase liquid in sandy porous
media Multiphase flow and chemical transport. Advances in Water Resources, 21(2), 145-157,
1998.
Ostendorf, D., R. Richards, and F. Beck, LNAPL Retention in sandy soil. Groundwater, 31(2), (March-
April): 285-291, 1993.
Oudot, J., A. Ambles, S. Bourgeois, C. Gatellier, and N. Sebyera, Title Hydrocarbon infiltration and
biodegradation in a landfarming experiment. Source Environmental Pollution: Series A : Ecological
and Biological: 17-40, 1989.
Oudot, J., Selective migration of low and medium molecular weight hydrocarbons in petroleum-
contaminated terrestrial environments. Oil & Chemical Pollution, 6(4), 251-261, 1990.
Ouyang, Y., R. S. Mansell, and R. D. Rhue, Flow of gasoline-in-water microemulsion through water-
saturated soil columns. Ground Water, 399-406, 1995.
Ouyang, Y., R. S. Mansell, and R. D. Rhue, Title Flow of gasoline-in-water microemulsion through water-
saturated soil columns. Groundwater, 399-406,May/June, 1995.
Overcamp, T. J., M. J. Hartz, T. M. Boley, and M. Sharma, Fate of spills of volatile hydrocarbons;
downward transport versus volatilization. United States English Report Record Level:
Monographic: 63 pp, 1994.
Ovsienko, S., S. Zatsepa, and A. Ivchenko, Oil Spreading on the Snow/Ice Surface. International
Symposium On Ice, Trondheim; Norway IAHR 94, 1074-1092, 1994.
Panday, S., and M. Y. Corapcioglu, Theory of phase-separate multicomponent contaminant transport in
frozen soils. Journal of Contaminant Hydrology, 16(3), 235-269, 1994.
Panday, S., Y, S. Wu, P.S. Huyakorn, and E. P. Springer, A three-dimensional multiphase flow model for
assessing NAPL contamination in porous and fractured media; 2, Porous medium simulation examples.
Journal of Contaminant Hydrology, 16(2), 131-156, 1994.
Panday, S., Y. S. Wu, P. S. Huyakorn, S. C. Wade, and Z. A. Saleem, A composite numerical model for
assessing subsurface transport of oily wastes and chemical constituents. Journal of Contaminant
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 21
Hydrology, 25(1-2), 39-62, 1997.
Pantazidou, M. P. D., Migration of nonaqueous liquids in partly saturated granular media (granul
ar
media).
Engine
e r i ng,
Ci v i l ;
Hydrol
o g y .
Berkele
y ,
Univers
ity of
Califor
nia, 52-
0 8 b ,
283pp,
1991
Paradis, S. M. E., A laboratory investigation of surfactant movement in oil saturated particulate media.
McGill University (Canada). Mai, 35-06, 155pp, 1996.
Parker, J., J. Kauarachchi, V. Kremesec, E. Hockman, Modeling free product recovery at hydrocarbon
spill sites. Proceeding Conference on Hydrocarbons in Groundwater, NWWA, 1990.
Parker, J. C., J. L. Zhu, T. G. Johnson, Modeling free product migration and recovery at hydrocarbon spill
sites. Ground Water, 32, 119-28, 1994.
Parker, L., A model for hysterectic constitutive relation governing multiphase flow 1. Saturation-pressure
relations. Water Resources Research, 23(12), 2187-2196, 1987.
Parker, L. K., A parametric model for constitutive properties governing multiphase flow in porous media.
Water Resources Research, 23(4), 618-624, 1987.
Pascal, H., Gravity flow of a non-Newtonian fluid sheet on an inclined plane. International Journal of
Engineering Science, 29(10), 1307-1313, 1991.
Paterson, D. A. and C. J. Apelt, Depth-Averaged Equations for Turbulent Free-Surface Flow. Technical
Report , NTIS Dec 88 (Received by NTIS in Jun 1990), 1988.
Pattle, R., Diffusion from an instantaneous point source with a concentration-dependent coefficient.
Quarterly Journal of Mechanics and Applied Math, XII (part 4), 1959.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 22
Pennell, K. D., L. M. Abriola, and W. J. Weber, Surfactant-enhanced solubilization of residual dodecane
in soil columns. 1. Experimental investigation. Environmental Science & Technology, 27, 2332-40,
1993.
Piatt, J. J., M. L. Brusseau, Rate-limited sorption of hydrophobic organic compounds by soils with well-
characterized organic matter. Environmental Science & Technology, 32(11), 1604-8, 1998.
Plante, A. F., Biodegradation of land applied oily food waste and associated changes in soils structural
properties, University of Guelph (Canada), 35-03,145pp,1996.

Powers, S. E., L. M. Abriola, and W. J. Weber, Jr., A phenomenological model for transient NAPL
dissolution processes. Eos, Transactions, American Geophysical Union, Agu 1992 Fall Meeting,
73(43), Suppl., 155, 1992.
Pruess, K., A Fickian diffusion model for the spreading of liquid plumes infiltrating in heterogeneous media,
1996.
Qin, F., Development and application of NMRI for determining fluid saturation distributions in porous
media (multiphase flow), Texas A&M University. Dai, 56-06b, 173 pp, 1995.
Raisbeck, J. M. and M. F. Mohtadi, The Environmental Impacts of Oil Spills on Land in the Arcatic
Regions. Water, Air, and Soil Pollution, 3(2), 195-208, 1975.
Reible, D., T. Illangasekare, D. Doshi, and M. Malhiet, Infiltration of Immiscible Contaminants in the
unsaturated Zone. Ground Water, 28(5), 685-692, 1990.
Rixey, W. G., The long-term dissolution characteristics of a residually trapped BTX mixture in soil, 1996.
Robinson, W. T., F. J. Biehl, and J. P. Martin, Stabilized Petroleum Waste Interaction with Silty Clay
Subgrade, 1989.
Robson, S. G. and G. J. Saulnier, Jr., Hydrogeochemistry and simulated solute transport, Piceance Basin,
northwestern Colorado. U. S. Geological Survey Professional Paper, 65pp, 1981.
Rohrer, W., H. O. Pfannkuch, Groundwater spreading of hydrocarbon spills with special emphasis on
monitor system design in glacial drift. Proceedings and Abstracts - Institute on Lake Superior
Geology, Annual Meeting, no. 22, 52, 1976.
Ross, P. J., J. Y. Parlange, and R. Haverkamp, Two-dimensional interaction of a wetting front with an
impervious layer: analytical and numerical solutions. Transport in Porous Media, 20(3), 251-263,
1995.
Rostron, B. J., and J. Toth, Numerical simulation of oil/oily-contaminant migration and entrapment in a
lenticular reservoir. Applied Hydrogeology, 1(2), 15-33, 1992.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 23
Roy, J. L., and W. B. McGill, Characterization of disaggregated nonwettable surface soils found at old
crude oil spill sites. Canadian Journal Of Soil Science, 78(2), 331-344, 1989.
Rubin, H., E. Mechrez, Z. Gerstl, Y. Chen, U. Mingelgrin, and B. Yaron, Transport of organic pollutants
in a multiphase system; Toxic organic chemicals in porous media. Source: Ecological Studies, 1989.
Rubin, H. B., H. Daniels, G. Rouve, U. Zoller, Migration of oil and petroleum pollutants in soils and
groundwater. Groundwater Contamination and Control Environmental Science and Pollution
Control Series, 11, 355-375, 1994.
Ryan, P., Y. Cohen, One-Dimensional Subsurface Transport of a Nonaqueous Phase Liquid Containing
Sparingly Water Soluble Organics: A front-tracking model. Water Resources Research, 27(7), 1487-
1500, 1991.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 24
Ryan, P. A., Theoretical studies of mass transport in the unsaturated soil zone (soil), University of
California, Los Angeles, 319, 1990.
Rye, H., A Multi-component Oil Spill Model for Calculation of Evaporation and Dissolution of
Condensate. Environmental modeling seminar. Trondheim; Norway, SINTEF, Environmental
Modelling Working Group, 179-189, 1994.
Sa, L., Physical and mathematical modeling of toluene diffusion and biodegradation in unsaturated sandy
soil, Oregon Graduate Institute Of Science & Technology, 284 pp, 1994.
Sale, T., and D. Applegate, Mobile NAPL recovery: conceptual, field, and mathematical considerations.
Ground Water, 35 418-26, May/June, 1997.
Saleem, Z. A., P. S. Huyakorn, Y. S. Wu, N.S. Park, and S. Panday, Efficient subsurface model for
multiphase flow and multiple species transport. Program with Abstracts - Geological Association of
Canada; Mineralogical Association of Canada; Canadian Geophysical Union, Joint Annual Meeting,,
Geological Association of Canada, Waterloo, ON, Canada, 1994.
Saripalli, K. P., Determination of specific NAPL-water interfacial areas of residual NAPLs in porous
media using the interfacial tracers technique. Journal of Contaminant Hydrology, 30, 375-391,
1998.
Schaefer, C. E., Jr., Diffusive mobility of aromatic hydrocarbons in unsaturated and saturated soils
(benzene, toluene, phenanthrine). Rutgers, State University of New Jersey. New Brunswick. 58-
11b, 195 pp, 1997.
Schaefer Wolfgang, T. R., Simulating transport and removal of xylene during remediation of a sandy
aquifer. Journal of Contaminant Hydrology, 19(3), 205-236, 1995.
Schroth, M. H., J. D. Istok, S. J. Ahearn, J. S. Selker, Geometry and position of light nonaqueous-phase
liquid lenses in water-wetted porous media. Journal of Contaminant Hydrology, 19(4), 269-287,
1995.
Schwille, F., Migration of Organic Fluids Immiscible with water in the unsaturated and saturated zones. :
31-35, 1985.
Seitinger, P., A. Baumgartner, and H. Schindlbauer, Behaviour of oil spills in soil and groundwater. Erdoel
Erdgas Kohle, 110(4), 211-215, 1994.
Sheng, J., Multiphase immiscible flow through porous media (groundwater, contamination), Virginia
Polytechnic Institute And State University,146 pp, 1986.
Shetty, S. and R. L. Cerro, Spreading of a liquid point source over a complex surface. Industrial &
Engineering Chemistry Research, 37, 626-35, 1998.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 25
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 26
Shetty, S. A. and R. L. Cerro, Spreading of liquid point sources over inclined solid surfaces. Industrial
& Engineering Chemistry Research, 34, 4078-86, 1995.
Shields, W. J., S. M. Brown, P. T. Kostecki, and E. J. Calabrese, Applicability of POSSM to petroleum
product spills. Petroleum Contaminated Soils; Volume I, Remediation Techniques,
Environmental Fate, Risk Assessment. Chelsea, MI, United States, Lewis Publ., 87-104, 1989.
Simmons, C. S., J. F. McBride, J. W. Cary, R. J. Lenhard, Organic liquid infiltration into unsaturated
porous media. Subsurface contamination by immiscible fluids. Proc. international conference, Calgary,
Balkema, 1990.
Sleep, B. E. and J. F. Sykes, The influence of infiltrating wetting fronts on transport of volatile organic
compounds in variably saturated porous media. Subsurface contamination by immiscible fluids. Proc.
international conference, Calgary, Balkema, 1990.
Smith, P., A similarity solution for slow viscous flow down and inclined plane. J. Fluid Mech, 58 (part 2),
275-288, 1973.
Smith, R. A., J. R. Slack, T. Wyant, K. J. Lanfear, The oil spill risk analysis model of the U. S. Geological
Survey. U. S. Geological Survey Professional Paper, Reston, VA, United States, U. S. Geological
Survey: p. 1227, 1982.
Soerens, T. S., D. A. Sabatini, and J. H. Harwell, Effects of flow bypassing and nonuniform NAPL
distribution on the mass transfer characteristics of NAPL dissolution. Water Resources Research
34(7), 1657-73, 1998.
Spaulding, M., A. Odula, V. Kolluru, A Hybrid Model to Predict the Entrainment and Subsurface
Transport of Oil. 16th Arctic and Marine Oil Spill Program technical seminar Environmental Protection
Conservation and Protection, Arctic and marine oil spill program technical seminar, environment;
technology development branch.environment Canada; 1993.
Squire, J. J., Effect of unsaturated zone soil moisture content on vapor phase pollutant propagation in
controlled laboratory experiments, University of Nevada, Las Vegas, 127 pp, 1996.
Srivastava, R. K., Measurement of in situ concentration during miscible displacement and its application
to study dispersion and viscous fingering, University of Waterloo (Canada), 307 pp, 1990.
Swanson, A. B., and D. S. Durnford, Laboratory experiments of LNAPL spreading and distribution at a
water table. Eos, Transactions, American Geophysical Union, 77(46) Suppl., 275, 1996.
Tardy, P. Q., M. L., P. Thiez, and E. Seva , Numerical model of hydrocarbon contaminant flow in soil
and aquifers. International Conference on Development and Application of Computer Techniques to
Environmental Studies., Computational Mechanics Publ, Southampton, Engl., 1996.
Terry, S. P., Effects of surface wastewater disposal on an oil field landscape. Environmental
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 27
Geosciences, 3(3), 115-121, 1996.
Tomasko, D., J. P. Butler, and R. J. Charbeneau, Modeling LNAPL transport in the vadose zone.
Groundwater Management, 349-354, 1992.
Unger, A. and P. A. Forsyth, Variable spatial and temporal weighting schemes for use in multi-phase
compositional problems. Advances in Water Resources, 19(1), 1-27, 1996.
Unger, A. J., P. A. Forsyth, and E. A. Sudicky, Influence of alternative dissolution models and subsurface
heterogeneity on DNAPL disappearance times. Journal of Contaminant Hydrology, 30(3-4), 217-
242, 1998.
Van Dijke, M. I. and S. Van der Zee, Analytical description of LNAPL redistribution with entrapment.
International Conference on Computational Methods in Water Resource., Cancun (Mexico),
Computational Mechanics Publications, Southampton, So40 7aa (Uk), 1996.
Van Geel, P. J. and J. F. Sykes, The importance of fluid entrapment, saturation hysteresis and residual
saturations on the distribution of a lighter-than-water non-aqueous phase liquid in a variably saturated
sand medium. Journal of Contaminant Hydrology, 25(3-4), 249-270, 1997.
Veldman, A. E. and M. E. Vogels, Free surface flow calculations on a eularian grid. International
Conference on Finite Element Methods in Flow Problems., Huntsville, AL, 1998.
Vidal, J., Gas/oil gravity drainage: Laboratory study of oil reimbibition and streaming between stacked
matrix blocks. the European Petroleum Conference, Cannes, Fr., Society of Petroleum Engineers
(Spe),P.O. Box 833836, Richardson, TX (USA), 1992.
Vizika, O., Effect of the spreading coefficient on the efficiency of oil recovery with gravity drainage.
Conference Symposium on Enhanced Oil Recovery presented at the 205th National Meeting of the
American Chemical Society, Denver, CO, USA, 1993.
Voinov, O. V., On spreading of a viscous liquid drop over surface due to capillary forces action.
Prikladnaya Matematika i Mekhanika, 59(5), p 767-775, 1995.
Waddill, D. W. and J. C. Parker, A semianalytical model to predict recovery of light, nonaqueousphase
liquids from unconfined aquifers. Ground Water, 35(2), 280-290, 1997.
Waddill, D. W. and J. C. Parker, Title Simulated recovery of light, nonaqueous phase liquid from
unconfined heterogeneous aquifers. Ground Water, 35(6), 938-947, 1997.
Walser, G. S., Vadose zone infiltration, mobilization and retention of non-aqueous phase liquids. Civil
Engineering, Boulder, University of Colorado At Boulder. 56-09b, 255 pp, 1995.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 28
Weaver, J., R. Charbeneau, and B. Lien, A screening model for NAPL transport in the vadose zone using
Green-Ampt and kinematic wave theory. Water Resources Research, 30(1), 93-105, 1994.
Weaver, J. W., Kinematic modeling of multiphase subsurface transport, The University of Texas at Austin.
227 pages, 1988.
Weaver, J. W. and R. J. Charbeneau, Hydrocarbon spill exposure assessment modelling. Proceedings of
Petroleum hydrocarbons and organic chemicals in ground water, prevention, detection, and restoration,
a conference and exposition, 1990.
Weaver, J. W., Application of the Hydrocarbon Spill Screening Model to Field Sites. Specialty conference
-- Non-aqueous phase liquids (NAPLs) in surface environment: assessment and remediation,
Washington; DC, ASCE, Environmental Engineering Division, Nov., 1996.
Weaver, W., J. R. Charbneeau, J. D. Tauxe, B. K. Lien, and J. B. Provost, The hydrocarbon spill
screening model (HSSM) Vol 1: Users Guide. Ada, OK, EPA, Kerr Enviro. Research Lab, office of
research and development, 210, 1994.
Weir, G. J. and W. M. Kissling, The influence of airflow on the vertical infiltration of water into soil. Water
Resources Research, 28(10), 2765-2772, 1992.
Weyer, Ed, Subsurface Contamination by Immiscible fluids (1). Modeling flow in three phase porous media
with nonwetting fluid entrapment. Rotterdam, Balkema, 1992.
Weyer, Ed, Subsurface Contamination by Immiscible Fluids (3). Organic liquid infiltration into unsaturated
porous media. Rotterdam, Balkema, 1992a.
Weyer, Ed, Subsurface Contmination by Immiscible fluids (2). Numerical modeling of
multiphase/multicomponent flow and transport in porous media: an overview. rotterdam, Balkema,
1992b.
White, M. D., M. Oostrom, and R. J. Lenhard, Modeling fluid flow and transport in variably saturated
porous media with the STOMP simulator. 1. Nonvolatile three-phase model description. Advances
in Water Resources, 18(6), 353-364, 1995.
Whittaker, M., S. J. T. Pollard, and A. E. Fallick, Characterization of refractory wastes at heavy oil-
contaminated sites; a review of conventional and novel analytical methods. Environmental
Technology, 16(11), 1009-1033, 1995.
Wiedemeier, T. H., P. R. Guest, R. L. Henry, and C. B. Keith, The use of Bioplume II to support
regulatory negotiations at a fuel spill site near Denver, Colorado. The 1993 meeting on Petroleum
hydrocarbons and organic chemicals in ground water; prevention, detection, and restoration, 1993.
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 29
Winnicky, K. L., On the permeability of frozen silt to organic contaminants, Carleton University (Canada).
34-06, 111 pages, 1995.
Wise, W. R., C. Chang, R. A. Klopp, and P. B. Bedient, Impact of recharge through residual oil upon
sampling of underlying ground water. Ground Water Monit. Rev., 1(2), 93-100, 1991.
Wittmaack, R., CORFLOW: A code for the numerical simulation of free-surface flow. Nuclear
Technology, 119(2), 158-180, 1997.
Wyman, B. C., A risk model for hazardous material transportation spills in calcasieu parish, Louisiana. Dai,
The University of Texas at Dallas. 50-11B, 219pp, 1989.
Xiang, Y., and S. Mishra, Probabilistic multiphase flow modeling using the limit-state method. Ground
Water, 35(5), 820-824, 1997.
Yortsos, Y. C., A. B. Huang, Linear-stability analysis of immiscible displacement: simple basic flow
profiles. SPE Reservoir Engineering, 1, 378-90, 1986.
Yortsos, Y. C., and F. J. Hickernell, Linear stability of immiscible displacement in porous media. SIAM
Journal on Applied Mathematics, 49, 730-48, 1989.
Zaidel, J., and D. Russo, Analytical models of steady state organic species transport in the vadose zone
with kinetically controlled volatilization and dissolution. Water Resources Research, 28(10), 3343-
3356, 1993.
Zaidel, J., and D. Russo, Title Analytical models of steady state organic species transport in the vadose
zone with kinetically controlled volatilization and dissolution. Source Water Resources Research,
3343-3356, 1993.
Zaidel, J., D. Russo, and G. Feldman, Theoretical analysis of the impact of vapor transport on the NAPL
distribution in dry soils. Advances in Water Resources, (June): 145-162, 1996.
Zaidel, J. R., D., Analytical models of steady state organic species transport in the vadose zone with
kinetically controlled volatilization and dissolution. Water Resources Research, 29(10), 3343-3356,
1993.
Zhou, D., and M. Blunt, Effect of spreading coefficient on the distribution of light non-aqueous phase liquid
in the subsurface. Journal of Contaminant Hydrology, 25(1-2), 1-19, 1997.
Zhou, J., Numerical and experimental analysis of diesel and JP-5 transport in unsaturated soils. Los
Angeles, University of California, 234pp, 1994.
Zoltai, S. C., and G. P. Kershaw, Large Volume Oil Spill on Land Surface: The Vozey Oil Field, Russia.
18th Arctic and Marine Oil Spill Program technical seminar, Arctic And Marine Oilspill Program
C:\ MODELS\ OSFM\ R06- 99.157.wpd Hydr oGeoLogi c, I nc. 12/ 24/ 03 A- 30
Technical Seminar 1995;, Environment Canada, 1995.

Вам также может понравиться