
Citation 
 Permanent Link:
 http://ufdc.ufl.edu/AA00004971/00001
Material Information
 Title:
 Numerical modeling of waveinduced currents using a parabolic wave equation
 Creator:
 Winer, Harley Stanford, 1944
 Publication Date:
 1988
 Language:
 English
 Physical Description:
 x, 129 leaves : ill. ; 28 cm.
Subjects
 Subjects / Keywords:
 Amplitude ( jstor )
Beaches ( jstor ) Boundary conditions ( jstor ) Breakwaters ( jstor ) Coastal currents ( jstor ) Momentum ( jstor ) Stress waves ( jstor ) Surface water ( jstor ) Velocity ( jstor ) Waves ( jstor ) Tidal currents  Mathematical models ( lcsh ) Wave equation ( lcsh ) Waves  Mathematical models ( lcsh ) City of Gainesville ( local )
 Genre:
 bibliography ( marcgt )
theses ( marcgt ) nonfiction ( marcgt )
Notes
 Thesis:
 Thesis (Ph. D.)University of Florida, 1988.
 Bibliography:
 Includes bibliographical references.
 General Note:
 Typescript.
 General Note:
 Vita.
 Statement of Responsibility:
 by Harley Stanford Winer.
Record Information
 Source Institution:
 University of Florida
 Holding Location:
 University of Florida
 Rights Management:
 Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for nonprofit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
 Resource Identifier:
 024843713 ( ALEPH )
AFM2500 ( NOTIS ) 20076927 ( OCLC ) AA00004971_00001 ( sobekcm )

Downloads 
This item has the following downloads:

Full Text 
NUMERICAL MODELING OF WAVEINDUCED CURRENTS USING A PARABOLIC
WAVE EQUATION
By
HARLEY STANFORD WINER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988
I BNPlsTY OF FLORIDI IBRARLES
NUMERICAL MODELING OF WAVEINDUCED CURRENTS USING A PARABOLIC
WAVE EQUATION
By
HARLEY STANFORD WINER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988
[UNIVERSITY OF FLORIDA LIBRARIES
ACKNOWLEDGEMENTS
The writing of a dissertation is a long process involving interaction with and support
from many people. The author would like to thank all those who in one way or another
offered encouragement, support, advice and helped in the problem solving process. The
contributions of those who are not mentioned by name are in no way diminished by the lack
of individual acknowledgement.
First, the author wishes to express deep appreciation to the members of his committee:
the chairman, Dr. Hsiang Wang, who enticed him to come to Florida and enroll in the
doctoral program, and the cochairman, Dr. James Kirby, who shared his knowledge of wave
equations and numerical modeling, along with Dr. Robert Dean, Dr. Bent Christensen and
Dr. Joseph Hammack. Each of these men has been both a teacher and a friend. The author
wishes to also thank Dr. Peter Sheng who at one time served on the supervisory committee
but due to a prior commitment was unable to be at the defense and officially be on the
committee.
The author would like to thank the staff of the Coastal and Oceanographic Engineering
Laboratory at the University of Florida and the staff of the department of Coastal and
Oceanographic Engineering who helped and aided the author in innumerable ways. Thanks
are also extended to Ms. Lillean Pieter who helped prepare the figures and Dr. Claudio
Neves, a fellow student, who manipulated the I^TjrX software bo as to produce a document
in thesis mode.
A special acknowledgement is given to the authors parents who encouraged and pushed
him to complete this doctorate. The author would also like to thank his teachers, friends
and fellow students who offered encouragement and interesting discussions.
Lastly the author thanks his soulmate and partner, Esther DeJong, who managed to
be patient with each extension of the final date. Baby cakes, its PAinally Done. I love you.
in
TABLE OF CONTENTS
ACKNOWLEDGEMENTS
LIST OF FIGURES vi
ABSTRACT ix
CHAPTERS
1 INTRODUCTION 1
1.1 Literature 4
1.2 General Description of the Model 8
2 CIRCULATION MODEL 16
2.1 Governing Equations 16
2.2 Radiation Stresses 24
3 WAVE MODEL 31
3.1 Governing Wave Equation 31
3.2 A Parabolic Approximation 36
3.3 Energy Dissipation Due to Wave Breaking 45
4 FINITE DIFFERENCE SCHEMES AND BOUNDARY CONDITIONS 49
4.1 Numerical Solution of the Governing Equations 49
4.2 Finite Differencing of the Parabolic Wave Equation 52
4.3 Boundary Conditions in the Circulation Model 56
4.4 Boundary Conditions in the Wave Model 59
5 RESULTS 63
5.1 Wave Setup 64
5.2 Longshore Currents 68
iv
70
5.3 Waves On a RipChannel . .
5.4 ShorePerpendicular Breakwater 76
5.5 Shore Parallel Emerged Breakwater 88
5.6 Comparison with Experimental Results of Gourlay 92
6 CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER STUDY ... 103
6.1 Improved Wave Formulations 104
6.2 Including Wave Reflection 105
6.3 Extending the Model to Include Sediment Transport 105
APPENDICES
A DERIVATION OF THE DEPTH AVERAGED EQUATIONS 107
B DERIVATION OF THE RADIATION STRESS TERMS 113
C LAG RANG LAN DERIVATION OF THE GOVERNING WAVE EQUATION . 117
D OBTAINING EQUATION (3.79) 120
BIBLIOGRAPHY 124
BIOGRAPHICAL SKETCH 129
v
LIST OF FIGURES
1.1 Computational domain of the model 8
1.2 Flow chart of computer program 11
1.3 Flow chart of circulation model 13
1.4 Flow chart of wave model 14
2.1 Longshore current profiles for different friction formulations for an 11
second, 1.028 meter wave at 17.28 degrees to a 1:25 plane beach 22
2.2 Experimental results for wave setup and setdown. Reprinted from
Bowen, Inman, and Simmons (1968), Journal of Geophysical Research,
vol. 73(8) page 2573, copyright by the American Geophysical Union. . 30
4.1 Definition sketch of grid nomenclature and coordinate system 56
4.2 Startup function according to Eq. 4.43 with Ci==30 and C2=3 60
4.3 Waves passing through a lateral boundary unaffected by the boundary. 61
5.1 Comparison of the numerical solution of Vemulakonda et al. (1985) for
setup with experimental data of Bowen (1968). Source: Vemulakonda
et al. (1985) 65
5.2 Comparison of numerical results from the model with experimental re
sults of Hansen and Svendsen (1979). Wave setup and wave heights for
a 2 second wave on a beach slope of 1:34 and incident wave height of 7
centimeters 67
5.3 Nondimensional longshore velocities vresus the nondimensional dis
tance offshore from the analytical solution of LonguetHiggins. Reprinted
from LonguetHiggins (1970b), Journal of Geophysical Research, vol.
75(33) page 6793, copyright by the American Geophysical Union 68
5.4 Comparison of experimental results of Visser (1982) and numerical result
for longshore currents due to a 1.1 second wave on a slope of 1:20 for
the deep water conditions H0 = .065m and 6a = 20 degrees 69
5.5 Top: Depth contours, in meters, using Eq. 5.7 with the coordinates as
shown. Bottom: Depth contours as used in the present model. Grid
spacing is 5 meters 71
vi
5.6 Velocity vectors from the results of Ebersole and Dalrymple for a wave
angle of 30 degrees. Scale for velocity vectors is 1 inch = 2.87 m/sec.
Source: Ebersole and Dalrymple (1979) 72
5.7 Velocity vectors obtained from the present model using the same input
conditions as for Ebersole and Dalrymple 73
5.8 Mean water surface elevation contours showing that the water surface is
lower above the trough. Values indicated are in millimeters 74
5.9 Depth contours in meters for the case of a bar with a gap 75
5.10 Velocity vectors for the case of waves over a bar with a gap. Dashed
lines indicate depth contours. Grid spacing equal to 10 meters 75
5.11 Depth contours for the case of normally incident waves upon a shoal.
Grid spacing is .05 meters. Contour interval is .02 meters 77
5.12 Velocity vectors for the case of waves breaking upon a shoal 78
5.13 Maximum velocity across the top of a shoal versus incident wave height
for waves over a shoal on a plane beach. Incident wave height less than
9.1 centimeters are nonbreaking waves. Incident wave heights greater
than 9.1 centimeters produced waves breaking over the shoal 79
5.14 Position of the groin in relation to the grid rows and columns 80
5.15 Results for the numerical model of Liu and Mei (1975). Deepwater
wave height equal to 1 meter, deepwater wave angle equal to 45 degrees.
Source: Liu and Mei (1975) 83
5.16 Results from the present numerical model using the same conditions as
were used by Liu and Mei. Offshore boundary condition of constant wa
ter surface level which precludes tangential flow at the offshore boundary. 84
5.17 Results from the present numerical model using the same conditions as
were used by Liu and Mei. Offshore boundary condition of no onoffshore
flow. Grid spacing equal to 20 meters 85
5.18 Experimental setup for the tests conducted by Hales (1980). Source:
Hales (1980) 86
5.19 Experimental measurements of the middepth averaged velocities obtain
by Hales. Measurements at one foot intervals. Source: Hales (1980). . 87
5.20 Ripcurrent velocities adjacent to the down wave side of the jetty. Com
parison of experimental and numerical results 88
5.21 Velocity vectors of the currents obtained numerically for the same con
ditions as for the Hales experiment. Grid spacing equal to one foot. . 89
5.22 Definition sketch of grids for the case of a shore parallel breakwater. . 90
vu
5.23 Wave crests in the lee of a semiinfinite breakwater in a region of constant
depth 91
5.24 Velocity vectors obtained from the model for a 100 meter long breakwater
100 meters offshore for a 1 meter, 6 second wave for increasing angles of
incidence 93
5.25 Stream function contour lines obtained from the Liu model for a 10
second 1 meter wave normally approaching a 700 meter long breakwater
350 meters offshore on a 1 on 50 slope. Source: Liu and Mei (1975). . 94
5.26 Flow lines obtained from the present model for a 10 second 1 meter wave
normally approaching a 700 meter long breakwater 350 meters offshore
on a 1 on 50 slope 95
5.27 Velocity vectors obtained from the present model for a 10 second 1 meter
wave normally approaching a 700 meter long breakwater 350 meters
offshore on a 1 on 50 slope 95
5.28 Setup contour lines obtained from the present model for a 10 second
1 meter wave normally approaching a 700 meter long breakwater 350
meters offshore on a 1 on 50 slope 96
5.29 Physical layout for the experiment of Gourlay. Source: Gourlay (1974),
Proceedings 14*^ Coastal Engineering Conference, copyright by the Amer
ican Society of Civil Engineers, reprinted with permission 97
5.30 Computational domain for the experiment of Gourlay. Grid spacing
equal to .1 meter 98
5.31 Comparison of experimental and numerical results for the experiment of
Gourlay: Wave Setup Contours. Top figure is from Gourlay (1974) Pro
ceedings 14^ Coastal Engineering Conference, copyright by the Amer
ican Society of Civil Engineers, reprinted with permission. The bottom
figure is the computer results 100
5.32 Results from the experiment of Gourlay: Contours of current veloci
ties and streamlines. Source: Gourlay (1974), Proceedings 14^ Coastal
Engineering Conference, copyright by the American Society of Civil En
gineers, reprinted with permission 101
5.33 Results from the numerical model: Streamlines of the flow 101
5.34 Results from the numerical model: Contour lines of the current veloci
ties 102
viii
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
NUMERICAL MODELING OF WAVEINDUCED CURRENTS USING A PARABOLIC
WAVE EQUATION
By
HARLEY STANFORD WINER
August 1988
Chairman: Dr. Hsiang Wang
Major Department: Aerospace Engineering, Mechanics and Engineering Science
Prediction of waves and currents in the nearshore region in the presence of proposed
breakwaters is needed in the design process in order to optimize the placement of the
structures. A numerical model is developed which solves for the wave field, the depth
averaged mean currents and the mean water surface elevation within a given computa
tional domain with the presence of breakwaters. The model iterates between a wave model
and a circulation model. The wave model uses a parabolic approximation to a combined
refractiondiffraction mild slope equation which includes currents. The circulation model
uses an alternating direction implicit method solution to the depth averaged equations of
momentum using the radiation stress terms obtained through solution of the wave model as
the forcing terms. Each of the terms in the depth averaged equations of momentum involve
assumptions that are discussed in the report. The model is constructed in modular form so
that improvements can easily be incorporated into the model.
The model was employed for several test cases such as rip channels, bar gaps, and
both shore perpendicular and shore parallel breakwaters. Good agreement was obtained
when the model was used to simulate a laboratory experiment of Michael R. Gourlay
(Wave setup and wave generated currents in the lee of a breakwater or headland,
Proc. 14*^ Coastal Eng. Conf., Copenhagen, 1974, pp. 19761995).
IX
Suggestions are made for improving the model and for including sediment transport in
the model.
CHAPTER 1
INTRODUCTION
As waves propagate they are transformed by variations in currents and depths, along
with interaction with obstructions. The major processes of wave transformation include
shoaling, reflection, refraction, diffraction and dissipation. Shoaling is the change of wave
energy, propagation speed and wave height due to changes in the water depth and/or current
field. Refraction is the change in the direction of wave propagation due to the spatial
gradient of the bottom contours and/or current field. Diffraction is the phenomenon of
wave energy being transmitted laterally along a wave crest. Dissipation is the process of
energy loss due to a multitude of possible causes such as bottom percolation, damping
due to viscous effects, turbulence, and wave breaking. As waves approach a shoreline and
break, momentum exchanges can produce currents and mean water level variations. These
currents can be in the form of longshore currents, rip currents, or circulation cells in the lee
of obstructions. They in turn modulate the incident wave field. The process can be quite
complicated depending upon the complexity of the bathymetry, the ambient current field,
and the presence of obstructions.
Coastal engineers and planners have many reasons for wanting to obtain accurate pre
dictions of waves and currents and water elevations in the coastal rone. For instance, pre
dictions of extreme conditions are needed for site planning and erosion mitigation studies.
Harbor authorities need wave predictions for structure design and for operational decisions.
This dissertation develops a numerical model for obtaining predictions of waves, currents
and mean surface elevations in a given domain with the offshore wave conditions and the
bathymetry as the input. The model uses a linear wave equation and depth integrated
equations of momentum and mass conservation. Comparison with experimental results
1
2
indicates that the model produces useful and valid predictions of the waves and currents in
the nearshore region in the presence of obstructions such as groins and breakwaters. The
predictions obtained through use of the computer model herein reported are not intended
to be the sole resource for the engineer. A good engineer or planner will seek answers from
multiple sources and cross check them, recognizing the limitations of each method used to
obtain the answer.
Predictions of waves and currents can be obtained for different cases analytically,
through experiments or by numerical solutions. Analytical solutions are obtained only
for simple cases involving idealized beaches using simplified assumptions which are not al
ways satisfied in nature. LonguetHiggins and Stewart (1963) give an analytic solution for
wave setup for steady state conditions on a plane beach with straight and parallel con
tours, assuming linear waves approaching the beach normally, and spilling breakers where
the wave height shoreward of the breaker line is a constant times the total depth of the
water column. LonguetHiggins (1970a) gives a cross shore profile of the longshore current
for a plane beach with straight and parallel contours, for linear waves approaching the beach
at an angle, spilling breakers, and linearized bottom friction, while neglecting the influence
wave setup has upon the total depth of the water column. In a companion paper, Longuet
Higgins (1970b) extended the solution to include lateral diffusion of the current momentum.
There are other analytical solutions for waves and currents but all are limited to very simple
cases.
The literature is replete with laboratory investigations of waves and currents. Physical
models have advantages and disadvantages. The advantages of experimentation are that
results can be obtained for complex situations where there is uncertainty about the exact
form of the governing equations or about the precise form of any of the component terms
of the equation. Any reasonable amount of detail can be included. Visualization of the
physical processes provides an opportunity to identify problems which otherwise might not
be evident.
3
However there are limitations to the use of laboratory experiments. There is the problem
of scale effects which means that all scaling requirements cannot be satisfied simultaneously;
thus, all features may not be interpreted properly to the prototype. For example, the
time scale dictated by Froude modelling laws in which gravitational forces dominate is
different from that obtained from Reynolds modelling laws in which viscous forces are most
important. Another disadvantage of physical models is the cost of labor to construct and
run large scale experiments. Also, because of the cost of construction, it is not always very
easy to change details in the model. Thus, it is not always possible to isolate one parameter
for study. For example, an investigation of the influence of the size of the gap between
segmented breakwaters might require many reconstructions of the rubble mound structure.
The large size of a model means that it is not easily stored once the initial experiment
is completed. The options are to either dismantle the model upon completion of testing
or to incur some maintenance costs. The significance of this is that the model cannot
always be easily retested if, for example, additional input information becomes available or
man or nature makes a substantial alteration of the prototype after the model has been
dismantled. Also the results of a model for one prototype are not always transferable to
other prototype situations. Often the available space for the physical model will require
the use of a distorted model in which the vertical scale is different from the horizontal
scale. This presents the additional problem that for a distorted model, when modelling
a situation where both refraction and diffraction are equally important, the laws of scale
modelling dictate different time scales for refraction and diffraction. For refraction the
time ratio is Tr \[Lv, where the subscript r is for the ratio of the value in the model
to the value in the prototype and T and Lv are time and vertical length respectively. For
diffraction Tr = ~n1" in shallow water and Tr = \/Ljjt in deep water, where Ljjr represents
y/Lvr
the horizontal length scale.
Numerical models do not have many of the disadvantages found with physical models.
Any mechanism for which the governing equations are known can be modeled numerically.
4
There are no limits in terms of scale effects. Numerical models, once developed, can be
adapted easily to many different prototype situations. There is no problem in the storage or
duplication of the model. Today computer facilities are far more available than laboratory
facilities. Computer models easily allow for sensitivity testing of the importance of an
individual variable and any specific feature of the model can be monitored.
Numerical models also have their disadvantages. In fact some of their advantages
are also their disadvantages. For example, being able to monitor any feature also means
that the computer will deliver only the information that is explicitly specified. Thus an
investigation of one aspect will not indicate problems with other aspects. Though any
mechanism for which the governing equation is known can be modeled, the model is only as
good as the assumptions involved in the equations. For example, the forces exerted on the
water column by the complex interaction between the fluid motion and the bottom are not
totally understood and are often greatly simplified. Another example is the use of depth
integrated twodimensional models of fluid flows when the reality is threedimensional. Both
of the above are incorporated in the present study. Knowing the correct equation does not
always make for successful computer modelling since the complexity of the equations may
make for difficulties (for example solving for the complete threedimensional flow field using
the NavierStokes equations). Computer time and cost for large programs can often be
significant.
1.1 Literature
This section is a brief history and literature review of the attempts of the coastal
engineering profession to obtain predictions of waves and currents in the nearshore region
given the offshore wave conditions.
The extension of Snells law which governs optical wave refraction to the analogous
problem of water waves over straight and parallel contours has been known for some time.
Johnson, OBrien, and Isaacs (1948) reported graphical methods for constructing wave
refraction diagrams. Two methods were reported, one producing wave fronts and the second
5
producing wave rays. Pierson (1951) showed that the method using wave fronts introduced
more error than the wave ray method. These methods proceeded upon the assumption that
the contours are locally straight. The wave rays obtained were important in that the wave
height was determined to vary inversely as the square root of the spacing between the wave
rays. Using these ray construction techniques OBrien (1950) was able to demonstrate that
the April 1930 destruction of the Long Beach Harbor breakwater on a seemingly calm day
most probably resulted from the focusing of long wave energy from a South Pacific storm.
Munk and Arthur (1952) showed a method to obtain the wave height directly from the
curvature of the ray and the rate of change of the depth contours. The wave ray tracing
techniques did not include the effects of currents and had the disadvantage (from a numerical
modeling point of view) that wave heights and directions were obtained at positions along
a ray and not at prespecified grid positions. Noda et al. (1974) solved this problem by
devising a numerical scheme that solved a wave energy equation and a statement for the
irrotationality of the wave number to obtain wave height and direction at grid points. This
scheme also included the effects of current refraction but neglected diffraction.
Numerical models of wave induced currents were developed by Noda et al. (1974) in
which time dependent and advective acceleration terms were ignored. Birkemeier and Dai
ry mple (1976) wrote a timemarching explicit model which neglected lateral mixing and
advective acceleration. Ebersole and Dalrymple (1979) added lateral mixing and the non
linear advective acceleration terms. Vemulakonda (1984) wrote an implicit model which
included lateral mixing and advective acceleration terms. All of the above used the numer
ical scheme of Noda et al. (1974) to obtain wave heights and wave angles. Each of these
models neglected diffraction.
Diffraction of water waves represents some very difficult mathematical problems. His
torically most solutions were based upon the assumption of a flat bottom, since to quote
Newman (1972),
The restriction to constant depth h is important.... Problems involving wave
propagation in a fluid of variable depth are of great interest, and the resulting
6
diffraction effects are among the most important of all ... but these problems
are less tractable than those involving a fluid of constant depth, (page 2)
Penney and Price (1944) showed that Sommerfelds solution of the optical diffraction prob
lem as described by Bateman (1944) is also a solution of the constant depth water wave
diffraction problem. Penney and Price (1952) solved the constant depth diffraction problem
for breakwaters assuming a semiinfinitely long, infinitesimally thin barrier. The solution
is in terms of Fresnel integrals which need to be evaluated numerically. The Shore Protec
tion Manual (SPM,1977) produced many graphical solutions based upon Penney and Price
(1952) that give the wave height and direction in the lee of breakwaters.
The problem of combined refraction and diffraction according to the SPM could be
handled by assuming a constant depth for a few wavelengths so as to obtain the diffraction
effects and then continue with the refraction effects. Perlin and Dean (1983) used essentially
the same procedure in a numerical model which calculated the longshore thrust in terms of
the breaking wave height and angle.
Liu and Mei (1975) solved for the wave field around shore parallel and shore perpendicu
lar breakwaters by using a curvilinear coordinate system where one coordinate corresponded
to the wave ray and the other to a curve of constant phase. Their solution was for an am
plitude modulation factor D(Â£,f) which accounts for diffraction (page 72), where Â£ and f
are the two coordinates. An equation for D(Â£,f) was obtained and a boundary value prob
lem was established by matching conditions in the incident zone and the shadow zones. For
constant depth the solution for Â£>(Â£,f) is in terms of Fresnel integrals. Upon obtaining the
wave field the radiations stresses were obtained and used to force a circulation model that
solved for the stream function and mean water surface level while ignoring the nonlinear
advective acceleration terms and the lateral diffusion of current momentum. Although Liu
and Mei produced results for the wave field, currents, and setup resulting from the diffrac
tion of waves by breakwaters, their study is not a wavecurrent interaction model since
there are no current terms in the wave equations and no feedback of current terms from the
circulation model.
7
Berkhoff (1972) derived an equation governing combined refraction and diffraction with
the assumption of a mild slope. Radder (1979) applied the splitting method of Tappert
(1977) to obtain a parabolic equation governing the propagation of the forward propagat
ing component of the wave field. Booij (1981) extended the work of Berkhoff to include
the effects of currents and also obtained a parabolic approximation. Kirby (1984) made
corrections to the equation of Booij.
Parabolic wave equations have recently been used in waveinduced circulation models.
Liu and Mei (1975) solved for the wave field by matching conditions between the shadow
zones and the illuminated zone in the lee of a structure. The wave information was then
used to force a circulation model which formulated the governing depth integrated time
averaged momentum equations in terms of a stream function. However this model did not
have current feedback in the wave equation. Yan (1987) and the present study both include
a current feedback for the parabolic wave equation and solve directly for the mean currents.
Prior to the use of parabolic wave equations to obtain the wave forcing in circulation
models, refraction wave models using a numerical scheme derived by Noda et al. (1974)
were used. Birkemeier and Dairymple (1976) produced an explicit numerical solution of
the depth averaged equations of momentum and continuity using the wave model of Noda
et al. (1974). Their model did not include the nonlinear advective acceleration terms
or any mixing. Ebersole and Dalrymple (1980) included the advective acceleration terms
and lateral mixing in an explicit model which also used the refraction scheme of Noda
et al. Vemulakonda (1984) Produced an implicit model which also was based upon the
wave refraction scheme of Noda et al. Yan (1987) used a parabolic approximation to the
mild slope equation which accounts for combined refraction and diffraction. The present
model Improves upon Yans model in numerical efficiency, by introducing structures and
most importantly by obtaining the radiation stress terms directly from the complex wave
amplitude.
8
Figure 1.1: Computational domain of the model.
1.2 General Description of the Model
This section is a description of the numerical model that has been constructed to calcu
late the wave and current field within a given domain. This domain usually is bordered by
a shoreline along one boundary and an offshore region along the opposite boundary. The
other boundaries are sufficiently far from the region of interest (i.e. a structure or shoal,
etc.) so as to not affect the region of interest. This will be discussed further in the section
on boundary conditions. The computational domain for the purposes of this investigation
consists of a rectilinear coordinate system. Figure 1.1 shows the computational domain.
The model consists of two parts, a circulation model and a wave model. The circulation
model computes the mean water surface level and the depth averaged mean currents using
the radiation stresses from the waves as the forcing terms and includes bottom friction,
advective acceleration terms and the lateral diffusion of current momentum. The wave
model determines the radiation stress terms by first solving for the complex amplitude as a
9
function of the total depth and the depth averaged mean currents, and then computing the
radiation stresses from the complex amplitudes. The model starts with initial conditions
of zero mean currents and zero mean water surface elevation. The offshore incident wave
height is initially zero and then is smoothly increased up to its given value using a hyperbolic
tangent curve. This is done so as to minimize any transients due to a shock startup.
The model iterates between the wave model and the circulation model until steady state
convergence is obtained. Each of the two parts of the model will be described separately.
The circulation model uses an alternating direction implicit (ADI) method of solution.
Three equations are solved for three unknowns. The unknowns are fj the mean water surface
level, and U and V, the two components of the depth averaged mean current. The equations
are the continuity equation and the x and ydirection momentum equations. Each iteration
of the circulation model solves for the three unknowns on the next time level. In other words
each iteration advances the values of fj, U and V by a time increment, At. For the purposes
of this investigation the model solves for steady state conditions using the time marching
character of the circulation model. The model can be used to solve for time dependent
wave fields, but this would require the substitution of a time dependent wave equation in
the wave model. The ADI method first solves, grid row by grid row, for U at the next time
level and an intermediate value of fj, by solving the x direction equation of motion and the
continuity equation using an implicit numerical scheme. This solution is in terms of the
known values of 17,V,and fj at the present time level. Once all the grid rows are solved in
the x direction, the y equation of motion and the continuity equation are solved, again by
an implicit scheme, grid column by grid column, for V and fj at the next time level in terms
of U at the new time level, fj at the intermediate level and V at the present level. For each
of the implicit schemes in the two directions, the governing equations yield a tridiagonal
matrix. The tridiagonal matrix and the method of solution will be discussed later.
The wave model also yields a tridiagonal matrix equation. The governing equation is
a parabolic equation for the complex amplitude A. The complex amplitude contains phase
10
information meaning among other things that y variations of the wave field are contained
within the complex amplitude. The term parabolic equation indicates that there are
first order derivatives in the x direction and second order derivatives in the y direction.
The method of solution is to solve for the amplitude on the 1+1 grid row in terms of the
known amplitude values on the I grid row. The solution scheme starts on the first grid row
using the prescribed offshore incident wave and then advances grid row by grid row to the
last grid row. To advance to the next grid row the amplitude values on the 1+1 grid row
are implicitly formulated using a CrankNicolson scheme. This will be further discussed in
section 4.2 which describes the finitedifferencing scheme for the parabolic wave equation.
Once the complex amplitudes are known the radiation stress terms are computed in terms
of the absolute value of the amplitude and the gradients of the amplitude. This is described
in section 2.2.
The circulation model and the wave model are the heart of the program; however there
is much more. In all there are some 27 subroutines in the program. These subroutines cover
everything from input and output to convergence checks and flooding of the shoreline. The
main program is compartmentalized into many subroutines so as to facilitate any changes
or future upgrading. For example, the bottom shear stress, a term of importance in the
equations of motion, currently uses a crude approximation. If a better approximation is
developed that is also computationally efficient, it will be an easy task to replace the present
subroutine that computes the bottom shear with the newer subroutine. There are many
places where the model can be improved with better assumptions and better approxima
tions. These will be discussed in detail in the final chapter outlining recommendations for
further study. The modular construction of the model also allows for ease in substituting
one wave model for another.
Figure 1.2 shows a flow chart of the computer program. The subroutine INPUT is called
to initialize the values of the unknowns and to input the water depth, the offshore wave
height and direction and the position of any structures. The three subroutines WAVE,
11
Figure 1.2: Flow chart of computer program.
12
SHEAR, and LATMDC compute the radiation stresses, the bottom shear stress, and the
lateral mixing coefficients. These terms are the forcing terms in the equations of motion
and are used in the subroutine CIRC, which is the circulation model, to solve for fj, U and V.
There is a decision tree which directs the program to the three subroutines WAVE, SHEAR,
and LATMIX at selected intervals. This is purely arbitrary and is done in the interest of
computer speed. It is assumed that the forcing terms do not change at a great rate and can
thus be updated periodically. This updating is more frequent during the startup time while
the offshore incident wave height is being increased. The CHECK subroutine is to check the
model for convergence. Since steady state conditions are assumed, this subroutine indicates
convergence when the absolute value of the percentage difference between the updated values
of U,V, and fj and the previous values is less than an arbitrarily small specified value. The
UPDATE subroutine simply updates the values of the unknowns for the next iteration. The
FLOOD subroutine allows for the addition or elimination of grid rows at the shoreline to
allow for beach flooding due to setup.
Figure 1.3 is a flow chart of the circulation model. The subroutines XCOEF and
YCOEF determine the coefficients of the tridiagonal matrix equation for the general cases
in the x and y directions. The subroutine TRIDA solves the tridiagonal matrix equation
using the doubleelimination scheme described by Carnahan, Luther, and Wilkes (1969).
Subroutines XCOJ1 and XCOJN are adaptations of XCOEF to determine the coefficients
of the tridiagonal matrix for the special case along the lateral boundary where J=1 or
J=N. Likewise YCOIM is an adaptation of YCOEF for the I=M grid row which borders
the shoreline. These adaptations are necessary because boundary conditions necessitate
different numerical formulation of some of the derivative terms. Similarly with the presence
of an emerged impermeable barrier bordering the shore side of the JJJ grid row, special
subroutines YCJJJ and YCJJJ1 are used on the grid rows I=JJJ and I=JJJ+1 so as to
properly include the boundary conditions for these grid rows, and for shore perpendicular
groins subroutines XCJGR and XCJGR1 incorporate the boundary conditions for a groin
13
Figure 1.3: Flow chart of circulation model.
14
Figure 1.4: Flow chart of wave model.
located between the J = JGR and the J = JGR+1 grid columns.
Figure 1.4 is a flow chart of the wave model. Subroutine WAVNUM determines the wave
number k, the group velocity Ct and the intrinsic frequency a at each of the grid centers.
These terms are used in the governing parabolic equation. Subroutine AMPCO determines
the coefficients of the tridiagonal matrix and CTRIDA solves the matrix equation for the
complex amplitude on the next grid row. CTRIDA is essentially the same as TRIDA except
that it has been modified to handle complex numbers. The subroutine WW is called by the
AMPCO subroutine in order to determine the dissipation coefficient following the work of
Dally (1980,1987).
Chapter 2 describes the circulation model in greater detail. Individual sections will
cover the governing equations and the formulation of each of the component terms. Chap
ter 3 describes the wave model, the derivation of the linear wave equation, the parabolic
15
approximation, the manner of including wave energy dissipation due to wave breaking and
the method of obtaining the radiation stress terms directly from the complex amplitude.
The fourth chapter gives details of the numerical methods used to obtain solution of the
governing equations and the boundary conditions employed. The fifth chapter presents the
results of applications of the model and comparisons with other studies both numerical and
experimental, as well as some analytical solutions to some idealized cases. The final chapter
summarizes the study and makes recommendations for future research.
CHAPTER 2
CIRCULATION MODEL
2.1 Governing Equations
The governing equations for the circulation model are the depth integrated time aver
aged horizontal equations of momentum
dU TTdU dU dfj 1 1
aT + uli + + 9al+ jDr ~ jDr
+
J_ (dstx asxv\ 1 drl =
pD \ dx dy ) p dy
dV TJdV SV dfj 1 1
aT + 0 + v 87+3 Jy + Td'1' Jdt
dx
+
_i_ = 0
pD \ dx dy ) p dx
(2.1)
(2.2)
and the continuity equation
(2.3)
where
U = x component of the mean current
V = y component of the mean current
fj = mean water surface elevation
p = water density
D = the total water depth = h0 + fj
h0 = local stillwater depth
rj = the lateral shear stress due to turbulent mixing
Tix = x component of the bottom shear stress
16
17
rky = y component of the bottom shear stress
t,x = x component of the surface shear stress
t,v = y component of the surface shear stress
and *^xy ^nd Svv are the radiation stress components which arise from the excess
momentum flux due to waves. It should be noted that U and V include the wave induced
mass transport.
These equations are obtained by integrating the local x and y momentum equations and
the continuity equation over the depth of the water column and then time averaging the
results. This is demonstrated in Appendix A, which follows closely the work of Ebersole and
Dalrymple (1979). For greater details of the derivation the reader is referred to Ebersole
and Dalrymple (1979) or Phillips (1969).
These are the same equations that have been used by other modelers such as Birke
meier and Dalrymple (1976), Ebersole and Dalrymple (1979), Vemulakonda (1984), and
Yan (1987). Birkemeier and Dalrymple used an explicit numerical scheme and neglected
the advective acceleration terms and the lateral mixing. Ebersole and Dalrymple added the
advective acceleration and lateral mixing terms but still used an explicit numerical scheme.
Vemulakonda introduced an implicit numerical formulation for these equations. Each of
these investigators used a refraction wave model based upon the work of Noda et al. (1974)
which solves for the wave height and wave angle at given grid points. Yan (1987) used a
combined refractiondiffraction wave equation to determine the complex wave amplitude
(magnitude and phase) and then solved for the wave angle in terms of the slope of the wa
ter surface. Each of these studies obtained the radiation stress components in terms of the
wave height and wave angle. The present study differs from the previous ones in that the
radiation stress terms are obtained directly from the complex amplitude and its gradients.
This will be discussed in section 2.2.
The use of Eqs. (2.12.3) involves many assumptions. The primary assumption is
that the flow field can be represented in two dimensions using depth and time averaged
18
equations. This assumption implies that vertical fluctuations of the flow which are zero in
the time average can be ignored. Many details of the flow may be missed by this assumption
of purely horizontal flow. Of course reducing the problem to a twodimensional flow has
the distinct advantage of reducing an essentially intractable problem to a simpler tractable
form.
Some of the terms in Eqs. (2.12.3) also involve certain assumptions. These terms
are the bottom friction, the surface wind stress and the lateral mixing. These are all
phenomenon that are not completely understood so that assumptions are made concerning
the complex forces that are represented. These assumptions usually involve the use of
empirical coefficients. In this study the wind forces are ignored so the assumptions involved
with the surface stress will not be discussed in detail here. It would be a relatively simple
task, though, to add a wind surface stress to the model, using a quadratic shear stress as
formulated by Birkemeier and Dairymple (1976) and also used by Ebersole and Dedrymple
(1979) following the work of Van Dorn (1953) in determining the value of the empirical
coefficient.
The interaction of an oscillatory fluid flow with a real bottom is very complex. Many
physical phenomena may be at work. With sand bottoms there can be percolation into the
sand and frictional effects resulting in wave energy dissipation. With mud bottoms where
the mud is essentially a dense fluid layer there can be damping and internal waves along
the mudwater interface. Wave energy losses due to bottom effects are not included in the
model but can easily be incorporated into the wave dissipation subroutine. Resuspension
and settling of particles also constitutes a complex phenomenon with density changes and
momentum transfers. All of these effects are glossed over in the model by assuming a rigid
impermeable bottom. As with all of the models discussed above the bottom friction term is
represented in the quadratic form. Noda et al. (1974), Birkemeier and Dalrymple (1976) and
Vemulakonda (1984) all used the weak current formulation as proposed by LonguetHiggins
(1970a) in which
19
Tbz = pF uorl U
(2.4)
Tiv = pF uori V
(2.5)
where F is a drag coefficient (of the order of .01) and uorj, is the time average over a wave
period of the absolute value of the wave orbital velocity at the bottom. From linear wave
theory it is found that
O IT
Uort = T sinh kh (2,6)
where H is the wave height and T is the wave period, k is the wavenumber, and h = D is
the total water depth. Ebersole and Dalrymple used a more complete quadratic formulation
Tb = pF *t
(2.7)
where the overbar represents the time average over a wave period and ut is comprised of both
the mean current and the wave orbital velocities. The total velocity vector tiÂ¡ is expressed
as
ui = (U + ticos#) + (V + sin0),7 (2.8)
so that its magnitude is given by
u* = Vu2 + V2 + 2 + 2U cos 9 + 2 V sin 6 (2.9)
The wave orbital velocity is expressed as
um cos at
(2.10)
where um is the maximum wave orbital velocity at the bottom which is found to be
xH
T sinh kh
The x and y components of the time averaged bottom friction are thus
rhx = ^ Jq {U+ mcos0coeat)\ut\d{at)
Hv = (p/ [V + sin icos at) tT{d(ot)
X Jo
(2.11)
(2.12)
(2.13)
where tÂ¡ is written as
20
tTj = \Ju2 + V2 + titeos2 at + 2Uum cos at cos 9 + 2Vtim cos at sin 0 (214)
The integrals in Eqs. (2.12) and (2.13) are then evaluated by a 16 point summation using
Simpsons Rule.
In the present study several different formulations are used and compared. The first is
the formulation given by Longuet Higgins (1970a) and used by several previous investiga
tors. The second is a modification of the LonguetHiggins formulation with partial inclusion
of current strength in the friction formula.
T* = pF \uorb\ U + pF\U\U
(2.15)
rbv = pF\uorb\V + pF\U\V
(2.16)
It is thought that this formulation has some value since in the lee of structures where there
is relatively weak wave activity, there may still be relatively strong currents generated by
the gradients of the setup outside of the sheltered region. The third formulation is similar
to that of Ebersole and Dalrymple in that the full quadratic expression is used. However the
wave orbital velocities at the bottom are expressed as gradients of the complex amplitude
and the resulting integral is determined using an eight point Gaussian integration. It is
found that an eight point Gaussian integration of a cosine squared integration over an
interval of 2x is accurate to within 1 percent.
The quadratic formulation of the bottom friction is used
Tb = pF uitT( (2.17)
where tT* is composed of both the mean current and the wave orbital velocities. The wave
orbital velocities are expressed as the real part of the gradient of the complex velocity
potential $ which is expressed as
. A(,y) cosh k(h0 + z) ,Mt
9 a cosh kh
a
(2.18)
21
where A is the complex amplitude, u/ is the absolute frequency, a is the intrinsic or relative
frequency following the current, and h0t h and k are as previously explained. Equations
governing $ and a parabolic approximation solution for A are presented in the next chapter.
The total velocity vector tT< is expressed as
ut =
= ^ *(Â§!)] ?+[v+*(!;)]
where SR indicates the real part, so that its magnitude is given by
Kl f'+ + (I) + 2V* (I?)+ [ (Â£)]+ [* (Â£)]
The x and y components of the time averaged bottom friction are thus
 'r?Â£[u+*{Â£)\
(2.19)
(2.20)
(2.21)
(2.22)
As in Ebersole and Dalrymple, the assumption is made that only the gradients of A are
retained and the gradients of the depth and spatially varying quantities such as k and a
are ignored. This is essentially the assumption of a flat bottom and very slowly varying
currents.
Figure 2.1 plots the magnitude of the longshore current as a function of the distance
offshore for the various friction formulations discussed above. These velocity profiles were
obtained numerically for an eleven second wave with a deep water wave height of 1.028
meters (1 meter at offshore grid of the computational domain, using a AX of 10 meters
and 30 grid rows) and a deep water wave angle of 17.28 degrees (10 degrees at the model
boundary) approaching a 1:25 plane beach. A friction factor, F, of .01 was used for each case.
Using a friction formulation that ignores the contribution of the waves results in a very weak
friction and thus a large longshore current. Using the Longuet Higgins formulation gives a
friction slightly less than the more proper time averaged formulation, while a combination of
the LonguetHiggins formulation and the strong current model (no wave contribution) yields
a slightly stronger friction. For practical numerical considerations the latter formulation is
VELOCITY (M/SEC)
22
in
Figure 2.1: Longshore current profiles for different friction formulations for an 11 second,
1.028 meter wave at 17.28 degrees to a 1:25 plane beach.
most often used since it requires less computation time and more importantly the Gaussian
integration involving the reed part of the gradient of complex amplitudes at times produces
stability problems.
The lateral mixing term in Eqs. (2.1) and (2.2) also involve several simplifying assump
tions. The first assumption is found in the derivation of the two depth integrated equations
of motion where it is assumed that the turbulent shear stress term pu'v1 is independent
of depth. The other assumption is that the process of momentum transfer due to turbulent
fluctuations can be represented by a product of a mixing length parameter and derivatives
of the mean current. This is what is done in the present study and also in previous studies.
Ebersole and Dalrymple (1979), Vemulakonda (1984), and Yan (1987) all used the formu
lation given by LonguetHiggins (1970b) in which he developed an analytical expression for
the cross shore distribution of the longshore current velocities using a linearized friction
formulation as described above, linear waves at an angle to the normal of a plane beach,
23
and a spilling breaker assumption. LonguetHiggins wrote
/ dU dV\ ..
n = p'23)
in which ex and ey Eire mixing length coefficients.
LonguetHiggins reasoned that the mixing length parameter ex must tend to zero as
the shoreline was approached. He assumed that ex is proportional to the distance from the
shoreline, X, multiplied by a velocity scale which he chose as the shallow water celerity
y/gh. Thus ex is written as
ex = NX \/gh (224)
where N is a dimensionless constant for which LonguetHiggins chose the limits as
0 < N < 0.016
(2.25)
Obviously, for physical reasons there needs to be some limit on the scale of these eddies.
Ebersole and Dalrymple arbitrarily chose the value of ex near the breaker line to be the
maximum value and for ex to be constant offshore from the breaker line. The same limit
upon ex is used in this study. The coefficient ev is assumed to be a constant and it is
arbitrarily assigned the maximum value of ex.
In the present study the further assumption is made in computing that the cross
derivatives are negligible (a common assumption in fluid mechanics). Therefore, the lateral
mixing terms become
(Â£)
(2.26)
/ av\
\Â£xdx)
(2.27)
in the x direction and
in the y direction. This means that the lateral mixing term is represented by a purely lateral
diffusion of momentum. That the lateral mixing term is a lateral diffusion term has some
significance in that the diffusion term is expressed explicitly and thus there results a time
step limitation to an implicit model. This will be discussed in section 4.1 dealing with the
finite differencing of the equations governing the circulation model.
24
2.2 Radiation Stresses
In a series of papers in the early 1960s LonguetHiggins and Stewart (1962, 1963, 1964)
introduced and developed the mathematics of the radiation stresses. These stresses which
can be conceptually thought of as the flux of excess momentum can be obtained through
integration of the momentum equations over the water column as is shown in Appendix A.
Using linear wave theory the radiation stress components are found to be
Szz = E n(cos2 6 + 1) ^
At
Svv = E n(sin2 0 + 1) ^
At m
Szv = En cos 0 sin 6
(2.28)
(2.29)
(2.30)
where E is the wave energy equal to pgH2, n is the ratio of the group velocity to the wave
celerity and 6 is the angle the wave rays make with the x axis.
If the wave height and wave angle are known at each grid point, the radiation stresses
and their gradients are readily determined for use in the governing equations. To determine
the wave height and angle at each grid point Birkemeier and Dalrymple (1976), Ebersole
and Dalrymple (1979) and Vemulakonda (1984) each used the following scheme which is
based upon the irrotationality of the wave number and an equation for the conservation of
wave energy. The irrotationality of the wave number determines changes in the wave angle
due to refraction while the energy equation determines changes in the wave height due to
shoaling and refraction. This scheme was first introduced by Noda et al. (1974).
The irrotationality of the wave number
V x = 0 (2.31)
where
k = \k\ cos Oi + fc sin Oj
(2.32)
is expanded in Cartesian coordinates to obtain
d0 .36
cos 6I sin 6 + sin 0
ox ay
fc dx
cos 6
1*1 dy
(2.33)
25
For waves on a current the dispersion relation is
oj = a + k
= yjg\k\ tanh \k\h + Jk cos 0 + Jb sin 0
(2.34)
Equation (2.34) is differentiated to eliminate and from Eq. (2.33). Using a
forward difference derivative in x and a backward difference derivative in y, 0, is obtained
as a function of cos 0, sin 0, U,V,k,h and their derivatives as well as 0 at the adjoining grids.
The sin 0 and cos 0 terms are expressed in terms of the surrounding four grids using a 2nd
order Taylor expansion. The solution is an iterative process alternating solution of the 0
equation with solution of the dispersion relationship (2.34) for k = it which is accomplished
by a NewtonRaphson method until acceptable convergence is obtained for k and 0.
Next an energy equation is solved for the wave height. A conservation of energy equation
is given by Phillips (1969 ed. Eq. 3.6.21) as
^+_{m + C(.)} + s.^ = 0 (2.35)
Assuming steady state conditions and expanding in Cartesian coordinates gives
(V + C,cos *)1 + (V + C, sin ) i a~ + A(U + C, cos S)
dU
dU
dV dV
+ + C'isin 0) + axx + fyx + fxv9vy
dy
dy
dx vv dy
0 (2.36)
where
xz ~E^ZX
1
VV Â£ Syv
ZV
*yz
n(cos2 0 + 1) i
2
n(sin2 0 + 1) 
2
n cos 0 sin 0
Defining
+ CB cos 0) + (V + C. sin
dy
(2.37)
where
dU dU dV dV
zxd'{'Tvx~d + Txv~d(Tvv~d.
(2.38)
a =
26
and noting that U,V,Cly and a are known quantities; taking forward difference derivatives
in x and backward difference derivatives in y and solving for the wave height, H,j, at the
iyj grid yields
Hij =
(V+C,in g)g,,y_l (U+C, cot 9)Hi+1j
aY ~ aX
V+C4ini U+C, cot 9
(2.39)
ay a:' LJjj
The computation is a row by row iteration proceeding from large i (offshore in the
coordinate system of the above cited authors) to small i. During the computation of each
Hij the breaking height is also calculated and if Hij exceeds this value, then the breaking
wave height is used. Several breaking height indices are possible. Miche (1944) gave as the
theoretical limiting wave steepness the condition
^r = .142tanh(fcj,hÂ¡,)
Lb
(2.40)
where the b subscript indicates breaking conditions. Le Mhaut and Koh (1967) indicated
from experimental data that the limiting steepness is
^r = .12tanh(fcÂ¡,h,)
Lb
(2.41)
or
ff* = ^.12tanh(Jfc6fc6) (2.42)
Kb
This is the breaking height criterion used by Noda et al. (1974). Perhaps the simplest
breaking height criterion is the spilling breaker assumption
Hb = .78 hb (2.43)
which is used by Vemulakonda (1984) and Vemulakonda et al. (1985).
The scheme described above of determining the wave angle 6 from the irrotationality
of the wave number (2.33) and the dispersion relation (2.34) and the wave height from
the energy equation (2.35) has its limitations. Principally this scheme does not take into
account the effects of diffraction. This limitation is surmounted by use of a parabolic wave
equation. However this presents another problem in that the wave angle is not determined
27
in a parabolic equation solution. Only the complex amplitude, which contains the mag
nitude and the phase of the wave, is obtained. Yans (1987) solution to this problem is
to obtain 0 directly from the complex amplitude by assuming that the slope of the water
surface resulting from the waves indicates the direction of wave propagation. This value
for 0 and twice the absolute value of the complex amplitude for the wave height are then
used in Eqs. (2.282.30) to obtain the radiation stresses. Hie value thus obtained for 0 is
then also used in the dispersion relationship. For plane waves this works fine; however, in
diffraction zones where shortcrested wave are encountered problems may arise, since the
instantaneous slope of the wave at a particular location can be other than in the direction
of wave propagation.
In this report the radiation stresses are obtained directly from the complex amplitude
and its gradients using equations developed by Mei (1972) and repeated in his book (1982).
The basis of the derivation is to use the definition of the velocity potential to substitute for
the wave induced velocities in the definition for the radiation stresses.
For an inviscid incompressible fluid the instantaneous equations of motion are
+ = Vp pgez (2.44)
V q = 0 (2.45)
where q represents the velocity components, p is the pressure and ex is the z or vertical unit
vector. Denoting the horizontal components of q as ua,a = 1,2 and the vertical component
of q as u>, the boundary conditions are
= 0 (2.46)
= tv (2.47)
on z = rj, where rÂ¡ is the instantaneous free surface, and the bottom boundary condition
dh
u, = , (2.48)
q is assumed to be composed of a time averaged part q and a fluctuating part q'\ i.e.
dr)
dr]
dt +Up'dxp
f= fl*. ,*) + *(*. .*,*)
(2.49)
28
Integrating a horizontal component of (2.44) vertically over the depth of the water column
from z = h to z = rÂ¡, using the boundary conditions (2.462.48) and then taking time
averages (the same procedure as in appendix A) the following definition of the radiation
stresses is obtained:
ro
Sap = p u'au'pdz + 6ap
J h
[i
Pdz~ +
h
+ y(i')
(2.50)
where the overbar is used to indicate the time averaged values and 6ap is the Kronicker
delta function
tap
for a = P
for a ^ f3
(2.51)
Mei (1972) was able to express Eq. (2.50) in terms of the general complex amplitude
function B by relating it to the velocity potential and the free surface height by
_ to cosh k[h0 + z) iut
a cosh kh
t) = B e~iut
(2.52)
(2.53)
An expression for p is obtained by integrating the vertical component of Eq. (2.44)
dw
dw
dw
dp
(2.54)
Integrating from any depth z to the surface, using Leibnitzs rule for interchanging the
order of differentiation and integration, employing the two surface boundary conditions,
Eqs. (2.46) and (2.47) and the continuity Eq. (2.45), Eq. (2.54) yields:
d fi d H
= pgin ~ z) +pJ wdz + pJ Upwdz pw]
(2.55)
Taking time averages gives
rfl du'w'
p[x,y,z) = pain z) + pj* gX/9 dz~p)2 (2 56)
Substitution of Eqs. (2.52),(2.53) and (2.56) into (2.50) and making the assumption that
horizontal derivatives of k, h and a can be neglected (this is essentially a mild slope as
sumption), the following is obtained after straightforward but tedious calculations which
29
are shown in Appendix B:
c P9 L (dBdB*) 1 f 2kh \
4  ydxa dxp J k2 \ sinh2fch/
ap
IBI
2 kh
sinh 2 kh
+
dxa dxp
(VhB2 *2B2)(2*hcoth2*h 1)] } (2.57)
where B* denotes the complex conjugate of B. In obtaining Eq. (2.57) use was made of the
flat bottom Helmholtz equation
Vj;Â£ + Jfc2Â£ = 0 (2.58)
The general complex amplitude function B is related to the complex amplitude A which
is solved for in the wave model by
B = Ae^flco,idz^
(2.59)
There are some problems with Eq. (2.57). One may question the assumption of a
flat bottom; however the inclusion of the slope and gradients of the wave number and the
intrinsic frequency leads to intractable mathematics. The use of the localized flat bottom
is also implicit in the derivation of LonguetHiggins and Stewart to obtain Eqs. (2.28 2.30)
since in carrying out the integrations involved it is assumed that horizontal derivatives of
the velocity potential are limited to derivatives of the phase function. For plane waves
using a spilling breaker assumption, results using Eq. (2.57) differ little from those using
Eqs. (2.282.30); however in diffraction zones Eq. (2.57) is superior for reasons mentioned
previously.
The major problem with the use of Eq. (2.57) is also a major problem associated with
the use of LonguetHigginss formulation. This problem is that in both formulations the
initiation of wave breaking implies the dissipation of energy and thus the forcing of set
up and longshore currents starting at the breaker line. The physical reality is somewhat
different in that in the initial stages of breaking, energy is not immediately dissipated but
rather is transformed from organized wave energy to turbulent energy and to a surface roller
(Svendsen 1984). There is a space lag between the initiation of breaking and the dissipation
30
30
23
20
1.3
l 0
03
0
03
6
4
2
>
0
2
4
eai.
Figure 2.2: Experimental results for wave setup and setdown. Reprinted from Bowen, In
man, and Simmons (1968), Journal of Geophysical Research, vol. 73(8) page 2573, copyright
by the American Geophysical Union.
of energy. This can be seen in figure 2.2 which gives the results for the experiments of Bowen
et al. (1968). Analytical and numerical solutions for the mean water surface elevation will
give a setup starting at the break point while the solid fine in the lower graph shows
that setup measured in the experiment begins shoreward of the break point. This will be
discussed further in Chapter 5.
CHAPTER 3
WAVE MODEL
3.1 Governing Wave Equation
Severed methods have been used to derive the governing equation for the propagation
of waves over varying topography and varying currents. Among the methods used are per
turbation techniques, multiple scales, a Lagrangian and the use of Greens second identity.
With the assumption of a mild slope and weak current conditions each of these methods
should yield the same equation. For illustrative purposes the method of using Greens sec
ond identity will be used here to derive a linear equation. A Lagrangian derivation of the
same equation is presented in Appendix C. The linear equation is used since second order
waves exhibit singularities as the water depth goes to zero.
A velocity potential t is assumed such that the water particle velocities are given by
V, where V is the threedimensional gradient operator
d d d 
(3.1)
and *,j, and k are the unit vectors in the x,y, and z directions, respectively. The velocities
are a superposition of steady currents and wave induced motion. The velocity potential is
given by
t = to + eft (3.2)
where to is defined such that its gradient yields U,V, and W which are the steady current
terms, and ft is the wave component of the velocity potential. An undefined scaling e
is used so as to separate the wave and steady current part of the velocity potential. The
wave part of the velocity potential is composed of two parts: /, which contains the depth
dependency and which is a function of the horizontal coordinates and time. The depth
31
32
dependency / is given by linear theory as
j cosh k(h0 + z)
cosh kh
(3.3)
where h0 is the depth of the bottom referenced to the still water line, and h is the total
water depth and k is the wave number vector. It should be noted that / is also a function
of the horizontal coordinates since k and h0 vary with horizontal position.
For incompressible, irrotational flow the governing equation for the velocity potential
is the Laplace equation
VV = 0 h0
where r? is the instantaneous water surface elevation.
Using a horizontal gradient operator V/, defined by
_ d d
Vkfc, + ^J
the Laplace equation is written as
+ *z 0
h0 < z < i)
(3.5)
(3.6)
The governing equation is complemented by the following boundary conditions. On the free
surface the kinematic free surface boundary condition is given by
VhV = 4>t
z = rÂ¡
The dynamic free surface boundary condition with zero atmospheric pressure
1,
t + ^(V)2 + gz = 0
z = rÂ¡
(3.7)
(3.8)
and the bottom boundary condition
 Vh = 4>t
z = h0
(3.9)
The derivation technique is to write Greens second identity for and / as defined by
Eqs. (3.2) and (3.3) respectively.
rf) rn
fztdz dz = [fz fz]lt
(3.10)
33
Then Eq. (3.6) is used to substitute for zz in the first term of Eq. (3.10) and the boundary
conditions are used to substitute for z at the mean free surface and at the bottom in the
third term of Greens identity. Equation (3.2) is used and terms of order e only are retained
in order to obtain the governing equation for .
In order to obtain z at fj it is assumed that
r] fj efj.
(3.11)
The kinematic free surface boundary condition, Eq. (3.7), is then expanded in a Taylor
series about z = fj,
J*=ff
d
dn
gj + V/,f7 4>M
= 0
J*=ff
(3.12)
solving for z /j using Eqs. 3.2 and 3.11 while retaining only terms of order e yields
dfÂ¡
Iff
A total horizontal derivative
dt
+ U Vhfj + Vh Vfcff fjWz
*=fl
(3.13)
D 9 ft T7
Di = ai + uVh
(3.14)
is introduced so that the first two terms on the right hand side of Eq. (3.13) can be written
as the total horizontal derivative of fj. The final manipulation of Eq. (3.13) is to recognize
that ^ at the surface is the horizontal divergence of the horizontal current (= V>, ).
Thus
M ff= 'VhT) + fjVh
(3.15)
fj is eliminated from Eq. (3.15) by expanding the dynamic free surface boundary condition,
Eq. (3.8), in a Taylor series about z = fj.
4>t + ^(V<)2 + gz
J*=rj
= 9V +
M=fl
4>t + ^4>?
, 1 3 ,
9 + 4>*t + 2^(v^)
= o
*=ff
Using Eq. 3.2 to substitute for the order t terms yield a solution for fj
(3.16)
(3.17)
34
Substituting into Eq. (3.15) gives
(*)
The first term of the Greens identity using Eq. (3.6) becomes
r f*ndz= r fvi+dz=r /v*. (v^* (3.19)
* ho ho ho
Now
Vh = + efVh4> +e4>Vhf (3.20)
and
^h = (/ V/1^) + eVhij> Vhf + zV2hf (321)
Retaining only terms of order z and dropping the last term since it is second order in the
slope the integral is equal to
r /Vfc.(/V^)dz r fVh4>Vhf dz (3.22)
J ho /lo
These two terms are handled using Leibnitzs rule of integration.
vaT /(/v)dz = r vhf(fvj)dz+r fvh(fvht)dz
* ho ho hp
+^hVf2 U + _a V/,<Â£ (3.23)
Using this relationship and recognizing that / 
depth of /2 is CCt/g, where C is the wave celerity, Ca is the wave group velocity and g is
the gravitational constant, the first term is equal to
 V, (9 Vj) + Vhfj + Vhh0f2 _fco
9
where the z has been dropped.
The second term of Eq. (3.10) is evaluated quite easily.
dz z
(3.24)
(3.25)
35
The first integral above is dropped since there are only steady current terms. The remaining
term, dropping the e, is
= r ft f**dz = k2 r f2dz = k2^i (3.26)
Jh0 Jho 9
The third term in Eq. (3.10) is
ft. W = ft. If,ft. \h0 (3.27)
The first term on the right hand side is equal to t* / since / 1^= 1 and has already been
determined. The second term on the right hand side of Eq. (3.27) is easily determined using
the bottom boundary condition Eq. (3.9)
ft.\ko=fVhtVhh0 \k. (3.28)
Using Eq. (3.20) this becomes
 ft. Uo= fU Vkh0 _fco +e/2 _fco Vht vhh0 + eftVkf Vfch0 _fco (3.29)
The first term is dropped since it not order e and the third term is clearly higher order in
the slope. Thus, again dropping the t
Mk=/!lVkfVA (3.30)
This term cancels the bottom boundary term resulting from the Leibnitz integration above.
The fourth term of Eq. (3.10) is also easily determined.
tf. *.= feBmcrah Jfch ^ l*.= ~k tanh kh Y* ^ (331)
Evaluating t at the mean water surface and retaining only terms of order e yields
t!.th=~i (332)
Collecting terms from (3.24),(3.26),(3.18),(3.30) and(3.32) yields the following linear
governing equation for t
+ (V* )^ Vk(CCtVht) + (a2 k2CC)t = 0 (3.33)
This equation was first derived by Kirby (1984).
36
3.2 A Parabolic Approximation
Equation (3.33) is a hyperbolic wave equation. It can be altered to an elliptic equation
by assuming that any time dependency is solely in the wave phase. Before doing this an
energy dissipation term DD is introduced as follows. It is assumed that Eq. (3.33) holds
for the case of no energy dissipation; then for the case with dissipation the left hand side is
balanced by any energy that is dissipated. The governing equation is then written as
+ (V* V*(CC,Vfc) + (
In the derivation of the governing hyperbolic equation using a Lagrangian as presented in
Appendix C the DD term can be viewed as a virtual work term.
Following Kirby (1983) the DD term is assumed to be of the form
Dd>
(3.35)
where w is an undefined coefficient indicating the strength of the dissipation. Eventually
the w term will be related to the energy dissipation due to wave breaking following the work
of Dally, Dean and Dalrymple (1984).
Using Eq. (3.35) in Eq. (3.34) and making the assumption that the only time dependency
is in the phase, i.e.
dd>
= tu) (3.36)
reduces the hyperbolic equation (3.34) to the elliptic form
2tuff + 0 Vh( vj) + (Vfc V)(U Vfc) v (cc,vfc)
+{[Vk U)} = rw<Â£ (3.37)
where only the phase contribution to the horizontal derivative of ^ is retained in obtaining
the term on the right hand side of (3.37).
Before deriving a parabolic approximation to Eq. (3.37), the need for such an approxi
mation should be discussed. There are two major computational drawbacks to numerically
37
solving an elliptic equation. The first is that solution is required simultaneously for each
grid. This means that for a large computational domain, say m by n grids, it is necessary
to invert anmxnbymxn matrix. If m and n are sufficiently large, problems may be
encountered in terms of the interned memory storage, and even if the available storage is
not a problem, the number of computations required to invert the matrix may make for
an exceedingly slow solution. In some situations this may be acceptable; however for the
purposes of a wave induced circulation model in which repeated solution for the wave field
is necessary the program would take a very long time to run. The other disadvantage in
volved in the numerical solution of an elliptic equation is that boundary conditions must be
specified at all of the boundaries. In many applications the only known boundary condition
is the offshore, or incident wave amplitude. For example, the domain of interest may be the
wave field in the lee of an obstruction where the only down wave condition is that the waves
freely pass through the down wave boundary. Or in the present study in which the wave
amplitude is known to go to zero at the shoreline a problem ensues in that the shoreline is
(for purposes of the circulation model) at the grid edge and not at the grid point.
A parabolic equation eliminates both of these problems. Since the solution proceeds grid
row by grid row where each solution uses the results of the previous grid row, no down wave
information is required. Since only one grid row is solved at a time, the solution requires
only that a tridiagonal matrix equation be solved to obtain values for the grid row. The only
required boundary conditions are the conditions on the first grid row (usually the offshore
boundary) and lateral boundary conditions which are usually a derivative condition that
allows for the wave to pass through the boundary without any reflection at the boundary.
There are several ways to transform the elliptic Eq. (3.37) to a parabolic equation. The
most direct is to make certain assumptions concerning the length scale of variations in the x
direction. At the heart of the parabolic approximation is the assumption that the direction
of wave propagation is essentially along the x axis. For waves propagating at an angle to
38
the x axis the proper form of 4> is
= g V) e(/ k co**+/ *,in <**) (3.38)
and the proper form of the dispersion relationship is
u> = a + kcosOU + A: sin 0 V (3.39)
where ff is the angle of the wave propagation relative to the x axis. The essence of a parabolic
approximation is to assume that the angle that the wave rays make with the x axis is small
so that the ksin0 term can be neglected in Eqs. (3.38) and (3.39) and cos 6 is assumed to
be unity. The manner of doing this varies with different authors. Yan (1987) retains the
complete dispersion relation (Eq. (3.39)) but deletes the y dependency in Eq. (3.38) in the
derivation of his governing wave equation. He solves for 6 by determining the slope of the
instantaneous water surface. For long crested waves this method is satisfactory although
requiring additional computations. However, this method may produce erroneous results in
diffraction zones where the presence of short crested waves will produce a horizontal com
ponent of the Blope normal vector that is different from the direction of wave propagation.
Kirbys (1983) method is to make the assumption that
=g^!Vle<(ftdz) (340)
o
so that the e* / *,in 6dv Gf phase function is absorbed into the amplitude function,
A.
Using Eq. (3.40) directly in Eq. (3.37) and further assuming that derivatives of Ax are
small compared to derivatives of the phase function (i.e. that ikAz ) will yield a
parabolic equation. For illustration the case of linear diffraction in the absence of currents
and in a region of constant depth is examined. Substitution of Eq. (3.2) into the Laplace
equation (3.6) yields the Helmholtz equation
vÂ£ + fc2 = o
(3.41)
This equation is also obtained directly from Eq. (3.37) for the stated assumptions of no
currents, no dissipation, and constant depth. Substitution of (3.40) into the Helmholtz
equation gives
Axx + 2ikAx + AVV = 0 (3.42)
and employing the above mentioned assumption concerning derivatives of Ax yields the
linear parabolic equation
2 ikAx + AVV = 0 (3.43)
Another method of obtaining a parabolic equation is to use a splitting method in which
it is assumed that 4> is composed of forward moving and backward moving components that
are uncoupled.
4> = + 4> (3.44)
Uncoupled equations for + and are developed and since the equations are uncoupled
only the equation for + is retained. In the following the work of Kirby (1983) and Booij
(1981) is closely followed.
The second order elliptic equation
d_ /1Â£Â£\
dx Vqf dx )
+ yS = 0
(3.45)
can be split exactly into equations governing forward and backward disturbances $+ an $
which are given by
d$+
dx
d$~
dx
+7$+
7~
Booij chose the relation
(3.46)
(3.47)
(3.48)
40
where Â£ is an as yet unknown operator on . Substituting into Eq. (3.45) and neglecting
derivatives of ( alone leads to the result
4>zz +
<Â¡>Z + = 0
(3.50)
The elliptic equation (3.37) is now put into the form of Eq. (3.50). Letting p = CCg
for notational convenience and expanding some of the terms of Eq. (3.37) in their spatial
coordinates gives
[p4>z)z + (p4>y)v + k7p4> + (u>2 a2 + iw(Vfc U))4> + iaw4>
~{U2k)z ~ {V2iv)v (UV]>z)v (UV4>v)z + 2jU Vj = 0 (3.51)
Booij chose to neglect terms involving the square of the mean current components assuming
that they are smaller than terms involving p and then arranged the above equation as
4>zz + P l[Pz + 2iuU)4>x + k2 ^1 + Jj'j 4> Q (3.52)
where
M4> = (w2 a2 + tw(Vfc ) + 0w)< + 2iuV<Â¡>v + [p4>v)v (3.53)
Kirby points out that this choice by Booij leads to his inability to derive the correct wave
action equation. Kirby instead does not drop any of the squares of the mean current
components until after the splitting is complete. By assuming that the waves are oriented
in the x direction the dispersion relation is written as
a = oj kU
(3.54)
leading to the result
(jj2 o2 2ukU {kU)2
(3.55)
Using this relationship Kirby then arranged Eq. (3.51) as
h, + (p v')'(p U*)J, +(1 + ^ ) i = 0
(3.56)
41
where
Ml = {2ukU + ij(yh) + i
+ {{pV')ly) +2tWV^ (3.57)
Comparison with (3.50) gives
 *(1+W^y) <3'58)
Y = (p^2) (359)
or solving for 7 and Â£
> = (1+Â¡?(^))5 (3'60)
rifr7*)i(l+p^^)* (3.61)
Booij points out that the splitting matrix approach employed by Radder (1979) is equivalent
to the choice
1R = *(1 + 2k2[pU2])
Â£r = ki(pU2)2
Using the results (3.60) and (3.61) in Eq. (3.46) along with the definition (3.49) gives the
parabolic equation
(1 +
M
k2{p U2)
By expanding the pseudooperators in the form
M
^ <Â£+j = tkk*(p U2)* ^1 +
M
k2(pU2)
v 2
J 4 ^+(3.62)
(
1 +
1 +
i)> (i +
M
k2(p U2)J \ 4k2{pU2)
M \* .. /_ 3 M
iYi+ (i+Â¡Â¡
)*+
k2(pU2)) \ 4 k2{p U2)j
the result of both expansions may be summarized in the general equation
(3.63)
(3.64)
fz{kib ( + kfrlu*)) +} = V)1 0 + toU>)) ** (3'65)
42
where
{\
Radder
i Boij
(3.66)
and where P2 = Pi + j for both approximations. Kirby chose the approximation with
Pi = 0 since for his derivation which was for weakly nonlinear waves it should correspond
to the information contained in the nonlinear Schrodinger equation. With Pi = 0 Eq. (3.65)
becomes
1
k(p u2)fc + {fc(p u2)}x 4>+ = ik2(p U2)j>+ + 'M4>+
(3.67)
where M+ is defined by Eq. (3.57). An equation for the complex amplitude A is then
obtained by making the substitution
(3.68)
which after dropping squares of the components of the mean current results in
(C, + U)A, + \ 'A + (Â£)VA, + \ (7)J A
+i [cc' () J + ?A = 0 (3'69)
Kirby then further modified the equation by a shift to a reference phase function using the
relation
A = A'ei(ifkdx) (3.70)
where k can be taken as an averaged wave number, resulting in
(Ct + U)A'X + i(ic k)(CÂ¡ + U)A' + I A'
The shift to a reference phase function is intended to incorporate the phase information
in the complex amplitude so that the instantaneous water surface can be easily recovered
in the form
q(x,y) = 3i{AV/Â£d1}
(3.72)
43
where 3? designates the real part. This is an important consideration especially when ob
taining the radiation stresses.
It is important that in obtaining Eq. (3.69) that squares of the products of the current
components be retained until the last step. The reason for this is seen in that terms such
as kUV and kU2 are combined with terms of ojV and uU using the dispersion relationship
to give terms of aU and &V.
Equation (3.71) has been successfully used by Kirby in the form given and with modifi
cations. The modifications are the addition of a nonlinear term which makes the equation
applicable for weakly nonlinear waves. However for present purposes Eq. (3.71) presents a
certain problem which will shortly be corrected. The problem is that for waves propagating
at an angle to the x axis it is not possible to obtain a correct form of the conservation
of wave action. The case of no current on a beach with straight and parallel contours is
given as an illustration. The equation for the conservation of wave action reduces to the
conservation of wave energy flux. In mathematical terms this is
^Cscos0a2j =0 (3.73)
where a is the wave amplitude. For waves approaching the beach in a direction normal to
the beach Eq. (3.71) reduces to
+(C,M' = 0 (3.74)
Multiplying Eq. (3.74) by the complex conjugate of A', and add to it the complex conjugate
equation of (3.74) multiplied by A', yields
[ct\A\2]x = 0 (3.75)
which is Eq. (3.73) for the special case of cos# = 1. However it is not possible to obtain
Eq. (3.73) from Eq. (3.71) for the general case of waves at angle 6 on a beach with straight
and parallel contours.
The importance of this inability to give the correct form of Eq. (3.73) is that if energy
flux outside the breaker line is not conserved then gradients of the Sxv radiation stress
44
will generate currents offshore of the breaker line in the circulation model. Although
such currents in general will be of a small magnitude, they can cause some numerical
computational problems. The problem though is easily corrected by using as the dispersion
relationship
u) = a + k cos 9 U (3.76)
and the substitution
_jA(x,y) j(f kcottdx)
(3.77)
Rather than using a splitting method, the procedure will be to go directly from Eq. (3.51)
using Eq. (3.77) and writing
u>2 a1 = 2uik cos 9U (k cos 6 U)"1
(3.78)
in place of Eq. (3.55). Dropping the (j)zx term and all terms containing products of the
current components that cannot be reduced as explained above results in the following
parabolic equation.
Cg cos 6 + U'
(Cg cos 6 + U)AZ +
At
A + V Av +
Â§a
cos20)A i CCg{^)yJ +^A = 0
Now adding a phase shift
t cotdzj icoafdz)
so that the free surface can be recovered by
rj(x)V) = A'ei(flcose'dz)
(3.79)
(3.80)
(3.81)
where 6 is the wave angle obtained by Snells law in the absence of any obstructions or y
variation of the depth or current. The following governing equation is thus obtained.
CgCosO + U'
(Cg cos 6 + U)A'X  i(k cos 9 k cos 9)(Ct cos 9 + U)A' +
At
A'
+VA'v + 2 (7) A> ~ 2kCi^ ~ Cs2 ~ 2
CC~)v
+ ^A' = 0 (3.82)
45
Details of the algebra and assumptions involved in obtaining Eq. (3.79) are given in Ap
pendix D. For computational purposes it will be assumed that cos# = cos#.
From Eq. (3.82) the correct form of the conservation of wave action can be obtained.
For the case of no y variation of the wave number Eq. (3.82) reduces to
(Cf cos # + U)A'X + L
Cgcos # + U
A' + VA[
l+i/
(3.83)
Multiplying Eq. (3.83) by the complex conjugate of A and adding to it A times the complex
conjugate equation of (3.83) yields
(C, cos# + U)
+
\V\A\2]
a
X
a
= 0
(3.84)
which is a closer approximation to the exact expression which should have a (C, sin + V)
term instead of V in the y derivative term.
3.3 Energy Dissipation Due to Wave Breaking
Kirby (1983) found a method of relating the dissipation coefficient w to the work of
Dally (1980) who proposed that the decay of energy flux in the surf zone was proportional
to the amount of excess energy flux. The excess energy flux is the difference between the
actual flux and a stable energy flux. In mathematical terms this relationship is expressed
as
(ECg)x = ^[ECg(ECg)t] (3.85)
where
EC,
E
c,
K
h
the energy flux
wave energy = ^pgH2
d(T
the rate of energy propagation which is
ok
constant
the water depth
46
(ECt), = the stable energy flux.
H = the local wave height
The stable energy flux is obtained by assuming it to be a function of the height of a wave
propagating over a flat bottom after breaking ceases, and that this height is a constant
times the water depth, = 7. The constants K and 7 were determined by analysis of
experimental data obtained from Horikawa and Kuo (1966).
Kirby related Dailys work to the dissipation term w by deriving a conservation equation
for the wave action j. Garrett (1967) and Bretherton and Garrett (1968) showed that for
waves propagating in a moving and variable medium, the quantity that is conserved is the
wave action (7) > which is the wave energy divided by the intrinsic frequency. Kirby derived
a wave action conservation equation from the linear hyperbolic wave Eq. (3.34) using the
definition of W as given by Eq. (3.35):
^ + (v* U)^ Vfc(CC,Vfcft + (
Kirby follows the approach of Jonsson (1981) assuming that can be represented as
1* = Reie (3.87)
where R and 8 are real. From linear theory it is obvious that
R = A. (3.88)
G
To lowest order the derivatives of are approximated as
0< = u (3.89)
Vhe = k (3.90)
Substituting into (3.86) gives the complex equation
n2p DR
BjT + (V*' V(CC,VkR) i(r,B + KrJt,)
1 {Vk (0o)R+2{o0) VhR + V (kCC,)R + 2kCC, Vkl
t<7w<> = 0 (3.91)
47
Since R and 0 are real, the real and imaginary parts of (3.91) must individually equal zero.
Taking the imaginary part and multiplying by R gives
V* (kCCg)R2 awR2 =
(3.92)
which using the substitution kCCa = aCt is written as
{(rR2)t + (aR2(0 + Cy)) = awR2
Using (3.88) R2 may be replaced by
Ri = = 2p
a2 a \a J
which yields the wave action conservation equation
Assuming a time steady wave field and neglecting currents, (3.95) becomes
(3.93)
(3.94)
(3.95)
[ECt)x = w E
(3.96)
Comparing (3.96) with (3.85) and noting that (Cg)t = Cg, w is written as
(3.97)
where H is the local wave height given by 2A.
In obtaining Eq. (3.97) Kirby neglected the effects of the currents. However, the same
expression (3.97) is obtained when currents are included in the derivation. Dally (1987)
extended his work to investigate wave breaking on currents and proposed the following
equation:
+ U) (f )]i = lEC* ~ (3 98)
For the case of a steady state wave field propagating in the x direction the conservation of
wave action Eq. (3.95) reduces to
(3.99)
48
Setting the right hand sides of Eqs. (3.98) and (3.99) equal to each other results in Eq. (3.97).
At first glance it may appear unusual that the same expression for the dissipation
coefficient results for the case of currents as for the case of no currents. This can be
explained in part by the fact that Dailys work was with waves where breaking was depth
limited. For the case of current limited breaking (for example a stopping current), another
expression is necessary. In this study Eq. (3.97) will be used, since in general the current
field will be moderately weak. Such an assumption of weak currents was invoked in the
derivation of the governing parabolic equation.
CHAPTER 4
FINITE DIFFERENCE SCHEMES AND BOUNDARY CONDITIONS
4.1 Numerical Solution of the Governing Equations
The finite differencing of the governing equations in the circulation model is derived
by a matrix analysis as suggested by Sheng and Butler (1982). A rational analysis of the
complete Eqs. (2.12.3) including all the nonlinear terms is not possible. However guidance
can be obtained through an analysis of the linearized equations. For this purpose the linear
equations for the special case of a flat bottom and without any of the forcing terms will be
examined.
dfj au av n
7T + D + D = 0
dt dx dy
au dfj
lt+3dz=
dX. + gdA = 0
at ydy
Let W be a vector of fj,U and V
W =
U
V
Using Eq. (4.4), the governing equations are written as
(4.1)
(4.2)
(4.3)
(4.4)
Wt + AWX + BWy = 0
where A and B are the following matrices
(4.5)
' 0
D
0 '
' 0
0
D '
A =
9
0
0
and
B =
0
0
0
0
0
0
. 9
0
0
(4.6)
To find an implicit finite difference scheme for the governing equations, Eq. (4.5) is
finite differenced as
jyn+i jyn
AT
+ AW?+1 + BW^1 = 0
(4.7)
49
50
This equation is solved for Wn+1 in terms of Wn
(1 + 2AI + 2A y)Wn+1 = Wn
(4.8)
where
1 AT
= 2AAX{
and
iBeZi
2 AY
(4.9)
4Aj;AvWn+i is added to the left hand side of the equation and 2AvWn is added and sub
tracted to the right hand side of Eq. (4.8) so that the equation can be factored to get
(1 + 2A.)(1 + 2Av)Wn+l = (1 2Av)Wn + 2AvWn (4.10)
An intermediate value W is introduced so that the above equation can be written as
two equations
(l + 2AI)W* = (l2A)Wn
(4.11)
(1 + 2Xv)Wn+1 = W* + 2A vWn
(4.12)
Note that by eliminating W* from the two Eqs. (4.11) and (4.12), Eq. (4.8) is recovered
with the addition of an error of 4A2AyWn+l which has been introduced.
Writing out the component equations of Eq. (4.11) gives
AT AT
v* = yn 9^Svfjn
(4.13)
(4.14)
(4.15)
(4.16)
While writing out the component equations of Eq. (4.12) gives
Vn+1 + = n* + D^6yvn
un+1 = u*
yn+1 c fÂ¡n+l y c =n
v +9AY6yV ~v +gAYW
(4.17)
(4.18)
(4.19)
51
Using Eq. (4.15) to eliminate V* in Eq. (4.19) and (4.18) to eliminate U* in Eq. (4.14) the
following are obtained. For the x direction
AT AT
r + daxSiU = DayS,V'
AT
Un+1 + g8tfÂ¡* = Un
and for the y direction
AT AT
r+1 + v'"+1 = +
AT
yn+1 + 9^Svfjn+1 = vn
(4.20)
(4.21)
(4.22)
(4.23)
The above equations are numerically stable for any combination of AT, AX, and AY.
However as a circulation model they are incomplete in that they do not contain any of the
forcing terms that drive the circulation. In particular in this model wave forcing represented
by gradients in the radiation stresses and bottom friction are of concern. Bottom friction is
especially important because at steady state conditions gradients in the radiation stresses
must be balanced by either gradients in the water surface elevation or by bottom friction
forces. In the absence of friction forces the model could yield continuously accelerating
currents. The method of adding the forcing terms is somewhat arbitrary in that the rational
derivation of the finite differencing scheme of Sheng and Butler cannot be expanded to
include these essentially non linear terms. The addition of these terms to the model also
limits the latitude of the ratio of jn a manner that is not analyzable. That both the
friction and the radiation stresses are nonlinear should be recognized. The nonlinearity
of the velocity friction term is quite evident since it is represented by the quadratic term
pF\U\U. The radiation stress terms, which represent excess momentum flux induced by
waves, are functions of the wave height squared, based on linear wave theory. In shallow
water the wave height H is a function of the total water depth, h0 + fj so that H2 is a
function of fj2.
All the forcing terms and the nonlinear acceleration terms are added at the present or
time level with the exception of the bottom friction term. The quadratic friction term
52
is finite differenced as pF\Un\Un+1 and the friction term as suggested by LonguetHiggins is
differenced as pF uorj Un+1, where uorj is in terms of values at the n time level. However,
when using the integral quadratic expression, it is evaluated entirely in terms of the present
or known values.
The nonlinear advective acceleration terms are added at the ntfl time level using central
difference derivatives. In the x direction momentum equation these terms are
7T + ^ilj i 1 [ir i 1/ i t/ i tr \ ^J+l 1
u,j 2AX h + v*iJ + vJ+l + ^y
(4.24)
and in the y direction equation they are
jWj + U<U + ViJ+l + (4.25)
In the model of Vemulakonda (1984) which is based upon the WIFM (Waterways Ex
periment Station Implicit Flooding Model) (Butler 1978) model, a three time level leapfrog
scheme is used. This gives the advantage that all of the forcing terms (and the nonlinear
acceleration terms) are added at the time centered level. However this has some disad
vantages in that additional computer memory and computational time are required and
leapfrog schemes can exhibit instabilities. In the present study a two time level scheme
is used and the wave forcing terms as well as the nonlinear acceleration terms are added
at the present time level. Since, for all of the applications in this report, the model is run
until steady state conditions are obtained, it is reasoned that there is no great necessity to
properly time center all of the terms. If the use of the model is extended to time dependent
applications, this may need to be corrected.
4.2 Finite Differencing of the Parabolic Wave Equation
Dividing Eq. (3.82) by (Cycosi + U) yields
f C, cos B+U j
A'x + i(k cos k cos 6) A1 + ^ ^
;A' +
+
2 (C,cos+ 17) (C,cos + 7) v
, i kCa(l cosJ ) A,
v 2 (Ct cos + U)
(Y.)
2(Ccos + 7) \a J
2(Ce cos + U)
CC,
+
w
2(Cy cos 0 + U)
A' = 0
(4.26)
53
The following notation is used to describe the finite differencing of Eq. (4.26): Fj
represents the value of F(i,j) where F is any term and i and j designate the grid row
and grid column. Because of the staggered grid system used for the circulation model as
previously explained, in which the velocity components are specified at the grid edges, the
velocity terms U and V are averaged across the grid so as to produce a grid centered term
for use in the wave model. Thus 17) in the wave model is actually \(Uj + ^)+1) and similarly
for V and other grid designations.
A CrankNicholson finite differencing scheme is used for Eq. (4.26). All terms not
involving an x derivative are an average of the value of the term on the grid row and on
the i + l grid row. Terms having y derivatives are centered at, j on the row and +1, j
(X'+1X'.)
on the + 1 grid row. The one term with an x derivative is written as 1 1 Thus all
terms are centered at + ,j. The finite differencing of Eq. (4.26) is written as
1
+ 2
4+14
AX
+ ' [(jfc+1 cos $<+1 Jfc)+1 cos 0,+1)A;+1 + (jfc* cos 0* k) cos 0*)Aj]
1 f [(C7g cos g + C/)/o]>1 ~ [(Cg cos g + f7)/cx]
h 2AX \ \(Ca cos 0 + U)/(r]fl + \(Ca cos 0 + U)/o
Va
(Ce cos 6 + U)
r
 4I
2AY
l}(4++4)
( Vo V a;+i a;.,]
^(<7,0)8 0 +t/)^. 2AY
1
+ 2
2(Cycos0 + U)
(Cg cos0 + C7)
<+' r(V/);i\ (V/a))t\
l
+
2(CÂ¡ cos 0 + 17)
)!(
2AY
w*)Ui PH1
2AY
4
t
4
kCg{l cosJ0)l*+1 A<+1 +
(Cg cos 0 + U)
kCÂ¡( 1 cos2 0)
(#),
(Cg cos 0 + U)
i+l
'P
'+!
Vj
(),
+
A>+1
V)
(Cg cos 0 + i/);+1 (Cg cos 0 + U))
/
1 w*+1 A,+1 w* A\ )
+4 {(Cycos0 + u)*1 + (Cgcos0 + u)) j 0 ^4'27)
54
where p = CC9 for notational convenience and the y derivative of p ^ j is written as
(rUi + pj) [(Â£)+, ~ (Â£))] (p> + p;~i) [(f) ($)'
2(AY)2
(4.28)
In order to evaluate w)+1 by the method to be explained in section 3.3 an intermediate
value of A)+1 is obtained by the explicit formulation
A,+1 A*
3 + i(k* cos S' k'j cos 0') A'j
AX
1 [{Ct cos 0 + V)/<,])+' \(Ct cos 0 + U)/a]
1 [ic
' AX 1 [(C, cos 0 + U)/af+1 + [(Ctf cos 0 + U)/a]
(
}
(Cg cos0 + U)/a
j+i
2AY
+
( V
[(v;+, (W;,!
A* *
kCg{\ cos2 0)
{2(Cecos0 + U) J .
2AY
A* 2
(Cscosi + i/)
A
n
V)
T +
}4
2 (Cg cos 0 + U)) _r 2[Cg cos 0 + U))
When Eq. (4.27) is written out using Eq. (4.28) and w*+1 obtained through use of
(4.29) there results for any particular values of i and j an equation for the three unknown
values A}**, A)+1, and A.\. The equation is put into the form
ai aYA + hi }+1 + Cj Ajt\ = dj (4.30)
where the coefficients aj, bj, and Cj are known quantities involving the properties of the
environment (depth, current, celerity, etc.) and the dj involves the environmental properties
and the known values of the amplitude on the grid row. When Eq. (4.30) is written for
all the j values, j equal to 1 through N, for a specified i value, the ensemble of N equations
takes the form of a tridiagonal matrix equation. For illustrative purposes let N be equal to
four. The resulting matrix equation is
r 4>+1i
aj ci
A+1
1 '
a2 &2 c2
A2+1
d2
as b$ Cj
A*3+1
S
a< 4 C4
4+1
. d* .
L 4+1J
(4.31)
55
Equation (4.31) contains four simultaneous equations with six unknowns. Two bound
ary conditions should be specified. At each boundary a functional relationship is established
between the grid inside the domain and the grid outside of the domain. For the wave prop
agation model there are two types of boundary conditions that are readily available for use
in a computer model. These are a totally reflective boundary condition which would be
used to model a laboratory experiment in a wave basin, and nonreflecting noninterfering
boundary which will allow waves to freely pass through the boundary. For the reflective
boundary, the mathematical statement of the boundary condition is ^ = 0. At the lower
boundary this is finite differenced as = 0 which gives the functional relationship
A1 = Ao. (4.32)
At the upper boundary the functional relationship for a reflective boundary is
An+i = An.
(4.33)
For the nonreflective boundary the mathematical statement is = imA where m is
the longshore component of the wavenumber. Since the lateral boundaries are sufficiently
far from any diffraction effects, only refraction and shoaling effects are present. Thus
using Snells Law m = k sin# = k0 sin0o. The non reflective boundary condition is finite
differenced at the lower boundary as
Ai Aq im , . ,
Â£y~~ =Y[Al + A] (434)
which gives
Aq =
fj.nAY)
1 +
tmAYA
2 )
A\
(4.35)
Similarly at the upper boundary the functional relationship used to eliminate Ajv+i is
(l+taAX)
Aw+1 = '(l imAY\AN (436)
The use of either pair of boundary conditions, (4.32) and (4.33) or (4.35) and (4.36) to
eliminate Aq and A;v+i will then yield an N by N tridiagonal matrix equation comprising N
56
Figure 4.1: Definition sketch of grid nomenclature and coordinate system.
equations for N unknowns. Returning to the illustrative example where N is equal to four,
the matrix Eq. (4.31) becomes
6i + 7\d\ ci
02 f>2 C%
O3 63
CS
' A\+1 '
di '
A'2+1
d2
4+1
3
[ 4+1J
. d< .
(4.37)
where 7\ and 7n represent the functional relationships as described above on the j=l and
j= N boundaries.
Equation (4.37) is solved easily for the unknown A)+1 using a double sweep method.
In the computer program the subroutine CTRIDA uses an algorithm of Carnahan, Luther,
and Wilkes (1969) to efficiently accomplish the double sweep solution of eq. (4.37).
4.3 Boundary Conditions in the Circulation Model
As shown in figure 4.3 a staggered grid system is used so that velocities are expressed
at the grid edge and fj and the wave properties are defined at the grid center. So that the
wave model will have full grids it is desirable to place the boundaries of the model at the
grid edges on all sides, with the exception of the offshore grid row where the wave conditions
are an input or boundary condition for the wave model. Figure 1.1 can also be referred to.
The offshore boundary condition that is used in the circulation model is that the mean
57
water surface level is not changed by any of the circulation within the model, mathematically
ETA(1,J) = 0 J = 1, N
(4.38)
This condition necessitates that the offshore boundary be sufficiently far from any setup or
setdown. Other possible offshore conditions are a noflow condition or a radiation condition.
The no flow condition would conserve water within the computational domain so that the
volume of water involved in set up would be compensated by an equal volume of setdown
offshore. In an open ocean environment the open ocean can be conceived of as an infinite
reservoir. However in modeling laboratory experiments a no flow condition as the offshore
boundary condition would be the most appropriate. A radiation boundary condition as
proposed by Butler and Sheng (1984) imposes the condition
dr) dr)
 + C  = 0
dt dx
(4.39)
This condition allows long waves to freely propagate out of the computational domain
without any reflection. This tends to eliminate any seiche within the computational domain
which is induced by the startup transients. This condition was applied at one time in the
present model but it did not preserve the offshore zero setup condition and was changed in
favor of the more rigid zero water surface elevation condition. If the model is adapted for
use with time varying wave trains (an adaptation that the model can easily accommodate)
then the radiation condition will become imperative.
At the shoreline the boundary condition is that there is no flow into or out of the beach.
Since the model allows for flooding of the beach the index of the shoreline changes. The
model allows for differential flooding along the shoreline, for example, due to differences of
wave intensity in the lee of structures. For each grid column in the y direction the last grid
is the IWET(J) grid, where IWET(J) is allowed to change with changes of the mean water
level. The numerical statement of the no flow boundary condition at the shoreline is
t/(IWET(J)+l,J) = 0
J=1,N
(4.40)
58
This boundary condition is used in other models, though not all models allow for differential
flooding of the shoreline. Yan (1987), whose model was used to investigate waves and
currents around a tidal or fluvial discharge, used the no flow condition away from the
discharge and defined Â£f(IWET(J)+l,J) as equal to the discharge per grid divided by the
water depth at the grids where discharge occurred.
Other conditions are also imposed at the shoreline boundary. These are that the wave
height and the longshore velocity are both zero. These conditions are imposed at the
shoreward grid edge of the IWET(J) grids even though the wave height and longshore
velocities are specified a half grid space form the shoreline. This is done by writing x
derivatives of the radiation stress terms and of the V velocities that need to be positioned
at the center of the IWET(J) grid in terms of three points. The three points are at the
IWET(J)1 and IWET(J) grids and the zero value at the shoreline at the edge of the
IWET(J) grid.
For the lateral boundary conditions several alternatives are available. Birkemeier and
Dalrymple (1976) and Ebersole and Dairymple (1979) both used a periodic or repeating
beach. This meant that computations were performed between the second and N grid,
and values outside of the computational domain are determined according to
Q(I,1) = Q(I,N) (4.41)
Q{I,N + 1) = <5(7,2) for all I (4.42)
for any quantity Q.
At one point this was used in the present model until the problem of what should have
been transient edge waves generated by wave obstructions during the startup were found
to be trapped in the model. The periodic boundary condition allowed for disturbances
propagating out of the domain at one lateral boundary to enter the domain at the other
boundary. For straight and parallel contours, the periodic beach condition worked well in
the present model but with the presence of breakwaters another condition had to be used.
For the lateral boundary conditions there are two choices, either closed or open bound
59
aries. Closed boundaries are easily applied by the condition of no flow across the boundary.
Such a choice of boundary conditions would accurately duplicate the physical conditions ex
isting in a laboratory basin (or for prototype conditions within a harbor). Open boundaries
are obtained by specifying derivative conditions. In the present study an open boundary
is obtained by specifying that there are no gradients of the depth and the wave proper
ties across the lateral boundaries. This necessitates that the lateral boundaries are placed
sufficiently far from any obstructions or any y gradients caused by the obstructions. An
alternative open boundary condition would be that there is no gradient of the flow. This
would, however, put the position of the boundary condition at the center of the bordering
grid rather than at the edge of the grid and could possibly introduce error at the boundary.
4.4 Boundary Conditions in the Wave Model
In the wave model only lateral boundary conditions are needed for computational pur
poses. The offshore condition is merely the specified wave amplitude and wave angle. No
condition is needed at the shoreline since the parabolic wave model marches grid row by
grid grow and needs only the lateral conditions.
The offshore condition is that each time the wave model is called the wave amplitude
on the first grid row is specified as
A(1,J) = C(t) ~~A0(J) J=1,N (4.43)
where A0(J) is defined in the input subroutine according to
A0(J) = e*osin 0oJaY (444)
where k0 is the wave number along the first grid row and 90 is the angle the wave ray
makes with the x axis. This offshore condition necessitates that the water depth be constant
along the offshore grid row. This is another reason for choosing the offshore condition in
the circulation model to be a constant zero. With the presence of wave obstructions and
current deflectors such as jetties it is possible that with other offshore conditions the mean
water surface elevation could vary along the offshore grid row. The use of one condition in
60
o
Figure 4.2: Startup function according to Eq. 4.43 with Ci=30 and Ci=3.
the wave model which is contrary to the condition in the circulation model can introduce
instabilities in the overall model. The interaction of the two models is more sensitive than
either of the models individually.
The C(t) term in equation 4.43 is a startup function which smoothly goes from zero
to unity so as to bring the incident wave height from zero to H0 and minimize any shock
loading of the model. Mathematically it is given by
C(t) = .5
t
1 + tanh( Ci)
w
(4.45)
where C\ and Ci are arbitrary constants to adjust the slope and offset of the curve. Fig
ure 4.2 shows equation 4.45 for C\ and Ci equal to 30 and 3 respectively.
The lateral boundary conditions are either open or reflective. The reflective condition
is expressed as
(4.46)
while the open condition is expressed as
ai =,mA <447)
where m is the longshore component of the wavenumber which in the absence of diffraction
effects remains constant
m = k0 sin 90 = k sin 9
(4.48)
61
Direction of
Wave Propagation
x
Figure 4.3: Waves passing through a lateral boundary unaffected by the boundary.
according to Snells law. Equation 4.46 or 4.47 are used to eliminate the amplitude outside
of the domain from the tridiagonal matrix equation. To illustrate equation 4.47 is finite
differenced at the lower lateral boundary as
A(I,1) A(I,0) A(I,1) + A(I,0)
AY im 2
Solving for A(I,0)
i imAY
A(I,0) = tVa(I,1) (4.50)
1 + im^
Thus A(I,0) is eliminated in terms of A(I,1). similarly at the upper boundary A(I,N+1) is
eliminated in terms of A(I,N). The same procedure is used with equation 4.46 which yields
the simple relationships
A(I,0) = A(I,1) (4.51)
and
A(I,N+1) = A(I,N)
(4.52)
62
It is important that the boundary conditions 4.46 and 4.47 be used as prescribed
above. The parabolic equation gives N equations containing the N+2 unknowns A(I+1,J)
for J=0,N+1. The two boundary conditions are thus used to reduce the number of un
knowns to a number equal to the number of equations. If equation 4.47 is written in terms
of A(I+1,1) and A(I+1,2) at the lower boundary and in terms of A(I+1,N1) and A(I+1,N)
at the upper boundary and the governing wave equation is written for J=2,N1, then errors
may be introduced at the boundaries. This is especially true for the case of a reflective
boundary. Figure 4.3 shows waves freely passing through the lateral boundaries without
any noticeable reflection for the case of a plane wave with a period of 10 seconds propagating
over constant depth at an angle of 10 degrees to the x axis.
CHAPTER 5
RESULTS
The numerical model combining the iterative interaction of the circulation model and
the wave model as described in the previous chapters was tested for several cases in which
either analytical or experimental results are available. The model was also run for cases
where previous numerical results are available.
The two cases involving analytical solutions are for setup induced by normally incident
waves on a plane beach and longshore currents generated by incident plane waves at an angle
on a plane beach. The first section of this chapter covers setup, comparing the results
with the analytical solutions of LonguetHiggins and Stewart (1964) and the experimental
results of Bowen et al. (1968) and Hansen and Svendsen (1979). The second section covers
longshore currents and compares results with the analytical solution of LonguetHiggins
(1970a,b) and experimental results of Visser (1982).
The model was run for the case of a ripchannel on a plane beach. Comparisons with the
numerical results of Birkemeier and Dairymple (1976) and Ebersole and Dalrymple (1979)
are given in the third section. A rip current was also generated for the case of normal waves
breaking on a shore parallel bar with a gap. No comparisons are available for this case.
The model was run for the cases of shore perpendicular and shore parallel breakwaters
represented in the model as infinitesimally thin impermeable barriers. For the shore per
pendicular barrier, the results are compared with the experimental results obtained at the
Waterways Experiment Station as reported by Hales (1980). For the shore parallel barrier
the results are compared with the numerical results of Liu and Mei (1975) who did not
include a wave current feedback mechanism.
In the final section the model was used to simulate the laboratory experiments of
63
64
Gourlay (1974) in which the shoreline in the lee of a breakwater was made to curve and
meet the breakwater as occurs with tombolo formation. The results of the model were in
close agreement with Gourlays experimental values.
5.1 Wave Setup
LonguetHiggins and Stewart (1964) gave a simple but elegant derivation for the wave
induced setup on a plane beach assuming linear shallow water theory. For the steady state
case of normally incident waves on a plane beach without any y variations and without any
currents the depth integrated equations of motions reduce to
gdi + as~=0
(5.1)
dx pD dx
In shallow water for normally incident waves the radiation stress term from linear wave
theory, Eq. (2.28), becomes
Sxx = E(n + l) = lE=^pgH2
(5.2)
A spilling breaker assumption is made so that the wave height if is a constant ratio k to
the water depth D = h0 + fj,
H = K(h0 + fj) (5.3)
Substituting equations (5.2) and (5.3) into Eq. (5.1) gives
3rj _ 3 dK2(h0 + fj)1 3 2 fdhQ dfj\
dx 16(h0 I fj) dx 8* V dx + dx)
Solving for gives the result
dfj /c2 dh0
dx 1 + /c2 dx
that the slope of the water surface inside the breaker line is a constant times the slope of
the beach.
Previous models are able to produce the results of the analytical expression but to
date none, including the present, have matched the experimental results. In figure 5.1 it
is seen that the slope of the mean water surface produced by the wave setup is nearly
65
DISTANCE FROM STILLWATER LINE
IN BEACH, x CMS
3.0
2.5
2.0
1.5
CMS
1.0
IS?
0.5
0
0.5
Figure 5.1: Comparison of the numerical solution of Vemulakonda et al. (1985) for setup
with experimental data of Bowen (1968). Source: Vemulakonda et al. (1985)
a constant. For the numerical results the setup starts at the breaking point, whereas in
the experimental results the water surface elevation is nearly constant for some distance
shoreward of the breaking point and then there is a change in the slope of the water surface
elevation. The numerical results follow the analytical solution. This is an area that needs
further investigation. The reason for the difference between the analytical and experimental
results is that the analytical formulation is predicated upon the radiation stress term being
a function of the wave height, so that as the wave height is reduced after the breaking point
so is the radiation stress term. As pointed out by Svendsen (1984) the radiation stress term
remains nearly constant after breaking for a distance roughly equal to five or six times the
breaking depth. This distance is termed by Svendsen as the outer or transition region in
which there is a rapid change of wave shape. In the inner region the wave shape according
to Svendsen remains nearly constant and is characterized by a surface roller or bore. By
incorporating the surface roller Svendsen is able to obtain close agreement for wave height
and setup to the experimental results of Hansen and Svendsen (1979). However this close
66
agreement is just for the inner region. For the outer region there is no viable solution other
than that the radiation stress term and the water surface elevation remain nearly constant.
Further there is no analytical separation of the outer and inner regions other than the a
posteri definition of where there is a sharp change in the slope of the water surface elevation.
This definition is not useful in a model that seeks to determine the water surface elevations.
Svendsens work shows promise and should be pursued; it is not however at this point
viable for inclusion in a numerical circulation model. For one it needs to be expanded for the
case of waves at a angle (this need will be discussed further in the next section). Another
limitation is that Svendsens model is for monotonically decreasing depths, which is a severe
restriction (for example excluding barred profiles).
Svendsen conceptually addresses the question concerning the spacial difference between
the initiation of the setup in the analytical and experimental results. The analytical model
assumes that once breaking commences energy dissipation also begins. Svendsen states that
the organized potential energy of the wave prior to breaking is transformed into forward
momentum in the outer region. In keeping with the definition of the radiation stresses as
the flux of excess momentum, i.e. momentum that would not be present in the absence of
the wave, it is easy to conceptualize that the radiation stress term should remain nearly
constant just after breaking. The problem is in assuming that the wave form in the breaking
zone is characterized the same as outside the breaking zone. The author, at present, offers
no solution other than to recognize the existence of the problem.
In the present study results for the setup on a plane beach with normally incident
waves differ somewhat slightly from the analytical solution as given by LonguetHiggins and
Stewart. This is because the present model does not use the spilling breaker assumption.
Instead wave energy is dissipated according to a model suggested by Dally (1980) which is
explained in section 3.3. Thus the slope of the mean water surface is not a constant times
the bottom slope. A result from the experiments of Hansen and Svendsen (1979) show this
clearly in figure 5.2. The figure shows a slight curvature in the wave height after breaking
67
mm
Figure 5.2: Comparison of numerical results from the model with experimental results of
Hansen and Svendsen (1979). Wave setup and wave heights for a 2 second wave on a beach
slope of 1:34 and incident wave height of 7 centimeters.
and in the slope of the setup. Although the problem of where the setup begins is still
present, the curvature of the mean water surface is the same for both cases. It is also quite
clear that the diminution of the wave height is not constant as is assumed by the spilling
breaker assumption.
In figure 5.2 the breaking wave height in the experiment is much greater than in the
model. This is because linear wave theory cannot adequately predict wave shoaling. Ven
gayil and Kirby (1986) were able to duplicate the shoaling recorded in the experiments of
Hansen and Svendsen (1979) by numerically solving coupled equations for the many har
monic components. Their results showed that in the shoaling process prior to wave breaking,
wave energy was transmitted to the higher frequency components. Their work however is
not adaptable for use in the current model since it is not known how to dissipate energy
68
Figure 5.3: Nondimensional longshore velocities vresus the nondimensional distance off
shore from the analytical solution of LonguetHiggins. Reprinted from LonguetHiggins
(1970b), Journal of Geophysical Research, vol. 75(33) page 6793, copyright by the Ameri
can Geophysical Union.
from the various frequency components once breaking is initiated.
5.2 Longshore Currents
LonguetHiggins (1970b) gives an analytical solution for the longshore current generated
by obliquely incident waves on a plane beach. The solution involves a spilling breaker
assumption and assumptions concerning the form of the bottom friction and the lateral
mixing. For steady state conditions the y equation of motion reduces to a balance between
the gradient of the Sxy term and the bottom friction with the lateral mixing serving to
distribute the current. Details of the solution can be found in the cited paper.
Figure 5.3 gives the results of the analytical solution of LonguetHiggins. In the figure
the ordinate V is the nondimensional velocity v/v0, where t; is the actual velocity and v0
is the velocity at the breaking point obtained for the case of no lateral mixing.
5jt. .isinfl .
v0 =(fffis)i (5.6)
O C
The abscissa X is the nondimensional distance z/n, where ij is the distance from the
69
Wave SetUp Line Breaker Line
Figure 5.4: Comparison of experimental results of Visser (1982) and numerical result for
longshore currents due to a 1.1 second wave on a slope of 1:20 for the deep water conditions
H0 = .065m and 60 = 20 degrees.
shore to the breaker line.
Previous models have reproduced the results of the Longuet Higgins solution by incor
porating all of the same assumptions (i.e. spilling breakers and bottom friction). Since the
present model does not contain the same assumptions comparison will not be made with the
analytical results of LonguetHiggins. Instead comparison is made with the experimental
results of Visser (1982). For both the analytical and the numerical results it should be noted
that the same inadequacy is present as was discussed in the previous section. In both cases
it is implied that wave energy dissipation commences as soon as the wave begins breaking.
Thus both analytical and numerical solutions initiate a longshore current starting at the
breaker line. It is reasonable to expect that the same holds for the y component of the
onshore flux of excess momentum as for the x component. That is, that the Sxv term also
remains nearly constant over a certain unspecified distance as the organized momentum flux
of the non breaking wave becomes turbulent momentum upon breaking. This is illustrated
in figure 5.4 which shows a comparison of the experimental results of Visser (1982) and the
model results for the same test conditions. The maximum current is nearly the same in
both the model and the experiment but the maximum is further offshore in the model. This
can be attributed to the initiation of a longshore current in the model at the breaker line.
70
5.3 Waves On a RipChannel
Both Birkemeier and Dalrymple (1976) and Ebersole and Dalrymple (1979) examined
the case of waves incident upon a shore with the presence of a ripchannel. The bottom
topography was given according to a formula first used by Noda et al. (1974)
h{x,y) = mx jl + Aes^)1/S sin10 ^(y ztan^) (5.7)
where (3 is the angle the ripchannel makes with the x axis which is, in their coordinate
system, directed offshore from the still water line, m is the beach slope, A is an amplitude
of the bottom variation and A is the length of the repeating beach. The origin of the
coordinates is at the intersection of the still water line and the trough as is shown in
figure 5.5 which gives the depth contours for the case where the slope is .025 and f) is equal
to 30 degrees.
Due to the restriction of the present model that there be no y gradients at the lateral
bound aries the above formulation is modified to be
h(x ) = / ^{i + AeW^sin^fivztan/?)} f (y xtanÂ£) < *
K [mi  j(y x tan fi)\ > x V '
Also shown in figure 5.5 are the contours obtained through use of the modified Eq. 5.8
for /? equal to 30 degrees and a .025 slope. The two formulas differ only in that the repeating
rip channels in the Noda formulation are removed. For application in the present model the
presence of the trough in the farthest offshore grid row is removed so that the depth in the
offshore grid row remains constant in compliance with the requirements of the boundary
conditions as discussed in sections 4.3 and 4.4.
Figure 5.6 shows the results obtained by Ebersole and Dalrymple for the case of a wave
angle of 30 degrees a wave height of 1.0 meters and a period of 4 seconds. Figure 5.7 gives
the results for the present model. Even though the scales of the two figures are different it
is evident that there is agreement between the two. The length scale of the eddy is about
30 meters for both cases. In the caption to the figure of his results, Ebersole indicated that
the lateral mixing coefficient may have been large. The present author does not believe
71
Figure 5.5: Top: Depth contours, in meters, using Eq. 5.7 with the coordinates as shown.
Bottom: Depth contours as used in the present model. Grid spacing is 5 meters.
72
Figure 5.6: Velocity vectors from the results of Ebersole and Dalrymple for a wave angle
of 30 degrees. Scale for velocity vectors is 1 inch = 2.87 m/sec. Source: Ebersole and
Dalrymple (1979).
73
Figure 5.7: Velocity vectors obtained from the present model using the same input condi
tions as for Ebersole and Dalrymple.
74
Figure 5.8: Mean water surface elevation contours showing that the water surface is lower
above the trough. Values indicated are in millimeters.
this to be the case. In the present study the lateral mixing coefficient was reduced to as
low a value as 2 meter2/second without substantially changing the magnitude of the eddy
adjacent to the rip trough. However Ebersole did find a significant difference between the
case of no lateral mixing and with lateral mixing. Results for the present model for the
case of no lateral mixing are not available since the model sometimes exhibits an instability
without the stabilizing influence of the lateral mixing terms. Figure 5.8 shows the mean
water surface elevation contours obtained from the model which show that the water mean
surface is lower above the trough.
The agreement between the results of the Ebersole model and the present model for
the case of waves incident upon a ripchannel should not be surprising since diffraction in
this case is small compared to current and depth refraction.
Another test result is the case of normally incident waves breaking on a bar in which
75
Figure 5.9: Depth contours in meters for the case of a bar with a gap.
Figure 5.10: Velocity vectors for the case of waves over a bar with a gap. Dashed lines
indicate depth contours. Grid spacing equal to 10 meters.
76
there is a gap in the bar. Figure 5.9 gives the contour lines for this test case. The topography
was created by adding an inverted parabola on to a plane beach with a slope of 1:40. The gap
was created by suppressing the bar with a symmetric smooth tangent function. Figure 5.10
shows the velocity vectors resulting from a 6 second wave with an incident wave height of 1
meter. The contours of the bottom topography are represented by dashed lines. The driving
force of the circulation is the return flow through the gap in the bar of the water that is
carried over the bar by the breaking process. These results may not be totally correct given
the arguments of section 5.1. If there were no gap in the bar the waves breaking on the bar
would push water shoreward until sufficient setup was produced. However with a gap in
the bar lateral flow is allowed and setup cannot be maintained.
Kirby (conversation with the author) has remarked that he has observed the initiation
of a very strong current in the direction of the waves produced by waves breaking on a sand
mound in a small wave basin. This was tested in the model using normally incident waves
on a slope of .02 with a shoal described by the following equation which is the shoal used
in the Delft experiments of Berkhoff et al. (1982).
M*y)
.45 x < 5.82m
.45 .02(5.82 + x) (5*5) + (I)2 > 10 (5.9)
.45 .02(5.82 + x) 5^1  {*)=' + .3 (5*5) + (g)1 < 1.0
Figure 5.11 shows the depth contours for this test case. Figure 5.12 shows the velocity
vectors obtained for the case of a barely breaking wave. For the case of a non breaking
wave the velocity vectors plotted to the same scale are not discernable. Figure 5.13 plots
the max velocity obtained in the shoal region versus the incident wave height. At a wave
height of about 10 centimeters there is a sharp change in the slope of the curve indicating
increased velocities once breaking commences.
5.4 ShorePerpendicular Breakwater
The model is tested here for the case of waves interacting with a shoreperpendicular
breakwater. The breakwater is represented as an infinitesimally thin impermeable barrier
located in between the J = Jbrk and J = Jbrk+i grid columns and extending from
77
i i i i i i i i i i i i i tit ri' i i i i i i i i i i i i i i i i i i i i
. 400 400
?4f)
40

TT
TT
1R0
1R0
TT
0R0
0R0
J L
~l 1 1
i i i i i i i i i i i i i
1 1 1 1 l i l~
Figure 5.11: Depth contours for the case of normally incident waves upon a shoal. Grid
spacing is .05 meters. Contour interval is .02 meters.
78
Figure 5.12: Velocity vectors for the case of waves breaking upon a shoal.
79
Figure 5.13: Maximum velocity across the top of a shoal versus incident wave height for
waves over a shoal on a plane beach. Incident wave height less than 9.1 centimeters are
nonbreaking waves. Incident wave heights greater than 9.1 centimeters produced waves
breaking over the shoal.
80
Figure 5.14: Position of the groin in relation to the grid rows and columns.
the shore boundary to the I = Ibrk grid row. Figure 5.4 illustrates the position of the
breakwater. Mathematically, this thin impermeable barrier behaves as an internal solid
boundary that prevents any activity on one side of the barrier from being transmitted to
the other side. This requires modifications to both wave and current models.
In the wave model an internal boundary condition is imposed so that at the breakwa
ter face wave energy is totally reflected. This is accomplished by imposing the reflection
condition
f* = 0
dy
(5.10)
on both sides of the breakwater from grid row I = Ibrk to the shoreline, or
A[I,Jbrk) = A[I,Jbrk+\) for all I > Ibrk (511)
When the wave equation is expressed in finite difference for J = Jbrk, >1 (/, Jbrk+i), which
is on the other side of the barrier, is eliminated by virtue Eq. 5.11. Likewise when the wave
equation is applied to the other side of the barrier A[I, Jbrk) is also eliminated. Further
81
reflective conditions are imposed at the barrier by assuming that the derivatives of any
property perpendicular to the breakwater has a zero value at the breakwater.
In the circulation model the interned boundary condition of impermeability at the bar
rier is imposed by setting the velocity component normal to the barrier equal to zero.
V(I,Jbrk+i) = 0 for all I > Ibrk (5.12)
To insure that currentinduced momentum is not diffused across the barrier, subroutines
were introduced in the computation for the x sweep for the two grid columns adjacent to
the barrier.
Two sample cases were run for comparison purposes. The first case is a breakwater
extending 400 meters from the shoreline on a beach with a composite slope of 1 on 10 for
the first one hundred meters and a slope of 1 on 100 beyond that point. This is a geometry
selected by Liu and Mei (1975). A deep water wave height of 1 meter and a deep water
wave angle of 45 degrees was used by Liu and Mei. Using linear shoaling and refraction
this gives a wave height of .84 meters and a wave angle of 31.15 degrees as the wave input
conditions at the offshore grid row 800 meters from shore in a depth of 16.95 meters.
The solution method of Liu and Mei was to solve for the wave field in the absence of
currents and mean water level changes and then use the wave field as a constant forcing in the
solution of the circulation. By ignoring the lateral mixing and advective acceleration terms
and assuming steady state conditions, and by expressing the two horizontal components of
the depth integrated flow as derivatives of a stream function ip, the two horizontal equations
of motion were reduced to one Poisson equation
vV = /(*,y) (5.13)
where the / term contains the forcing due to water level gradients, friction and wave forcing.
This equation represents a boundary value problem with two unknowns ip and fj. Upon
specifying boundary values an initial guess for the water surface fj is made and then Eq. 5.13
is solved for ip. With this new solution for ip the value of rj is updated by solving steady
82
state equations of motion. This process is repeated until convergence.
Figures 5.15 and 5.16 show the results for Liu and Mei and the present study respec
tively. Figure 5.15 shows the contours of the stream function whereas figure 5.16 shows
flow lines which give the direction of the flow but not the intensity of the flow. The present
model shows a separation eddy at the tip of the breakwater which is absent in Lius results.
Both models show a series of rip cells in the up wave side of the jetty which result from
the interaction of the incident and reflected waves. There are not as many rip cells in the
present model as the weaker ones farther upstream are buried by the stronger longshore
current. Neither the Liu model nor the present one show a rip current at the downwave side
of the jetty. In figure 5.16 the flow lines in the offshore region are directed in the onshore
 offshore direction which is different from the stream function contour lines in figure 5.15.
This is because of a difference in the offshore boundary condition in the two models. In
the Liu model the offshore condition is a constant xp which means that the flow must be
tangential to the offshore boundary. In the example used in the present model the offshore
condition is a constant water surface (fj = 0) which forces the tangential velocity to be zero
at the offshore boundary. An attempt to change the offshore condition to a zero normal
flow condition did not accurately conserve water within the computational domain, even
though convergence was achieved. Figure 5.17 shows the flow lines resulting from using no
onshore flow at the offshore boundary.
The second case is to simulate the experiments conducted at the Waterways Experiment
Station and reported by Hales (1980). A jetty perpendicular to the shoreline extended 15
feet out from the still water line on a plane beach with a slope of 1 on 20. The plane beach
extended beyond the jetty tip until a depth of one foot was reached and further offshore
the laboratory basin was a constant depth of one foot. Waves were sent at an angle by
use of a moveable wave paddle and wave guides which were set up to follow the wave rays
at the ends of the wave paddle. Figure 5.18 shows the experimental setup for these tests.
These wave guides in essence made the experiments to be in a closed basin and greatly
83
Figure 5.15: Results for the numerical model of Liu and Mei (1975). Deepwater wave height
equal to 1 meter, deepwater wave angle equal to 45 degrees. Source: Liu and Mei (1975).
84
Figure 5.16: Results from the present numerical model using the same conditions as were
used by Liu and Mei. Offshore boundary condition of constant water surface level which
precludes tangential flow at the offshore boundary.
85
Figure 5.17: Results from the present numerical model using the same conditions as were
used by Liu and Mei. Offshore boundary condition of no onoffshore flow. Grid spacing
equal to 20 meters.
86
Figure 5.18: Experimented setup for the tests conducted by Hales (1980). Source: Hales
(1980).
87
/
i
ce
J
L
L
L
S'
A
a
A
A
A
i
l
a
A
A A
A *
A A
>
1.0 FT/SEC
A
RVERPGE MIDDEPTH VELOCITIES
COMBINED REFRACTION QND DIFFRACTION
ANGLE OF INCIDENCE = 30 DEGREE?
WAVE PERIOD = 1.0 SECONDS
WAVE HEIGHT NEAR GENERATOR = 0.139 FEFT
Figure 5.19: Experimental measurements of the middepth averaged velocities obtain by
Hales. Measurements at one foot intervals. Source: Hales (1980).
affected some of the circulation patterns away from the jetty. Figure 5.19 shows the average
middepth velocities recorded on the down wave side of the jetty. Note the strong offshore
current next to the boundary. Since the boundaries are less than two jetty lengths from the
jetty, one must ask whether the circulation near the jetty was also affected. The numerical
simulation of the Hales experiments was performed using open lateral boundaries, since
to fully represent the experimental conditions in which the lateral boundaries are neither
straight (i.e. curving wave rays) nor perpendicular to the shoreline boundary would require
mapping techniques that are not yet included in the capability of the model.
Figure 5.20 shows a comparison of the experimented and numerical results obtained for
the magnitude of the rip current adjacent to the down wave side of the jetty for the case of
a 1 second wave with an angle of incidence of 30 degrees and a one foot depth wave height
88
Figure 5.20: Ripcurrent velocities adjacent to the down wave side of the jetty. Comparison
of experimental and numerical results.
of .139 ft. One would expect the depth averaged velocity to be somewhere between the
middepth velocity and the bottom velocity. The model appears to slightly underpredict
the rip current. In figure 5.19 a circulation cell in the lee of the groin is barely discernable.
In figure 5.21 the circulation cell is very obvious.
5.5 Shore Parallel Emerged Breakwater
In order to solve for the wave and current field resulting from the interaction of waves
and an emerged breakwater several boundary conditions at the breakwater are incorporated
into the model in addition to the boundary conditions already discussed in sections 4.3
and 4.4. The breakwater is represented as being infinitesimally thin and impermeable.
Figure 5.22 shows the position of the breakwater on the shoreward side of the I = Ibrk grid
row. The impermeability condition implies that neither waves nor currents pass through
the breakwater. The no current condition is easily imposed in the circulation model by
making Cf(lBjyr+i> J) equal to zero for the entire length of the breakwater. In addition the
circulation model is modified so that momentum diffusion and derivatives do not take place
89
Figure 5.21: Velocity vectors of the currents obtained numerically for the same conditions
as for the Hales experiment. Grid spacing equal to one foot.

Full Text 
PAGE 1
180(5,&$/ 02'(/,1* 2) :$9(,1'8&(' &855(176 86,1* $ 3$5$%2/,& :$9( (48$7,21 %\ +$5/(< 67$1)25' :,1(5 $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( 6&+22/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$ >81,9(56,7< 2) )/25,'$ /,%5$5,(6
PAGE 2
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fV SDUHQWV ZKR HQFRXUDJHG DQG SXVKHG KLP WR FRPSOHWH WKLV GRFWRUDWH 7KH DXWKRU ZRXOG DOVR OLNH WR WKDQN KLV WHDFKHUV IULHQGV DQG IHOORZ VWXGHQWV ZKR RIIHUHG HQFRXUDJHPHQW DQG LQWHUHVWLQJ GLVFXVVLRQV
PAGE 3
/DVWO\ WKH DXWKRU WKDQNV KLV VRXOPDWH DQG SDUWQHU (VWKHU 'HRQJ ZKR PDQDJHG WR EH SDWLHQW ZLWK HDFK H[WHQVLRQ RI WKH ILQDO GDWH %DE\ FDNHV LWfV 3$LQDOO\ 'RQH ORYH \RX LQ
PAGE 4
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
PAGE 5
:DYHV 2Q D 5LS&KDQQHO 6KRUH3HUSHQGLFXODU %UHDNZDWHU 6KRUH 3DUDOOHO (PHUJHG %UHDNZDWHU &RPSDULVRQ ZLWK ([SHULPHQWDO 5HVXOWV RI *RXUOD\ &21&/86,216 $1' 5(&200(1'$7,216 )25 )857+(5 678'< ,PSURYHG :DYH )RUPXODWLRQV ,QFOXGLQJ :DYH 5HIOHFWLRQ ([WHQGLQJ WKH 0RGHO WR ,QFOXGH 6HGLPHQW 7UDQVSRUW $33(1',&(6 $ '(5,9$7,21 2) 7+( '(37+ $9(5$*(' (48$7,216 % '(5,9$7,21 2) 7+( 5$',$7,21 675(66 7(506 & /$* 5$1* /$1 '(5,9$7,21 2) 7+( *29(51,1* :$9( (48$7,21 2%7$,1,1* (48$7,21 f %,%/,2*5$3+< %,2*5$3+,&$/ 6.(7&+ Y
PAGE 6
/,67 2) ),*85(6 &RPSXWDWLRQDO GRPDLQ RI WKH PRGHO )ORZ FKDUW RI FRPSXWHU SURJUDP )ORZ FKDUW RI FLUFXODWLRQ PRGHO )ORZ FKDUW RI ZDYH PRGHO /RQJVKRUH FXUUHQW SURILOHV IRU GLIIHUHQW IULFWLRQ IRUPXODWLRQV IRU DQ VHFRQG PHWHU ZDYH DW GHJUHHV WR D SODQH EHDFK ([SHULPHQWDO UHVXOWV IRU ZDYH VHWXS DQG VHWGRZQ 5HSULQWHG IURP %RZHQ ,QPDQ DQG 6LPPRQV f RXUQDO RI *HRSK\VLFDO 5HVHDUFK YRO f SDJH FRS\ULJKW E\ WKH $PHULFDQ *HRSK\VLFDO 8QLRQ 'HILQLWLRQ VNHWFK RI JULG QRPHQFODWXUH DQG FRRUGLQDWH V\VWHP 6WDUWXS IXQFWLRQ DFFRUGLQJ WR (T ZLWK &L DQG & :DYHV SDVVLQJ WKURXJK D ODWHUDO ERXQGDU\ XQDIIHFWHG E\ WKH ERXQGDU\ &RPSDULVRQ RI WKH QXPHULFDO VROXWLRQ RI 9HPXODNRQGD HW DO f IRU VHWXS ZLWK H[SHULPHQWDO GDWD RI %RZHQ f 6RXUFH 9HPXODNRQGD HW DO f &RPSDULVRQ RI QXPHULFDO UHVXOWV IURP WKH PRGHO ZLWK H[SHULPHQWDO UHn VXOWV RI +DQVHQ DQG 6YHQGVHQ f :DYH VHWXS DQG ZDYH KHLJKWV IRU D VHFRQG ZDYH RQ D EHDFK VORSH RI DQG LQFLGHQW ZDYH KHLJKW RI FHQWLPHWHUV 1RQGLPHQVLRQDO ORQJVKRUH YHORFLWLHV YUHVXV WKH QRQGLPHQVLRQDO GLVn WDQFH RIIVKRUH IURP WKH DQDO\WLFDO VROXWLRQ RI /RQJXHW+LJJLQV 5HSULQWHG IURP /RQJXHW+LJJLQV Ef RXUQDO RI *HRSK\VLFDO 5HVHDUFK YRO f SDJH FRS\ULJKW E\ WKH $PHULFDQ *HRSK\VLFDO 8QLRQ &RPSDULVRQ RI H[SHULPHQWDO UHVXOWV RI 9LVVHU f DQG QXPHULFDO UHVXOW IRU ORQJVKRUH FXUUHQWV GXH WR D VHFRQG ZDYH RQ D VORSH RI IRU WKH GHHS ZDWHU FRQGLWLRQV + P DQG D GHJUHHV 7RS 'HSWK FRQWRXUV LQ PHWHUV XVLQJ (T ZLWK WKH FRRUGLQDWHV DV VKRZQ %RWWRP 'HSWK FRQWRXUV DV XVHG LQ WKH SUHVHQW PRGHO *ULG VSDFLQJ LV PHWHUV YL
PAGE 7
9HORFLW\ YHFWRUV IURP WKH UHVXOWV RI (EHUVROH DQG 'DOU\PSOH IRU D ZDYH DQJOH RI GHJUHHV 6FDOH IRU YHORFLW\ YHFWRUV LV LQFK PVHF 6RXUFH (EHUVROH DQG 'DOU\PSOH f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f 'HHSZDWHU ZDYH KHLJKW HTXDO WR PHWHU GHHSZDWHU ZDYH DQJOH HTXDO WR GHJUHHV 6RXUFH /LX DQG 0HL f 5HVXOWV IURP WKH SUHVHQW QXPHULFDO PRGHO XVLQJ WKH VDPH FRQGLWLRQV DV ZHUH XVHG E\ /LX DQG 0HL 2IIVKRUH ERXQGDU\ FRQGLWLRQ RI FRQVWDQW ZDn WHU VXUIDFH OHYHO ZKLFK SUHFOXGHV WDQJHQWLDO IORZ DW WKH RIIVKRUH ERXQGDU\ 5HVXOWV IURP WKH SUHVHQW QXPHULFDO PRGHO XVLQJ WKH VDPH FRQGLWLRQV DV ZHUH XVHG E\ /LX DQG 0HL 2IIVKRUH ERXQGDU\ FRQGLWLRQ RI QR RQRIIVKRUH IORZ *ULG VSDFLQJ HTXDO WR PHWHUV ([SHULPHQWDO VHWXS IRU WKH WHVWV FRQGXFWHG E\ +DOHV f 6RXUFH +DOHV f ([SHULPHQWDO PHDVXUHPHQWV RI WKH PLGGHSWK DYHUDJHG YHORFLWLHV REWDLQ E\ +DOHV 0HDVXUHPHQWV DW RQH IRRW LQWHUYDOV 6RXUFH +DOHV f 5LSFXUUHQW YHORFLWLHV DGMDFHQW WR WKH GRZQ ZDYH VLGH RI WKH MHWW\ &RPn SDULVRQ RI H[SHULPHQWDO DQG QXPHULFDO UHVXOWV 9HORFLW\ YHFWRUV RI WKH FXUUHQWV REWDLQHG QXPHULFDOO\ IRU WKH VDPH FRQn GLWLRQV DV IRU WKH +DOHV H[SHULPHQW *ULG VSDFLQJ HTXDO WR RQH IRRW 'HILQLWLRQ VNHWFK RI JULGV IRU WKH FDVH RI D VKRUH SDUDOOHO EUHDNZDWHU YX
PAGE 8
:DYH FUHVWV LQ WKH OHH RI D VHPLLQILQLWH EUHDNZDWHU LQ D UHJLRQ RI FRQVWDQW GHSWK 9HORFLW\ YHFWRUV REWDLQHG IURP WKH PRGHO IRU D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH IRU D PHWHU VHFRQG ZDYH IRU LQFUHDVLQJ DQJOHV RI LQFLGHQFH 6WUHDP IXQFWLRQ FRQWRXU OLQHV REWDLQHG IURP WKH /LX PRGHO IRU D VHFRQG PHWHU ZDYH QRUPDOO\ DSSURDFKLQJ D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH RQ D RQ VORSH 6RXUFH /LX DQG 0HL f )ORZ OLQHV REWDLQHG IURP WKH SUHVHQW PRGHO IRU D VHFRQG PHWHU ZDYH QRUPDOO\ DSSURDFKLQJ D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH RQ D RQ VORSH 9HORFLW\ YHFWRUV REWDLQHG IURP WKH SUHVHQW PRGHO IRU D VHFRQG PHWHU ZDYH QRUPDOO\ DSSURDFKLQJ D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH RQ D RQ VORSH 6HWXS FRQWRXU OLQHV REWDLQHG IURP WKH SUHVHQW PRGHO IRU D VHFRQG PHWHU ZDYH QRUPDOO\ DSSURDFKLQJ D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH RQ D RQ VORSH 3K\VLFDO OD\RXW IRU WKH H[SHULPHQW RI *RXUOD\ 6RXUFH *RXUOD\ f 3URFHHGLQJV rA &RDVWDO (QJLQHHULQJ &RQIHUHQFH FRS\ULJKW E\ WKH $PHUn LFDQ 6RFLHW\ RI &LYLO (QJLQHHUV UHSULQWHG ZLWK SHUPLVVLRQ &RPSXWDWLRQDO GRPDLQ IRU WKH H[SHULPHQW RI *RXUOD\ *ULG VSDFLQJ HTXDO WR PHWHU &RPSDULVRQ RI H[SHULPHQWDO DQG QXPHULFDO UHVXOWV IRU WKH H[SHULPHQW RI *RXUOD\ :DYH 6HWXS &RQWRXUV 7RS ILJXUH LV IURP *RXUOD\ f 3URn FHHGLQJV A &RDVWDO (QJLQHHULQJ &RQIHUHQFH FRS\ULJKW E\ WKH $PHUn LFDQ 6RFLHW\ RI &LYLO (QJLQHHUV UHSULQWHG ZLWK SHUPLVVLRQ 7KH ERWWRP ILJXUH LV WKH FRPSXWHU UHVXOWV 5HVXOWV IURP WKH H[SHULPHQW RI *RXUOD\ &RQWRXUV RI FXUUHQW YHORFLn WLHV DQG VWUHDPOLQHV 6RXUFH *RXUOD\ f 3URFHHGLQJV A &RDVWDO (QJLQHHULQJ &RQIHUHQFH FRS\ULJKW E\ WKH $PHULFDQ 6RFLHW\ RI &LYLO (Qn JLQHHUV UHSULQWHG ZLWK SHUPLVVLRQ 5HVXOWV IURP WKH QXPHULFDO PRGHO 6WUHDPOLQHV RI WKH IORZ 5HVXOWV IURP WKH QXPHULFDO PRGHO &RQWRXU OLQHV RI WKH FXUUHQW YHORFLn WLHV YLLL
PAGE 9
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n WLRQDO GRPDLQ ZLWK WKH SUHVHQFH RI EUHDNZDWHUV 7KH PRGHO LWHUDWHV EHWZHHQ D ZDYH PRGHO DQG D FLUFXODWLRQ PRGHO 7KH ZDYH PRGHO XVHV D SDUDEROLF DSSUR[LPDWLRQ WR D FRPELQHG UHIUDFWLRQGLIIUDFWLRQ PLOG VORSH HTXDWLRQ ZKLFK LQFOXGHV FXUUHQWV 7KH FLUFXODWLRQ PRGHO XVHV DQ DOWHUQDWLQJ GLUHFWLRQ LPSOLFLW PHWKRG VROXWLRQ WR WKH GHSWK DYHUDJHG HTXDWLRQV RI PRPHQWXP XVLQJ WKH UDGLDWLRQ VWUHVV WHUPV REWDLQHG WKURXJK VROXWLRQ RI WKH ZDYH PRGHO DV WKH IRUFLQJ WHUPV (DFK RI WKH WHUPV LQ WKH GHSWK DYHUDJHG HTXDWLRQV RI PRPHQWXP LQYROYH DVVXPSWLRQV WKDW DUH GLVFXVVHG LQ WKH UHSRUW 7KH PRGHO LV FRQVWUXFWHG LQ PRGXODU IRUP VR WKDW LPSURYHPHQWV FDQ HDVLO\ EH LQFRUSRUDWHG LQWR WKH PRGHO 7KH PRGHO ZDV HPSOR\HG IRU VHYHUDO WHVW FDVHV VXFK DV ULS FKDQQHOV EDU JDSV DQG ERWK VKRUH SHUSHQGLFXODU DQG VKRUH SDUDOOHO EUHDNZDWHUV *RRG DJUHHPHQW ZDV REWDLQHG ZKHQ WKH PRGHO ZDV XVHG WR VLPXODWH D ODERUDWRU\ H[SHULPHQW RI 0LFKDHO 5 *RXUOD\ f:DYH VHWXS DQG ZDYH JHQHUDWHG FXUUHQWV LQ WKH OHH RI D EUHDNZDWHU RU KHDGODQGf 3URF rA &RDVWDO (QJ &RQI &RSHQKDJHQ SS f ,;
PAGE 10
6XJJHVWLRQV DUH PDGH IRU LPSURYLQJ WKH PRGHO DQG IRU LQFOXGLQJ VHGLPHQW WUDQVSRUW LQ WKH PRGHO
PAGE 11
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n GLFWLRQV RI ZDYHV DQG FXUUHQWV DQG ZDWHU HOHYDWLRQV LQ WKH FRDVWDO URQH )RU LQVWDQFH SUHn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
PAGE 12
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n ZD\V VDWLVILHG LQ QDWXUH /RQJXHW+LJJLQV DQG 6WHZDUW f JLYH DQ DQDO\WLF VROXWLRQ IRU ZDYH VHWXS IRU VWHDG\ VWDWH FRQGLWLRQV RQ D SODQH EHDFK ZLWK VWUDLJKW DQG SDUDOOHO FRQn WRXUV DVVXPLQJ OLQHDU ZDYHV DSSURDFKLQJ WKH EHDFK QRUPDOO\ DQG VSLOOLQJ EUHDNHUV ZKHUH WKH ZDYH KHLJKW VKRUHZDUG RI WKH EUHDNHU OLQH LV D FRQVWDQW WLPHV WKH WRWDO GHSWK RI WKH ZDWHU FROXPQ /RQJXHW+LJJLQV Df JLYHV D FURVV VKRUH SURILOH RI WKH ORQJVKRUH FXUUHQW IRU D SODQH EHDFK ZLWK VWUDLJKW DQG SDUDOOHO FRQWRXUV IRU OLQHDU ZDYHV DSSURDFKLQJ WKH EHDFK DW DQ DQJOH VSLOOLQJ EUHDNHUV DQG OLQHDUL]HG ERWWRP IULFWLRQ ZKLOH QHJOHFWLQJ WKH LQIOXHQFH ZDYH VHWXS KDV XSRQ WKH WRWDO GHSWK RI WKH ZDWHU FROXPQ ,Q D FRPSDQLRQ SDSHU /RQJXHW +LJJLQV Ef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
PAGE 13
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fÂ§ ?>/Y ZKHUH WKH VXEVFULSW U LV IRU WKH UDWLR RI WKH YDOXH LQ WKH PRGHO WR WKH YDOXH LQ WKH SURWRW\SH DQG 7 DQG /Y DUH WLPH DQG YHUWLFDO OHQJWK UHVSHFWLYHO\ )RU GLIIUDFWLRQ 7U aQ LQ VKDOORZ ZDWHU DQG 7U ?/MMW LQ GHHS ZDWHU ZKHUH /MMU UHSUHVHQWV \/YU WKH KRUL]RQWDO OHQJWK VFDOH 1XPHULFDO PRGHOV GR QRW KDYH PDQ\ RI WKH GLVDGYDQWDJHV IRXQG ZLWK SK\VLFDO PRGHOV $Q\ PHFKDQLVP IRU ZKLFK WKH JRYHUQLQJ HTXDWLRQV DUH NQRZQ FDQ EH PRGHOHG QXPHULFDOO\
PAGE 14
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f &RPSXWHU WLPH DQG FRVW IRU ODUJH SURJUDPV FDQ RIWHQ EH VLJQLILFDQW /LWHUDWXUH 7KLV VHFWLRQ LV D EULHI KLVWRU\ DQG OLWHUDWXUH UHYLHZ RI WKH DWWHPSWV RI WKH FRDVWDO HQJLQHHULQJ SURIHVVLRQ WR REWDLQ SUHGLFWLRQV RI ZDYHV DQG FXUUHQWV LQ WKH QHDUVKRUH UHJLRQ JLYHQ WKH RIIVKRUH ZDYH FRQGLWLRQV 7KH H[WHQVLRQ RI 6QHOOfV ODZ ZKLFK JRYHUQV RSWLFDO ZDYH UHIUDFWLRQ WR WKH DQDORJRXV SUREOHP RI ZDWHU ZDYHV RYHU VWUDLJKW DQG SDUDOOHO FRQWRXUV KDV EHHQ NQRZQ IRU VRPH WLPH RKQVRQ 2f%ULHQ DQG ,VDDFV f UHSRUWHG JUDSKLFDO PHWKRGV IRU FRQVWUXFWLQJ ZDYH UHIUDFWLRQ GLDJUDPV 7ZR PHWKRGV ZHUH UHSRUWHG RQH SURGXFLQJ ZDYH IURQWV DQG WKH VHFRQG
PAGE 15
SURGXFLQJ ZDYH UD\V 3LHUVRQ f VKRZHG WKDW WKH PHWKRG XVLQJ ZDYH IURQWV LQWURGXFHG PRUH HUURU WKDQ WKH ZDYH UD\ PHWKRG 7KHVH PHWKRGV SURFHHGHG XSRQ WKH DVVXPSWLRQ WKDW WKH FRQWRXUV DUH ORFDOO\ VWUDLJKW 7KH ZDYH UD\V REWDLQHG ZHUH LPSRUWDQW LQ WKDW WKH ZDYH KHLJKW ZDV GHWHUPLQHG WR YDU\ LQYHUVHO\ DV WKH VTXDUH URRW RI WKH VSDFLQJ EHWZHHQ WKH ZDYH UD\V 8VLQJ WKHVH UD\ FRQVWUXFWLRQ WHFKQLTXHV 2f%ULHQ f ZDV DEOH WR GHPRQVWUDWH WKDW WKH $SULO GHVWUXFWLRQ RI WKH /RQJ %HDFK +DUERU EUHDNZDWHU RQ D VHHPLQJO\ FDOP GD\ PRVW SUREDEO\ UHVXOWHG IURP WKH IRFXVLQJ RI ORQJ ZDYH HQHUJ\ IURP D 6RXWK 3DFLILF VWRUP 0XQN DQG $UWKXU f VKRZHG D PHWKRG WR REWDLQ WKH ZDYH KHLJKW GLUHFWO\ IURP WKH FXUYDWXUH RI WKH UD\ DQG WKH UDWH RI FKDQJH RI WKH GHSWK FRQWRXUV 7KH ZDYH UD\ WUDFLQJ WHFKQLTXHV GLG QRW LQFOXGH WKH HIIHFWV RI FXUUHQWV DQG KDG WKH GLVDGYDQWDJH IURP D QXPHULFDO PRGHOLQJ SRLQW RI YLHZf WKDW ZDYH KHLJKWV DQG GLUHFWLRQV ZHUH REWDLQHG DW SRVLWLRQV DORQJ D UD\ DQG QRW DW SUHVSHFLILHG JULG SRVLWLRQV 1RGD HW DO f VROYHG WKLV SUREOHP E\ GHYLVLQJ D QXPHULFDO VFKHPH WKDW VROYHG D ZDYH HQHUJ\ HTXDWLRQ DQG D VWDWHPHQW IRU WKH LUURWDWLRQDOLW\ RI WKH ZDYH QXPEHU WR REWDLQ ZDYH KHLJKW DQG GLUHFWLRQ DW JULG SRLQWV 7KLV VFKHPH DOVR LQFOXGHG WKH HIIHFWV RI FXUUHQW UHIUDFWLRQ EXW QHJOHFWHG GLIIUDFWLRQ 1XPHULFDO PRGHOV RI ZDYH LQGXFHG FXUUHQWV ZHUH GHYHORSHG E\ 1RGD HW DO f LQ ZKLFK WLPH GHSHQGHQW DQG DGYHFWLYH DFFHOHUDWLRQ WHUPV ZHUH LJQRUHG %LUNHPHLHU DQG 'DLn U\ PSOH f ZURWH D WLPHPDUFKLQJ H[SOLFLW PRGHO ZKLFK QHJOHFWHG ODWHUDO PL[LQJ DQG DGYHFWLYH DFFHOHUDWLRQ (EHUVROH DQG 'DOU\PSOH f DGGHG ODWHUDO PL[LQJ DQG WKH QRQn OLQHDU DGYHFWLYH DFFHOHUDWLRQ WHUPV 9HPXODNRQGD f ZURWH DQ LPSOLFLW PRGHO ZKLFK LQFOXGHG ODWHUDO PL[LQJ DQG DGYHFWLYH DFFHOHUDWLRQ WHUPV $OO RI WKH DERYH XVHG WKH QXPHUn LFDO VFKHPH RI 1RGD HW DO f WR REWDLQ ZDYH KHLJKWV DQG ZDYH DQJOHV (DFK RI WKHVH PRGHOV QHJOHFWHG GLIIUDFWLRQ 'LIIUDFWLRQ RI ZDWHU ZDYHV UHSUHVHQWV VRPH YHU\ GLIILFXOW PDWKHPDWLFDO SUREOHPV +LVn WRULFDOO\ PRVW VROXWLRQV ZHUH EDVHG XSRQ WKH DVVXPSWLRQ RI D IODW ERWWRP VLQFH WR TXRWH 1HZPDQ f 7KH UHVWULFWLRQ WR FRQVWDQW GHSWK K LV LPSRUWDQW 3UREOHPV LQYROYLQJ ZDYH SURSDJDWLRQ LQ D IOXLG RI YDULDEOH GHSWK DUH RI JUHDW LQWHUHVW DQG WKH UHVXOWLQJ
PAGE 16
GLIIUDFWLRQ HIIHFWV DUH DPRQJ WKH PRVW LPSRUWDQW RI DOO EXW WKHVH SUREOHPV DUH OHVV WUDFWDEOH WKDQ WKRVH LQYROYLQJ D IOXLG RI FRQVWDQW GHSWK SDJH f 3HQQH\ DQG 3ULFH f VKRZHG WKDW 6RPPHUIHOGfV VROXWLRQ RI WKH RSWLFDO GLIIUDFWLRQ SUREn OHP DV GHVFULEHG E\ %DWHPDQ f LV DOVR D VROXWLRQ RI WKH FRQVWDQW GHSWK ZDWHU ZDYH GLIIUDFWLRQ SUREOHP 3HQQH\ DQG 3ULFH f VROYHG WKH FRQVWDQW GHSWK GLIIUDFWLRQ SUREOHP IRU EUHDNZDWHUV DVVXPLQJ D VHPLLQILQLWHO\ ORQJ LQILQLWHVLPDOO\ WKLQ EDUULHU 7KH VROXWLRQ LV LQ WHUPV RI )UHVQHO LQWHJUDOV ZKLFK QHHG WR EH HYDOXDWHG QXPHULFDOO\ 7KH 6KRUH 3URWHFn WLRQ 0DQXDO 630f SURGXFHG PDQ\ JUDSKLFDO VROXWLRQV EDVHG XSRQ 3HQQH\ DQG 3ULFH f WKDW JLYH WKH ZDYH KHLJKW DQG GLUHFWLRQ LQ WKH OHH RI EUHDNZDWHUV 7KH SUREOHP RI FRPELQHG UHIUDFWLRQ DQG GLIIUDFWLRQ DFFRUGLQJ WR WKH 630 FRXOG EH KDQGOHG E\ DVVXPLQJ D FRQVWDQW GHSWK IRU D IHZ ZDYHOHQJWKV VR DV WR REWDLQ WKH GLIIUDFWLRQ HIIHFWV DQG WKHQ FRQWLQXH ZLWK WKH UHIUDFWLRQ HIIHFWV 3HUOLQ DQG 'HDQ f XVHG HVVHQWLDOO\ WKH VDPH SURFHGXUH LQ D QXPHULFDO PRGHO ZKLFK FDOFXODWHG WKH ORQJVKRUH WKUXVW LQ WHUPV RI WKH EUHDNLQJ ZDYH KHLJKW DQG DQJOH /LX DQG 0HL f VROYHG IRU WKH ZDYH ILHOG DURXQG VKRUH SDUDOOHO DQG VKRUH SHUSHQGLFXn ODU EUHDNZDWHUV E\ XVLQJ D FXUYLOLQHDU FRRUGLQDWH V\VWHP ZKHUH RQH FRRUGLQDWH FRUUHVSRQGHG WR WKH ZDYH UD\ DQG WKH RWKHU WR D FXUYH RI FRQVWDQW SKDVH 7KHLU VROXWLRQ ZDV IRU fDQ DPn SOLWXGH PRGXODWLRQ IDFWRU 'eIf ZKLFK DFFRXQWV IRU GLIIUDFWLRQf SDJH f ZKHUH e DQG I DUH WKH WZR FRRUGLQDWHV $Q HTXDWLRQ IRU 'eIf ZDV REWDLQHG DQG D ERXQGDU\ YDOXH SUREn OHP ZDV HVWDEOLVKHG E\ PDWFKLQJ FRQGLWLRQV LQ WKH LQFLGHQW ]RQH DQG WKH VKDGRZ ]RQHV )RU FRQVWDQW GHSWK WKH VROXWLRQ IRU e!eIf LV LQ WHUPV RI )UHVQHO LQWHJUDOV 8SRQ REWDLQLQJ WKH ZDYH ILHOG WKH UDGLDWLRQV VWUHVVHV ZHUH REWDLQHG DQG XVHG WR IRUFH D FLUFXODWLRQ PRGHO WKDW VROYHG IRU WKH VWUHDP IXQFWLRQ DQG PHDQ ZDWHU VXUIDFH OHYHO ZKLOH LJQRULQJ WKH QRQOLQHDU DGYHFWLYH DFFHOHUDWLRQ WHUPV DQG WKH ODWHUDO GLIIXVLRQ RI FXUUHQW PRPHQWXP $OWKRXJK /LX DQG 0HL SURGXFHG UHVXOWV IRU WKH ZDYH ILHOG FXUUHQWV DQG VHWXS UHVXOWLQJ IURP WKH GLIIUDFn WLRQ RI ZDYHV E\ EUHDNZDWHUV WKHLU VWXG\ LV QRW D ZDYHFXUUHQW LQWHUDFWLRQ PRGHO VLQFH WKHUH DUH QR FXUUHQW WHUPV LQ WKH ZDYH HTXDWLRQV DQG QR IHHGEDFN RI FXUUHQW WHUPV IURP WKH FLUFXODWLRQ PRGHO
PAGE 17
%HUNKRII f GHULYHG DQ HTXDWLRQ JRYHUQLQJ FRPELQHG UHIUDFWLRQ DQG GLIIUDFWLRQ ZLWK WKH DVVXPSWLRQ RI D PLOG VORSH 5DGGHU f DSSOLHG WKH VSOLWWLQJ PHWKRG RI 7DSSHUW f WR REWDLQ D SDUDEROLF HTXDWLRQ JRYHUQLQJ WKH SURSDJDWLRQ RI WKH IRUZDUG SURSDJDWn LQJ FRPSRQHQW RI WKH ZDYH ILHOG %RRLM f H[WHQGHG WKH ZRUN RI %HUNKRII WR LQFOXGH WKH HIIHFWV RI FXUUHQWV DQG DOVR REWDLQHG D SDUDEROLF DSSUR[LPDWLRQ .LUE\ f PDGH FRUUHFWLRQV WR WKH HTXDWLRQ RI %RRLM 3DUDEROLF ZDYH HTXDWLRQV KDYH UHFHQWO\ EHHQ XVHG LQ ZDYHLQGXFHG FLUFXODWLRQ PRGHOV /LX DQG 0HL f VROYHG IRU WKH ZDYH ILHOG E\ PDWFKLQJ FRQGLWLRQV EHWZHHQ WKH VKDGRZ ]RQHV DQG WKH LOOXPLQDWHG ]RQH LQ WKH OHH RI D VWUXFWXUH 7KH ZDYH LQIRUPDWLRQ ZDV WKHQ XVHG WR IRUFH D FLUFXODWLRQ PRGHO ZKLFK IRUPXODWHG WKH JRYHUQLQJ GHSWK LQWHJUDWHG WLPH DYHUDJHG PRPHQWXP HTXDWLRQV LQ WHUPV RI D VWUHDP IXQFWLRQ +RZHYHU WKLV PRGHO GLG QRW KDYH FXUUHQW IHHGEDFN LQ WKH ZDYH HTXDWLRQ
PAGE 18
)LJXUH &RPSXWDWLRQDO GRPDLQ RI WKH PRGHO *HQHUDO 'HVFULSWLRQ RI WKH 0RGHO 7KLV VHFWLRQ LV D GHVFULSWLRQ RI WKH QXPHULFDO PRGHO WKDW KDV EHHQ FRQVWUXFWHG WR FDOFXn ODWH WKH ZDYH DQG FXUUHQW ILHOG ZLWKLQ D JLYHQ GRPDLQ 7KLV GRPDLQ XVXDOO\ LV ERUGHUHG E\ D VKRUHOLQH DORQJ RQH ERXQGDU\ DQG DQ RIIVKRUH UHJLRQ DORQJ WKH RSSRVLWH ERXQGDU\ 7KH RWKHU ERXQGDULHV DUH VXIILFLHQWO\ IDU IURP WKH UHJLRQ RI LQWHUHVW LH D VWUXFWXUH RU VKRDO HWFf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
PAGE 19
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f PHWKRG RI VROXWLRQ 7KUHH HTXDWLRQV DUH VROYHG IRU WKUHH XQNQRZQV 7KH XQNQRZQV DUH IM WKH PHDQ ZDWHU VXUIDFH OHYHO DQG 8 DQG 9 WKH WZR FRPSRQHQWV RI WKH GHSWK DYHUDJHG PHDQ FXUUHQW 7KH HTXDWLRQV DUH WKH FRQWLQXLW\ HTXDWLRQ DQG WKH [ DQG \GLUHFWLRQ PRPHQWXP HTXDWLRQV (DFK LWHUDWLRQ RI WKH FLUFXODWLRQ PRGHO VROYHV IRU WKH WKUHH XQNQRZQV RQ WKH QH[W WLPH OHYHO ,Q RWKHU ZRUGV HDFK LWHUDWLRQ DGYDQFHV WKH YDOXHV RI IM 8 DQG 9 E\ D WLPH LQFUHPHQW $W )RU WKH SXUSRVHV RI WKLV LQYHVWLJDWLRQ WKH PRGHO VROYHV IRU VWHDG\ VWDWH FRQGLWLRQV XVLQJ WKH WLPH PDUFKLQJ FKDUDFWHU RI WKH FLUFXODWLRQ PRGHO 7KH PRGHO FDQ EH XVHG WR VROYH IRU WLPH GHSHQGHQW ZDYH ILHOGV EXW WKLV ZRXOG UHTXLUH WKH VXEVWLWXWLRQ RI D WLPH GHSHQGHQW ZDYH HTXDWLRQ LQ WKH ZDYH PRGHO 7KH $', PHWKRG ILUVW VROYHV JULG URZ E\ JULG URZ IRU 8 DW WKH QH[W WLPH OHYHO DQG DQ LQWHUPHGLDWH YDOXH RI IM E\ VROYLQJ WKH [ GLUHFWLRQ HTXDWLRQ RI PRWLRQ DQG WKH FRQWLQXLW\ HTXDWLRQ XVLQJ DQ LPSOLFLW QXPHULFDO VFKHPH 7KLV VROXWLRQ LV LQ WHUPV RI WKH NQRZQ YDOXHV RI 9DQG IM DW WKH SUHVHQW WLPH OHYHO 2QFH DOO WKH JULG URZV DUH VROYHG LQ WKH [ GLUHFWLRQ WKH \ HTXDWLRQ RI PRWLRQ DQG WKH FRQWLQXLW\ HTXDWLRQ DUH VROYHG DJDLQ E\ DQ LPSOLFLW VFKHPH JULG FROXPQ E\ JULG FROXPQ IRU 9 DQG IM DW WKH QH[W WLPH OHYHO LQ WHUPV RI 8 DW WKH QHZ WLPH OHYHO IM DW WKH LQWHUPHGLDWH OHYHO DQG 9 DW WKH SUHVHQW OHYHO )RU HDFK RI WKH LPSOLFLW VFKHPHV LQ WKH WZR GLUHFWLRQV WKH JRYHUQLQJ HTXDWLRQV \LHOG D WULGLDJRQDO PDWUL[ 7KH WULGLDJRQDO PDWUL[ DQG WKH PHWKRG RI VROXWLRQ ZLOO EH GLVFXVVHG ODWHU 7KH ZDYH PRGHO DOVR \LHOGV D WULGLDJRQDO PDWUL[ HTXDWLRQ 7KH JRYHUQLQJ HTXDWLRQ LV D SDUDEROLF HTXDWLRQ IRU WKH FRPSOH[ DPSOLWXGH $ 7KH FRPSOH[ DPSOLWXGH FRQWDLQV SKDVH
PAGE 20
LQIRUPDWLRQ PHDQLQJ DPRQJ RWKHU WKLQJV WKDW \ YDULDWLRQV RI WKH ZDYH ILHOG DUH FRQWDLQHG ZLWKLQ WKH FRPSOH[ DPSOLWXGH 7KH WHUP fSDUDEROLF HTXDWLRQf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n WLRQV 7KHVH ZLOO EH GLVFXVVHG LQ GHWDLO LQ WKH ILQDO FKDSWHU RXWOLQLQJ UHFRPPHQGDWLRQV IRU IXUWKHU VWXG\ 7KH PRGXODU FRQVWUXFWLRQ RI WKH PRGHO DOVR DOORZV IRU HDVH LQ VXEVWLWXWLQJ RQH ZDYH PRGHO IRU DQRWKHU )LJXUH VKRZV D IORZ FKDUW RI WKH FRPSXWHU SURJUDP 7KH VXEURXWLQH ,1387 LV FDOOHG WR LQLWLDOL]H WKH YDOXHV RI WKH XQNQRZQV DQG WR LQSXW WKH ZDWHU GHSWK WKH RIIVKRUH ZDYH KHLJKW DQG GLUHFWLRQ DQG WKH SRVLWLRQ RI DQ\ VWUXFWXUHV 7KH WKUHH VXEURXWLQHV :$9(
PAGE 21
)LJXUH )ORZ FKDUW RI FRPSXWHU SURJUDP
PAGE 22
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f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
PAGE 23
)LJXUH )ORZ FKDUW RI FLUFXODWLRQ PRGHO
PAGE 24
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f &KDSWHU GHVFULEHV WKH FLUFXODWLRQ PRGHO LQ JUHDWHU GHWDLO ,QGLYLGXDO VHFWLRQV ZLOO FRYHU WKH JRYHUQLQJ HTXDWLRQV DQG WKH IRUPXODWLRQ RI HDFK RI WKH FRPSRQHQW WHUPV &KDSn WHU GHVFULEHV WKH ZDYH PRGHO WKH GHULYDWLRQ RI WKH OLQHDU ZDYH HTXDWLRQ WKH SDUDEROLF
PAGE 25
DSSUR[LPDWLRQ WKH PDQQHU RI LQFOXGLQJ ZDYH HQHUJ\ GLVVLSDWLRQ GXH WR ZDYH EUHDNLQJ DQG WKH PHWKRG RI REWDLQLQJ WKH UDGLDWLRQ VWUHVV WHUPV GLUHFWO\ IURP WKH FRPSOH[ DPSOLWXGH 7KH IRXUWK FKDSWHU JLYHV GHWDLOV RI WKH QXPHULFDO PHWKRGV XVHG WR REWDLQ VROXWLRQ RI WKH JRYHUQLQJ HTXDWLRQV DQG WKH ERXQGDU\ FRQGLWLRQV HPSOR\HG 7KH ILIWK FKDSWHU SUHVHQWV WKH UHVXOWV RI DSSOLFDWLRQV RI WKH PRGHO DQG FRPSDULVRQV ZLWK RWKHU VWXGLHV ERWK QXPHULFDO DQG H[SHULPHQWDO DV ZHOO DV VRPH DQDO\WLFDO VROXWLRQV WR VRPH LGHDOL]HG FDVHV 7KH ILQDO FKDSWHU VXPPDUL]HV WKH VWXG\ DQG PDNHV UHFRPPHQGDWLRQV IRU IXWXUH UHVHDUFK
PAGE 26
&+$37(5 &,5&8/$7,21 02'(/ *RYHUQLQJ (TXDWLRQV 7KH JRYHUQLQJ HTXDWLRQV IRU WKH FLUFXODWLRQ PRGHO DUH WKH GHSWK LQWHJUDWHG WLPH DYHUn DJHG KRUL]RQWDO HTXDWLRQV RI PRPHQWXP G8 77G8 fG8 GIM D7 XOL DO M'Uf a M'U B GVW[ DV[Y? GUO S' ? G[ G\ f S G\ G9 7G9 69 GIM D7 Y \ 7Gnn GW G[ BLB S' ? G[ G\ f S G[ f f DQG WKH FRQWLQXLW\ HTXDWLRQ f ZKHUH 8 [ FRPSRQHQW RI WKH PHDQ FXUUHQW 9 \ FRPSRQHQW RI WKH PHDQ FXUUHQW IM PHDQ ZDWHU VXUIDFH HOHYDWLRQ S ZDWHU GHQVLW\ WKH WRWDO ZDWHU GHSWK K IM K ORFDO VWLOOZDWHU GHSWK UM WKH ODWHUDO VKHDU VWUHVV GXH WR WXUEXOHQW PL[LQJ 7L[ [ FRPSRQHQW RI WKH ERWWRP VKHDU VWUHVV
PAGE 27
UN\ \ FRPSRQHQW RI WKH ERWWRP VKHDU VWUHVV W[ [ FRPSRQHQW RI WKH VXUIDFH VKHDU VWUHVV WY \ FRPSRQHQW RI WKH VXUIDFH VKHDU VWUHVV DQG rA[\ AQG 6YY DUH WKH UDGLDWLRQ VWUHVV FRPSRQHQWV ZKLFK DULVH IURP WKH H[FHVV PRPHQWXP IOX[ GXH WR ZDYHV ,W VKRXOG EH QRWHG WKDW 8 DQG 9 LQFOXGH WKH ZDYH LQGXFHG PDVV WUDQVSRUW 7KHVH HTXDWLRQV DUH REWDLQHG E\ LQWHJUDWLQJ WKH ORFDO [ DQG \ PRPHQWXP HTXDWLRQV DQG WKH FRQWLQXLW\ HTXDWLRQ RYHU WKH GHSWK RI WKH ZDWHU FROXPQ DQG WKHQ WLPH DYHUDJLQJ WKH UHVXOWV 7KLV LV GHPRQVWUDWHG LQ $SSHQGL[ $ ZKLFK IROORZV FORVHO\ WKH ZRUN RI (EHUVROH DQG 'DOU\PSOH f )RU JUHDWHU GHWDLOV RI WKH GHULYDWLRQ WKH UHDGHU LV UHIHUUHG WR (EHUVROH DQG 'DOU\PSOH f RU 3KLOOLSV f 7KHVH DUH WKH VDPH HTXDWLRQV WKDW KDYH EHHQ XVHG E\ RWKHU PRGHOHUV VXFK DV %LUNH PHLHU DQG 'DOU\PSOH f (EHUVROH DQG 'DOU\PSOH f 9HPXODNRQGD f DQG
PAGE 28
HTXDWLRQV 7KLV DVVXPSWLRQ LPSOLHV WKDW YHUWLFDO IOXFWXDWLRQV RI WKH IORZ ZKLFK DUH ]HUR LQ WKH WLPH DYHUDJH FDQ EH LJQRUHG 0DQ\ GHWDLOV RI WKH IORZ PD\ EH PLVVHG E\ WKLV DVVXPSWLRQ RI SXUHO\ KRUL]RQWDO IORZ 2I FRXUVH UHGXFLQJ WKH SUREOHP WR D WZRGLPHQVLRQDO IORZ KDV WKH GLVWLQFW DGYDQWDJH RI UHGXFLQJ DQ HVVHQWLDOO\ LQWUDFWDEOH SUREOHP WR D VLPSOHU WUDFWDEOH IRUP 6RPH RI WKH WHUPV LQ (TV f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f DQG DOVR XVHG E\ (EHUVROH DQG 'HGU\PSOH f IROORZLQJ WKH ZRUN RI 9DQ 'RUQ f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f %LUNHPHLHU DQG 'DOU\PSOH f DQG 9HPXODNRQGD f DOO XVHG WKH ZHDN FXUUHQW IRUPXODWLRQ DV SURSRVHG E\ /RQJXHW+LJJLQV
PAGE 29
Df LQ ZKLFK 7E] S) _XRUO_ 8 f 7LY S) _XRUL_ 9 f ZKHUH ) LV D GUDJ FRHIILFLHQW RI WKH RUGHU RI f DQG _XRUM_ LV WKH WLPH DYHUDJH RYHU D ZDYH SHULRG RI WKH DEVROXWH YDOXH RI WKH ZDYH RUELWDO YHORFLW\ DW WKH ERWWRP )URP OLQHDU ZDYH WKHRU\ LW LV IRXQG WKDW 2 ,7 _8RUW_ 7 VLQK NK f ZKHUH + LV WKH ZDYH KHLJKW DQG 7 LV WKH ZDYH SHULRG N LV WKH ZDYHQXPEHU DQG K LV WKH WRWDO ZDWHU GHSWK (EHUVROH DQG 'DOU\PSOH XVHG D PRUH FRPSOHWH TXDGUDWLF IRUPXODWLRQ 7E S) r_W_ f ZKHUH WKH RYHUEDU UHSUHVHQWV WKH WLPH DYHUDJH RYHU D ZDYH SHULRG DQG XW LV FRPSULVHG RI ERWK WKH PHDQ FXUUHQW DQG WKH ZDYH RUELWDO YHORFLWLHV 7KH WRWDO YHORFLW\ YHFWRU WLc LV H[SUHVVHG DV XL 8 WLFRVf} 9 VLQf f VR WKDW LWV PDJQLWXGH LV JLYHQ E\ _Xr_ 9X 9 8Âœ FRV 9 VLQ f 7KH ZDYH RUELWDO YHORFLW\ LV H[SUHVVHG DV fÂ§ XP FRV DW f ZKHUH XP LV WKH PD[LPXP ZDYH RUELWDO YHORFLW\ DW WKH ERWWRP ZKLFK LV IRXQG WR EH [+ fÂ§ 7 VLQK NK 7KH [ DQG \ FRPSRQHQWV RI WKH WLPH DYHUDJHG ERWWRP IULFWLRQ DUH WKXV UK[ A T ^8 PFRVFRHDWf?XW?G^DWf +Y S >9 VLQ LFRV DWf f _W7^_GRWf Â; R f f f
PAGE 30
ZKHUH _Wc_ LV ZULWWHQ DV _W7M_ ?X 9 WLWHRV DW 8XP FRV DW FRV 9WLP FRV DW VLQ f 7KH LQWHJUDOV LQ (TV f DQG f DUH WKHQ HYDOXDWHG E\ D SRLQW VXPPDWLRQ XVLQJ 6LPSVRQfV 5XOH ,Q WKH SUHVHQW VWXG\ VHYHUDO GLIIHUHQW IRUPXODWLRQV DUH XVHG DQG FRPSDUHG 7KH ILUVW LV WKH IRUPXODWLRQ JLYHQ E\ /RQJXHW +LJJLQV Df DQG XVHG E\ VHYHUDO SUHYLRXV LQYHVWLJDn WRUV 7KH VHFRQG LV D PRGLILFDWLRQ RI WKH /RQJXHW+LJJLQV IRUPXODWLRQ ZLWK SDUWLDO LQFOXVLRQ RI FXUUHQW VWUHQJWK LQ WKH IULFWLRQ IRUPXOD 7r S) ?XRUE? 8 S)?8?8 f UEY S)?XRUE?9 S)?8?9 f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f ZKHUH W7r LV FRPSRVHG RI ERWK WKH PHDQ FXUUHQW DQG WKH ZDYH RUELWDO YHORFLWLHV 7KH ZDYH RUELWDO YHORFLWLHV DUH H[SUHVVHG DV WKH UHDO SDUW RI WKH JUDGLHQW RI WKH FRPSOH[ YHORFLW\ SRWHQWLDO ZKLFK LV H[SUHVVHG DV $}\f FRVK NK ]f 0W D FRVK NK D f
PAGE 31
ZKHUH $ LV WKH FRPSOH[ DPSOLWXGH X LV WKH DEVROXWH IUHTXHQF\ D LV WKH LQWULQVLF RU UHODWLYH IUHTXHQF\ IROORZLQJ WKH FXUUHQW DQG KW K DQG N DUH DV SUHYLRXVO\ H[SODLQHG (TXDWLRQV JRYHUQLQJ DQG D SDUDEROLF DSSUR[LPDWLRQ VROXWLRQ IRU $ DUH SUHVHQWHG LQ WKH QH[W FKDSWHU 7KH WRWDO YHORFLW\ YHFWRU W7 LV H[SUHVVHG DV XW A rif@ ">Yrf@ ZKHUH 65 LQGLFDWHV WKH UHDO SDUW VR WKDW LWV PDJQLWXGH LV JLYHQ E\ .O f In ,f 9r ,"f >m ef@f >r ef@f 7KH [ DQG \ FRPSRQHQWV RI WKH WLPH DYHUDJHG ERWWRP IULFWLRQ DUH WKXV nU"e>Xr^ef? f f f f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f DQG D GHHS ZDWHU ZDYH DQJOH RI GHJUHHV GHJUHHV DW WKH PRGHO ERXQGDU\f DSSURDFKLQJ D SODQH EHDFK $ IULFWLRQ IDFWRU ) RI ZDV XVHG IRU HDFK FDVH 8VLQJ D IULFWLRQ IRUPXODWLRQ WKDW LJQRUHV WKH FRQWULEXWLRQ RI WKH ZDYHV UHVXOWV LQ D YHU\ ZHDN IULFWLRQ DQG WKXV D ODUJH ORQJVKRUH FXUUHQW 8VLQJ WKH /RQJXHW +LJJLQV IRUPXODWLRQ JLYHV D IULFWLRQ VOLJKWO\ OHVV WKDQ WKH PRUH SURSHU WLPH DYHUDJHG IRUPXODWLRQ ZKLOH D FRPELQDWLRQ RI WKH /RQJXHW+LJJLQV IRUPXODWLRQ DQG WKH VWURQJ FXUUHQW PRGHO QR ZDYH FRQWULEXWLRQf \LHOGV D VOLJKWO\ VWURQJHU IULFWLRQ )RU SUDFWLFDO QXPHULFDO FRQVLGHUDWLRQV WKH ODWWHU IRUPXODWLRQ LV
PAGE 32
9(/2&,7< 06(&f LQ )LJXUH /RQJVKRUH FXUUHQW SURILOHV IRU GLIIHUHQW IULFWLRQ IRUPXODWLRQV IRU DQ VHFRQG PHWHU ZDYH DW GHJUHHV WR D SODQH EHDFK PRVW RIWHQ XVHG VLQFH LW UHTXLUHV OHVV FRPSXWDWLRQ WLPH DQG PRUH LPSRUWDQWO\ WKH *DXVVLDQ LQWHJUDWLRQ LQYROYLQJ WKH UHHG SDUW RI WKH JUDGLHQW RI FRPSOH[ DPSOLWXGHV DW WLPHV SURGXFHV VWDELOLW\ SUREOHPV 7KH ODWHUDO PL[LQJ WHUP LQ (TV f DQG f DOVR LQYROYH VHYHUDO VLPSOLI\LQJ DVVXPSn WLRQV 7KH ILUVW DVVXPSWLRQ LV IRXQG LQ WKH GHULYDWLRQ RI WKH WZR GHSWK LQWHJUDWHG HTXDWLRQV RI PRWLRQ ZKHUH LW LV DVVXPHG WKDW WKH WXUEXOHQW VKHDU VWUHVV WHUP SXnY LV LQGHSHQGHQW RI GHSWK 7KH RWKHU DVVXPSWLRQ LV WKDW WKH SURFHVV RI PRPHQWXP WUDQVIHU GXH WR WXUEXOHQW IOXFWXDWLRQV FDQ EH UHSUHVHQWHG E\ D SURGXFW RI D PL[LQJ OHQJWK SDUDPHWHU DQG GHULYDWLYHV RI WKH PHDQ FXUUHQW 7KLV LV ZKDW LV GRQH LQ WKH SUHVHQW VWXG\ DQG DOVR LQ SUHYLRXV VWXGLHV (EHUVROH DQG 'DOU\PSOH f 9HPXODNRQGD f DQG
PAGE 33
DQG D VSLOOLQJ EUHDNHU DVVXPSWLRQ /RQJXHW+LJJLQV ZURWH G8 G9? Q Snf LQ ZKLFK H[ DQG H\ (LUH PL[LQJ OHQJWK FRHIILFLHQWV /RQJXHW+LJJLQV UHDVRQHG WKDW WKH PL[LQJ OHQJWK SDUDPHWHU H[ PXVW WHQG WR ]HUR DV WKH VKRUHOLQH ZDV DSSURDFKHG +H DVVXPHG WKDW H[ LV SURSRUWLRQDO WR WKH GLVWDQFH IURP WKH VKRUHOLQH ; PXOWLSOLHG E\ D YHORFLW\ VFDOH ZKLFK KH FKRVH DV WKH VKDOORZ ZDWHU FHOHULW\ \JK 7KXV H[ LV ZULWWHQ DV H[ 1; ?JK f ZKHUH 1 LV D GLPHQVLRQOHVV FRQVWDQW IRU ZKLFK /RQJXHW+LJJLQV FKRVH WKH OLPLWV DV 1 f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f 7KHUHIRUH WKH ODWHUDO PL[LQJ WHUPV EHFRPH Â‘ef f DY? ?e[G[f f LQ WKH [ GLUHFWLRQ DQG LQ WKH \ GLUHFWLRQ 7KLV PHDQV WKDW WKH ODWHUDO PL[LQJ WHUP LV UHSUHVHQWHG E\ D SXUHO\ ODWHUDO GLIIXVLRQ RI PRPHQWXP 7KDW WKH ODWHUDO PL[LQJ WHUP LV D ODWHUDO GLIIXVLRQ WHUP KDV VRPH VLJQLILFDQFH LQ WKDW WKH GLIIXVLRQ WHUP LV H[SUHVVHG H[SOLFLWO\ DQG WKXV WKHUH UHVXOWV D WLPH VWHS OLPLWDWLRQ WR DQ LPSOLFLW PRGHO 7KLV ZLOO EH GLVFXVVHG LQ VHFWLRQ GHDOLQJ ZLWK WKH ILQLWH GLIIHUHQFLQJ RI WKH HTXDWLRQV JRYHUQLQJ WKH FLUFXODWLRQ PRGHO
PAGE 34
5DGLDWLRQ 6WUHVVHV ,Q D VHULHV RI SDSHUV LQ WKH HDUO\ V /RQJXHW+LJJLQV DQG 6WHZDUW f LQWURGXFHG DQG GHYHORSHG WKH PDWKHPDWLFV RI WKH UDGLDWLRQ VWUHVVHV 7KHVH VWUHVVHV ZKLFK FDQ EH FRQFHSWXDOO\ WKRXJKW RI DV WKH fIOX[ RI H[FHVV PRPHQWXPf FDQ EH REWDLQHG WKURXJK LQWHJUDWLRQ RI WKH PRPHQWXP HTXDWLRQV RYHU WKH ZDWHU FROXPQ DV LV VKRZQ LQ $SSHQGL[ $ 8VLQJ OLQHDU ZDYH WKHRU\ WKH UDGLDWLRQ VWUHVV FRPSRQHQWV DUH IRXQG WR EH 6]] ( QFRV f fÂ§ A $W 6YY ( QVLQ f fÂ§ A $W P 6]Y (Q FRV VLQ f f f ZKHUH ( LV WKH ZDYH HQHUJ\ HTXDO WR _SJ+ Q LV WKH UDWLR RI WKH JURXS YHORFLW\ WR WKH ZDYH FHOHULW\ DQG LV WKH DQJOH WKH ZDYH UD\V PDNH ZLWK WKH [ D[LV ,I WKH ZDYH KHLJKW DQG ZDYH DQJOH DUH NQRZQ DW HDFK JULG SRLQW WKH UDGLDWLRQ VWUHVVHV DQG WKHLU JUDGLHQWV DUH UHDGLO\ GHWHUPLQHG IRU XVH LQ WKH JRYHUQLQJ HTXDWLRQV 7R GHWHUPLQH WKH ZDYH KHLJKW DQG DQJOH DW HDFK JULG SRLQW %LUNHPHLHU DQG 'DOU\PSOH f (EHUVROH DQG 'DOU\PSOH f DQG 9HPXODNRQGD f HDFK XVHG WKH IROORZLQJ VFKHPH ZKLFK LV EDVHG XSRQ WKH LUURWDWLRQDOLW\ RI WKH ZDYH QXPEHU DQG DQ HTXDWLRQ IRU WKH FRQVHUYDWLRQ RI ZDYH HQHUJ\ 7KH LUURWDWLRQDOLW\ RI WKH ZDYH QXPEHU GHWHUPLQHV FKDQJHV LQ WKH ZDYH DQJOH GXH WR UHIUDFWLRQ ZKLOH WKH HQHUJ\ HTXDWLRQ GHWHUPLQHV FKDQJHV LQ WKH ZDYH KHLJKW GXH WR VKRDOLQJ DQG UHIUDFWLRQ 7KLV VFKHPH ZDV ILUVW LQWURGXFHG E\ 1RGD HW DO f 7KH LUURWDWLRQDOLW\ RI WKH ZDYH QXPEHU 9 [ f ZKHUH N ?N? FRV 2L _IF_ VLQ 2M f LV H[SDQGHG LQ &DUWHVLDQ FRRUGLQDWHV WR REWDLQ G FRV fÂ§, VLQ fÂ§ VLQ R[ D\ _IF_ G[ fÂ§ FRV r G\ f
PAGE 35
)RU ZDYHV RQ D FXUUHQW WKH GLVSHUVLRQ UHODWLRQ LV RM D N \MJ?N? WDQK ?N?K _N_ FRV _E_ VLQ f (TXDWLRQ f LV GLIIHUHQWLDWHG WR HOLPLQDWH DQG IURP (T f 8VLQJ D IRUZDUG GLIIHUHQFH GHULYDWLYH LQ [ DQG D EDFNZDUG GLIIHUHQFH GHULYDWLYH LQ \ Â LV REWDLQHG DV D IXQFWLRQ RI FRV VLQ 89NK DQG WKHLU GHULYDWLYHV DV ZHOO DV DW WKH DGMRLQLQJ JULGV 7KH VLQ DQG FRV WHUPV DUH H[SUHVVHG LQ WHUPV RI WKH VXUURXQGLQJ IRXU JULGV XVLQJ D QG RUGHU 7D\ORU H[SDQVLRQ 7KH VROXWLRQ LV DQ LWHUDWLYH SURFHVV DOWHUQDWLQJ VROXWLRQ RI WKH HTXDWLRQ ZLWK VROXWLRQ RI WKH GLVSHUVLRQ UHODWLRQVKLS f IRU N _LW_ ZKLFK LV DFFRPSOLVKHG E\ D 1HZWRQ5DSKVRQ PHWKRG XQWLO DFFHSWDEOH FRQYHUJHQFH LV REWDLQHG IRU N DQG 1H[W DQ HQHUJ\ HTXDWLRQ LV VROYHG IRU WKH ZDYH KHLJKW $ FRQVHUYDWLRQ RI HQHUJ\ HTXDWLRQ LV JLYHQ E\ 3KLOOLSV HG (T f DV A_B^P &f` V_A f $VVXPLQJ VWHDG\ VWDWH FRQGLWLRQV DQG H[SDQGLQJ LQ &DUWHVLDQ FRRUGLQDWHV JLYHV 9 &FRV rf__ 9 & VLQ mf L Da $8 & FRV 6f G8 G8 G9 G9 &nLVLQ f D[[fÂ§ I\[fÂ§ I[YfÂ§Y\ fÂ§ G\ G\ G[ YY G\ f ZKHUH r[] fÂ§ a(A=; r99 fÂ§ e 6\Y f=9 fÂ§ r\] fÂ§ QFRV f fÂ§ L QVLQ f fÂ§ Q FRV VLQ 'HILQLQJ &% FRV f fÂ§ 9 & VLQ G\ f ZKHUH G8 G8 G9 G9 ][GÂn^n7Y[aG 7[YaGÂ7YYaGÂ f D
PAGE 36
DQG QRWLQJ WKDW 89&O\ DQG D DUH NQRZQ TXDQWLWLHV WDNLQJ IRUZDUG GLIIHUHQFH GHULYDWLYHV LQ [ DQG EDFNZDUG GLIIHUHQFH GHULYDWLYHV LQ \ DQG VROYLQJ IRU WKH ZDYH KHLJKW +M DW WKH L\M JULG \LHOGV +LM 9&}LQ JfJ\BO B 8& FRW f+LM D< a D; 9&mLQL 8& FRW f D\ Dn /MM 7KH FRPSXWDWLRQ LV D URZ E\ URZ LWHUDWLRQ SURFHHGLQJ IURP ODUJH L RIIVKRUH LQ WKH FRRUGLQDWH V\VWHP RI WKH DERYH FLWHG DXWKRUVf WR VPDOO L 'XULQJ WKH FRPSXWDWLRQ RI HDFK +LM WKH EUHDNLQJ KHLJKW LV DOVR FDOFXODWHG DQG LI +LM H[FHHGV WKLV YDOXH WKHQ WKH EUHDNLQJ ZDYH KHLJKW LV XVHG 6HYHUDO EUHDNLQJ KHLJKW LQGLFHV DUH SRVVLEOH 0LFKH f JDYH DV WKH WKHRUHWLFDO OLPLWLQJ ZDYH VWHHSQHVV WKH FRQGLWLRQ AU WDQKIFMKcf /E f ZKHUH WKH E VXEVFULSW LQGLFDWHV EUHDNLQJ FRQGLWLRQV /H 0KDXW DQG .RK f LQGLFDWHG IURP H[SHULPHQWDO GDWD WKDW WKH OLPLWLQJ VWHHSQHVV LV AU WDQKIFcKf /E f RU IIr AWDQKIFIFf f .E 7KLV LV WKH EUHDNLQJ KHLJKW FULWHULRQ XVHG E\ 1RGD HW DO f 3HUKDSV WKH VLPSOHVW EUHDNLQJ KHLJKW FULWHULRQ LV WKH VSLOOLQJ EUHDNHU DVVXPSWLRQ +E KE f ZKLFK LV XVHG E\ 9HPXODNRQGD f DQG 9HPXODNRQGD HW DO f 7KH VFKHPH GHVFULEHG DERYH RI GHWHUPLQLQJ WKH ZDYH DQJOH IURP WKH LUURWDWLRQDOLW\ RI WKH ZDYH QXPEHU f DQG WKH GLVSHUVLRQ UHODWLRQ f DQG WKH ZDYH KHLJKW IURP WKH HQHUJ\ HTXDWLRQ f KDV LWV OLPLWDWLRQV 3ULQFLSDOO\ WKLV VFKHPH GRHV QRW WDNH LQWR DFFRXQW WKH HIIHFWV RI GLIIUDFWLRQ 7KLV OLPLWDWLRQ LV VXUPRXQWHG E\ XVH RI D SDUDEROLF ZDYH HTXDWLRQ +RZHYHU WKLV SUHVHQWV DQRWKHU SUREOHP LQ WKDW WKH ZDYH DQJOH LV QRW GHWHUPLQHG
PAGE 37
LQ D SDUDEROLF HTXDWLRQ VROXWLRQ 2QO\ WKH FRPSOH[ DPSOLWXGH ZKLFK FRQWDLQV WKH PDJn QLWXGH DQG WKH SKDVH RI WKH ZDYH LV REWDLQHG
PAGE 38
,QWHJUDWLQJ D KRUL]RQWDO FRPSRQHQW RI f YHUWLFDOO\ RYHU WKH GHSWK RI WKH ZDWHU FROXPQ IURP ] K WR ] Uc XVLQJ WKH ERXQGDU\ FRQGLWLRQV f DQG WKHQ WDNLQJ WLPH DYHUDJHV WKH VDPH SURFHGXUH DV LQ DSSHQGL[ $f WKH IROORZLQJ GHILQLWLRQ RI WKH UDGLDWLRQ VWUHVVHV LV REWDLQHG UR 6DS S XnDXnSG] DS fÂ§ K >L 3G]a K Â \Lnff f ZKHUH WKH RYHUEDU LV XVHG WR LQGLFDWH WKH WLPH DYHUDJHG YDOXHV DQG DS LV WKH .URQLFNHU GHOWD IXQFWLRQ WDS IRU D 3 IRU D A I f 0HL f ZDV DEOH WR H[SUHVV (T f LQ WHUPV RI WKH JHQHUDO FRPSOH[ DPSOLWXGH IXQFWLRQ % E\ UHODWLQJ LW WR WKH YHORFLW\ SRWHQWLDO DQG WKH IUHH VXUIDFH KHLJKW E\ B WR FRVK N>K ]f LXW D FRVK NK Wf % HaLXW f f $Q H[SUHVVLRQ IRU S LV REWDLQHG E\ LQWHJUDWLQJ WKH YHUWLFDO FRPSRQHQW RI (T f GZ GZ GZ GS f ,QWHJUDWLQJ IURP DQ\ GHSWK ] WR WKH VXUIDFH XVLQJ /HLEQLW]fV UXOH IRU LQWHUFKDQJLQJ WKH RUGHU RI GLIIHUHQWLDWLRQ DQG LQWHJUDWLRQ HPSOR\LQJ WKH WZR VXUIDFH ERXQGDU\ FRQGLWLRQV (TV f DQG f DQG WKH FRQWLQXLW\ (T f (T f \LHOGV G IL G + SJLQ a ]f SfÂ§ZG] SfÂ§8SZG] SZ@ f 7DNLQJ WLPH DYHUDJHV JLYHV UIO GXnZn S>[\]f SDLQ ]f SMr J; G]aSmf f 6XEVWLWXWLRQ RI (TV ff DQG f LQWR f DQG PDNLQJ WKH DVVXPSWLRQ WKDW KRUL]RQWDO GHULYDWLYHV RI N K DQG D FDQ EH QHJOHFWHG WKLV LV HVVHQWLDOO\ D PLOG VORSH DVn VXPSWLRQf WKH IROORZLQJ LV REWDLQHG DIWHU VWUDLJKWIRUZDUG EXW WHGLRXV FDOFXODWLRQV ZKLFK
PAGE 39
DUH VKRZQ LQ $SSHQGL[ % F 3 / G%G%rf I NK ? \G[D G[S N ? VLQKIFK rDS ,%, NK VLQK NK G[D G[S _9K%_ r_%_fÂrKFRWKrK f@ ` f ZKHUH %r GHQRWHV WKH FRPSOH[ FRQMXJDWH RI % ,Q REWDLQLQJ (T f XVH ZDV PDGH RI WKH IODW ERWWRP +HOPKROW] HTXDWLRQ 9Me IFe f 7KH JHQHUDO FRPSOH[ DPSOLWXGH IXQFWLRQ % LV UHODWHG WR WKH FRPSOH[ DPSOLWXGH $ ZKLFK LV VROYHG IRU LQ WKH ZDYH PRGHO E\ % $HAIOFRLG]A f 7KHUH DUH VRPH SUREOHPV ZLWK (T f 2QH PD\ TXHVWLRQ WKH DVVXPSWLRQ RI D IODW ERWWRP KRZHYHU WKH LQFOXVLRQ RI WKH VORSH DQG JUDGLHQWV RI WKH ZDYH QXPEHU DQG WKH LQWULQVLF IUHTXHQF\ OHDGV WR LQWUDFWDEOH PDWKHPDWLFV 7KH XVH RI WKH ORFDOL]HG IODW ERWWRP LV DOVR LPSOLFLW LQ WKH GHULYDWLRQ RI /RQJXHW+LJJLQV DQG 6WHZDUW WR REWDLQ (TV f VLQFH LQ FDUU\LQJ RXW WKH LQWHJUDWLRQV LQYROYHG LW LV DVVXPHG WKDW KRUL]RQWDO GHULYDWLYHV RI WKH YHORFLW\ SRWHQWLDO DUH OLPLWHG WR GHULYDWLYHV RI WKH SKDVH IXQFWLRQ )RU SODQH ZDYHV XVLQJ D VSLOOLQJ EUHDNHU DVVXPSWLRQ UHVXOWV XVLQJ (T f GLIIHU OLWWOH IURP WKRVH XVLQJ (TV f KRZHYHU LQ GLIIUDFWLRQ ]RQHV (T f LV VXSHULRU IRU UHDVRQV PHQWLRQHG SUHYLRXVO\ 7KH PDMRU SUREOHP ZLWK WKH XVH RI (T f LV DOVR D PDMRU SUREOHP DVVRFLDWHG ZLWK WKH XVH RI /RQJXHW+LJJLQVfV IRUPXODWLRQ 7KLV SUREOHP LV WKDW LQ ERWK IRUPXODWLRQV WKH LQLWLDWLRQ RI ZDYH EUHDNLQJ LPSOLHV WKH GLVVLSDWLRQ RI HQHUJ\ DQG WKXV WKH IRUFLQJ RI VHWn XS DQG ORQJVKRUH FXUUHQWV VWDUWLQJ DW WKH EUHDNHU OLQH 7KH SK\VLFDO UHDOLW\ LV VRPHZKDW GLIIHUHQW LQ WKDW LQ WKH LQLWLDO VWDJHV RI EUHDNLQJ HQHUJ\ LV QRW LPPHGLDWHO\ GLVVLSDWHG EXW UDWKHU LV WUDQVIRUPHG IURP RUJDQL]HG ZDYH HQHUJ\ WR WXUEXOHQW HQHUJ\ DQG WR D VXUIDFH UROOHU 6YHQGVHQ f 7KHUH LV D VSDFH ODJ EHWZHHQ WKH LQLWLDWLRQ RI EUHDNLQJ DQG WKH GLVVLSDWLRQ
PAGE 40
O f HDL )LJXUH ([SHULPHQWDO UHVXOWV IRU ZDYH VHWXS DQG VHWGRZQ 5HSULQWHG IURP %RZHQ ,Qn PDQ DQG 6LPPRQV f RXUQDO RI *HRSK\VLFDO 5HVHDUFK YRO f SDJH FRS\ULJKW E\ WKH $PHULFDQ *HRSK\VLFDO 8QLRQ RI HQHUJ\ 7KLV FDQ EH VHHQ LQ ILJXUH ZKLFK JLYHV WKH UHVXOWV IRU WKH H[SHULPHQWV RI %RZHQ HW DO f $QDO\WLFDO DQG QXPHULFDO VROXWLRQV IRU WKH PHDQ ZDWHU VXUIDFH HOHYDWLRQ ZLOO JLYH D VHWXS VWDUWLQJ DW WKH EUHDN SRLQW ZKLOH WKH VROLG ILQH LQ WKH ORZHU JUDSK VKRZV WKDW VHWXS PHDVXUHG LQ WKH H[SHULPHQW EHJLQV VKRUHZDUG RI WKH EUHDN SRLQW 7KLV ZLOO EH GLVFXVVHG IXUWKHU LQ &KDSWHU
PAGE 41
&+$37(5 :$9( 02'(/ *RYHUQLQJ :DYH (TXDWLRQ 6HYHUHG PHWKRGV KDYH EHHQ XVHG WR GHULYH WKH JRYHUQLQJ HTXDWLRQ IRU WKH SURSDJDWLRQ RI ZDYHV RYHU YDU\LQJ WRSRJUDSK\ DQG YDU\LQJ FXUUHQWV $PRQJ WKH PHWKRGV XVHG DUH SHUn WXUEDWLRQ WHFKQLTXHV PXOWLSOH VFDOHV D /DJUDQJLDQ DQG WKH XVH RI *UHHQfV VHFRQG LGHQWLW\ :LWK WKH DVVXPSWLRQ RI D PLOG VORSH DQG ZHDN FXUUHQW FRQGLWLRQV HDFK RI WKHVH PHWKRGV VKRXOG \LHOG WKH VDPH HTXDWLRQ )RU LOOXVWUDWLYH SXUSRVHV WKH PHWKRG RI XVLQJ *UHHQfV VHFn RQG LGHQWLW\ ZLOO EH XVHG KHUH WR GHULYH D OLQHDU HTXDWLRQ $ /DJUDQJLDQ GHULYDWLRQ RI WKH VDPH HTXDWLRQ LV SUHVHQWHG LQ $SSHQGL[ & 7KH OLQHDU HTXDWLRQ LV XVHG VLQFH VHFRQG RUGHU ZDYHV H[KLELW VLQJXODULWLHV DV WKH ZDWHU GHSWK JRHV WR ]HUR $ YHORFLW\ SRWHQWLDO W LV DVVXPHG VXFK WKDW WKH ZDWHU SDUWLFOH YHORFLWLHV DUH JLYHQ E\ 9I! ZKHUH 9 LV WKH WKUHHGLPHQVLRQDO JUDGLHQW RSHUDWRU G G r G f DQG rM DQG N DUH WKH XQLW YHFWRUV LQ WKH [\ DQG ] GLUHFWLRQV UHVSHFWLYHO\ 7KH YHORFLWLHV DUH D VXSHUSRVLWLRQ RI VWHDG\ FXUUHQWV DQG ZDYH LQGXFHG PRWLRQ 7KH YHORFLW\ SRWHQWLDO LV JLYHQ E\ W WR HIW f ZKHUH WR LV GHILQHG VXFK WKDW LWV JUDGLHQW \LHOGV 89 DQG : ZKLFK DUH WKH VWHDG\ FXUUHQW WHUPV DQG IW LV WKH ZDYH FRPSRQHQW RI WKH YHORFLW\ SRWHQWLDO $Q XQGHILQHG VFDOLQJ H LV XVHG VR DV WR VHSDUDWH WKH ZDYH DQG VWHDG\ FXUUHQW SDUW RI WKH YHORFLW\ SRWHQWLDO 7KH ZDYH SDUW RI WKH YHORFLW\ SRWHQWLDO LV FRPSRVHG RI WZR SDUWV ZKLFK FRQWDLQV WKH GHSWK GHSHQGHQF\ DQG ZKLFK LV D IXQFWLRQ RI WKH KRUL]RQWDO FRRUGLQDWHV DQG WLPH 7KH GHSWK
PAGE 42
GHSHQGHQF\ LV JLYHQ E\ OLQHDU WKHRU\ DV M B FRVK NK ]f FRVK NK f ZKHUH K LV WKH GHSWK RI WKH ERWWRP UHIHUHQFHG WR WKH VWLOO ZDWHU OLQH DQG K LV WKH WRWDO ZDWHU GHSWK DQG N LV WKH ZDYH QXPEHU YHFWRU ,W VKRXOG EH QRWHG WKDW LV DOVR D IXQFWLRQ RI WKH KRUL]RQWDO FRRUGLQDWHV VLQFH N DQG K YDU\ ZLWK KRUL]RQWDO SRVLWLRQ )RU LQFRPSUHVVLEOH LUURWDWLRQDO IORZ WKH JRYHUQLQJ HTXDWLRQ IRU WKH YHORFLW\ SRWHQWLDO LV WKH /DSODFH HTXDWLRQ 99 K]Uc f ZKHUH U" LV WKH LQVWDQWDQHRXV ZDWHU VXUIDFH HOHYDWLRQ 8VLQJ D KRUL]RQWDO JUDGLHQW RSHUDWRU 9 GHILQHG E\ B G } G m 9NIF AWKH /DSODFH HTXDWLRQ LV ZULWWHQ DV W!r] fÂ§ K ] Lf f f 7KH JRYHUQLQJ HTXDWLRQ LV FRPSOHPHQWHG E\ WKH IROORZLQJ ERXQGDU\ FRQGLWLRQV 2Q WKH IUHH VXUIDFH WKH NLQHPDWLF IUHH VXUIDFH ERXQGDU\ FRQGLWLRQ LV JLYHQ E\ f 9K9 !W ] Uc 7KH G\QDPLF IUHH VXUIDFH ERXQGDU\ FRQGLWLRQ ZLWK ]HUR DWPRVSKHULF SUHVVXUH W!W A9Âf J] ] Uc f f DQG WKH ERWWRP ERXQGDU\ FRQGLWLRQ 9KI! f !W ] K f 7KH GHULYDWLRQ WHFKQLTXH LV WR ZULWH *UHHQfV VHFRQG LGHQWLW\ IRU I! DQG DV GHILQHG E\ (TV f DQG f UHVSHFWLYHO\ UIf UQ II!]WG] Â G] >IW!] I!I]@OW f
PAGE 43
7KHQ (T f LV XVHG WR VXEVWLWXWH IRU I!]] LQ WKH ILUVW WHUP RI (T f DQG WKH ERXQGDU\ FRQGLWLRQV DUH XVHG WR VXEVWLWXWH IRU I!] DW WKH PHDQ IUHH VXUIDFH DQG DW WKH ERWWRP LQ WKH WKLUG WHUP RI *UHHQfV LGHQWLW\ (TXDWLRQ f LV XVHG DQG WHUPV RI RUGHU H RQO\ DUH UHWDLQHG LQ RUGHU WR REWDLQ WKH JRYHUQLQJ HTXDWLRQ IRU M! ,Q RUGHU WR REWDLQ I!] DW IM LW LV DVVXPHG WKDW U@ fÂ§ IM HIM f 7KH NLQHPDWLF IUHH VXUIDFH ERXQGDU\ FRQGLWLRQ (T f LV WKHQ H[SDQGHG LQ D 7D\ORU VHULHV DERXW ] IM r II f G GQ JM f 9I !0 r II f VROYLQJ IRU I!] _M XVLQJ (TV DQG ZKLOH UHWDLQLQJ RQO\ WHUPV RI RUGHU H \LHOGV GIc ,IIfÂ§ $ WRWDO KRUL]RQWDO GHULYDWLYH GW 8 f 9KIM 9KW! f 9IFII IM:] r IO f IW 7 'L DL Xf9K f LV LQWURGXFHG VR WKDW WKH ILUVW WZR WHUPV RQ WKH ULJKW KDQG VLGH RI (T f FDQ EH ZULWWHQ DV WKH WRWDO KRUL]RQWDO GHULYDWLYH RI IM 7KH ILQDO PDQLSXODWLRQ RI (T f LV WR UHFRJQL]H WKDW fÂ§ A DW WKH VXUIDFH LV WKH KRUL]RQWDO GLYHUJHQFH RI WKH KRUL]RQWDO FXUUHQW 9! f f 7KXV W!0 _II n n9K7f IM9K f f IM LV HOLPLQDWHG IURP (T f E\ H[SDQGLQJ WKH G\QDPLF IUHH VXUIDFH ERXQGDU\ FRQGLWLRQ (T f LQ D 7D\ORU VHULHV DERXW ] IM !W A9Âf J] r UM 9 0 IO !W A!" !rW AYAf R r II 8VLQJ (T WR VXEVWLWXWH IRU I! WKH RUGHU W WHUPV \LHOG D VROXWLRQ IRU IM f f
PAGE 44
6XEVWLWXWLQJ LQWR (T f JLYHV rfmf 7KH ILUVW WHUP RI WKH *UHHQfV LGHQWLW\ XVLQJ (T f EHFRPHV U IW!rQG] U IYLG] U Yr YAr f r f KR r f KR r KR 1RZ 9KW! HI9K! H!9KI f DQG AK f f 9Af H9KLM!f 9KI ]M!9KI f 5HWDLQLQJ RQO\ WHUPV RI RUGHU ] DQG GURSSLQJ WKH ODVW WHUP VLQFH LW LV VHFRQG RUGHU LQ WKH VORSH WKH LQWHJUDO LV HTXDO WR U 9IF9AfG] U I9K!9KI G] f fKR r fOR 7KHVH WZR WHUPV DUH KDQGOHG XVLQJ /HLEQLW]fV UXOH RI LQWHJUDWLRQ YD7 YfÂfG] U YKIIYMfG]U IYKIYKWfG] r KR r KR r KS AK9I 8 _BDf 9e f 8VLQJ WKLV UHODWLRQVKLS DQG UHFRJQL]LQJ WKDW _M LV HTXDO WR RQH DQG WKH LQWHJUDO RYHU WKH GHSWK RI LV &&WJ ZKHUH & LV WKH ZDYH FHOHULW\ &D LV WKH ZDYH JURXS YHORFLW\ DQG J LV WKH JUDYLWDWLRQDO FRQVWDQW WKH ILUVW WHUP LV HTXDO WR 9 f fÂ§ 9Mf 9KIM 9KKI _BIFR ZKHUH WKH ] KDV EHHQ GURSSHG 7KH VHFRQG WHUP RI (T f LV HYDOXDWHG TXLWH HDVLO\ G] fÂ§ ] f f
PAGE 45
7KH ILUVW LQWHJUDO DERYH LV GURSSHG VLQFH WKHUH DUH RQO\ VWHDG\ FXUUHQW WHUPV 7KH UHPDLQLQJ WHUP GURSSLQJ WKH H LV U IW IrrG] NMM! U IG] NAL f K KR 7KH WKLUG WHUP LQ (T f LV IW : IW ,IIW ?K f 7KH ILUVW WHUP RQ WKH ULJKW KDQG VLGH LV HTXDO WR Wr __ VLQFH A DQG KDV DOUHDG\ EHHQ GHWHUPLQHG 7KH VHFRQG WHUP RQ WKH ULJKW KDQG VLGH RI (T f LV HDVLO\ GHWHUPLQHG XVLQJ WKH ERWWRP ERXQGDU\ FRQGLWLRQ (T f IW?NR I9KW9KK ?N f 8VLQJ (T f WKLV EHFRPHV IW 8R I8 f 9NK _BIFR H _BIFR 9KW f YKK HIW9NI Â‘ 9IFK _BIFR f 7KH ILUVW WHUP LV GURSSHG VLQFH LW QRW RUGHU H DQG WKH WKLUG WHUP LV FOHDUO\ KLJKHU RUGHU LQ WKH VORSH 7KXV DJDLQ GURSSLQJ WKH W 0N _O9NI9$ f 7KLV WHUP FDQFHOV WKH ERWWRP ERXQGDU\ WHUP UHVXOWLQJ IURP WKH /HLEQLW] LQWHJUDWLRQ DERYH 7KH IRXUWK WHUP RI (T f LV DOVR HDVLO\ GHWHUPLQHG f WI _r fÂ§ÂIH%PFUDK IFK A Or aW!N WDQK NK f
PAGE 46
$ 3DUDEROLF $SSUR[LPDWLRQ (TXDWLRQ f LV D K\SHUEROLF ZDYH HTXDWLRQ ,W FDQ EH DOWHUHG WR DQ HOOLSWLF HTXDWLRQ E\ DVVXPLQJ WKDW DQ\ WLPH GHSHQGHQF\ LV VROHO\ LQ WKH ZDYH SKDVH %HIRUH GRLQJ WKLV DQ HQHUJ\ GLVVLSDWLRQ WHUP '' LV LQWURGXFHG DV IROORZV ,W LV DVVXPHG WKDW (T f KROGV IRU WKH FDVH RI QR HQHUJ\ GLVVLSDWLRQ WKHQ IRU WKH FDVH ZLWK GLVVLSDWLRQ WKH OHIW KDQG VLGH LV EDODQFHG E\ DQ\ HQHUJ\ WKDW LV GLVVLSDWHG 7KH JRYHUQLQJ HTXDWLRQ LV WKHQ ZULWWHQ DV 9r f 9r&&9IFÂf U Nn&&0 '' f ,Q WKH GHULYDWLRQ RI WKH JRYHUQLQJ K\SHUEROLF HTXDWLRQ XVLQJ D /DJUDQJLDQ DV SUHVHQWHG LQ $SSHQGL[ & WKH '' WHUP FDQ EH YLHZHG DV D YLUWXDO ZRUN WHUP )ROORZLQJ .LUE\ f WKH '' WHUP LV DVVXPHG WR EH RI WKH IRUP 'G! f ZKHUH Z LV DQ XQGHILQHG FRHIILFLHQW LQGLFDWLQJ WKH VWUHQJWK RI WKH GLVVLSDWLRQ (YHQWXDOO\ WKH Z WHUP ZLOO EH UHODWHG WR WKH HQHUJ\ GLVVLSDWLRQ GXH WR ZDYH EUHDNLQJ IROORZLQJ WKH ZRUN RI 'DOO\ 'HDQ DQG 'DOU\PSOH f 8VLQJ (T f LQ (T f DQG PDNLQJ WKH DVVXPSWLRQ WKDW WKH RQO\ WLPH GHSHQGHQF\ LV LQ WKH SKDVH LH GG! fÂ§WXfL! f UHGXFHV WKH K\SHUEROLF HTXDWLRQ f WR WKH HOOLSWLF IRUP WXII Â‘ f 9K f YMf 9IF f 9f8 f 9IFÂf Yf f FFYIFÂf ^U X N&&W LX!>9N f 8f`M! UZe f ZKHUH RQO\ WKH SKDVH FRQWULEXWLRQ WR WKH KRUL]RQWDO GHULYDWLYH RI A LV UHWDLQHG LQ REWDLQLQJ WKH WHUP RQ WKH ULJKW KDQG VLGH RI f %HIRUH GHULYLQJ D SDUDEROLF DSSUR[LPDWLRQ WR (T f WKH QHHG IRU VXFK DQ DSSUR[Ln PDWLRQ VKRXOG EH GLVFXVVHG 7KHUH DUH WZR PDMRU FRPSXWDWLRQDO GUDZEDFNV WR QXPHULFDOO\
PAGE 47
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n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f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f DQG ODWHUDO ERXQGDU\ FRQGLWLRQV ZKLFK DUH XVXDOO\ D GHULYDWLYH FRQGLWLRQ WKDW DOORZV IRU WKH ZDYH WR SDVV WKURXJK WKH ERXQGDU\ ZLWKRXW DQ\ UHIOHFWLRQ DW WKH ERXQGDU\ 7KHUH DUH VHYHUDO ZD\V WR WUDQVIRUP WKH HOOLSWLF (T f WR D SDUDEROLF HTXDWLRQ 7KH PRVW GLUHFW LV WR PDNH FHUWDLQ DVVXPSWLRQV FRQFHUQLQJ WKH OHQJWK VFDOH RI YDULDWLRQV LQ WKH [ GLUHFWLRQ $W WKH KHDUW RI WKH SDUDEROLF DSSUR[LPDWLRQ LV WKH DVVXPSWLRQ WKDW WKH GLUHFWLRQ RI ZDYH SURSDJDWLRQ LV HVVHQWLDOO\ DORQJ WKH [ D[LV )RU ZDYHV SURSDJDWLQJ DW DQ DQJOH WR
PAGE 48
WKH [ D[LV WKH SURSHU IRUP RI LV Â __J 9f Hm N FRrmÂ‘r} rLQ } rrf f DQG WKH SURSHU IRUP RI WKH GLVSHUVLRQ UHODWLRQVKLS LV X! D NFRV28 $ VLQ 9 f ZKHUH II LV WKH DQJOH RI WKH ZDYH SURSDJDWLRQ UHODWLYH WR WKH [ D[LV 7KH HVVHQFH RI D SDUDEROLF DSSUR[LPDWLRQ LV WR DVVXPH WKDW WKH DQJOH WKDW WKH ZDYH UD\V PDNH ZLWK WKH [ D[LV LV VPDOO VR WKDW WKH NVLQ WHUP FDQ EH QHJOHFWHG LQ (TV f DQG f DQG FRV LV DVVXPHG WR EH XQLW\ 7KH PDQQHU RI GRLQJ WKLV YDULHV ZLWK GLIIHUHQW DXWKRUV
PAGE 49
7KLV HTXDWLRQ LV DOVR REWDLQHG GLUHFWO\ IURP (T f IRU WKH VWDWHG DVVXPSWLRQV RI QR FXUUHQWV QR GLVVLSDWLRQ DQG FRQVWDQW GHSWK 6XEVWLWXWLRQ RI f LQWR WKH +HOPKROW] HTXDWLRQ JLYHV $[[ LN$[ $99 f DQG HPSOR\LQJ WKH DERYH PHQWLRQHG DVVXPSWLRQ FRQFHUQLQJ GHULYDWLYHV RI $[ \LHOGV WKH OLQHDU SDUDEROLF HTXDWLRQ LN$[ $99 f $QRWKHU PHWKRG RI REWDLQLQJ D SDUDEROLF HTXDWLRQ LV WR XVH D VSOLWWLQJ PHWKRG LQ ZKLFK LW LV DVVXPHG WKDW LV FRPSRVHG RI IRUZDUG PRYLQJ DQG EDFNZDUG PRYLQJ FRPSRQHQWV WKDW DUH XQFRXSOHG ! f 8QFRXSOHG HTXDWLRQV IRU I! DQG DUH GHYHORSHG DQG VLQFH WKH HTXDWLRQV DUH XQFRXSOHG RQO\ WKH HTXDWLRQ IRU I! LV UHWDLQHG ,Q WKH IROORZLQJ WKH ZRUN RI .LUE\ f DQG %RRLM f LV FORVHO\ IROORZHG 7KH VHFRQG RUGHU HOOLSWLF HTXDWLRQ GB ee? G[ 9TI G[ f \6} f FDQ EH VSOLW H[DFWO\ LQWR HTXDWLRQV JRYHUQLQJ IRUZDUG DQG EDFNZDUG GLVWXUEDQFHV DQ ZKLFK DUH JLYHQ E\ G G[ Ga G[ Â fÂ§},!a %RRLM FKRVH WKH UHODWLRQ f f f
PAGE 50
ZKHUH e LV DQ DV \HW XQNQRZQ RSHUDWRU RQ I! 6XEVWLWXWLQJ LQWR (T f DQG QHJOHFWLQJ GHULYDWLYHV RI DORQH OHDGV WR WKH UHVXOW !]] c!= f 7KH HOOLSWLF HTXDWLRQ f LV QRZ SXW LQWR WKH IRUP RI (T f /HWWLQJ S &&J IRU QRWDWLRQDO FRQYHQLHQFH DQG H[SDQGLQJ VRPH RI WKH WHUPV RI (T f LQ WKHLU VSDWLDO FRRUGLQDWHV JLYHV >S!]f] S!\fY NS! X! D LZ9IF f 8ff! LDZ! a^8Nf] a ^9LYfY 89@!]fY 89!Yf] M8 9M f %RRLM FKRVH WR QHJOHFW WHUPV LQYROYLQJ WKH VTXDUH RI WKH PHDQ FXUUHQW FRPSRQHQWV DVVXPLQJ WKDW WKH\ DUH VPDOOHU WKDQ WHUPV LQYROYLQJ S DQG WKHQ DUUDQJHG WKH DERYH HTXDWLRQ DV !]] 3 O>3] LX8f![ N A MnM fÂ§ 4 f ZKHUH 0! Z D WZ9IF f f mZfÂ LX9c!Y >S!YfY f .LUE\ SRLQWV RXW WKDW WKLV FKRLFH E\ %RRLM OHDGV WR KLV LQDELOLW\ WR GHULYH WKH FRUUHFW ZDYH DFWLRQ HTXDWLRQ .LUE\ LQVWHDG GRHV QRW GURS DQ\ RI WKH VTXDUHV RI WKH PHDQ FXUUHQW FRPSRQHQWV XQWLO DIWHU WKH VSOLWWLQJ LV FRPSOHWH %\ DVVXPLQJ WKDW WKH ZDYHV DUH RULHQWHG LQ WKH [ GLUHFWLRQ WKH GLVSHUVLRQ UHODWLRQ LV ZULWWHQ DV D RM fÂ§ N8 f OHDGLQJ WR WKH UHVXOW MM fÂ§ R fÂ§ XN8 ^N8f f 8VLQJ WKLV UHODWLRQVKLS .LUE\ WKHQ DUUDQJHG (T f DV K S YnfnS 8rffA f L f
PAGE 51
ZKHUH 0O ^XN8 LM\Kf LUY`O^89ÂYf]^89Â]fY ^^S9nfO\f W:9A f &RPSDULVRQ ZLWK f JLYHV rf:A\f nf < SAf f RU VROYLQJ IRU DQG e f! fc"Aff nf Â ULIUmrfLOSAAfr f %RRLM SRLQWV RXW WKDW WKH VSOLWWLQJ PDWUL[ DSSURDFK HPSOR\HG E\ 5DGGHU f LV HTXLYDOHQW WR WKH FKRLFH 5 r N>S8@f eU NLS8f 8VLQJ WKH UHVXOWV f DQG f LQ (T f DORQJ ZLWK WKH GHILQLWLRQ f JLYHV WKH SDUDEROLF HTXDWLRQ 0 N^S 8f %\ H[SDQGLQJ WKH SVHXGRRSHUDWRUV LQ WKH IRUP 0 A eM WNNrS 8fr A 0 NS8f Y Af Lf! r L 0 NS fÂ§ 8f? N^S8f 0 ?r B 0 L
PAGE 52
ZKHUH f^? 5DGGHU L %RLM f DQG ZKHUH 3 3L M IRU ERWK DSSUR[LPDWLRQV .LUE\ FKRVH WKH DSSUR[LPDWLRQ ZLWK 3L VLQFH IRU KLV GHULYDWLRQ ZKLFK ZDV IRU ZHDNO\ QRQOLQHDU ZDYHV LW VKRXOG FRUUHVSRQG WR WKH LQIRUPDWLRQ FRQWDLQHG LQ WKH QRQOLQHDU 6FKURGLQJHU HTXDWLRQ :LWK 3L (T f EHFRPHV NS XfIF ^IFS Xf`[ LNS 8fM! n0! f ZKHUH 0I! LV GHILQHG E\ (T f $Q HTXDWLRQ IRU WKH FRPSOH[ DPSOLWXGH $ LV WKHQ REWDLQHG E\ PDNLQJ WKH VXEVWLWXWLRQ f ZKLFK DIWHU GURSSLQJ VTXDUHV RI WKH FRPSRQHQWV RI WKH PHDQ FXUUHQW UHVXOWV LQ & 8f$ ? n$ ef9$ ? f$ L >FFn mf f "$ nf .LUE\ WKHQ IXUWKHU PRGLILHG WKH HTXDWLRQ E\ D VKLIW WR D UHIHUHQFH SKDVH IXQFWLRQ XVLQJ WKH UHODWLRQ $ $nHLLfING[f f ZKHUH N FDQ EH WDNHQ DV DQ DYHUDJHG ZDYH QXPEHU UHVXOWLQJ LQ &W 8f$n; LLF Nf&c 8f$n $n 7KH VKLIW WR D UHIHUHQFH SKDVH IXQFWLRQ LV LQWHQGHG WR LQFRUSRUDWH WKH SKDVH LQIRUPDWLRQ LQ WKH FRPSOH[ DPSOLWXGH VR WKDW WKH LQVWDQWDQHRXV ZDWHU VXUIDFH FDQ EH HDVLO\ UHFRYHUHG LQ WKH IRUP T[\f L^$9eG` f
PAGE 53
ZKHUH GHVLJQDWHV WKH UHDO SDUW 7KLV LV DQ LPSRUWDQW FRQVLGHUDWLRQ HVSHFLDOO\ ZKHQ REn WDLQLQJ WKH UDGLDWLRQ VWUHVVHV ,W LV LPSRUWDQW WKDW LQ REWDLQLQJ (T f WKDW VTXDUHV RI WKH SURGXFWV RI WKH FXUUHQW FRPSRQHQWV EH UHWDLQHG XQWLO WKH ODVW VWHS 7KH UHDVRQ IRU WKLV LV VHHQ LQ WKDW WHUPV VXFK DV N89 DQG N8 DUH FRPELQHG ZLWK WHUPV RI RM9 DQG X8 XVLQJ WKH GLVSHUVLRQ UHODWLRQVKLS WR JLYH WHUPV RI D8 DQG t9 (TXDWLRQ f KDV EHHQ VXFFHVVIXOO\ XVHG E\ .LUE\ LQ WKH IRUP JLYHQ DQG ZLWK PRGLILn FDWLRQV 7KH PRGLILFDWLRQV DUH WKH DGGLWLRQ RI D QRQOLQHDU WHUP ZKLFK PDNHV WKH HTXDWLRQ DSSOLFDEOH IRU ZHDNO\ QRQOLQHDU ZDYHV +RZHYHU IRU SUHVHQW SXUSRVHV (T f SUHVHQWV D FHUWDLQ SUREOHP ZKLFK ZLOO VKRUWO\ EH FRUUHFWHG 7KH SUREOHP LV WKDW IRU ZDYHV SURSDJDWLQJ DW DQ DQJOH WR WKH [ D[LV LW LV QRW SRVVLEOH WR REWDLQ D FRUUHFW IRUP RI WKH FRQVHUYDWLRQ RI ZDYH DFWLRQ 7KH FDVH RI QR FXUUHQW RQ D EHDFK ZLWK VWUDLJKW DQG SDUDOOHO FRQWRXUV LV JLYHQ DV DQ LOOXVWUDWLRQ 7KH HTXDWLRQ IRU WKH FRQVHUYDWLRQ RI ZDYH DFWLRQ UHGXFHV WR WKH FRQVHUYDWLRQ RI ZDYH HQHUJ\ IOX[ ,Q PDWKHPDWLFDO WHUPV WKLV LV A&VFRVDM f ZKHUH D LV WKH ZDYH DPSOLWXGH )RU ZDYHV DSSURDFKLQJ WKH EHDFK LQ D GLUHFWLRQ QRUPDO WR WKH EHDFK (T f UHGXFHV WR Â&0n f 0XOWLSO\LQJ (T f E\ WKH FRPSOH[ FRQMXJDWH RI $n DQG DGG WR LW WKH FRPSOH[ FRQMXJDWH HTXDWLRQ RI f PXOWLSOLHG E\ $n \LHOGV >FW?$?@[ f ZKLFK LV (T f IRU WKH VSHFLDO FDVH RI FRV +RZHYHU LW LV QRW SRVVLEOH WR REWDLQ (T f IURP (T f IRU WKH JHQHUDO FDVH RI ZDYHV DW DQJOH RQ D EHDFK ZLWK VWUDLJKW DQG SDUDOOHO FRQWRXUV 7KH LPSRUWDQFH RI WKLV LQDELOLW\ WR JLYH WKH FRUUHFW IRUP RI (T f LV WKDW LI HQHUJ\ IOX[ RXWVLGH WKH EUHDNHU OLQH LV QRW FRQVHUYHG WKHQ JUDGLHQWV RI WKH 6[Y UDGLDWLRQ VWUHVV
PAGE 54
ZLOO JHQHUDWH FXUUHQWV RIIVKRUH RI WKH EUHDNHU OLQH LQ WKH FLUFXODWLRQ PRGHO $OWKRXJK VXFK FXUUHQWV LQ JHQHUDO ZLOO EH RI D VPDOO PDJQLWXGH WKH\ FDQ FDXVH VRPH QXPHULFDO FRPSXWDWLRQDO SUREOHPV 7KH SUREOHP WKRXJK LV HDVLO\ FRUUHFWHG E\ XVLQJ DV WKH GLVSHUVLRQ UHODWLRQVKLS Xf D N FRV 8 f DQG WKH VXEVWLWXWLRQ BM$[\f MI NFRWWG[f f 5DWKHU WKDQ XVLQJ D VSOLWWLQJ PHWKRG WKH SURFHGXUH ZLOO EH WR JR GLUHFWO\ IURP (T f XVLQJ (T f DQG ZULWLQJ X! fÂ§ D XLN FRV 8 fÂ§ N FRV 8f f LQ SODFH RI (T f 'URSSLQJ WKH Mf][ WHUP DQG DOO WHUPV FRQWDLQLQJ SURGXFWV RI WKH FXUUHQW FRPSRQHQWV WKDW FDQQRW EH UHGXFHG DV H[SODLQHG DERYH UHVXOWV LQ WKH IROORZLQJ SDUDEROLF HTXDWLRQ &J FRV 8n &J FRV 8f$= fÂ§ $W $ 9 $Y iD FRVf$ L &&J^Af\A$ 1RZ DGGLQJ D SKDVH VKLIW W FRWG]M LFRDIG]f VR WKDW WKH IUHH VXUIDFH FDQ EH UHFRYHUHG E\ UM[f9f $nHLIOFRVHnG]f f f f ZKHUH LV WKH ZDYH DQJOH REWDLQHG E\ 6QHOOfV ODZ LQ WKH DEVHQFH RI DQ\ REVWUXFWLRQV RU \ YDULDWLRQ RI WKH GHSWK RU FXUUHQW 7KH IROORZLQJ JRYHUQLQJ HTXDWLRQ LV WKXV REWDLQHG &J&RV2 8n &J FRV 8f$n; LN FRV N FRV f&W FRV 8f$n fÂ§ $W $n 9$nY f $! a N&LA a &rV a &&ÂƒaÂfY A$n f
PAGE 55
'HWDLOV RI WKH DOJHEUD DQG DVVXPSWLRQV LQYROYHG LQ REWDLQLQJ (T f DUH JLYHQ LQ $Sn SHQGL[ )RU FRPSXWDWLRQDO SXUSRVHV LW ZLOO EH DVVXPHG WKDW FRV FRV )URP (T f WKH FRUUHFW IRUP RI WKH FRQVHUYDWLRQ RI ZDYH DFWLRQ FDQ EH REWDLQHG )RU WKH FDVH RI QR \ YDULDWLRQ RI WKH ZDYH QXPEHU (T f UHGXFHV WR &I FRV 8f$n; fÂ§ / &JFRV 8 $n 9$> OLp f 0XOWLSO\LQJ (T f E\ WKH FRPSOH[ FRQMXJDWH RI $ DQG DGGLQJ WR LW $ WLPHV WKH FRPSOH[ FRQMXJDWH HTXDWLRQ RI f \LHOGV & FRV 8f ?9?$?@ D ; D f ZKLFK LV D FORVHU DSSUR[LPDWLRQ WR WKH H[DFW H[SUHVVLRQ ZKLFK VKRXOG KDYH D & VLQ r 9f WHUP LQVWHDG RI 9 LQ WKH \ GHULYDWLYH WHUP (QHUJ\ 'LVVLSDWLRQ 'XH WR :DYH %UHDNLQJ .LUE\ f IRXQG D PHWKRG RI UHODWLQJ WKH GLVVLSDWLRQ FRHIILFLHQW Z WR WKH ZRUN RI 'DOO\ f ZKR SURSRVHG WKDW WKH GHFD\ RI HQHUJ\ IOX[ LQ WKH VXUI ]RQH ZDV SURSRUWLRQDO WR WKH DPRXQW RI H[FHVV HQHUJ\ IOX[ 7KH H[FHVV HQHUJ\ IOX[ LV WKH GLIIHUHQFH EHWZHHQ WKH DFWXDO IOX[ DQG D VWDEOH HQHUJ\ IOX[ ,Q PDWKHPDWLFDO WHUPV WKLV UHODWLRQVKLS LV H[SUHVVHG DV (&Jf[ A>(&J(&JfW@ f ZKHUH (& ( F K WKH HQHUJ\ IOX[ ZDYH HQHUJ\ ASJ+ G7 WKH UDWH RI HQHUJ\ SURSDJDWLRQ ZKLFK LV fÂ§ RN FRQVWDQW WKH ZDWHU GHSWK
PAGE 56
(&Wf WKH fVWDEOHf HQHUJ\ IOX[ + WKH ORFDO ZDYH KHLJKW 7KH VWDEOH HQHUJ\ IOX[ LV REWDLQHG E\ DVVXPLQJ LW WR EH D IXQFWLRQ RI WKH KHLJKW RI D ZDYH SURSDJDWLQJ RYHU D IODW ERWWRP DIWHU EUHDNLQJ FHDVHV DQG WKDW WKLV KHLJKW LV D FRQVWDQW WLPHV WKH ZDWHU GHSWK 7KH FRQVWDQWV DQG ZHUH GHWHUPLQHG E\ DQDO\VLV RI H[SHULPHQWDO GDWD REWDLQHG IURP +RULNDZD DQG .XR f .LUE\ UHODWHG 'DLO\fV ZRUN WR WKH GLVVLSDWLRQ WHUP Z E\ GHULYLQJ D FRQVHUYDWLRQ HTXDWLRQ IRU WKH ZDYH DFWLRQ M *DUUHWW f DQG %UHWKHUWRQ DQG *DUUHWW f VKRZHG WKDW IRU ZDYHV SURSDJDWLQJ LQ D PRYLQJ DQG YDULDEOH PHGLXP WKH TXDQWLW\ WKDW LV FRQVHUYHG LV WKH ZDYH DFWLRQ f ZKLFK LV WKH ZDYH HQHUJ\ GLYLGHG E\ WKH LQWULQVLF IUHTXHQF\ .LUE\ GHULYHG D ZDYH DFWLRQ FRQVHUYDWLRQ HTXDWLRQ IURP WKH OLQHDU K\SHUEROLF ZDYH (T f XVLQJ WKH GHILQLWLRQ RI : DV JLYHQ E\ (T f A Yr f 8fA 9IF&&9IFIW U Nn&&0 DZr f .LUE\ IROORZV WKH DSSURDFK RI RQVVRQ f DVVXPLQJ WKDW r FDQ EH UHSUHVHQWHG DV r 5HLH f ZKHUH 5 DQG DUH UHDO )URP OLQHDU WKHRU\ LW LV REYLRXV WKDW 5 _$_ f 7R ORZHVW RUGHU WKH GHULYDWLYHV RI k DUH DSSUR[LPDWHG DV X f 9KH N f 6XEVWLWXWLQJ LQWR f JLYHV WKH FRPSOH[ HTXDWLRQ QS '5 %M7 9rn 9}&&9N5f LU% .UWf ^9N f Rf5^Rf Â‘ 9K5 9f f N&&f5 N&& Â‘ 9NO WZÂ! f
PAGE 57
6LQFH 5 DQG DUH UHDO WKH UHDO DQG LPDJLQDU\ SDUWV RI f PXVW LQGLYLGXDOO\ HTXDO ]HUR 7DNLQJ WKH LPDJLQDU\ SDUW DQG PXOWLSO\LQJ E\ 5 JLYHV 9r f N&&Jf5 DZ5 f ZKLFK XVLQJ WKH VXEVWLWXWLRQ N&&D D&W LV ZULWWHQ DV ^U5fW f D5 &\ff DZ5 8VLQJ f 5 PD\ EH UHSODFHG E\ 5L S D D ?D ZKLFK \LHOGV WKH ZDYH DFWLRQ FRQVHUYDWLRQ HTXDWLRQ $VVXPLQJ D WLPH VWHDG\ ZDYH ILHOG DQG QHJOHFWLQJ FXUUHQWV f EHFRPHV f f f >(&Wf[ Z ( f &RPSDULQJ f ZLWK f DQG QRWLQJ WKDW &JfW &J Z LV ZULWWHQ DV f ZKHUH + LV WKH ORFDO ZDYH KHLJKW JLYHQ E\ _$_ ,Q REWDLQLQJ (T f .LUE\ QHJOHFWHG WKH HIIHFWV RI WKH FXUUHQWV +RZHYHU WKH VDPH H[SUHVVLRQ f LV REWDLQHG ZKHQ FXUUHQWV DUH LQFOXGHG LQ WKH GHULYDWLRQ 'DOO\ f H[WHQGHG KLV ZRUN WR LQYHVWLJDWH ZDYH EUHDNLQJ RQ FXUUHQWV DQG SURSRVHG WKH IROORZLQJ HTXDWLRQ 8f I f@L O(&r a f )RU WKH FDVH RI D VWHDG\ VWDWH ZDYH ILHOG SURSDJDWLQJ LQ WKH [ GLUHFWLRQ WKH FRQVHUYDWLRQ RI ZDYH DFWLRQ (T f UHGXFHV WR f
PAGE 58
6HWWLQJ WKH ULJKW KDQG VLGHV RI (TV f DQG f HTXDO WR HDFK RWKHU UHVXOWV LQ (T f $W ILUVW JODQFH LW PD\ DSSHDU XQXVXDO WKDW WKH VDPH H[SUHVVLRQ IRU WKH GLVVLSDWLRQ FRHIILFLHQW UHVXOWV IRU WKH FDVH RI FXUUHQWV DV IRU WKH FDVH RI QR FXUUHQWV 7KLV FDQ EH H[SODLQHG LQ SDUW E\ WKH IDFW WKDW 'DLO\fV ZRUN ZDV ZLWK ZDYHV ZKHUH EUHDNLQJ ZDV GHSWK OLPLWHG )RU WKH FDVH RI FXUUHQW OLPLWHG EUHDNLQJ IRU H[DPSOH D VWRSSLQJ FXUUHQWf DQRWKHU H[SUHVVLRQ LV QHFHVVDU\ ,Q WKLV VWXG\ (T f ZLOO EH XVHG VLQFH LQ JHQHUDO WKH FXUUHQW ILHOG ZLOO EH PRGHUDWHO\ ZHDN 6XFK DQ DVVXPSWLRQ RI ZHDN FXUUHQWV ZDV LQYRNHG LQ WKH GHULYDWLRQ RI WKH JRYHUQLQJ SDUDEROLF HTXDWLRQ
PAGE 59
&+$37(5 ),1,7( ',))(5(1&( 6&+(0(6 $1' %281'$5< &21',7,216 1XPHULFDO 6ROXWLRQ RI WKH *RYHUQLQJ (TXDWLRQV 7KH ILQLWH GLIIHUHQFLQJ RI WKH JRYHUQLQJ HTXDWLRQV LQ WKH FLUFXODWLRQ PRGHO LV GHULYHG E\ D PDWUL[ DQDO\VLV DV VXJJHVWHG E\ 6KHQJ DQG %XWOHU f $ UDWLRQDO DQDO\VLV RI WKH FRPSOHWH (TV f LQFOXGLQJ DOO WKH QRQOLQHDU WHUPV LV QRW SRVVLEOH +RZHYHU JXLGDQFH FDQ EH REWDLQHG WKURXJK DQ DQDO\VLV RI WKH OLQHDUL]HG HTXDWLRQV )RU WKLV SXUSRVH WKH OLQHDU HTXDWLRQV IRU WKH VSHFLDO FDVH RI D IODW ERWWRP DQG ZLWKRXW DQ\ RI WKH IRUFLQJ WHUPV ZLOO EH H[DPLQHG GIM DX DY Q 7 'fÂ§ 'fÂ§ GW G[ G\ DX GIM OWG] r G; JG$ DW \G\ /HW : EH D YHFWRU RI IM8 DQG 9 : 8 9 8VLQJ (T f WKH JRYHUQLQJ HTXDWLRQV DUH ZULWWHQ DV f f f f :W $:; %:\ ZKHUH $ DQG % DUH WKH IROORZLQJ PDWULFHV f n n n n $ DQG % f 7R ILQG DQ LPSOLFLW ILQLWH GLIIHUHQFH VFKHPH IRU WKH JRYHUQLQJ HTXDWLRQV (T f LV ILQLWH GLIIHUHQFHG DV M\QL B M\Q $7 $:" %:A f
PAGE 60
7KLV HTXDWLRQ LV VROYHG IRU :Q LQ WHUPV RI :Q $, $ \f:Q :Q f ZKHUH $7 $$;^ DQG L%H=L $< f $M$Y:QL LV DGGHG WR WKH OHIW KDQG VLGH RI WKH HTXDWLRQ DQG $Y:Q LV DGGHG DQG VXEn WUDFWHG WR WKH ULJKW KDQG VLGH RI (T f VR WKDW WKH HTXDWLRQ FDQ EH IDFWRUHG WR JHW $f $Yf:QO $Yf:Q $Y:Q f $Q LQWHUPHGLDWH YDOXH : r LV LQWURGXFHG VR WKDW WKH DERYH HTXDWLRQ FDQ EH ZULWWHQ DV WZR HTXDWLRQV O $,f:r O$ff:Q f ;Yf:Q :r $ Y:Q f 1RWH WKDW E\ HOLPLQDWLQJ :r IURP WKH WZR (TV f DQG f (T f LV UHFRYHUHG ZLWK WKH DGGLWLRQ RI DQ HUURU RI $$\:QO ZKLFK KDV EHHQ LQWURGXFHG :ULWLQJ RXW WKH FRPSRQHQW HTXDWLRQV RI (T f JLYHV $7 $7 Yr \Q A6YIMQ f f f f :KLOH ZULWLQJ RXW WKH FRPSRQHQW HTXDWLRQV RI (T f JLYHV 9Q Qr 'A\YQ XQ Xr \Q F IcQO B \} F Q Y $<\9 aY J$<: f f f
PAGE 61
8VLQJ (T f WR HOLPLQDWH 9r LQ (T f DQG f WR HOLPLQDWH 8r LQ (T f WKH IROORZLQJ DUH REWDLQHG )RU WKH [ GLUHFWLRQ $7 $7 U GD[6L8 'D\69fn $7 8Q JfÂ§WIcr 8Q DQG IRU WKH \ GLUHFWLRQ $7 $7 U Yn mf $7 \Q A6YIMQ YQ f f f f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
PAGE 62
LV ILQLWH GLIIHUHQFHG DV S)?8Q?8Q DQG WKH IULFWLRQ WHUP DV VXJJHVWHG E\ /RQJXHW+LJJLQV LV GLIIHUHQFHG DV S) _XRUM_ 8Q ZKHUH _XRUM_ LV LQ WHUPV RI YDOXHV DW WKH Q WLPH OHYHO +RZHYHU ZKHQ XVLQJ WKH LQWHJUDO TXDGUDWLF H[SUHVVLRQ LW LV HYDOXDWHG HQWLUHO\ LQ WHUPV RI WKH SUHVHQW RU NQRZQ YDOXHV 7KH QRQOLQHDU DGYHFWLYH DFFHOHUDWLRQ WHUPV DUH DGGHG DW WKH QWIO WLPH OHYHO XVLQJ FHQWUDO GLIIHUHQFH GHULYDWLYHV ,Q WKH [ GLUHFWLRQ PRPHQWXP HTXDWLRQ WKHVH WHUPV DUH 7 ALfÂ§OM L >LU L L W L WU ? AfO fÂ§ XM $; K YrLY}O A\ f DQG LQ WKH \ GLUHFWLRQ HTXDWLRQ WKH\ DUH M:M 88 9LO f ,Q WKH PRGHO RI 9HPXODNRQGD f ZKLFK LV EDVHG XSRQ WKH :,)0 :DWHUZD\V ([n SHULPHQW 6WDWLRQ ,PSOLFLW )ORRGLQJ 0RGHOf %XWOHU f PRGHO D WKUHH WLPH OHYHO OHDSIURJ VFKHPH LV XVHG 7KLV JLYHV WKH DGYDQWDJH WKDW DOO RI WKH IRUFLQJ WHUPV DQG WKH QRQOLQHDU DFFHOHUDWLRQ WHUPVf DUH DGGHG DW WKH WLPH FHQWHUHG OHYHO +RZHYHU WKLV KDV VRPH GLVDGn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f E\ &\FRVL 8f \LHOGV I & FRV %8 M $n[ LN FRV Â• fÂ§ N FRV f $ A A $n &FRV f &FRV Âf Y L N&DO fÂ§ FRVÂf $ Y &W FRV 8f
PAGE 63
7KH IROORZLQJ QRWDWLRQ LV XVHG WR GHVFULEH WKH ILQLWH GLIIHUHQFLQJ RI (T f )M UHSUHVHQWV WKH YDOXH RI )LMf ZKHUH ) LV DQ\ WHUP DQG L DQG M GHVLJQDWH WKH JULG URZ DQG JULG FROXPQ %HFDXVH RI WKH VWDJJHUHG JULG V\VWHP XVHG IRU WKH FLUFXODWLRQ PRGHO DV SUHYLRXVO\ H[SODLQHG LQ ZKLFK WKH YHORFLW\ FRPSRQHQWV DUH VSHFLILHG DW WKH JULG HGJHV WKH YHORFLW\ WHUPV 8 DQG 9 DUH DYHUDJHG DFURVV WKH JULG VR DV WR SURGXFH D JULG FHQWHUHG WHUP IRU XVH LQ WKH ZDYH PRGHO 7KXV f LQ WKH ZDYH PRGHO LV DFWXDOO\ ?8M Aff DQG VLPLODUO\ IRU 9 DQG RWKHU JULG GHVLJQDWLRQV $ &UDQN1LFKROVRQ ILQLWH GLIIHUHQFLQJ VFKHPH LV XVHG IRU (T f $OO WHUPV QRW LQYROYLQJ DQ [ GHULYDWLYH DUH DQ DYHUDJH RI WKH YDOXH RI WKH WHUP RQ WKH r JULG URZ DQG RQ WKH L O JULG URZ 7HUPV KDYLQJ \ GHULYDWLYHV DUH FHQWHUHG DW} M RQ WKH } URZ DQG r M ;nfÂ§;nf RQ WKH m JULG URZ 7KH RQH WHUP ZLWK DQ [ GHULYDWLYH LV ZULWWHQ DV fÂ§ f 7KXV DOO WHUPV DUH FHQWHUHG DW } _M 7KH ILQLWH GLIIHUHQFLQJ RI (T f LV ZULWWHQ DV $; n >MIFf FRV IFf FRV f$ MIFr FRV r Nf FRV rf$M@ I >&J FRV J &fR@m! a >&J FRV J IfF[@m K $; ? ?&D FRV 8fU@IO ?&D FRV 8fR 9D &H FRV 8f U , $< O`ff 9R 9 _DL D@ Af WfA $< &\FRV 8f &J FRV &f n U9}fL? 9DffW? O &c FRV f f $< Zrf8L 3+ $< W N&J^O FRVfOr $ &J FRV 8f N&c fÂ§ FRV f Â‘f &J FRV 8f LO n3 mn 9M ff $! 9f &J FRV Lf &J FRV 8ff Â Zr $ Zr $? f ^&\FRV Xfr &JFRV Xff M Anf
PAGE 64
ZKHUH S && IRU QRWDWLRQDO FRQYHQLHQFH DQG WKH \ GHULYDWLYH RI S A M LV ZULWWHQ DV f U8L SMf >eff a eff@ S! SaLf >Iff fnf $^&W FRV 9f@fn ?&W FRV 8fD@ >LFf n $; >& FRV 8fDI >&WI FRV 8fD@ _`m &J FRV 8fD ML $< 9 >Y} : $r r N&J^? fÂ§ FRV f ^&HFRV 8f $< $r &VFRVL Lf $Â Q 9f 7 ` &J FRV 8ff BU >&J FRV 8ff r :KHQ (T f LV ZULWWHQ RXW XVLQJ (T f DQG Zr REWDLQHG WKURXJK XVH RI f WKHUH UHVXOWV IRU DQ\ SDUWLFXODU YDOXHV RI L DQG M DQ HTXDWLRQ IRU WKH WKUHH XQNQRZQ YDOXHV $`rr $f DQG $ÂÂ? 7KH HTXDWLRQ LV SXW LQWR WKH IRUP DL D<$ KL Â` &M $MW? GM f ZKHUH WKH FRHIILFLHQWV DM EM DQG &M DUH NQRZQ TXDQWLWLHV LQYROYLQJ WKH SURSHUWLHV RI WKH HQYLURQPHQW GHSWK FXUUHQW FHOHULW\ HWFf DQG WKH GM LQYROYHV WKH HQYLURQPHQWDO SURSHUWLHV DQG WKH NQRZQ YDOXHV RI WKH DPSOLWXGH RQ WKH } JULG URZ :KHQ (T f LV ZULWWHQ IRU DOO WKH M YDOXHV M HTXDO WR WKURXJK 1 IRU D VSHFLILHG L YDOXH WKH HQVHPEOH RI 1 HTXDWLRQV WDNHV WKH IRUP RI D WULGLDJRQDO PDWUL[ HTXDWLRQ )RU LOOXVWUDWLYH SXUSRVHV OHW 1 EH HTXDO WR IRXU 7KH UHVXOWLQJ PDWUL[ HTXDWLRQ LV U !L DM FL $f f Â n D t F $ G DV E &M $r Â6 D Â & Gr / f
PAGE 65
(TXDWLRQ f FRQWDLQV IRXU VLPXOWDQHRXV HTXDWLRQV ZLWK VL[ XQNQRZQV 7ZR ERXQGn DU\ FRQGLWLRQV VKRXOG EH VSHFLILHG $W HDFK ERXQGDU\ D IXQFWLRQDO UHODWLRQVKLS LV HVWDEOLVKHG EHWZHHQ WKH JULG LQVLGH WKH GRPDLQ DQG WKH JULG RXWVLGH RI WKH GRPDLQ )RU WKH ZDYH SURSn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f $W WKH XSSHU ERXQGDU\ WKH IXQFWLRQDO UHODWLRQVKLS IRU D UHIOHFWLYH ERXQGDU\ LV $QL $Q f )RU WKH QRQUHIOHFWLYH ERXQGDU\ WKH PDWKHPDWLFDO VWDWHPHQW LV fÂ§LP$ ZKHUH P LV WKH ORQJVKRUH FRPSRQHQW RI WKH ZDYHQXPEHU 6LQFH WKH ODWHUDO ERXQGDULHV DUH VXIILFLHQWO\ IDU IURP DQ\ GLIIUDFWLRQ HIIHFWV RQO\ UHIUDFWLRQ DQG VKRDOLQJ HIIHFWV DUH SUHVHQW 7KXV XVLQJ 6QHOOfV /DZ P N VLQ N VLQR 7KH QRQ UHIOHFWLYH ERXQGDU\ FRQGLWLRQ LV ILQLWH GLIIHUHQFHG DW WKH ORZHU ERXQGDU\ DV $L fÂ§ $T LP fÂ§e\aa <>$O $r@ f ZKLFK JLYHV $T IMQ$
PAGE 66
)LJXUH 'HILQLWLRQ VNHWFK RI JULG QRPHQFODWXUH DQG FRRUGLQDWH V\VWHP HTXDWLRQV IRU 1 XQNQRZQV 5HWXUQLQJ WR WKH LOOXVWUDWLYH H[DPSOH ZKHUH 1 LV HTXDO WR IRXU WKH PDWUL[ (T f EHFRPHV L ?G? FL I! &b 2 &6 n $? n f GL n $n G Â > G f ZKHUH ? DQG Q UHSUHVHQW WKH IXQFWLRQDO UHODWLRQVKLSV DV GHVFULEHG DERYH RQ WKH M O DQG M 1 ERXQGDULHV (TXDWLRQ f LV VROYHG HDVLO\ IRU WKH XQNQRZQ $f XVLQJ D GRXEOH VZHHS PHWKRG ,Q WKH FRPSXWHU SURJUDP WKH VXEURXWLQH &75,'$ XVHV DQ DOJRULWKP RI &DUQDKDQ /XWKHU DQG :LONHV f WR HIILFLHQWO\ DFFRPSOLVK WKH GRXEOH VZHHS VROXWLRQ RI HT f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
PAGE 67
ZDWHU VXUIDFH OHYHO LV QRW FKDQJHG E\ DQ\ RI WKH FLUFXODWLRQ ZLWKLQ WKH PRGHO PDWKHPDWLFDOO\ (7$f 1 f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f LPSRVHV WKH FRQGLWLRQ GUf GUf fÂ§ & fÂ§ GW G[ f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f WKHQ WKH UDGLDWLRQ FRQGLWLRQ ZLOO EHFRPH LPSHUDWLYH $W WKH VKRUHOLQH WKH ERXQGDU\ FRQGLWLRQ LV WKDW WKHUH LV QR IORZ LQWR RU RXW RI WKH EHDFK 6LQFH WKH PRGHO DOORZV IRU IORRGLQJ RI WKH EHDFK WKH LQGH[ RI WKH VKRUHOLQH FKDQJHV 7KH PRGHO DOORZV IRU GLIIHUHQWLDO IORRGLQJ DORQJ WKH VKRUHOLQH IRU H[DPSOH GXH WR GLIIHUHQFHV RI ZDYH LQWHQVLW\ LQ WKH OHH RI VWUXFWXUHV )RU HDFK JULG FROXPQ LQ WKH \ GLUHFWLRQ WKH ODVW JULG LV WKH ,:(7f JULG ZKHUH ,:(7f LV DOORZHG WR FKDQJH ZLWK FKDQJHV RI WKH PHDQ ZDWHU OHYHO 7KH QXPHULFDO VWDWHPHQW RI WKH QR IORZ ERXQGDU\ FRQGLWLRQ DW WKH VKRUHOLQH LV W,:(7fOf 1 f
PAGE 68
7KLV ERXQGDU\ FRQGLWLRQ LV XVHG LQ RWKHU PRGHOV WKRXJK QRW DOO PRGHOV DOORZ IRU GLIIHUHQWLDO IORRGLQJ RI WKH VKRUHOLQH
PAGE 69
DULHV &ORVHG ERXQGDULHV DUH HDVLO\ DSSOLHG E\ WKH FRQGLWLRQ RI QR IORZ DFURVV WKH ERXQGDU\ 6XFK D FKRLFH RI ERXQGDU\ FRQGLWLRQV ZRXOG DFFXUDWHO\ GXSOLFDWH WKH SK\VLFDO FRQGLWLRQV H[n LVWLQJ LQ D ODERUDWRU\ EDVLQ RU IRU SURWRW\SH FRQGLWLRQV ZLWKLQ D KDUERUf 2SHQ ERXQGDULHV DUH REWDLQHG E\ VSHFLI\LQJ GHULYDWLYH FRQGLWLRQV ,Q WKH SUHVHQW VWXG\ DQ RSHQ ERXQGDU\ LV REWDLQHG E\ VSHFLI\LQJ WKDW WKHUH DUH QR JUDGLHQWV RI WKH GHSWK DQG WKH ZDYH SURSHUn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n SRVHV 7KH RIIVKRUH FRQGLWLRQ LV PHUHO\ WKH VSHFLILHG ZDYH DPSOLWXGH DQG ZDYH DQJOH 1R FRQGLWLRQ LV QHHGHG DW WKH VKRUHOLQH VLQFH WKH SDUDEROLF ZDYH PRGHO PDUFKHV JULG URZ E\ JULG JURZ DQG QHHGV RQO\ WKH ODWHUDO FRQGLWLRQV 7KH RIIVKRUH FRQGLWLRQ LV WKDW HDFK WLPH WKH ZDYH PRGHO LV FDOOHG WKH ZDYH DPSOLWXGH RQ WKH ILUVW JULG URZ LV VSHFLILHG DV $f &Wf aa$f 1 f ZKHUH $f LV GHILQHG LQ WKH LQSXW VXEURXWLQH DFFRUGLQJ WR $f HfrRVLQ RD< f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
PAGE 70
R )LJXUH 6WDUWXS IXQFWLRQ DFFRUGLQJ WR (T ZLWK &L DQG &L WKH ZDYH PRGHO ZKLFK LV FRQWUDU\ WR WKH FRQGLWLRQ LQ WKH FLUFXODWLRQ PRGHO FDQ LQWURGXFH LQVWDELOLWLHV LQ WKH RYHUDOO PRGHO 7KH LQWHUDFWLRQ RI WKH WZR PRGHOV LV PRUH VHQVLWLYH WKDQ HLWKHU RI WKH PRGHOV LQGLYLGXDOO\ 7KH &Wf WHUP LQ HTXDWLRQ LV D VWDUWXS IXQFWLRQ ZKLFK VPRRWKO\ JRHV IURP ]HUR WR XQLW\ VR DV WR EULQJ WKH LQFLGHQW ZDYH KHLJKW IURP ]HUR WR + DQG PLQLPL]H DQ\ VKRFN ORDGLQJ RI WKH PRGHO 0DWKHPDWLFDOO\ LW LV JLYHQ E\ &Wf W WDQKfÂ§ &Lf Z f ZKHUH &? DQG &L DUH DUELWUDU\ FRQVWDQWV WR DGMXVW WKH VORSH DQG RIIVHW RI WKH FXUYH )LJn XUH VKRZV HTXDWLRQ IRU &? DQG &L HTXDO WR DQG UHVSHFWLYHO\ 7KH ODWHUDO ERXQGDU\ FRQGLWLRQV DUH HLWKHU RSHQ RU UHIOHFWLYH 7KH UHIOHFWLYH FRQGLWLRQ LV H[SUHVVHG DV f ZKLOH WKH RSHQ FRQGLWLRQ LV H[SUHVVHG DV Â‘DL P$ f ZKHUH P LV WKH ORQJVKRUH FRPSRQHQW RI WKH ZDYHQXPEHU ZKLFK LQ WKH DEVHQFH RI GLIIUDFWLRQ HIIHFWV UHPDLQV FRQVWDQW P N VLQ Â‘ N VLQ f
PAGE 71
'LUHFWLRQ RI :DYH 3URSDJDWLRQ [ )LJXUH :DYHV SDVVLQJ WKURXJK D ODWHUDO ERXQGDU\ XQDIIHFWHG E\ WKH ERXQGDU\ DFFRUGLQJ WR 6QHOOfV ODZ (TXDWLRQ RU DUH XVHG WR HOLPLQDWH WKH DPSOLWXGH RXWVLGH RI WKH GRPDLQ IURP WKH WULGLDJRQDO PDWUL[ HTXDWLRQ 7R LOOXVWUDWH HTXDWLRQ LV ILQLWH GLIIHUHQFHG DW WKH ORZHU ODWHUDO ERXQGDU\ DV $,f $,f B $,f $,f $< LP 6ROYLQJ IRU $,f L B LP$< $,f W9D,f f LPA 7KXV $,f LV HOLPLQDWHG LQ WHUPV RI $,f VLPLODUO\ DW WKH XSSHU ERXQGDU\ $,1f LV HOLPLQDWHG LQ WHUPV RI $,1f 7KH VDPH SURFHGXUH LV XVHG ZLWK HTXDWLRQ ZKLFK \LHOGV WKH VLPSOH UHODWLRQVKLSV $,f $,f f DQG $,1f $,1f f
PAGE 72
,W LV LPSRUWDQW WKDW WKH ERXQGDU\ FRQGLWLRQV DQG EH XVHG DV SUHVFULEHG DERYH 7KH SDUDEROLF HTXDWLRQ JLYHV 1 HTXDWLRQV FRQWDLQLQJ WKH 1 XQNQRZQV $,f IRU 1 7KH WZR ERXQGDU\ FRQGLWLRQV DUH WKXV XVHG WR UHGXFH WKH QXPEHU RI XQn NQRZQV WR D QXPEHU HTXDO WR WKH QXPEHU RI HTXDWLRQV ,I HTXDWLRQ LV ZULWWHQ LQ WHUPV RI $,f DQG $,f DW WKH ORZHU ERXQGDU\ DQG LQ WHUPV RI $,1f DQG $,1f DW WKH XSSHU ERXQGDU\ DQG WKH JRYHUQLQJ ZDYH HTXDWLRQ LV ZULWWHQ IRU 1 WKHQ HUURUV PD\ EH LQWURGXFHG DW WKH ERXQGDULHV 7KLV LV HVSHFLDOO\ WUXH IRU WKH FDVH RI D UHIOHFWLYH ERXQGDU\ )LJXUH VKRZV ZDYHV IUHHO\ SDVVLQJ WKURXJK WKH ODWHUDO ERXQGDULHV ZLWKRXW DQ\ QRWLFHDEOH UHIOHFWLRQ IRU WKH FDVH RI D SODQH ZDYH ZLWK D SHULRG RI VHFRQGV SURSDJDWLQJ RYHU FRQVWDQW GHSWK DW DQ DQJOH RI GHJUHHV WR WKH [ D[LV
PAGE 73
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f DQG WKH H[SHULPHQWDO UHVXOWV RI %RZHQ HW DO f DQG +DQVHQ DQG 6YHQGVHQ f 7KH VHFRQG VHFWLRQ FRYHUV ORQJVKRUH FXUUHQWV DQG FRPSDUHV UHVXOWV ZLWK WKH DQDO\WLFDO VROXWLRQ RI /RQJXHW+LJJLQV DEf DQG H[SHULPHQWDO UHVXOWV RI 9LVVHU f 7KH PRGHO ZDV UXQ IRU WKH FDVH RI D ULSFKDQQHO RQ D SODQH EHDFK &RPSDULVRQV ZLWK WKH QXPHULFDO UHVXOWV RI %LUNHPHLHU DQG 'DLU\PSOH f DQG (EHUVROH DQG 'DOU\PSOH f DUH JLYHQ LQ WKH WKLUG VHFWLRQ $ ULS FXUUHQW ZDV DOVR JHQHUDWHG IRU WKH FDVH RI QRUPDO ZDYHV EUHDNLQJ RQ D VKRUH SDUDOOHO EDU ZLWK D JDS 1R FRPSDULVRQV DUH DYDLODEOH IRU WKLV FDVH 7KH PRGHO ZDV UXQ IRU WKH FDVHV RI VKRUH SHUSHQGLFXODU DQG VKRUH SDUDOOHO EUHDNZDWHUV UHSUHVHQWHG LQ WKH PRGHO DV LQILQLWHVLPDOO\ WKLQ LPSHUPHDEOH EDUULHUV )RU WKH VKRUH SHUn SHQGLFXODU EDUULHU WKH UHVXOWV DUH FRPSDUHG ZLWK WKH H[SHULPHQWDO UHVXOWV REWDLQHG DW WKH :DWHUZD\V ([SHULPHQW 6WDWLRQ DV UHSRUWHG E\ +DOHV f )RU WKH VKRUH SDUDOOHO EDUULHU WKH UHVXOWV DUH FRPSDUHG ZLWK WKH QXPHULFDO UHVXOWV RI /LX DQG 0HL f ZKR GLG QRW LQFOXGH D ZDYH FXUUHQW IHHGEDFN PHFKDQLVP ,Q WKH ILQDO VHFWLRQ WKH PRGHO ZDV XVHG WR VLPXODWH WKH ODERUDWRU\ H[SHULPHQWV RI
PAGE 74
*RXUOD\ f LQ ZKLFK WKH VKRUHOLQH LQ WKH OHH RI D EUHDNZDWHU ZDV PDGH WR FXUYH DQG PHHW WKH EUHDNZDWHU DV RFFXUV ZLWK WRPEROR IRUPDWLRQ 7KH UHVXOWV RI WKH PRGHO ZHUH LQ FORVH DJUHHPHQW ZLWK *RXUOD\fV H[SHULPHQWDO YDOXHV :DYH 6HWXS /RQJXHW+LJJLQV DQG 6WHZDUW f JDYH D VLPSOH EXW HOHJDQW GHULYDWLRQ IRU WKH ZDYH LQGXFHG VHWXS RQ D SODQH EHDFK DVVXPLQJ OLQHDU VKDOORZ ZDWHU WKHRU\ )RU WKH VWHDG\ VWDWH FDVH RI QRUPDOO\ LQFLGHQW ZDYHV RQ D SODQH EHDFK ZLWKRXW DQ\ \ YDULDWLRQV DQG ZLWKRXW DQ\ FXUUHQWV WKH GHSWK LQWHJUDWHG HTXDWLRQV RI PRWLRQV UHGXFH WR JGL sDVa f G[ S' G[ ,Q VKDOORZ ZDWHU IRU QRUPDOO\ LQFLGHQW ZDYHV WKH UDGLDWLRQ VWUHVV WHUP IURP OLQHDU ZDYH WKHRU\ (T f EHFRPHV 6[[ (Q Of O( ASJ+ f $ VSLOOLQJ EUHDNHU DVVXPSWLRQ LV PDGH VR WKDW WKH ZDYH KHLJKW LI LV D FRQVWDQW UDWLR N WR WKH ZDWHU GHSWK K IM + .K IMf f 6XEVWLWXWLQJ HTXDWLRQV f DQG f LQWR (T f JLYHV UM B B G.K IMf B IGK4 GIM? G[ K IMf G[ r 9 G[ G[f 6ROYLQJ IRU JLYHV WKH UHVXOW GIM B _F GK G[ _F G[ WKDW WKH VORSH RI WKH ZDWHU VXUIDFH LQVLGH WKH EUHDNHU OLQH LV D FRQVWDQW WLPHV WKH VORSH RI WKH EHDFK 3UHYLRXV PRGHOV DUH DEOH WR SURGXFH WKH UHVXOWV RI WKH DQDO\WLFDO H[SUHVVLRQ EXW WR GDWH QRQH LQFOXGLQJ WKH SUHVHQW KDYH PDWFKHG WKH H[SHULPHQWDO UHVXOWV ,Q ILJXUH LW LV VHHQ WKDW WKH VORSH RI WKH PHDQ ZDWHU VXUIDFH SURGXFHG E\ WKH ZDYH VHWXS LV QHDUO\
PAGE 75
',67$1&( )520 67,//:$7(5 /,1( ,1 %($&+ [ &06 &06 ,6" )LJXUH &RPSDULVRQ RI WKH QXPHULFDO VROXWLRQ RI 9HPXODNRQGD HW DO f IRU VHWXS ZLWK H[SHULPHQWDO GDWD RI %RZHQ f 6RXUFH 9HPXODNRQGD HW DO f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f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f +RZHYHU WKLV FORVH
PAGE 76
DJUHHPHQW LV MXVW IRU WKH LQQHU UHJLRQ )RU WKH RXWHU UHJLRQ WKHUH LV QR YLDEOH VROXWLRQ RWKHU WKDQ WKDW WKH UDGLDWLRQ VWUHVV WHUP DQG WKH ZDWHU VXUIDFH HOHYDWLRQ UHPDLQ QHDUO\ FRQVWDQW )XUWKHU WKHUH LV QR DQDO\WLFDO VHSDUDWLRQ RI WKH RXWHU DQG LQQHU UHJLRQV RWKHU WKDQ WKH D SRVWHUL GHILQLWLRQ RI ZKHUH WKHUH LV D VKDUS FKDQJH LQ WKH VORSH RI WKH ZDWHU VXUIDFH HOHYDWLRQ 7KLV GHILQLWLRQ LV QRW XVHIXO LQ D PRGHO WKDW VHHNV WR GHWHUPLQH WKH ZDWHU VXUIDFH HOHYDWLRQV 6YHQGVHQfV ZRUN VKRZV SURPLVH DQG VKRXOG EH SXUVXHG LW LV QRW KRZHYHU DW WKLV SRLQW YLDEOH IRU LQFOXVLRQ LQ D QXPHULFDO FLUFXODWLRQ PRGHO )RU RQH LW QHHGV WR EH H[SDQGHG IRU WKH FDVH RI ZDYHV DW D DQJOH WKLV QHHG ZLOO EH GLVFXVVHG IXUWKHU LQ WKH QH[W VHFWLRQf $QRWKHU OLPLWDWLRQ LV WKDW 6YHQGVHQfV PRGHO LV IRU PRQRWRQLFDOO\ GHFUHDVLQJ GHSWKV ZKLFK LV D VHYHUH UHVWULFWLRQ IRU H[DPSOH H[FOXGLQJ EDUUHG SURILOHVf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f ZKLFK LV H[SODLQHG LQ VHFWLRQ 7KXV WKH VORSH RI WKH PHDQ ZDWHU VXUIDFH LV QRW D FRQVWDQW WLPHV WKH ERWWRP VORSH $ UHVXOW IURP WKH H[SHULPHQWV RI +DQVHQ DQG 6YHQGVHQ f VKRZ WKLV FOHDUO\ LQ ILJXUH 7KH ILJXUH VKRZV D VOLJKW FXUYDWXUH LQ WKH ZDYH KHLJKW DIWHU EUHDNLQJ
PAGE 77
PP )LJXUH &RPSDULVRQ RI QXPHULFDO UHVXOWV IURP WKH PRGHO ZLWK H[SHULPHQWDO UHVXOWV RI +DQVHQ DQG 6YHQGVHQ f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f ZHUH DEOH WR GXSOLFDWH WKH VKRDOLQJ UHFRUGHG LQ WKH H[SHULPHQWV RI +DQVHQ DQG 6YHQGVHQ f E\ QXPHULFDOO\ VROYLQJ FRXSOHG HTXDWLRQV IRU WKH PDQ\ KDUn PRQLF FRPSRQHQWV 7KHLU UHVXOWV VKRZHG WKDW LQ WKH VKRDOLQJ SURFHVV SULRU WR ZDYH EUHDNLQJ ZDYH HQHUJ\ ZDV WUDQVPLWWHG WR WKH KLJKHU IUHTXHQF\ FRPSRQHQWV 7KHLU ZRUN KRZHYHU LV QRW DGDSWDEOH IRU XVH LQ WKH FXUUHQW PRGHO VLQFH LW LV QRW NQRZQ KRZ WR GLVVLSDWH HQHUJ\
PAGE 78
)LJXUH 1RQGLPHQVLRQDO ORQJVKRUH YHORFLWLHV YUHVXV WKH QRQGLPHQVLRQDO GLVWDQFH RIIn VKRUH IURP WKH DQDO\WLFDO VROXWLRQ RI /RQJXHW+LJJLQV 5HSULQWHG IURP /RQJXHW+LJJLQV Ef RXUQDO RI *HRSK\VLFDO 5HVHDUFK YRO f SDJH FRS\ULJKW E\ WKH $PHULn FDQ *HRSK\VLFDO 8QLRQ IURP WKH YDULRXV IUHTXHQF\ FRPSRQHQWV RQFH EUHDNLQJ LV LQLWLDWHG /RQJVKRUH &XUUHQWV /RQJXHW+LJJLQV Ef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ÂIIILVfL f 2 & 7KH DEVFLVVD ; LV WKH QRQGLPHQVLRQDO GLVWDQFH ]Q ZKHUH LM LV WKH GLVWDQFH IURP WKH
PAGE 79
:DYH 6HW8S /LQH %UHDNHU /LQH )LJXUH &RPSDULVRQ RI H[SHULPHQWDO UHVXOWV RI 9LVVHU f DQG QXPHULFDO UHVXOW IRU ORQJVKRUH FXUUHQWV GXH WR D VHFRQG ZDYH RQ D VORSH RI IRU WKH GHHS ZDWHU FRQGLWLRQV + P DQG GHJUHHV VKRUH WR WKH EUHDNHU OLQH 3UHYLRXV PRGHOV KDYH UHSURGXFHG WKH UHVXOWV RI WKH /RQJXHW +LJJLQV VROXWLRQ E\ LQFRUn SRUDWLQJ DOO RI WKH VDPH DVVXPSWLRQV LH VSLOOLQJ EUHDNHUV DQG ERWWRP IULFWLRQf 6LQFH WKH SUHVHQW PRGHO GRHV QRW FRQWDLQ WKH VDPH DVVXPSWLRQV FRPSDULVRQ ZLOO QRW EH PDGH ZLWK WKH DQDO\WLFDO UHVXOWV RI /RQJXHW+LJJLQV ,QVWHDG FRPSDULVRQ LV PDGH ZLWK WKH H[SHULPHQWDO UHVXOWV RI 9LVVHU f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f DQG WKH PRGHO UHVXOWV IRU WKH VDPH WHVW FRQGLWLRQV 7KH PD[LPXP FXUUHQW LV QHDUO\ WKH VDPH LQ ERWK WKH PRGHO DQG WKH H[SHULPHQW EXW WKH PD[LPXP LV IXUWKHU RIIVKRUH LQ WKH PRGHO 7KLV FDQ EH DWWULEXWHG WR WKH LQLWLDWLRQ RI D ORQJVKRUH FXUUHQW LQ WKH PRGHO DW WKH EUHDNHU OLQH
PAGE 80
:DYHV 2Q D 5LS&KDQQHO %RWK %LUNHPHLHU DQG 'DOU\PSOH f DQG (EHUVROH DQG 'DOU\PSOH f H[DPLQHG WKH FDVH RI ZDYHV LQFLGHQW XSRQ D VKRUH ZLWK WKH SUHVHQFH RI D ULSFKDQQHO 7KH ERWWRP WRSRJUDSK\ ZDV JLYHQ DFFRUGLQJ WR D IRUPXOD ILUVW XVHG E\ 1RGD HW DO f K^[\f P[ MO $HVAf6 VLQ A\ ]WDQAf_ f ZKHUH LV WKH DQJOH WKH ULSFKDQQHO PDNHV ZLWK WKH [ D[LV ZKLFK LV LQ WKHLU FRRUGLQDWH V\VWHP GLUHFWHG RIIVKRUH IURP WKH VWLOO ZDWHU OLQH P LV WKH EHDFK VORSH $ LV DQ DPSOLWXGH RI WKH ERWWRP YDULDWLRQ DQG $ LV WKH OHQJWK RI WKH UHSHDWLQJ EHDFK 7KH RULJLQ RI WKH FRRUGLQDWHV LV DW WKH LQWHUVHFWLRQ RI WKH VWLOO ZDWHU OLQH DQG WKH WURXJK DV LV VKRZQ LQ ILJXUH ZKLFK JLYHV WKH GHSWK FRQWRXUV IRU WKH FDVH ZKHUH WKH VORSH LV DQG If LV HTXDO WR GHJUHHV 'XH WR WKH UHVWULFWLRQ RI WKH SUHVHQW PRGHO WKDW WKHUH EH QR \ JUDGLHQWV DW WKH ODWHUDO ERXQG DULHV WKH DERYH IRUPXODWLRQ LV PRGLILHG WR EH K[ f A^L $H:AVLQAILY]WDQ"f` _I \ [WDQef_ r >PL M\ fÂ§ [ WDQ ILf? [ 9 f n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
PAGE 81
)LJXUH 7RS 'HSWK FRQWRXUV LQ PHWHUV XVLQJ (T ZLWK WKH FRRUGLQDWHV DV VKRZQ %RWWRP 'HSWK FRQWRXUV DV XVHG LQ WKH SUHVHQW PRGHO *ULG VSDFLQJ LV PHWHUV
PAGE 82
)LJXUH 9HORFLW\ YHFWRUV IURP WKH UHVXOWV RI (EHUVROH DQG 'DOU\PSOH IRU D ZDYH DQJOH RI GHJUHHV 6FDOH IRU YHORFLW\ YHFWRUV LV LQFK PVHF 6RXUFH (EHUVROH DQG 'DOU\PSOH f
PAGE 83
)LJXUH 9HORFLW\ YHFWRUV REWDLQHG IURP WKH SUHVHQW PRGHO XVLQJ WKH VDPH LQSXW FRQGLn WLRQV DV IRU (EHUVROH DQG 'DOU\PSOH
PAGE 84
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
PAGE 85
)LJXUH 'HSWK FRQWRXUV LQ PHWHUV IRU WKH FDVH RI D EDU ZLWK D JDS )LJXUH 9HORFLW\ YHFWRUV IRU WKH FDVH RI ZDYHV RYHU D EDU ZLWK D JDS 'DVKHG OLQHV LQGLFDWH GHSWK FRQWRXUV *ULG VSDFLQJ HTXDO WR PHWHUV
PAGE 86
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f KDV UHPDUNHG WKDW KH KDV REVHUYHG WKH LQLWLDWLRQ RI D YHU\ VWURQJ FXUUHQW LQ WKH GLUHFWLRQ RI WKH ZDYHV SURGXFHG E\ ZDYHV EUHDNLQJ RQ D VDQG PRXQG LQ D VPDOO ZDYH EDVLQ 7KLV ZDV WHVWHG LQ WKH PRGHO XVLQJ QRUPDOO\ LQFLGHQW ZDYHV RQ D VORSH RI ZLWK D VKRDO GHVFULEHG E\ WKH IROORZLQJ HTXDWLRQ ZKLFK LV WKH VKRDO XVHG LQ WKH 'HOIW H[SHULPHQWV RI %HUNKRII HW DO f 0r}\f [ fÂ§P [f rff ,f f [f A ^rf n rff Jf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
PAGE 87
L L L L L L L L L L L L L WLW ULn L L L L L L L L L L L L L L L L L L L L f fÂ§ "If r 77 fÂ§ fÂ§ fÂ§ fÂ§ fÂ§ 77 5 5 fÂ§ fÂ§ 77 fÂ§ fÂ§ fÂ§ 5 5 fÂ§ fÂ§ / fÂ§ fÂ§ fÂ§ fÂ§ aO L L L L L L L L L L L L L O L Oa )LJXUH 'HSWK FRQWRXUV IRU WKH FDVH RI QRUPDOO\ LQFLGHQW ZDYHV XSRQ D VKRDO *ULG VSDFLQJ LV PHWHUV &RQWRXU LQWHUYDO LV PHWHUV
PAGE 88
)LJXUH 9HORFLW\ YHFWRUV IRU WKH FDVH RI ZDYHV EUHDNLQJ XSRQ D VKRDO
PAGE 89
)LJXUH 0D[LPXP YHORFLW\ DFURVV WKH WRS RI D VKRDO YHUVXV LQFLGHQW ZDYH KHLJKW IRU ZDYHV RYHU D VKRDO RQ D SODQH EHDFK ,QFLGHQW ZDYH KHLJKW OHVV WKDQ FHQWLPHWHUV DUH QRQEUHDNLQJ ZDYHV ,QFLGHQW ZDYH KHLJKWV JUHDWHU WKDQ FHQWLPHWHUV SURGXFHG ZDYHV EUHDNLQJ RYHU WKH VKRDO
PAGE 90
)LJXUH 3RVLWLRQ RI WKH JURLQ LQ UHODWLRQ WR WKH JULG URZV DQG FROXPQV WKH VKRUH ERXQGDU\ WR WKH ,EUN JULG URZ )LJXUH LOOXVWUDWHV WKH SRVLWLRQ RI WKH EUHDNZDWHU 0DWKHPDWLFDOO\ WKLV WKLQ LPSHUPHDEOH EDUULHU EHKDYHV DV DQ LQWHUQDO VROLG ERXQGDU\ WKDW SUHYHQWV DQ\ DFWLYLW\ RQ RQH VLGH RI WKH EDUULHU IURP EHLQJ WUDQVPLWWHG WR WKH RWKHU VLGH 7KLV UHTXLUHV PRGLILFDWLRQV WR ERWK ZDYH DQG FXUUHQW PRGHOV ,Q WKH ZDYH PRGHO DQ LQWHUQDO ERXQGDU\ FRQGLWLRQ LV LPSRVHG VR WKDW DW WKH EUHDNZDn WHU IDFH ZDYH HQHUJ\ LV WRWDOO\ UHIOHFWHG 7KLV LV DFFRPSOLVKHG E\ LPSRVLQJ WKH UHIOHFWLRQ FRQGLWLRQ Ir G\ f RQ ERWK VLGHV RI WKH EUHDNZDWHU IURP JULG URZ ,EUN WR WKH VKRUHOLQH RU $>,EUNf $>,EUN?f IRU DOO ,EUN f :KHQ WKH ZDYH HTXDWLRQ LV H[SUHVVHG LQ ILQLWH GLIIHUHQFH IRU EUN EUNLf ZKLFK LV RQ WKH RWKHU VLGH RI WKH EDUULHU LV HOLPLQDWHG E\ YLUWXH (T /LNHZLVH ZKHQ WKH ZDYH HTXDWLRQ LV DSSOLHG WR WKH RWKHU VLGH RI WKH EDUULHU $>, EUNf LV DOVR HOLPLQDWHG )XUWKHU
PAGE 91
UHIOHFWLYH FRQGLWLRQV DUH LPSRVHG DW WKH EDUULHU E\ DVVXPLQJ WKDW WKH GHULYDWLYHV RI DQ\ SURSHUW\ SHUSHQGLFXODU WR WKH EUHDNZDWHU KDV D ]HUR YDOXH DW WKH EUHDNZDWHU ,Q WKH FLUFXODWLRQ PRGHO WKH LQWHUQHG ERXQGDU\ FRQGLWLRQ RI LPSHUPHDELOLW\ DW WKH EDUn ULHU LV LPSRVHG E\ VHWWLQJ WKH YHORFLW\ FRPSRQHQW QRUPDO WR WKH EDUULHU HTXDO WR ]HUR 9,EUNLf IRU DOO ,EUN f 7R LQVXUH WKDW FXUUHQWLQGXFHG PRPHQWXP LV QRW GLIIXVHG DFURVV WKH EDUULHU VXEURXWLQHV ZHUH LQWURGXFHG LQ WKH FRPSXWDWLRQ IRU WKH [ VZHHS IRU WKH WZR JULG FROXPQV DGMDFHQW WR WKH EDUULHU 7ZR VDPSOH FDVHV ZHUH UXQ IRU FRPSDULVRQ SXUSRVHV 7KH ILUVW FDVH LV D EUHDNZDWHU H[WHQGLQJ PHWHUV IURP WKH VKRUHOLQH RQ D EHDFK ZLWK D FRPSRVLWH VORSH RI RQ IRU WKH ILUVW RQH KXQGUHG PHWHUV DQG D VORSH RI RQ EH\RQG WKDW SRLQW 7KLV LV D JHRPHWU\ VHOHFWHG E\ /LX DQG 0HL f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r\f f ZKHUH WKH WHUP FRQWDLQV WKH IRUFLQJ GXH WR ZDWHU OHYHO JUDGLHQWV IULFWLRQ DQG ZDYH IRUFLQJ 7KLV HTXDWLRQ UHSUHVHQWV D ERXQGDU\ YDOXH SUREOHP ZLWK WZR XQNQRZQV LS DQG IM 8SRQ VSHFLI\LQJ ERXQGDU\ YDOXHV DQ LQLWLDO JXHVV IRU WKH ZDWHU VXUIDFH IM LV PDGH DQG WKHQ (T LV VROYHG IRU LS :LWK WKLV QHZ VROXWLRQ IRU LS WKH YDOXH RI UM LV XSGDWHG E\ VROYLQJ VWHDG\
PAGE 92
VWDWH HTXDWLRQV RI PRWLRQ 7KLV SURFHVV LV UHSHDWHG XQWLO FRQYHUJHQFH )LJXUHV DQG VKRZ WKH UHVXOWV IRU /LX DQG 0HL DQG WKH SUHVHQW VWXG\ UHVSHFn WLYHO\ )LJXUH VKRZV WKH FRQWRXUV RI WKH VWUHDP IXQFWLRQ ZKHUHDV ILJXUH VKRZV IORZ OLQHV ZKLFK JLYH WKH GLUHFWLRQ RI WKH IORZ EXW QRW WKH LQWHQVLW\ RI WKH IORZ 7KH SUHVHQW PRGHO VKRZV D VHSDUDWLRQ HGG\ DW WKH WLS RI WKH EUHDNZDWHU ZKLFK LV DEVHQW LQ /LXf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f ZKLFK IRUFHV WKH WDQJHQWLDO YHORFLW\ WR EH ]HUR DW WKH RIIVKRUH ERXQGDU\ $Q DWWHPSW WR FKDQJH WKH RIIVKRUH FRQGLWLRQ WR D ]HUR QRUPDO IORZ FRQGLWLRQ GLG QRW DFFXUDWHO\ FRQVHUYH ZDWHU ZLWKLQ WKH FRPSXWDWLRQDO GRPDLQ HYHQ WKRXJK FRQYHUJHQFH ZDV DFKLHYHG )LJXUH VKRZV WKH IORZ OLQHV UHVXOWLQJ IURP XVLQJ QR RQVKRUH IORZ DW WKH RIIVKRUH ERXQGDU\ 7KH VHFRQG FDVH LV WR VLPXODWH WKH H[SHULPHQWV FRQGXFWHG DW WKH :DWHUZD\V ([SHULPHQW 6WDWLRQ DQG UHSRUWHG E\ +DOHV f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
PAGE 93
)LJXUH 5HVXOWV IRU WKH QXPHULFDO PRGHO RI /LX DQG 0HL f 'HHSZDWHU ZDYH KHLJKW HTXDO WR PHWHU GHHSZDWHU ZDYH DQJOH HTXDO WR GHJUHHV 6RXUFH /LX DQG 0HL f
PAGE 94
)LJXUH 5HVXOWV IURP WKH SUHVHQW QXPHULFDO PRGHO XVLQJ WKH VDPH FRQGLWLRQV DV ZHUH XVHG E\ /LX DQG 0HL 2IIVKRUH ERXQGDU\ FRQGLWLRQ RI FRQVWDQW ZDWHU VXUIDFH OHYHO ZKLFK SUHFOXGHV WDQJHQWLDO IORZ DW WKH RIIVKRUH ERXQGDU\
PAGE 95
)LJXUH 5HVXOWV IURP WKH SUHVHQW QXPHULFDO PRGHO XVLQJ WKH VDPH FRQGLWLRQV DV ZHUH XVHG E\ /LX DQG 0HL 2IIVKRUH ERXQGDU\ FRQGLWLRQ RI QR RQRIIVKRUH IORZ *ULG VSDFLQJ HTXDO WR PHWHUV
PAGE 96
)LJXUH ([SHULPHQWHG VHWXS IRU WKH WHVWV FRQGXFWHG E\ +DOHV f 6RXUFH +DOHV f
PAGE 97
Z L FH / / / 6n $ D $ $ $ L O D $ $ $ $ r $ $ )76(& $ 59(53*( 0,''(37+ 9(/2&,7,(6 &20%,1(' 5()5$&7,21 41' ',))5$&7,21 $1*/( 2) ,1&,'(1&( '(*5((" :$9( 3(5,2' 6(&21'6 :$9( +(,*+7 1($5 *(1(5$725 )()7 )LJXUH ([SHULPHQWDO PHDVXUHPHQWV RI WKH PLGGHSWK DYHUDJHG YHORFLWLHV REWDLQ E\ +DOHV 0HDVXUHPHQWV DW RQH IRRW LQWHUYDOV 6RXUFH +DOHV f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f QRU SHUSHQGLFXODU WR WKH VKRUHOLQH ERXQGDU\ ZRXOG UHTXLUH PDSSLQJ WHFKQLTXHV WKDW DUH QRW \HW LQFOXGHG LQ WKH FDSDELOLW\ RI WKH PRGHO )LJXUH VKRZV D FRPSDULVRQ RI WKH H[SHULPHQWHG DQG QXPHULFDO UHVXOWV REWDLQHG IRU WKH PDJQLWXGH RI WKH ULS FXUUHQW DGMDFHQW WR WKH GRZQ ZDYH VLGH RI WKH MHWW\ IRU WKH FDVH RI D VHFRQG ZDYH ZLWK DQ DQJOH RI LQFLGHQFH RI GHJUHHV DQG D RQH IRRW GHSWK ZDYH KHLJKW
PAGE 98
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f HTXDO WR ]HUR IRU WKH HQWLUH OHQJWK RI WKH EUHDNZDWHU ,Q DGGLWLRQ WKH FLUFXODWLRQ PRGHO LV PRGLILHG VR WKDW PRPHQWXP GLIIXVLRQ DQG GHULYDWLYHV GR QRW WDNH SODFH
PAGE 99
)LJXUH 9HORFLW\ YHFWRUV RI WKH FXUUHQWV REWDLQHG QXPHULFDOO\ IRU WKH VDPH FRQGLWLRQV DV IRU WKH +DOHV H[SHULPHQW *ULG VSDFLQJ HTXDO WR RQH IRRW
PAGE 100
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
PAGE 101
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
PAGE 102
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f UHSRUWHG UHVXOWV RI ODERUDWRU\ H[SHULPHQWV LQ ZKLFK FXUUHQWV ZHUH JHQn HUDWHG E\ WKH GLIIUDFWHG ZDYHV LQ WKH OHH RI D EUHDNZDWHU +LV H[SHULPHQW ZDV GHVLJQHG WR GHPRQVWUDWH WKDW ORQJVKRUH YDULDWLRQ RI WKH ZDYH KHLJKW FRXOG JHQHUDWH ORQJVKRUH FXUUHQWV 7KH H[SHULPHQWDO ZDYH EDVLQ ZDV WKXV FRQVWUXFWHG VR WKDW WKH VKRUHOLQH ZDV HYHU\ZKHUH SDUn DOOHO ZLWK WKH GLIIUDFWHG ZDYH FUHVWV )LJXUH VKRZV WKH ODERUDWRU\ VHW XS IRU *RXUOD\fV H[SHULPHQWV $ RQ FRQFUHWH EHDFK LV SDUDOOHO WR WKH LQFRPLQJ ZDYH FUHVWV LQ WKH H[n SRVHG ]RQH DQG LQ WKH VKDGRZ ]RQH WKH VORSH LV DOVR WR EXW WKH EHDFK LV FXUYHG ZLWK D FRQVWDQW UDGLXV FHQWHUHG RQ WKH EUHDNZDWHU WLS *RXUOD\fV UHDVRQLQJ LQ FKRRVLQJ WKLV DUUDQJHPHQW ZDV WKDW fVLQFH WKH GLIIUDFWHG ZDYH FUHVWV KDYH WKHRUHWLFDO FXUYHV ZKLFK FDQ EH DSSUR[LPDWHG E\ D FLUFOH WKH LQIOXHQFH RI ZDYHV EUHDNLQJ DW DQ DQJOH WR WKH EHDFK ZDV DOPRVW FRPSOHWHO\ HOLPLQDWHGf SDJH f
PAGE 103
R )LJXUH 9HORFLW\ YHFWRUV REWDLQHG IURP WKH PRGHO IRU D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH IRU D PHWHU VHFRQG ZDYH IRU LQFUHDVLQJ DQJOHV RI LQFLGHQFH
PAGE 104
)LJXUH 6WUHDP IXQFWLRQ FRQWRXU OLQHV REWDLQHG IURP WKH /LX PRGHO IRU D VHFRQG PHWHU ZDYH QRUPDOO\ DSSURDFKLQJ D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH RQ D RQ VORSH 6RXUFH /LX DQG 0HL f
PAGE 105
)LJXUH )ORZ OLQHV REWDLQHG IURP WKH SUHVHQW PRGHO IRU D VHFRQG PHWHU ZDYH QRUPDOO\ DSSURDFKLQJ D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH RQ D RQ VORSH )LJXUH 9HORFLW\ YHFWRUV REWDLQHG IURP WKH SUHVHQW PRGHO IRU D VHFRQG PHWHU ZDYH QRUPDOO\ DSSURDFKLQJ D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH RQ D RQ VORSH
PAGE 106
)LJXUH 6HWXS FRQWRXU OLQHV REWDLQHG IURP WKH SUHVHQW PRGHO IRU D VHFRQG PHWHU ZDYH QRUPDOO\ DSSURDFKLQJ D PHWHU ORQJ EUHDNZDWHU PHWHUV RIIVKRUH RQ D RQ VORSH $Q H[SHULPHQW RI *RXUOD\ ZDV QXPHULFDOO\ GXSOLFDWHG E\ PDNLQJ VHYHUDO DGDSWDWLRQV WR WKH RYHUDOO VFKHPH )LJXUH VKRZV WKH QXPHULFDO FRPSXWDWLRQDO JULGV XVHG LQ WKH SURn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
PAGE 107
)LJXUH 3K\VLFDO OD\RXW IRU WKH H[SHULPHQW RI *RXUOD\ 6RXUFH *RXUOD\ f 3URn FHHGLQJV A &RDVWDO (QJLQHHULQJ &RQIHUHQFH FRS\ULJKW E\ WKH $PHULFDQ 6RFLHW\ RI &LYLO (QJLQHHUV UHSULQWHG ZLWK SHUPLVVLRQ
PAGE 108
)LJXUH &RPSXWDWLRQ VLO GRPDLQ IRU WKH H[SHULPHQW RI *RXUOD\ *ULG VSDFLQJ HTXDO WR PHWHU
PAGE 109
,Q WKH ZDYH PRGHO SDUW RI WKH SURJUDP WKH SUREOHP RI ZHW DQG GU\ JULGV LV KDQGOHG XVLQJ D WHFKQLTXH SURSRVHG E\ .LUE\ DQG 'DOU\PSOH f E\ WUHDWLQJ WKH GU\ JULGV DV WKRXJK WKH\ KDYH D YHU\ VPDOO VD\ RQH PLOOLPHWHUf GHSWK RI ZDWHU 7KLV WHFKQLTXH DOORZV IRU NHHSLQJ D XQLIRUP FRPSXWDWLRQDO GRPDLQ IRU WKH ZDYH PRGHO ZKLOH DYRLGLQJ DQ\ QXPHULFDO SUREOHPV VXFK DV GLYLGLQJ E\ ]HUR GHSWKV *RXUOD\f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n WLFXODU LV YHU\ ZHOO VKRZQ LQ WKH QXPHULFDO UHVXOWV 7KH GLIIHUHQFHV DUH WKDW DV PHQWLRQHG HDUOLHU WKHUH LV QR VWDJQDWLRQ HGG\ VKRZQ LQ ILJXUH DQG WKH FHQWHU RI WKH SULPDU\ HGG\ LV VOLJKWO\ IDUWKHU EHKLQG WKH EUHDNZDWHU LQ WKH SK\VLFDO H[SHULPHQW WKDQ LQ WKH PRGHO UHn VXOWV 6LPLODUO\ DV VHHQ LQ ILJXUH WKH PD[LPXP FXUUHQWV REWDLQHG LQ WKH PRGHO DUH QRW
PAGE 110
L !" )LJXUH &RPSDULVRQ RI H[SHULPHQWDO DQG QXPHULFDO UHVXOWV IRU WKH H[SHULPHQW RI *RXUOD\ :DYH 6HWXS &RQWRXUV 7RS ILJXUH LV IURP *RXUOD\ f 3URFHHGLQJV A &RDVWDO (QJLQHHULQJ &RQIHUHQFH FRS\ULJKW E\ WKH $PHULFDQ 6RFLHW\ RI &LYLO (QJLQHHUV UHSULQWHG ZLWK SHUPLVVLRQ 7KH ERWWRP ILJXUH LV WKH FRPSXWHU UHVXOWV
PAGE 111
)LJXUH 5HVXOWV IURP WKH H[SHULPHQW RI *RXUOD\ &RQWRXUV RI FXUUHQW YHORFLWLHV DQG VWUHDPOLQHV 6RXUFH *RXUOD\ f 3URFHHGLQJV A &RDVWDO (QJLQHHULQJ &RQIHUHQFH FRS\ULJKW E\ WKH $PHULFDQ 6RFLHW\ RI &LYLO (QJLQHHUV UHSULQWHG ZLWK SHUPLVVLRQ )LJXUH 5HVXOWV IURP WKH QXPHULFDO PRGHO 6WUHDPOLQHV RI WKH IORZ
PAGE 112
)LJXUH 5HVXOWV IURP WKH QXPHULFDO PRGHO &RQWRXU OLQHV RI WKH FXUUHQW YHORFLWLHV DV IDU EHKLQG WKH EUHDNZDWHU DV WKH\ ZHUH LQ WKH H[SHULPHQW %RWK RI WKHVH VSDFLDO GLVSDUn LWLHV EHWZHHQ QXPHULFDO DQG SK\VLFDO UHVXOWV PD\ UHVXOW IURP WKH IDLOXUH RI WKH SDUDEROLF DSSUR[LPDWLRQ WR DGHTXDWHO\ SURSDJDWH ZDYH HQHUJ\ IXOO\ LQWR WKH OHH RI WKH EUHDNZDWHU ,Q FRPSDULQJ WKH YHORFLWLHV REWDLQHG LQ WKH PRGHO WR WKH H[SHULPHQWDOO\ PHDVXUHG YDOXHV LW VKRXOG EH QRWHG WKDW *RXUOD\fV UHVXOWV DUH IRU VXUIDFH YHORFLWLHV ZKHUHDV WKH PRGHO UHVXOWV DUH IRU GHSWKDYHUDJHG FXUUHQWV 7KLV PD\ EH D IDFWRU LQ H[SODLQLQJ WKDW WKH PD[LPXP LQ ILJXUH IRU WKH H[SHULPHQWDO UHVXOWV LV ODUJHU WKDQ WKH PD[LPXP LQ ILJXUH IRU WKH QXPHULFDO UHVXOWV 7KH GLIIHUHQFHV LQ WKH H[SHULPHQWDO DQG QXPHULFDO UHVXOWV ZKLFK ZHUH PHQWLRQHG DERYH VKRXOG EH NHSW LQ SHUVSHFWLYH 2YHUDOO WKH DJUHHPHQW LV JRRG ERWK TXDOLWDWLYHO\ DQG TXDQn WLWDWLYHO\
PAGE 113
&+$37(5 &21&/86,216 $1' 5(&200(1'$7,216 )25 )857+(5 678'< $ QXPHULFDO VROXWLRQ IRU WKH FLUFXODWLRQ DQG ZDYHV LQ FRDVWDO UHJLRQV KDV EHHQ SUHVHQWHG 7KH VROXWLRQ PHWKRG XVHV DQ DOWHUQDWLQJ GLUHFWLRQ LPSOLFLW PHWKRG RI VROYLQJ WKH GHSWK LQWHJUDWHG HTXDWLRQV RI PRPHQWXP DQG FRQWLQXLW\ :DYH IRUFLQJ WHUPV DUH REWDLQHG WKURXJK D SDUDEROLF DSSUR[LPDWLRQ WR WKH OLQHDU PLOG VORSH HTXDWLRQ IRU ZDYH SURSDJDWLRQ RYHU YDU\LQJ GHSWKV DQG FXUUHQWV 7KH QXPHULFDO VROXWLRQ LQYROYHV DQ LWHUDWLRQ EHWZHHQ WKH SDUDEROLF ZDYH VROXWLRQ DQG WKH $', FLUFXODWLRQ VROXWLRQ 7KH PRGHO ZDV UXQ IRU VHYHUDO WHVW FDVHV 7KHVH LQFOXGH FDVHV ZLWK YDU\LQJ WRSRJUDn SK\ VXFK DV ULS FKDQQHOV DQG EDU JDSV 7KH PRGHO ZDV DOVR WHVWHG ZLWK VKRUH SDUDOOHO DQG VKRUH SHUSHQGLFXODU EUHDNZDWHUV UHSUHVHQWHG DV LQILQLWHVLPDOO\ WKLQ LPSHUPHDEOH EDUULHUV &RPSDULVRQV ZHUH PDGH ZLWK DQDO\WLFDO DQG H[SHULPHQWDO UHVXOWV 7KHVH FRPSDULVRQV ZHUH IDYRUDEOH ,Q SDUWLFXODU FRPSDULVRQ RI WKH UHVXOWV IURP WKH PRGHO LQ VLPXODWLQJ WKH H[SHULn PHQWV RI *RXUOD\ f VKRZ WKH SUHGLFWLYH FDSDELOLW\ RI WKH PRGHO 7KH PRGHO ZDV DOVR UXQ IRU DQ LGHDOL]HG VKRUH SDUDOOHO EUHDNZDWHU YDU\LQJ WKH DVSHFW UDWLR RI WKH OHQJWK RI WKH EUHDNZDWHU WR WKH GLVWDQFH RIIVKRUH RI WKH EUHDNZDWHU WR GHWHUPLQH WKH UHODWLYH LPSRUWDQFH RI WKLV SDUDPHWHU 7KH SULPDU\ FRQFOXVLRQ RI WKLV VWXG\ LV WKDW LW LV SRVVLEOH WR JHW JRRG SUHGLFWLYH UHn VXOWV XVLQJ D QXPHULFDO PRGHO WKDW LQFRUSRUDWHV VRPH VLPSOLILFDWLRQV RI WKH SK\VLFDO UHDOLW\ ,Q WKH SUHVHQW VWXG\ WKH WKUHHGLPHQVLRQDO IORZV WKDW DUH HQFRXQWHUHG LQ QDWXUH DUH UHSn UHVHQWHG E\ WZRGLPHQVLRQDO GHSWK DYHUDJHG YHORFLWLHV 6LPSOLI\LQJ DVVXPSWLRQV DUH DOVR PDGH FRQFHUQLQJ ERWWRP IULFWLRQ PRPHQWXP WUDQVIHU RU PL[LQJ ZDYH EUHDNLQJ DQG ZDYH HQHUJ\ GLVVLSDWLRQ )RU HDFK RI WKHVH WRSLFV LW LV SRVVLEOH WR DUJXH WKDW QDWXUH LV JUHDWO\ GLIn IHUHQW DQG PRUH FRPSOH[ WKDQ WKH DVVXPSWLRQV HPERGLHG LQ WKLV UHSRUW
PAGE 114
SDUWLFXODU WKH FRPSDULVRQ WR WKH ODERUDWRU\ H[SHULPHQWV RI *RXUOD\f FRQILUP WKH SUHGLFWLYH FDSDELOLW\ RI WKH QXPHULFDO PRGHO 7KH VLPSOLILFDWLRQV PHQWLRQHG DERYH PDNH WKLV SRVVLn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f KDYH IRXQG D PHWKRG WR GHWHUPLQH D QRQ]HUR ZDYH KHLJKW LQ WKH OHH RI D EUHDNZDWHU +RZHYHU WKHLU VROXWLRQ LV LQ WKH IRUP RI D )RXULHU LQWHJUDO ZKLFK LV QRW VXLWHG WR WKH FRPSOH[ DPSOLWXGH IRUP UHTXLUHG LQ WKH SUHVHQW PRGHO
PAGE 115
,QFOXGLQJ :DYH 5HIOHFWLRQ 7KH DXWKRU KDV PDGH XQVXFFHVVIXO DWWHPSWV WR LQFOXGH ZDYH UHIOHFWLRQ LQ WKH PRGHO .LUE\ f XVHG FRXSOHG HTXDWLRQV JRYHUQLQJ WKH LQWHUDFWLRQ RI D IRUZDUG DQG D EDFNZDUG UHIOHFWHGf SURSDJDWLQJ ZDYH WKDW DUH FRXSOHG WKURXJK D VORSH LQWHUDFWLRQ +H ZDV DEOH WR REWDLQ WKH UHIOHFWLRQ DQG WUDQVPLVVLRQ FRHIILFLHQWV RYHU VXEPHUJHG EUHDNZDWHUV WKDW DJUHH ZHOO ZLWK WKH H[SHULPHQWV RI 6HHOLJ f 2QFH WKH DUUD\V RI WKH FRPSOH[ DPSOLWXGHV RI WKH WZR ZDYHV LV GHWHUPLQHG WKH UDGLDWLRQ VWUHVV WHUPV FDQ EH GHWHUPLQHG IURP HTXDWLRQV WKDW KDYH EHHQ GHULYHG IROORZLQJ WKH VDPH SURFHGXUH DV XVHG E\ 0HL f EXW GHILQLQJ I! DV EHLQJ FRPSULVHG RI WZR FRPSRQHQWV ,Q IDFW WKLV SURFHGXUH FDQ DOVR EH XVHG WR REWDLQ WKH UDGLDWLRQ VWUHVV WHUPV IRU DQ\ QXPEHU RI FRPSRQHQWV 7KH RQO\ OLPLW EHLQJ WKH DPRXQW RI FRPSXWDWLRQ UHTXLUHG DV WKH QXPEHU RI FRPSRQHQWV LQFUHDVHV 6PLWK f IROORZLQJ WKH ZRUN RI 0LOHV f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
PAGE 116
DV D IXQFWLRQ RI NQRZQ YDULDEOHV DQG WKHQ VROYH D VHSDUDWH PRGHO IRU VHGLPHQW EDODQFH DQG WKHQ XSGDWH WKH GHSWKV DQG SURFHHG LWHUDWLQJ EHWZHHQ WKH VHGLPHQW EDODQFH PRGHO DQG WKH SUHVHQW ZDYHFXUUHQW PRGHO PDUFKLQJ LQ WLPH 6LQFH LW KDV DOUHDG\ EHHQ GHPRQVWUDWHG WKDW WKH ZDYH FXUUHQW PRGHO JLYHV DGHTXDWH UHVXOWV LQ WKH OHH RI D VKRUH SDUDOOHO EUHDNZDWHU D JRRG WHVW IRU VXFK D VHGLPHQW WUDQVSRUW PRGHO ZROG EH WR VHH LI LW FRXOG \LHOG D WRPEROR IRUPDWLRQ LQ WKH OHH RI D VKRUH SDUDOOHO EUHDNZDWHU
PAGE 117
$33(1',; $ '(5,9$7,21 2) 7+( '(37+ $9(5$*(' (48$7,216 ,Q WKLV DSSHQGL[ D GHULYDWLRQ RI WKH JRYHUQLQJ HTXDWLRQV IRU WKH FLUFXODWLRQ PRGHO LV SUHVHQWHG 7KLV DSSHQGL[ FORVHO\ IROORZV WKH ZRUN RI (EHUVROH DQG 'DOU\PSOH f %HIRUH SURFHHGLQJ ZLWK WKH GHULYDWLRQV RI WKH GHSWK LQWHJUDWHG WLPH DYHUDJHG HTXDWLRQV RI PRWLRQ DQG WKH FRQWLQXLW\ HTXDWLRQ WKH /HLEQLW] 5XOH IRU WDNLQJ WKH GHULYDWLYH RI DQ LQWHJUDO LV VWDWHG VLQFH LW LV LQYRNHG WKURXJKRXW WKH GHULYDWLRQ 7\ &,n9f & Gb7G[ P\fn\fGe $! 7KH /HLEQLW] 5XOH FDQ DOVR EH XVHG LQ UHYHUVH WR WDNH GHULYDWLYHV RXWVLGH RI DQ LQWHJUDO 7KH NLQHPDWLF ERXQGDU\ FRQGLWLRQV DUH DOVR UHSURGXFHG VLQFH WKH XVH RI WKH /HLEQLW] 5XOH ZLOO \LHOG ERWWRP DQG VXUIDFH WHUPV WKDW ZLOO EH HOLPLQDWHG WKURXJK XVH RI WKHVH ERXQGDU\ FRQGLWLRQV 2Q WKH IUHH VXUIDFH WKH ERXQGDU\ FRQGLWLRQ LV GUf GUf GUf Zf D XfA YÂ£c $f DQG DW WKH ERWWRP WKH ERXQGDU\ FRQGLWLRQ $f ,Q WKH FLUFXODWLRQ PRGHO LW LV DVVXPHG WKDW WKH VWLOO ZDWHU GHSWK GRHV QRW YDU\ ZLWK WLPH VR WKH ILUVW WHUP RI (T $f LV GURSSHG DFWXDOO\ IRU WLPH YDU\LQJ GHSWK HVVHQWLDOO\ WKH VDPH HTXDWLRQV DUH REWDLQHGf 7KH FRQWLQXLW\ HTXDWLRQ RU WKH FRQVHUYDWLRQ RI PDVV HTXDWLRQ DV LW LV DER FDOOHGf LQ WKUHH GLPHQVLRQV LV G!Xf GSYf GSZf GW G[ G\ G] $f
PAGE 118
,QWHJUDWLQJ RYHU WKH GHSWK IURP ] fÂ§ fÂ§K WR ] Uc DQG XVLQJ WKH /HLEQLW] 5XOH WR WDNH GHULYDWLYHV RXWVLGH WKH LQWHJUDG WKH FRQWLQXLW\ HTXDWLRQ EHFRPHV G Ir GU@ G Q DW/SLfSnLUW7]/nrLO KR I ? GUf U ? GK G Q :V:NMDL/nf V "rfmIDfrr $f 7KH WHUPV ZLWK DQ RYHUEUDFH DUH S WLPHV WKH IUHH VXUIDFH NLQHPDWLF ERXQGDU\ FRQGLWLRQ DQG WKH WHUPV ZLWK DQ XQGHUEUDFH DUH S WLPHV WKH ERWWRP ERXQGDU\ FRQGLWLRQ (OLPLQDWLQJ ERWK RI WKHVH IURP (T $f JLYHV G >L GIL GUr 3G]]a SXG] fÂ§ SYG] GW K G[MN G\ Kf $f (TXDWLRQ $f LV QRZ WLPH DYHUDJHG 7LPH DYHUDJLQJ PHDQV DYHUDJLQJ RYHU D ZDYH SHULRG 7KH WLPH DYHUDJH RI D WHUP LV GHQRWHG E\ DQ RYHUEDU 7KXV A L U7 8 )GW $f ZKHUH ) UHSUHVHQWV DQ\ WHUP DQG 7 LV WKH ZDYH SHULRG /HWWLQJ X DQG Y EH FRPSULVHG RI D PHDQ IORZ D ZDYH LQGXFHG IORZ DQG DQ DUELWUDU\ RU UDQGRP WXUEXOHQFH FRPSRQHQW X Xn 9 9 9 9 $f $f 7KH WXUEXOHQW IOXFWXDWLRQV DUH DVVXPHG WR EH RI D KLJK IUHTXHQF\ UHODWLYH WR WKH ZDYH IUHTXHQF\ VR WKDW WKHLU WLPH DYHUDJHV RYHU WKH ZDYH SHULRG DUH ]HUR %\ VXEVWLWXWLQJ WKH DERYH H[SUHVVLRQV IRU X DQG Y LQWR (T $f DQG WDNLQJ WLPH DYHUDJHV WKH FRQWLQXLW\ HTXDWLRQ LV ^S^. }"f` A ^3LKR }"f` 3rG] $f $OOf
PAGE 119
7KH WLPH DYHUDJHV RI WKH YHUWLFDOO\ LQWHJUDWHG ZDYH LQGXFHG YHORFLWLHV DQG Y DUH QRW ]HUR 7KHVH TXDQWLWLHV DUH NQRZQ DV WKH ZDYH LQGXFHG PDVV IOX[ 'HILQLQJ 8 X X DQG 9 Y $f ZKHUH X IW U SG] DQG R IMfN S^K r"f WKH WLPH DYHUDJHG GHSWK LQWHJUDWHG FRQWLQXLW\ HTXDWLRQ LV bc iÂXKR WLf AYK Icf IW Y fÂ§ a7L UU SL! G] S^K ULfKR $f $f 7KH DVVXPSWLRQ WKDW WKH GHQVLW\ LV FRQVWDQW LQ ERWK WLPH DQG VSDFH KDV XVHG LQ REWDLQLQJ ERWK (T $f DQG (T $f 7KH DVVXPSWLRQ RI D FRQVWDQW ERWWRP KDV DOVR EHHQ XVHG LQ REWDLQLQJ $f ,I WKH ERWWRP LV QRW FRQVWDQW LQ WLPH D WHUP ZRXOG EH LQFOXGHG LQ (T $f 7KH [ DQG \ PRPHQWXP HTXDWLRQV DUH KDQGOHG LQ WKH VDPH PDQQHU DV WKH FRQWLQXLW\ HTXDWLRQ DERYH 7KH [ DQG \ GLUHFWLRQ PRPHQWXP HTXDWLRQV DUH GX GX GXY GXZ GS UGU[[ GUY[ GU[[ 7W DLU ,7 Zr D7` GY W GXY W GY W GYZ B GS GU[Y GUYY GU=@ GW G[ G\ G] ]LU $ ` $f $f S G\ S G[ G\ G] 7KH VDPH SURFHGXUH LV IROORZHG IRU ERWK WKH [ DQG \ HTXDWLRQV 2QO\ (T $f ZLOO EH GHDOW ZLWK KHUH ,QWHJUDWLQJ WKH OHIW KDQG VLGH RYHU WKH GHSWK DQG DSSO\LQJ WKH /HLEQLW] 5XOH WR WDNH WKH GHULYDWLYH RXWVLGH WKH LQWHJUDO JLYHV W 7LS? G IW GU@ GK /+6! aD7 LI G[MK QG[ r G[ G\ ; GUf GKR XYG] XWff fÂ§ XYfIFR fÂ§ K G\ G\ 8:fO a 8:fKR $f
PAGE 120
7KH WHUPV LQ (T $f WKDW DUH GHVLJQDWHG E\ DQ RYHUEUDFH DUH 8I WLPHV WKH NLQHPDWLF IUHH VXUIDFH FRQGLWLRQ DQG WKRVH GHVLJQDWHG E\ DQ XQGHUEUDFH DUH XBM WLPHV WKH ERWWRP ERXQGDU\ FRQGLWLRQ 7KHVH DUH ERWK GURSSHG OHDYLQJ f G U GU GU /+6f r X XG]Ga[ X XG]UGa\ X XY G] A 6XEVWLWXWLQJ WKH GHILQLWLRQV JLYHQ IRU X DQG Y LQ (TQ $f DQG $f DQG UHFRJQL]LQJ WKDW WKH WXUEXOHQW IOXFWXDWLQJ FRPSRQHQWV DUH RI D KLJK IUHTXHQF\ UHODWLYH WR WKH ZDYH IUHTXHQF\ VR WKDW WKH WLPH DYHUDJHV RI LQWHJUDOV LQYROYLQJ WKH SURGXFW RI D WXUEXOHQW FRPSRQHQW DQG D ZDYH LQGXFHG FRPSRQHQW DUH ]HUR JLYHV GaU GaU G aU GaU /+6 fÂ§ XG] fÂ§ G] fÂ§ X"G]fÂ§ WIG] $f GWMaK GW@. G[.f G[MK f G aUa7a= G aU G aU G aU e XnfG] fÂ§ XX G] fÂ§ XYG] fÂ§ XLLG] $f G[KR G[MK G\K G\KR G aU G aU G aU YG]fÂ§ XLfG] fÂ§ XnYnG] $f R\MKR R\K G\ K XVLQJ WKH GHILQLWLRQV RI (TV $f DQG $f WKLV EHFRPHV ZY^KLfUY/UG]\?ZAf/UGf/UL] $f ZKHUH WKH WZR WHUPV GHVLJQDWHG E\ XQGHUEUDFHV UHVXOW IURP EHLQJ DGGHG WR FRPSOHWH WKH 8 DQG 89 WHUPV 7KH ULJKW KDQG VLGH DIWHU LQWHJUDWLQJ RYHU WKH GHSWK DVVXPLQJ D FRQVWDQW GHQVLW\ DQG DVVXPLQJ DQ LQYLVFLG IOXLG VXFK WKDW QR KRUL]RQWDO YLVFRXV VWUHVVHV H[LVW LH UW[ DQG 7\[ DUH ]HUR LV 5+6 LU SLU Ds S KR G[ S fÂ§K G] $f 8VLQJ WKH /HLEQLW] 5XOH DQG DYHUDJLQJ RYHU D ZDYH SHULRG WKH ULJKW KDQG VLGH EHFRPHV G U GQ GKQ 5+6 fÂ§fÂ§ SG]SSKRfÂ§7[[ 7LLLU $f SG[MK S G[ S G[ S S Y n
PAGE 121
,OO 7=;T DQG U0;BNR DUH WKH WLPH DYHUDJHG VXUIDFH DQG ERWWRP VKHDU VWUHVVHV $VVXPLQJ WKDW WKH IUHH VXUIDFH SUHVVXUH SQ LV ]HUR WKH G\QDPLF IUHH VXUIDFH ERXQGDU\ FRQGLWLRQf WKH ULJKW KDQG VLGH LV BBB G GKa 5+6 fÂ§fÂ§ SG] SKfÂ§s 7Â7cL $f 3 G[MKF S G[ S S f 7KH PHDQ SUHVVXUH DW WKH ERWWRP FDQ EH H[SUHVVHG DV WKH VXP RI WKH ZDYH LQGXFHG RU G\QDPLF SUHVVXUH DQG WKH K\GURVWDWLF SUHVVXUH 3K 3G9QNR SJ>. W@f $f VR WKDW WKH SURGXFW 3KID FDQ EH PDQLSXODWHG WR UHDG GK GK B GIM m Â Â£V$n! "f mV $f 7HUPV IURP ERWK WKH ULJKW DQG OHIW KDQG VLGHV DUH FRPELQHG DV WKH UDGLDWLRQ VWUHVVHV 'HILQLQJ Vf rnf L] }r mf SLL]< aU! L 6Y6KAL] r G] @BK3LG] Vf B Âf c0. WI Af^O r A $f $f $ f DQG PDNLQJ WKH DVVXPSWLRQV WKDW WKH SURGXFW RI WKH G\QDPLF SUHVVXUH DW WKH ERWWRP DQG WKH ERWWRP VORSH LV QHJOLJLEOH WKDW WKH McI?>XnfG] WHUP FDQ EH LJQRUHG DQG WKDW WKH ODWHUDO IULFWLRQ WM FDXVHG E\ PRPHQWXP IOX[HV GXH WR WXUEXOHQW IOXFWXDWLRQV LV LQGHSHQGHQW RI GHSWK DQG WKDW UM LV GHILQHG E\ WM fÂ§SXnYn WKH [ GLUHFWLRQ HTXDWLRQ RI PRPHQWXP LV ZULWWHQ DV }UO A G UQU/ OG6[[ G6 ;9 A m DU8 nrf r} 7\89A [ Â‘ $f f_smr/,Uf LQ R 2; S G\ S S $f %\ VXEWUDFWLQJ WKH GHSWK LQWHJUDWHG FRQWLQXLW\ HTXDWLRQ DQG GLYLGLQJ E\ WKH WRWDO GHSWK WKH ILQDO IRUP RI WKH [ PRPHQWXP HTXDWLRQ LV 8 G8B G8B GW [ G\ S^K IMf G67 G6 [Y G[ G\ $f
PAGE 122
GIM GUL == = fÂ§U Q[ R $f G[ SG\ S^K S^K Ucf 7KH VDPH SURFHGXUH JLYHV DV WKH GHSWK LQWHJUDWHG WLPH DYHUDJHG \ PRPHQWXP HTXDWLRQ DY DY fDY L DV K 8 K 9 fÂ§ GW G[ G\ SK IMf G[ [\ DV YY G\ GUc GUL G\U SG[ SK Icf7nY S. Yf ?U}Y $f $f
PAGE 123
$33(1',; % '(5,9$7,21 2) 7+( 5$',$7,21 675(66 7(506 7KH EDVLV RI WKH GHULYDWLRQ LV WR XVH WKH GHILQLWLRQ RI WKH YHORFLW\ SRWHQWLDO WR VXEVWLWXWH IRU WKH ZDYH LQGXFHG YHORFLWLHV LQ WKH GHILQLWLRQ IRU WKH UDGLDWLRQ VWUHVVHV GI! B GI! 8ILaG= Z ]S G] ZKHUH I! LV H[SUHVVHG LQ WHUPV RI WKH JHQHUDO FRPSOH[ DPSOLWXGH IXQFWLRQ % DV W! B BLO fFRVK NK ]f BLXW 9 D % FRVK NK FFf ZKHUH FF UHSUHVHQWV WKH FRPSOH[ FRQMXJDWH WHUP $OVR Ucn LV H[SUHVVHG DV Un L%HLXW FFf ZKHUH WKH SULPH LV XVHG WR GHQRWH ZDYH LQGXFHG TXDQWLWLHV )RU FRQYHQLHQFH (T f RI 6HFWLRQ LV UHSURGXFHG EHORZ 6DS S/K XnD8nSG] D >BK "G] Kr Af \ IDnf $Q H[SUHVVLRQ IRU S LV REWDLQHG LQ 6HFWLRQ DQG LV UHSURGXFHG EHORZ 2O/: G] SZn\ 6XEVWLWXWLQJ (T %f LQWR (T %f JLYHV 6DS 3 XDXnSG] DS f ff O R %Of %f %f %f %f UL f Fr Ur GXnZn ? A SJUf ]fG] A ,S fÂ§/fÂ§G]f G] n n B r n r r n S>Zn< G] \2. \ IDnf m2 / B[ 3 %f
PAGE 124
7KH WHUPV RI (T %f DUH QRZ HYDOXDWHG WHUP E\ WHUP 7HUP B LJ G% FRVK N>K ]f BLXW LJ G%n FRVK NK ]f LXfW XnDXnS fÂ§ B 8f ? D G[D FRVK NK J FRVK NK [f D G[' FRVK NK %f % G%A G%AG%B B G%BG%BHLXLW B G%n G%r D FRVK NK G[D G[S G[D GLS G[D GLS G[D GLS fÂ§UfÂ§U SJ FRVK NK ]f S8r8H $VLQK NK SJ FRVK NK ]f IFVLQK NK G% G%n G%n G% G[D GLS G[D GLS Â‘ G%BG@U; GLS f %f %f Â SXnDXnSG] r fÂ‘ R SJ f G% N VLQK NK AG[4 G[S f SJ r G% LIFÂQK NK ?G[D G[S f? \G[D G[S N ? VL FRVK NK ]f G] f KII A VLQK NK IMf N>K IMff NK ? VLQK NK 7KH VHFRQG WHUP BK SJ^Y ]fG] SJ _nU?] KA a 33 7KH WKLUG WHUP I ,I rII G% ZRXLYALR aU rÂ‘ BOXf n\YsV 7 r\ ZW I f A [I f r r & f FRVK NK ]f BLXfW LJ G%n FRVK NK ]f FRVK NK D G[c FRVK NK LJN SVLQK NK ]f BOXOW A LIIr VLQK N>K ]f M FRVK NK J VLQK $K ]f f }}n / A VLQKIFK G[I n [I Z L B LJN VLQK N^K ]f BLXW LJN VLQK N>K ]f :$ 9 R FRVK NK D FRVK NK G% G%n f fÂ§% fÂ§% ,/n6: %,2f %OOf %f %f %f G[c B BGB G[ G] G% G[I G%n I %A%M B "B/ eeE fÂ§V IfÂ§ G[I [I G[I >fIF VLQK IFLf IMf VLQK NK UfB UI! VLQK NK ]f VLQK NK FRVK N>K IMf FRVKIFK ]f %f
PAGE 125
BSJ G ?G% % G%rAM Ur FRVK N>K Ucf FRVK NKf ]f G[c >G[I G[ Kf ?IFVLQK$K Ic 3, G[ 3 G BBQ W [ } Ucf N VLQK N^K IMff fÂ§ G% %r G%r E? &E0A Af / %LQKN^K IMf ? G[ G[ G[ ? VLQK NK IMfn A IF VLQK N^K IMf U nNKFRHKNK VLQKIFK G% G%r Â‘W >G[ G[ %? IF VLQK IFK $ %?QKNKf G% G%r G[ct%fGA% ([SDQGLQJ WKH GHULYDWLYH WHUP LQ (T %f G[ G% G%r f fÂ§%r fÂ§% G[ G[ G% G%r G% G%r U%r % G[A G[A G[ G[ I 5HFRJQL]LQJ WKDW I LV D IUHH LQGH[ VR WKDW 9r G% G[Â 6 9Â% IF% G%r G[" 9M>%r fÂ§N%r %f %f ZKLFK UHVXOWV IURP XVH RI WKH +HOPKROW] HTXDWLRQ ZKLFK JRYHUQV GLIIUDFWLRQ LQ ZDWHU RI FRQVWDQW GHSWK DQG AiU _Yf%_ %f VR WKDW WKH H[SDQGHG GHULYDWLYH WHUP %f EHFRPHV G%f G%r G%G%r G[ % G[ % G[I G[I _9r%_ IF_%_ %f DQG WKH UHVXOW LV O AaÂ=f 7 _9f6_ ) @ %f 7KH DXWKRU ZLVKHV WR H[SUHVV VRPH UHVHUYDWLRQV FRQFHUQLQJ WKH YDOLGLW\ RI (T %f 7KH VKDOORZ ZDWHU DV\PSWRWH IRU WKLV H[SUHVVLRQ LV ]HUR VLQFH IRU VPDOO NKIMf FRWK WK IMf NKILf L IFKAfFRWKIF}AfO ,Q GHHSZDWHU KRZHYHU FRWK$KAf f VR WKDW NK IMf FRWKÂL IMf NK ,I WKH GHULYDWLYH WHUPV DUH QRQ]HUR WKH 6]]
PAGE 126
WHUP ZLOO EH SURSRUWLRQDO WR WKH GHSWK 7KLV ZRXOG PHDQ WKDW GHHS ZDWHU QRQVKRDOLQJ DQG QRQGLVVLSDWLYH ZDYHV LQ YDU\LQJ GHSWKV ZRXOG SURGXFH JUDGLHQWV LQ WKH UDGLDWLRQ VWUHVVHV )RU SODQH ZDYHV WKLV LV QRW D SUREOHP VLQFH WKH FRPSOH[ DPSOLWXGH WHUPV DUH HDVLO\ VKRZQ WR EH ]HUR 7KLV LV VHHQ E\ VXEVWLWXWLRQ RI % DHN[ DQG %r DHaWN[ ZKHUH D _%_ 7KH IRXUWK WHUP LV VWUDLJKW IRUZDUG LJN SVLQK NK ]f BLXW LJN VLQK NK ]f f FRVK NK fÂ§%n 7 FRVK NK B }nJIF 9 R Â‘r! ]f B %A B %HA D FRVK NK n 7 B JN VLQK NK ]f Q JN VLQK N^K ]f @ a D FRVK N>K UMf a VLQKNK %% %f %f %f SZnfG] SJN?%? Â r fIWS VLQK NK ]f VLQKÂK a 9r VLQKIFL! 7KH ILIWK WHUP QHHGV QR HYDOXDWLRQ DQG LV VHHQ WR FDQFHO H[DFWO\ WKH VHFRQG WHUP 7KH VL[WK DQG ODVW WHUP RI (T %f LV Ucnf DQG LV HYDOXDWHG HDVLO\ Ucn %9ff UMnf %H: %%r %rHZWf 6, : a%%n ZKLFK FDQFHOV SDUW RI WKH UHVXOW IURP WKH IRXUWK WHUP $GGLQJ WKH DERYH UHVXOWV IRU DOO WKH WHUPV RI (T %f \LHOGV 3 %f %f %f %f F 3Âa%G%r? A NK ? ?LD GLS \ $ 9 VLQKNKf ADIf _fÂ§NK VLQK NK G[D GLS _9IFe_ N?%?fANKFRWK NK f@ %f
PAGE 127
$33(1',; & /$*5$1*,$1 '(5,9$7,21 2) 7+( *29(51,1* :$9( (48$7,21 7KH JRYHUQLQJ HTXDWLRQ IRU ZDYH PRWLRQ FDQ EH GHULYHG WKURXJK XVH RI D YDULDWLRQDO SULQFLSOH ZKHUH D /DJUDJLDQ LV WKH IXQFWLRQ WKDW LV YDULHG /XNH f VKRZHG WKDW D /DJUDQJLDQ IRU ZDYHV LV JLYHQ E\ / fÂ§ K r= G] &Of .LUE\ f XVHG D WLPH DYHUDJHG /DJUDQJLDQ WR GHULYH D ZDYH HTXDWLRQ )RU WKH IROORZLQJ GHULYDWLRQ HTXDWLRQ &Of ZLOO EH XVHG 7KH WKLUG WHUP LQ (T &Of LV GHWHUPLQHG VWUDLJKW IRUZDUG ZKLOH WKH ILUVW WZR WHUPV QHHG WR EH H[SDQGHG DERXW IM WKH PHDQ ZDWHU OLQH LQ D 7D\ORU VHULHV !W :fD G] >Uc Uff!W _U IO IMf &f :H DVVXPH WKDW WM DQG I! FDQ EH H[SUHVVHG DV rf 9 r9L I! W!R HIW!L &f ZKHUH F LV DQ RUGHULQJ SDUDPHWHU DQG LV WKH GHSWK GHSHQGHQF\ DV GHILQHG LQ (T f )RU WKH VDNH RI VLPSOLFLW\ ZH ZLOO DVVXPH WKDW WKH JUDGLHQW RI IR UHSUHVHQWV WKH VWHDG\ KRUL]RQWDO FXUUHQWV 7KLV PHDQV WKDW IR LV QRW D IXQFWLRQ RI WLPH DQG WKDW : ZLOO QRW EH LQFOXGHG DV LW ZDV LQ WKH GHULYDWLRQ LQ &KDSWHU 7KLV LV EHFDXVH WKH : WHUP LV QRW H[SDQGHG DERXW ] IM DV LW ZDV LQ WKH *UHHQfV IXQFWLRQ GHULYDWLRQ DQG LV QRW QHHGHG LQ RUGHU WR REWDLQ WKH 9IF f 8fMe WHUP ZKLFK DV ZLOO EH VHHQ LV REWDLQHG PRUH GLUHFWO\ LQ WKH SUHVHQW GHULYDWLRQ ,QFOXGLQJ : KHUH ZLOO RQO\ SURGXFH WHUPV WKDW DUH RI KLJKHU RUGHU LQ WKH PHDQ FXUUHQW
PAGE 128
,QWURGXFLQJ &f LQWR &f \LHOGV ÂJYL JK / A IG]AK Wcf H9.I!R9KW!L > I G] H9KIRW!L ^ 9G] > I0G]? K 8Kf .F f \Kf IK >9f@G]` FUML ^ fÂ§ H9NI!RW!LI _m\ Wn 7KH /DJUDQJLDQ / LV DVVXPHG WR EH RI WKH IRUP &f ,W WKHQ IROORZV WKDW / fÂ§ /T I/L H=m I" / ÂR mGL8 eeerf DQG &f &f Â JI!P fÂ§ÂL fÂ§YAR9A ^ U Yr G] U I]G]f P 9ARf Ff IO 8IFR frf I >I9KI r@ 9LAKW!RW!LI[ _Ff r fIWR 7R REWDLQ WKH GHVLUHG OLQHDU JRYHUQLQJ HTXDWLRQ LV YDULHG ZLWK UHVSHFW WR I!L DQG WKHQ LV YDULHG ZLWK UHVSHFW WR Uc? LQ RUGHU WR HOLPLQDWH Uc? 7KH YDULDWLRQDO SULQFLSOH KROGV WKDW WKH /DJUDQJLDQ LV FRQVWDQW XQGHU VPDOO YDULDWLRQV RI WKH XQNQRZQ SDUDPHWHUV ,Q PDWKHPDWLFDO WHUPV DQG '/ '! '/ 'Uf &f FLRf 6LQFH WKH RUGHULQJ SDUDPHWHU f LV DUELWUDU\ WKH DERYH KROGV IRU DOO /f } f f f ,Q WKH Â WHUP I!L LV H[SUHVVHG LQ WHUPV RI LWV GHULYDWLYHV 7KLV UHTXLUHV LQWHJUDWLRQ E\ SDUWV WR
PAGE 129
REWDLQ WKH YDULDWLRQ ZLWK UHVSHFW WR I!? 8SRQ FDUU\LQJ WKLV RXW LW LV IRXQG WKDW 3s G/LB G G8 B G8 [ 'K GID GWnGWIXf KOG9A[f$SSO\LQJ &OOf WR &f DQG UHFRJQL]LQJ WKDW WKH YDULDWLRQ GI!? LV DUELWUDU\ \LHOGV DIWHU FDUU\LQJ RXW WKH LQWHJUDWLRQV LQ Â && mB ,UA&7 YrfÂ§0 nf QX 9IF9r0Lf &f ZKLFK LV UHZULWWHQ DV aMAa AK^&&J9KI!Lf I!LHU N&&Wf &f 9DU\LQJ /L ZLWK UHVSHFW WR LML \LHOGV r W! KI!L &f ZKLFK FDQ EH ZULWWHQ DV r 9L SUU J 'W &f ZKLFK XSRQ VXEVWLWXWLRQ LQWR &f \LHOGV WKH JRYHUQLQJ HTXDWLRQ AU 9A&&S9A[f D rr&&Ifr ZKLFK XSRQ PDNLQJ WKH VXEVWLWXWLRQ WKDW 9AR EHFRPHV Yr r fAU a 9DFF9IFrf U IFfFFfA R FLf 'W &f
PAGE 130
$33(1',; 2%7$,1,1* (48$7,21 f ,Q WKLV $SSHQGL[ ZH VKRZ WKH DOJHEUD LQYROYHG LQ REWDLQLQJ HTXDWLRQ f IURP HTXDn WLRQ f XVLQJ WKH DVVXPSWLRQ WKDW WKH GLVSHUVLRQ UHODWLRQVKLS LV ZULWWHQ DV X fÂ§ D N FRV 28 :H VWDUW E\ XVLQJ WKH IROORZLQJ IRUP RI c! ^J L,\f &rn r &2 L G]f DQG WKHQ PDNH D SKDVH VKLIW E\ XVLQJ WKH VXEVWLWXWLRQ FR}I Â[LFR} G]f VR WKDW LQ WKH ILQDO IRUP B LJ $n^"n \f H} Â &2 Â G]f 'Of 'f 'f 'f DQG DOO \ GHSHQGHQF\ LV VROHO\ LQ WKH FRPSOH[ DPSOLWXGH $n )RU UHIHUHQFH HTXDWLRQ f LV UHSURGXFHG EHORZ 3!]f] SW!YfY NSc! >Z D LnZ9IF f 8f@ LDZ! 8!]f]^Y!YfY^XY![fY^XYLYf] LMMYK! R Gf :ULWLQJ RXW DOO WKH FRPSRQHQWV RI HTXDWLRQ 'f DQG UHDUUDQJLQJ WHUPV JLYHV ^S Xf[ ^S Xf![[ >S 9f!Y@Y NS! RM Rfa! RM8[! LX9Y! LFUZM! 89Â[f\ >89L!Yf[ LX8L![ LX9A!\ 'f 7KH GHULYDWLYHV RI I! GHILQHG E\ HTXDWLRQ DUH DV IROORZV !] fÂ§LJ >Af LN FRV A HrINHRWG]f
PAGE 131
O!][ fÂ§ LIFFR A WEFRVfD fÂ§ N FRV fÂ§ HrA IF FRr rG[f D D ^![fY fÂ§ n, ^Df HrArFRLUf_ LN FRV LN FRV f\ AFrfFRLGrf,, + ,, ^!]fY 7KH DERYH IRUPV RI WKH GHULYDWLYHV UHFRJQL]H WKDW WKH SKDVH IXQFWLRQ Hf NFRWHG[f A GHULYDWLYHV VLQFH N N[ \f +RZHYHU WKH \ GHULYDWLYHV RI WKH SKDVH IXQFWLRQ LQWHJUDO DUH QRW HYDOXDWHG VLQFH WKH SKDVH VKLIW HTXDWLRQ 'f ZLOO HOLPLQDWH DQ\ \ GHSHQGDQFH IURP WKH SKDVH IXQFWLRQ DQG SODFH DOO \ GHSHQGDQFH LQ WKH FRPSOH[ DPSOLWXGH 7KH f ?r ;; WHUP LV QHJOHFWHG IROORZLQJ WKH DVVXPSWLRQ H[SODLQHG LQ VHFWLRQ 6XEVWLWXWLQJ WKH DERYH GHULYDWLYHV LQWR HTXDWLRQ 'f JLYHV U 1 3a8f; f HLILFRrWG[f LN FRV S fÂ§ 8f[fÂ§HnINFROG]f ?U [ R LNFRVS8f f HI LNFRVf[S 8fHANFRWG]n! N FRV >S 8f Â£ÂIN f}mrrf S \A IÂ£HLINHRrWG[ff ? R ,9 9} \ NSHLILFRWLG]f Z DfHfIWFRHWG[f LX8[HAIWFRWG[A 2 2 A Â‘ V LX9YM0 rff G]f LZ$HAI r FRf A ^89fY Hf r FRf G]f 89 G[f fÂ§LN FRV 8Y9fÂ§Hf N FRr G[A fÂ§LN FRV 89Y fÂ§rFRrG]A LNFRV fY89HAI NFRW GO! D D LNFR%89 LNFRV f\89HLNFRW G]!! I $ r1 >89f; AHrNFRLLOff LX!XAHrHRUI[f RMNFRH8AINFRWG]A WX9 'f
PAGE 132
'URSSLQJ WHUPV WKDW FRQWDLQ SURGXFWV RI GHULYDWLYHV RI M DQG GHULYDWLYHV RI WKH FXUUHQW DV ZHOO DV SURGXFWV RI WKH FXUUHQW DQG VHFRQG GHULYDWLYHV RI e LH WHUPV WKDW DUH GHVLJQDWHG E\ DQ RYHUEUDFHf HTXDWLRQ 'f LV UHDUUDQJHG DV ,} >IFFRVS fÂ§ 8f X(M fÂ§ } >IFFRVS fÂ§ (f XO@ A }:L >aIF FRV %S 8f NS X D XIFFRVOM A fÂ§ LN FRV fY89 fÂ§ >9\ Z N FRV 8f IF FRV 8f\9 N FRV (9@ A ` Hrn N &2L >3DHAIF&2Off >.ZLIFFRV(f@ Jf ,W LV HDVLO\ VKRZQ WKDW NFRV^S 8f MM8 R^&fFRV 8f %\ ZULWLQJ WKH GLVSHUVLRQ UHODWLRQVKLS DV X! N FRV 28 LW LV DOVR HDVLO\ VKRZQ WKDW RU D R!N FRV 8 fÂ§ N FRV %8\ )URP HTXDWLRQ LW LV DSSDUHQW WKDW IRU D FRQVWDQW R N FRV f\ D Y 7KXV HTXDWLRQ 'f EHFRPHV ^m >U& FRV 8f@[ e LU&W FRV% 8f A >IFSO FRV f@ LZ$ W9\$ HLINFRWG]n! L9D AHLINFDG]A _S ZKHUH D LNFRV fY89e WHUP KDV EHHQ GURSSHG $W WKLV SRLQW D SKDVH VKLIW LV LQWURGXFHG Y 'f ',2f 'f 'f 'f A B $nHrI LFR}I G[I N FRm G]f 'f
PAGE 133
ZKHUH N DQG DUH D UHIHUHQFH ZDYH QXPEHU DQG ZDYH DQJOH )RU FRPSXWDWLRQDO SXUSRVHV N ZLOO EH DQ DYHUDJHG ZDYH QXPEHU DFURVV D JULG URZ DQG ZLOO EH WDNHQ DV WKH ZDYH DQJOH LQ WKH DEVHQFH RI FXUUHQWV XVLQJ 6QHOOfV ODZ )XUWKHU LW ZLOO EH DVVXPHG VROHO\ IRU FRPSXWDWLRQDO SXUSRVHV WKDW LV HTXDO WR 7KLV JLYHV }>D&WFRV 8f@[ M WII&L FRV 8f MM &W FRV 8f FRV N Nf$ mZ$ L9Y$ }9 fÂ§f LDY9 fÂ§ 'f ?R Y R ([SDQGLQJ WKH GHULYDWLYHV RI e DQG ZLWK D IHZ PLQRU PDQLSXODWLRQV RI WKH WHUPV WKH JRYHUQLQJ HTXDWLRQ f LV REWDLQHG &M&RV &2$[ nÂ&I&RV &2FRVN Nf$ ? ?&}FRV% 8 ?9 $Y N&WO FRV f$ ,&&J ] Z $ 'f
PAGE 134
%,%/,2*5$3+< $UWKXU 56 f f5HIUDFWLRQ RI VKDOORZ ZDWHU ZDYHV WKH FRPELQHG HIIHFWV RI FXUUHQWV DQG XQGHUZDWHU WRSRJUDSK\f 7UDQV $PHU *HRSK\VLFDO 8QLRQ f %DWHPDQ + f 3DUWLDO 'LIIHUHQWLDO (TXDWLRQV RI 0DWKHPDWLFDO 3K\VLFV 'RYHU 3XEn OLFDWLRQV 1HZ
PAGE 135
'DOO\ :5 f f$ QXPHULFDO PRGHO IRU EHDFK SURILOH HYROXWLRQf 06 7KHVLV 8QLn YHUVLW\ RI 'HODZDUH 1HZDUN 'DOO\ :5 f f:DYH WUDQVIRUPDWLRQ LQ WKH VXUI URQHf 7HFKQLFDO 5HSRUW 1R 'HSW RI &RDVWDO DQG 2FHDQRJUDSKLF (QJLQHHULQJ 8QLYHUVLW\ RI )ORULGD *DLQHVYLOOH 'DOO\ :5 'HDQ 5* DQG 'DOU\PSOH 5$ f f $ PRGHO IRU EUHDNHU GHFD\ RQ EHDFKHVf 3URF rA &RDVWDO (QJ &RQI +RXVWRQ 'HDQ 5* DQG 'DOU\PSOH 5$ f :DWHU :DYH 0HFKDQLFV IRU 6FLHQWLVWV DQG (QJLQHHUV 3UHQWLFH+DOO (QJOHZRRG &OLIInV 1HZ HUVH\ (EHUVROH %$ DQG 'DOU\PSOH 5$ f f$ QXPHULFDO PRGHO IRU QHDUVKRUH FLUFXODWLRQ LQFOXGLQJ FRQYHFWLYH DFFHOHUDWLRQ DQG ODWHUDO PL[LQJf 2FHDQ (QJLQHHULQJ 5HSRUW 1R 'HSW RI &LYLO (QJ 8QLYHUVLW\ RI 'HODZDUH 1HZDUN 'HODZDUH (EHUVROH %$ DQG 'DOU\PSOH 5$ f f1XPHULFDO PRGHOLQJ RI QHDUVKRUH FLUFXODn WLRQf 3URF A &RDVWHG (QJ &RQI 6\GQH\ *DUUHWW &5 f f'LVFXVVLRQ WKH DGLDEDWLF LQYDULDQW IRU ZDYH SURSDJDWLRQ LQ D QRQXQLIRUP PRYLQJ PHGLXPf 3URF 5R\ 6RF $ *RXUOD\ 05 f f:DYH VHWXS DQG ZDYH JHQHUDWHG FXUUHQWV LQ WKH OHH RI D EUHDNn ZDWHU RU KHDGODQGf 3URF rA &RDVWDO (QJ &RQI &RSHQKDJHQ +DOHV /= f f(URVLRQ FRQWURO RI VFRXU GXULQJ FRQVWUXFWLRQ UHSRUW ([SHULPHQWDO PHDVXUHPHQWV RI UHIUDFWLRQ GLIIUDFWLRQ DQG FXUUHQW SDWWHUQV QHDU MHWWLHVf 86 $UP\ (QJLQHHU :DWHUZD\V ([SHULPHQW 6WDWLRQ 9LFNHVEXUJ 0LVVLVVLSSL +DQVHQ % DQG 6YHQGVHQ ,$ f f5HJXODU ZDYHV LQ VKRDOLQJ ZDWHU ([SHULPHQn WDO GDWDf 6HULHV 3DSHU ,QVW RI +\GURG\QDPLFV DQG +\GUDXOLF (QJLQHHULQJ 7HFKQLFDO 8QLYHUVLW\ RI 'HQPDUN /\QJE\ +DUULVRQ : DQG :LOVRQ :6 f f'HULYDWLRQ RI QXPHULFDO FDOFXODWLRQ RI ZDYH UHIUDFWLRQf 7HFKQLFDO PHPRUDQGXP 1R 86 $UP\ &RDVWDO (QJLQHHULQJ 5Hn VHDUFK &HQWHU )W %HOYRLU +RULNDZD DQG .XR &7 f f$ VWXG\ RI ZDYH WUDQVIRUPDWLRQ LQVLGH WKH VXUI ]RQHf 3URF A &RDVWDO (QJ &RQI 7RN\R RKQVRQ : f f7KH UHIUDFWLRQ RI VXUIDFH ZDYHV E\ FXUUHQWVf 7UDQV $P *HRSK\V 8QLRQ RKQVRQ : 2f%ULHQ 03 DQG ,VDDFV ' f f*UDSKLFDO FRQVWUXFWLRQ RI ZDYH UHIUDFWLRQ GLDJUDPVf 7HFK 5HSRUW 1R 86 +\GURJUDSKLF 2IILFH :DVKLQJWRQ '& RQVVRQ ,* f f%RRLMfV FXUUUHQWZDYH HTXDWLRQ DQG WKH UD\ DSSUR[LPDWLRQf 3URJ 5HSRUW SS ,QVW +\GURG\Q DQG +\GUDXOLF (QJ 7HFKQLFDO 8QLYHUVLW\ RI 'HQPDUN /\QJE\
PAGE 136
.HOOHU % f f6XUIDFH ZDYHV RQ ZDWHU RI QRQXQLIRUP GHSWKf )OXLG 0HFK .LUE\ 7 f f 3URSDJDWLRQ RI ZHDNO\QRQOLQHDU VXUIDFH ZDWHU ZDYHV LQ UHJLRQV ZLWK YDU\LQJ GHSWK DQG FXUUHQWf 215 7HFK 5HSRUW 5HVHDUFK 5HS &( 'HSW RI &LYLO (QJ 8QLYHUVLW\ RI 'HODZDUH 1HZDUN .LUE\ 7 f f$ QRWH RQ OLQHDU VXUIDFH ZDYH FXUUHQW LQWHUDFWLRQ RYHU VORZO\ YDU\LQJ WRSRJUDSK\f *HRSK\J 5HV &f .LUE\ 7 f f $ PRGHO IRU JUDGXDO UHIOHFWLRQ RI ZHDNO\ WZRGLPHQVLRQDO ZDYHV LQ D VORZO\ YDU\LQJ GRPDLQ ZLWK FXUUHQWV JRYHUQLQJ HTXDWLRQVf 7HFK 5HSRUW 1R 'HSW RI &RDVWHG DQG 2FHDQRJUDSKLF (QJLQHHULQJ 8QLYHUVLW\ RI )ORULGD *DLQHVYLOOH .LUE\ 7 DQG 'DOU\PSOH 5$ f f1XPHULFDO PRGHOLQJ RI WKH QHDUVKRUH UHJLRQf 215 7HFK 5HSRUW 5HVHDUFK 5HSRUW &( 'HSW RI &LYLO (QJ 8QLYHUVLW\ RI 'HODZDUH 1HZDUN .LUE\ 7 DQG 'DOU\PSOH 5$ f f0RGHOLQJ ZDYHV LQ VXUI]RQHV DQG DURXQG LVn ODQGVf :DWHUZD\ 3RUW &RDVWDO DQG 2FHDQ (QJLQHHULQJ /H 0KDXW % DQG .RK 5&< f f2Q WKH EUHDNLQJ RI ZDYHV DUULYLQJ DW DQ DQJOH WR WKH VKRUHf +\GUDXOLF 5HV f /LX 3/) DQG 0HL && f f(IIHFW RI D EUHDNZDWHU RQ QHDUVKRUH FXUUHQWV GXH WR EUHDNLQJ ZDYHVf 7HFKQLFDO 0HPRUDQGXP 1R &RDVWDO (QJLQHHULQJ 5HVHDUFK &HQWHU )RUW %HOYRLU 9D /RQJXHW+LJJLQV 06 Df f/RQJVKRUH FXUUHQWV JHQHUDWHG E\ REOLTXHO\ LQFLGHQW VHD ZDYHV ,f *HRSK\V 5HV f /RQJXHW+LJJLQV 06 Ef f/RQJVKRUH FXUUHQWV JHQHUDWHG E\ REOLTXHO\ LQFLGHQW VHD ZDYHV ,,f *HRSK\V 5HV f /RQJXHW+LJJLQV 06 DQG 6WHZDUW 5: f f5DGLDWLRQ VWUHVV DQG PDVV WUDQVSRUW LQ JUDYLW\ ZDYHV ZLWK DSSOLFDWLRQ WR fVXUI EHDWVff )OXLG 0HFK /RQJXHW+LJJLQV 06 DQG 6WHZDUW 5: f f$ QRWH RQ ZDYH VHWXSf 0DU 5HV f /RQJXHW+LJJLQV 06 DQG 6WHZDUW 5: f f5DGLDWLRQ VWUHVVHV LQ ZDWHU ZDYHV D SK\VLFDO GLVFXVVLRQ ZLWK DSSOLFDWLRQVf 'HHS6HD 5HVHDUFK /XNH & f f$ YDULDWLRQDO SULQFLSOH IRU D IOXLG ZLWK D IUHH VXUIDFHf )OXLG 0HFK f 0HL && f f$ QRWH RQ WKH DYHUDJHG PRPHQWXP EDODQFH LQ WZRGLPHQVLRQDO ZDWHU ZDYHVf 0DU 5HV f 0HL && f 7KH $SSOLHG '\QDPLFV RI 2FHDQ 6XUIDFH :DYHV :LOH\,QWHUVFLHQFH 1HZ
PAGE 137
0LFKH 5 f f0RXYHPHQW RQGXODWRLUHV GH OD PHU HQ SURIRQGHXU FRQVWDDQWWWH RX GFURLVVDQWHf $QQ GHV 3RQW% HW &KDXVVVV 0LOHV : f f6XUIDFH ZDYH VFDWWHULQJ PDWUL[ IRU D VKHOI )OXLG 0HFK 0LOHV : f f2Q VXUIDFHZDYH GLIIUDFWLRQ E\ D WUHQFKf )OXLG 0HFK 0XQN :+ DQG $UWKHU 56 f f:DYH LQWHQVLW\ DORQJ D UHIUDFWHG UD\f LQ *UDYLW\ :DYHV 1%6 &LUF 1HZPDQ 1 f f'LIIUDFWLRQ RI ZDWHU ZDYHVf $SSOLHG 0HFK 5HYLHZ f 1RGD ( 6RQX &5XSHUW 9& DQG &ROOLQV , f f1HDUVKRUH FLUFXODWLRQ XQGHU VHD EUHH]H FRQGLWLRQV DQG ZDYHFXUUHQW LQWHUDFWLRQV LQ WKH VXUI ]RQHf 7HWUD 7HFK 5HSRUW 7& 2f%ULHQ 03 f f:DYH UHIUDFWLRQ DW /RQJ %HDVFK DQG 6DQWD %DUEDUD &DOLIRUQLDf %HDFK (URVLRQ %RDUG %XOOHWLQ f 3HQQH\ :* DQG 3ULFH $7 f f'LIIUDFWLRQ RI VHD ZDYHV E\ EUHDNZDWHUVf 7HFKQLn FDO +LVWRU\ 5HSRUW 1R 'LUHFWRUDWH 0LVH :HDSRQV 'HYHORSPHQW :DVKLQJWRQ '& 3HQQH\ :* DQG 3ULFH $7 f f7KH GLIIUDFWLRQ WKHRU\ RI VHD ZDYHV DQG WKH VKHOWHU DIIRUGHG E\ EUHDNZDWHUVf 3KLO 7UDQV 5R\ 6RF $ 3HUOLQ 0 DQG 'HDQ 5* f f$ QXPHULFDO PRGHO WR VLPXODWH VHGLPHQW WUDQVSRUW LQ WKH YLFLQLW\ RI FRDVWDO VWUXFWXUHVf 0LVH 5HSRUW 1R 86 $UP\ &RDVWDO (QJLQHHULQJ 5HVHDUFK &HQWHU )W %HOYRLU 9LUJLQLD 3KLOOLSV 20 f 7KH '\QDPLFV RI WKH 8SSHU 2FHDQ &DPEULGJH 8QLYHUVLW\ 3UHVV 3LHUVRQ :f f 7KH LQWHUSUHWDWLRQ RI FURVVHG RUWKRJRQDOV LQ WKH ZDYH UHIUDFWLRQ SKHQRPHQDf 7HFK 0HP 1R %HDFK (URVLRQ %RDUG :DVKLQJWRQ '& 3XWQDP $ DQG $UWKXU 56 f f'LIIUDFWLRQ RI ZDWHU ZDYHV E\ EUHDNZDWHUVf 7UDQV $P *HRSK\V 8QLRQ 5DGGHU $& f f2Q WKH SDUDEROLF HTXDWLRQ IRU ZDWHU ZDYH SURSDJDWLRQf )OXLG 0HFK f 6HHOLJ :1 f f7ZRGLPHQVLRQDO WHVWV RI ZDYH WUDQVPLVVLRQ DQG UHIOHFWLRQ FKDUn DFWHULVWLFV RI ODERUDWRU\ EUHDNZDWHUVf 7HFKQLFDO 5HSRUW 1R 86 $UP\ &RDVWDO (QJLQHHULQJ 5HVHDUFK &HQWHU )W %HOYRLU 9LUJLQLD 6KHQJ <3 DQG %XWOHU +/ f f0RGHOLQJ RI FRDVWDO FXUUHQWV DQG VHGLPHQW WUDQVn SRUWf 3URF A &RDVWDO (QJ &RQI &DSH 7RZQ 6KRUH 3URWHFWLRQ 0DQXDO f 86 $UP\ &RDVWDO (QJLQHHULQJ 5HVHDUFK &HQWHU )W %HOYRLU 9LUJLQLD
PAGE 138
6PLWK f f2Q VXUIDFH ZDYHV FURVVLQJ D VWHS ZLWK KRUL]RQWDO VKHDUf )OXLG 0HFK 6YHQGVHQ ,$ f f:DYH KHLJKWV DQG VHWXS LQ D VXUI ]RQHf &RDVWDO (QJ f 7DSSHUW )' f f7KH SDUDEROLF DSSUR[LPDWLRQ PHWKRGf ,Q :DYH 3URSDJDWLRQ DQG 8QGHUZDWHU $FRXVWLFV HGLWHG E\ % .HOOHU DQG 6 3DSDGDNLVf &KDSWHU 6SULQJHU 9HUODJ %HUOLQ 9DQ 'RUQ : f f:LQG VWUHVV RQ DQ DUWLILFLDO SRQGf 0DU 5HV f 9HPXODNRQGD 65 f f(URVLRQ FRQWURO RI VFRXU GXULQJ FRQVWUXFWLRQ UHSRUW &XUn UHQW D ZDYHLQGXFHG FXUUHQW PRGHOf 86 $UP\ (QJLQHHU :DWHUZD\V ([SHULn PHQW 6WDWLRQ 9LFNHVEXUJ 0LVVLVVLSSL 9HPXODNRQGD 65 6ZDLQ $ +RXVWRQ 5 )DUUDU 3' &KRX /: (EHUVROH %$ f f&RDVWDO DQG LQOHW SURFHVVHV QXPHULFDO PRGHOLQJ V\VWHP IRU 2UHJRQ ,Qn OHW 1RUWK &DUROLQDf 7HFK 5HSRUW &(5& 86 $UP\ (QJLQHHU :DWHUZD\V ([SHULPHQW 6WDWLRQ 9LFNHVEXUJ 0LVVLVVLSSL 9HQJD\LO 3 DQG .LUE\ 7 f f6KRDOLQJ DQG UHIOHFWLRQ RI QRQOLQHDU VKDOORZ ZDWHU ZDYHVf 7HFK 5HSRUW 1R &RDVWDO DQG 2FHDQRJUDSKLF (QJLQHHULQJ 'HSDUWn PHQW 8QLYHUVLW\ RI )ORULGD *DLQHVYLOOH 9LVVHU 3f f7KH SURSHU ORQJVKRUH FXUUHQW LQ D ZDYH EDVLQf 5HSRUW 1R 'HSW RI &LYLO (QJLQHHULQJ 'HOIW 8QLYHUVLW\ 'HOIW :KLWKDP *% f /LQHDU DQG 1RQOLQHDU :DYHV :LOH\ 1HZ
PAGE 139
%,2*5$3+,&$/ 6.(7&+ +DUOH\ :LQHU ZDV ERUQ LQ 5RDQRNH 9LUJLQLD RQ 0D\ DQG UDLVHG LQ &XPEHUODQG 0DU\ODQG 7KHUH KH DWWHQGHG SXEOLF VFKRROV URDPHG WKH PRXQWDLQV LQ KLV EDFN \DUG DQG H[FHOOHG LQ GRZQKLOO VNLLQJ $IWHU KLJK VFKRRO KH DWWHQGHG 'DUWPRXWK &ROOHJH ZKHUH KLV ILUVW DWWHPSW DW REWDLQLQJ D FROOHJH GHJUHH HQGHG D IHZ \HDUV VKRUW RI FRPSOHWLRQ 7KLV ZDV IROORZHG E\ VHYHUDO \HDUV RI EHLQJ fRQ WKH URDGf LQ WKH %RKHPLDQ WUDGLWLRQ RI DFN .HURXDF DORQJ ZLWK PDQ\ \HDUV RI SURWHVWLQJ WKH 9LHWQDP ZDU ,Q WKH DXWKRU GLG KLV SDUW WR HQG WKH ZDU E\ EHLQJ LQFDUFHUDWHG LQ WKH )HGHUDO &RUUHFWLRQDO ,QVWLWXWLRQ DW 'DQEXU\ &RQQHFWLFXW IRU UHVLVWLQJ WKH GUDIW 7KH DXWKRUf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f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
PAGE 140
, FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ LJ &KDLUPDQ c 3URIHVVRU RI $HURVSDFH (QJLQHHULQJ 0HFKDQLFV DQG (QJLQHHULQJ 6FLHQFH FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ I A DPHV 7 .LUE\ &RFKDLUUKÂ£Q $VVRFLDWH 3URIHVVRU RI &RDVWDO DQG 2FHDQRJUDSKLF (QJLQHHULQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DULG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH P VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ WIn9" %HQW $ &KULVWHQVHQ 3URIHVVRU RI &LYLO (QJLQHHULQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWR RI 3KLORVRSK\ 5REHUW 'HDQ *UDGXDWH 5HVHDUFK 3URIHVVRU RI &RDVWDO DQG 2FHDQRJUDSKLF (QJLQHHULQJ
PAGE 141
, FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV FI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 3URIHVVRU RI $HURVSDFH (QJLQHHULQJ 0HFKDQLFV DQG (QJLQHHULQJ 6FLHQFH 7KLV GLVVHUWDWLRQ ZDV VXEPLWWHG WR WKH *UDGXDWH )DFXOW\ RI WKH &ROOHJH RI (QJLQHHULQJ DQG WR WKH *UDGXDWH 6FKRRO DQG ZDV DFFHSWHG DV SDUWLDO IXOILOOPHQW RI WKH UHTXLUHPHQWV IRU WKH GHJUHH FI 'RFWRU RI 3KLORVRSK\ $XJXVW 6 'HDQ *UDGXDWH 6FKRRO
PAGE 142
81,9(56,7< 2) )/25,'$
PAGE 143
SDUWLFXODU WKH FRPSDULVRQ WR WKH ODERUDWRU\ H[SHULPHQWV RI *RXUOD\f FRQILUP WKH SUHGLFWLYH FDSDELOLW\ RI WKH QXPHULFDO PRGHO 7KH VLPSOLILFDWLRQV PHQWLRQHG DERYH PDNH WKLV SRVVLn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f KDYH IRXQG D PHWKRG WR GHWHUPLQH D QRQ]HUR ZDYH KHLJKW LQ WKH OHH RI D EUHDNZDWHU +RZHYHU WKHLU VROXWLRQ LV LQ WKH IRUP RI D )RXULHU LQWHJUDO ZKLFK LV QRW VXLWHG WR WKH FRPSOH[ DPSOLWXGH IRUP UHTXLUHG LQ WKH SUHVHQW PRGHO
PAGE 144
,QFOXGLQJ :DYH 5HIOHFWLRQ 7KH DXWKRU KDV PDGH XQVXFFHVVIXO DWWHPSWV WR LQFOXGH ZDYH UHIOHFWLRQ LQ WKH PRGHO .LUE\ f XVHG FRXSOHG HTXDWLRQV JRYHUQLQJ WKH LQWHUDFWLRQ RI D IRUZDUG DQG D EDFNZDUG UHIOHFWHGf SURSDJDWLQJ ZDYH WKDW DUH FRXSOHG WKURXJK D VORSH LQWHUDFWLRQ +H ZDV DEOH WR REWDLQ WKH UHIOHFWLRQ DQG WUDQVPLVVLRQ FRHIILFLHQWV RYHU VXEPHUJHG EUHDNZDWHUV WKDW DJUHH ZHOO ZLWK WKH H[SHULPHQWV RI 6HHOLJ f 2QFH WKH DUUD\V RI WKH FRPSOH[ DPSOLWXGHV RI WKH WZR ZDYHV LV GHWHUPLQHG WKH UDGLDWLRQ VWUHVV WHUPV FDQ EH GHWHUPLQHG IURP HTXDWLRQV WKDW KDYH EHHQ GHULYHG IROORZLQJ WKH VDPH SURFHGXUH DV XVHG E\ 0HL f EXW GHILQLQJ I! DV EHLQJ FRPSULVHG RI WZR FRPSRQHQWV ,Q IDFW WKLV SURFHGXUH FDQ DOVR EH XVHG WR REWDLQ WKH UDGLDWLRQ VWUHVV WHUPV IRU DQ\ QXPEHU RI FRPSRQHQWV 7KH RQO\ OLPLW EHLQJ WKH DPRXQW RI FRPSXWDWLRQ UHTXLUHG DV WKH QXPEHU RI FRPSRQHQWV LQFUHDVHV 6PLWK f IROORZLQJ WKH ZRUN RI 0LOHV f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
PAGE 145
DV D IXQFWLRQ RI NQRZQ YDULDEOHV DQG WKHQ VROYH D VHSDUDWH PRGHO IRU VHGLPHQW EDODQFH DQG WKHQ XSGDWH WKH GHSWKV DQG SURFHHG LWHUDWLQJ EHWZHHQ WKH VHGLPHQW EDODQFH PRGHO DQG WKH SUHVHQW ZDYHFXUUHQW PRGHO PDUFKLQJ LQ WLPH 6LQFH LW KDV DOUHDG\ EHHQ GHPRQVWUDWHG WKDW WKH ZDYH FXUUHQW PRGHO JLYHV DGHTXDWH UHVXOWV LQ WKH OHH RI D VKRUH SDUDOOHO EUHDNZDWHU D JRRG WHVW IRU VXFK D VHGLPHQW WUDQVSRUW PRGHO ZROG EH WR VHH LI LW FRXOG \LHOG D WRPEROR IRUPDWLRQ LQ WKH OHH RI D VKRUH SDUDOOHO EUHDNZDWHU
PAGE 146
$33(1',; $ '(5,9$7,21 2) 7+( '(37+ $9(5$*(' (48$7,216 ,Q WKLV DSSHQGL[ D GHULYDWLRQ RI WKH JRYHUQLQJ HTXDWLRQV IRU WKH FLUFXODWLRQ PRGHO LV SUHVHQWHG 7KLV DSSHQGL[ FORVHO\ IROORZV WKH ZRUN RI (EHUVROH DQG 'DOU\PSOH f %HIRUH SURFHHGLQJ ZLWK WKH GHULYDWLRQV RI WKH GHSWK LQWHJUDWHG WLPH DYHUDJHG HTXDWLRQV RI PRWLRQ DQG WKH FRQWLQXLW\ HTXDWLRQ WKH /HLEQLW] 5XOH IRU WDNLQJ WKH GHULYDWLYH RI DQ LQWHJUDO LV VWDWHG VLQFH LW LV LQYRNHG WKURXJKRXW WKH GHULYDWLRQ 7\ &,n9f & Gb7G[ P\fn\fGe $! 7KH /HLEQLW] 5XOH FDQ DOVR EH XVHG LQ UHYHUVH WR WDNH GHULYDWLYHV RXWVLGH RI DQ LQWHJUDO 7KH NLQHPDWLF ERXQGDU\ FRQGLWLRQV DUH DOVR UHSURGXFHG VLQFH WKH XVH RI WKH /HLEQLW] 5XOH ZLOO \LHOG ERWWRP DQG VXUIDFH WHUPV WKDW ZLOO EH HOLPLQDWHG WKURXJK XVH RI WKHVH ERXQGDU\ FRQGLWLRQV 2Q WKH IUHH VXUIDFH WKH ERXQGDU\ FRQGLWLRQ LV GUf GUf GUf Zf D XfA YÂ£c $f DQG DW WKH ERWWRP WKH ERXQGDU\ FRQGLWLRQ $f ,Q WKH FLUFXODWLRQ PRGHO LW LV DVVXPHG WKDW WKH VWLOO ZDWHU GHSWK GRHV QRW YDU\ ZLWK WLPH VR WKH ILUVW WHUP RI (T $f LV GURSSHG DFWXDOO\ IRU WLPH YDU\LQJ GHSWK HVVHQWLDOO\ WKH VDPH HTXDWLRQV DUH REWDLQHGf 7KH FRQWLQXLW\ HTXDWLRQ RU WKH FRQVHUYDWLRQ RI PDVV HTXDWLRQ DV LW LV DER FDOOHGf LQ WKUHH GLPHQVLRQV LV G!Xf GSYf GSZf GW G[ G\ G] $f
PAGE 147
,QWHJUDWLQJ RYHU WKH GHSWK IURP ] fÂ§ fÂ§K WR ] Uc DQG XVLQJ WKH /HLEQLW] 5XOH WR WDNH GHULYDWLYHV RXWVLGH WKH LQWHJUDG WKH FRQWLQXLW\ HTXDWLRQ EHFRPHV G Ir GU@ G Q DW/SLfSnLUW7]/nrLO KR I ? GUf U ? GK G Q :V:NMDL/nf V "rfmIDfrr $f 7KH WHUPV ZLWK DQ RYHUEUDFH DUH S WLPHV WKH IUHH VXUIDFH NLQHPDWLF ERXQGDU\ FRQGLWLRQ DQG WKH WHUPV ZLWK DQ XQGHUEUDFH DUH S WLPHV WKH ERWWRP ERXQGDU\ FRQGLWLRQ (OLPLQDWLQJ ERWK RI WKHVH IURP (T $f JLYHV G >L GIL GUr 3G]]a SXG] fÂ§ SYG] GW K G[MN G\ Kf $f (TXDWLRQ $f LV QRZ WLPH DYHUDJHG 7LPH DYHUDJLQJ PHDQV DYHUDJLQJ RYHU D ZDYH SHULRG 7KH WLPH DYHUDJH RI D WHUP LV GHQRWHG E\ DQ RYHUEDU 7KXV A L U7 8 )GW $f ZKHUH ) UHSUHVHQWV DQ\ WHUP DQG 7 LV WKH ZDYH SHULRG /HWWLQJ X DQG Y EH FRPSULVHG RI D PHDQ IORZ D ZDYH LQGXFHG IORZ DQG DQ DUELWUDU\ RU UDQGRP WXUEXOHQFH FRPSRQHQW X Xn 9 9 9 9 $f $f 7KH WXUEXOHQW IOXFWXDWLRQV DUH DVVXPHG WR EH RI D KLJK IUHTXHQF\ UHODWLYH WR WKH ZDYH IUHTXHQF\ VR WKDW WKHLU WLPH DYHUDJHV RYHU WKH ZDYH SHULRG DUH ]HUR %\ VXEVWLWXWLQJ WKH DERYH H[SUHVVLRQV IRU X DQG Y LQWR (T $f DQG WDNLQJ WLPH DYHUDJHV WKH FRQWLQXLW\ HTXDWLRQ LV ^S^. }"f` A ^3LKR }"f` 3rG] $f $OOf
PAGE 148
7KH WLPH DYHUDJHV RI WKH YHUWLFDOO\ LQWHJUDWHG ZDYH LQGXFHG YHORFLWLHV DQG Y DUH QRW ]HUR 7KHVH TXDQWLWLHV DUH NQRZQ DV WKH ZDYH LQGXFHG PDVV IOX[ 'HILQLQJ 8 X X DQG 9 Y $f ZKHUH X IW U SG] DQG R IMfN S^K r"f WKH WLPH DYHUDJHG GHSWK LQWHJUDWHG FRQWLQXLW\ HTXDWLRQ LV bc iÂXKR WLf AYK Icf IW Y fÂ§ a7L UU SL! G] S^K ULfKR $f $f 7KH DVVXPSWLRQ WKDW WKH GHQVLW\ LV FRQVWDQW LQ ERWK WLPH DQG VSDFH KDV XVHG LQ REWDLQLQJ ERWK (T $f DQG (T $f 7KH DVVXPSWLRQ RI D FRQVWDQW ERWWRP KDV DOVR EHHQ XVHG LQ REWDLQLQJ $f ,I WKH ERWWRP LV QRW FRQVWDQW LQ WLPH D WHUP ZRXOG EH LQFOXGHG LQ (T $f 7KH [ DQG \ PRPHQWXP HTXDWLRQV DUH KDQGOHG LQ WKH VDPH PDQQHU DV WKH FRQWLQXLW\ HTXDWLRQ DERYH 7KH [ DQG \ GLUHFWLRQ PRPHQWXP HTXDWLRQV DUH GX GX GXY GXZ GS UGU[[ GUY[ GU[[ 7W DLU ,7 Zr D7` GY W GXY W GY W GYZ B GS GU[Y GUYY GU=@ GW G[ G\ G] ]LU $ ` $f $f S G\ S G[ G\ G] 7KH VDPH SURFHGXUH LV IROORZHG IRU ERWK WKH [ DQG \ HTXDWLRQV 2QO\ (T $f ZLOO EH GHDOW ZLWK KHUH ,QWHJUDWLQJ WKH OHIW KDQG VLGH RYHU WKH GHSWK DQG DSSO\LQJ WKH /HLEQLW] 5XOH WR WDNH WKH GHULYDWLYH RXWVLGH WKH LQWHJUDO JLYHV W 7LS? G IW GU@ GK /+6! aD7 LI G[MK QG[ r G[ G\ ; GUf GKR XYG] XWff fÂ§ XYfIFR fÂ§ K G\ G\ 8:fO a 8:fKR $f
PAGE 149
7KH WHUPV LQ (T $f WKDW DUH GHVLJQDWHG E\ DQ RYHUEUDFH DUH 8I WLPHV WKH NLQHPDWLF IUHH VXUIDFH FRQGLWLRQ DQG WKRVH GHVLJQDWHG E\ DQ XQGHUEUDFH DUH XBM WLPHV WKH ERWWRP ERXQGDU\ FRQGLWLRQ 7KHVH DUH ERWK GURSSHG OHDYLQJ f G U GU GU /+6f r X XG]Ga[ X XG]UGa\ X XY G] A 6XEVWLWXWLQJ WKH GHILQLWLRQV JLYHQ IRU X DQG Y LQ (TQ $f DQG $f DQG UHFRJQL]LQJ WKDW WKH WXUEXOHQW IOXFWXDWLQJ FRPSRQHQWV DUH RI D KLJK IUHTXHQF\ UHODWLYH WR WKH ZDYH IUHTXHQF\ VR WKDW WKH WLPH DYHUDJHV RI LQWHJUDOV LQYROYLQJ WKH SURGXFW RI D WXUEXOHQW FRPSRQHQW DQG D ZDYH LQGXFHG FRPSRQHQW DUH ]HUR JLYHV GaU GaU G aU GaU /+6 fÂ§ XG] fÂ§ G] fÂ§ X"G]fÂ§ WIG] $f GWMaK GW@. G[.f G[MK f G aUa7a= G aU G aU G aU e XnfG] fÂ§ XX G] fÂ§ XYG] fÂ§ XLLG] $f G[KR G[MK G\K G\KR G aU G aU G aU YG]fÂ§ XLfG] fÂ§ XnYnG] $f R\MKR R\K G\ K XVLQJ WKH GHILQLWLRQV RI (TV $f DQG $f WKLV EHFRPHV ZY^KLfUY/UG]\?ZAf/UGf/UL] $f ZKHUH WKH WZR WHUPV GHVLJQDWHG E\ XQGHUEUDFHV UHVXOW IURP EHLQJ DGGHG WR FRPSOHWH WKH 8 DQG 89 WHUPV 7KH ULJKW KDQG VLGH DIWHU LQWHJUDWLQJ RYHU WKH GHSWK DVVXPLQJ D FRQVWDQW GHQVLW\ DQG DVVXPLQJ DQ LQYLVFLG IOXLG VXFK WKDW QR KRUL]RQWDO YLVFRXV VWUHVVHV H[LVW LH UW[ DQG 7\[ DUH ]HUR LV 5+6 LU SLU Ds S KR G[ S fÂ§K G] $f 8VLQJ WKH /HLEQLW] 5XOH DQG DYHUDJLQJ RYHU D ZDYH SHULRG WKH ULJKW KDQG VLGH EHFRPHV G U GQ GKQ 5+6 fÂ§fÂ§ SG]SSKRfÂ§7[[ 7LLLU $f SG[MK S G[ S G[ S S Y n
PAGE 150
,OO 7=;T DQG U0;BNR DUH WKH WLPH DYHUDJHG VXUIDFH DQG ERWWRP VKHDU VWUHVVHV $VVXPLQJ WKDW WKH IUHH VXUIDFH SUHVVXUH SQ LV ]HUR WKH G\QDPLF IUHH VXUIDFH ERXQGDU\ FRQGLWLRQf WKH ULJKW KDQG VLGH LV BBB G GKa 5+6 fÂ§fÂ§ SG] SKfÂ§s 7Â7cL $f 3 G[MKF S G[ S S f 7KH PHDQ SUHVVXUH DW WKH ERWWRP FDQ EH H[SUHVVHG DV WKH VXP RI WKH ZDYH LQGXFHG RU G\QDPLF SUHVVXUH DQG WKH K\GURVWDWLF SUHVVXUH 3K 3G9QNR SJ>. W@f $f VR WKDW WKH SURGXFW 3KID FDQ EH PDQLSXODWHG WR UHDG GK GK B GIM m Â Â£V$n! "f mV $f 7HUPV IURP ERWK WKH ULJKW DQG OHIW KDQG VLGHV DUH FRPELQHG DV WKH UDGLDWLRQ VWUHVVHV 'HILQLQJ Vf rnf L] }r mf SLL]< aU! L 6Y6KAL] r G] @BK3LG] Vf B Âf c0. WI Af^O r A $f $f $ f DQG PDNLQJ WKH DVVXPSWLRQV WKDW WKH SURGXFW RI WKH G\QDPLF SUHVVXUH DW WKH ERWWRP DQG WKH ERWWRP VORSH LV QHJOLJLEOH WKDW WKH McI?>XnfG] WHUP FDQ EH LJQRUHG DQG WKDW WKH ODWHUDO IULFWLRQ WM FDXVHG E\ PRPHQWXP IOX[HV GXH WR WXUEXOHQW IOXFWXDWLRQV LV LQGHSHQGHQW RI GHSWK DQG WKDW UM LV GHILQHG E\ WM fÂ§SXnYn WKH [ GLUHFWLRQ HTXDWLRQ RI PRPHQWXP LV ZULWWHQ DV }UO A G UQU/ OG6[[ G6 ;9 A m DU8 nrf r} 7\89A [ Â‘ $f f_smr/,Uf LQ R 2; S G\ S S $f %\ VXEWUDFWLQJ WKH GHSWK LQWHJUDWHG FRQWLQXLW\ HTXDWLRQ DQG GLYLGLQJ E\ WKH WRWDO GHSWK WKH ILQDO IRUP RI WKH [ PRPHQWXP HTXDWLRQ LV 8 G8B G8B GW [ G\ S^K IMf G67 G6 [Y G[ G\ $f
PAGE 151
GIM GUL == = fÂ§U Q[ R $f G[ SG\ S^K S^K Ucf 7KH VDPH SURFHGXUH JLYHV DV WKH GHSWK LQWHJUDWHG WLPH DYHUDJHG \ PRPHQWXP HTXDWLRQ DY DY fDY L DV K 8 K 9 fÂ§ GW G[ G\ SK IMf G[ [\ DV YY G\ GUc GUL G\U SG[ SK Icf7nY S. Yf ?U}Y $f $f
PAGE 152
$33(1',; % '(5,9$7,21 2) 7+( 5$',$7,21 675(66 7(506 7KH EDVLV RI WKH GHULYDWLRQ LV WR XVH WKH GHILQLWLRQ RI WKH YHORFLW\ SRWHQWLDO WR VXEVWLWXWH IRU WKH ZDYH LQGXFHG YHORFLWLHV LQ WKH GHILQLWLRQ IRU WKH UDGLDWLRQ VWUHVVHV GI! B GI! 8ILaG= Z ]S G] ZKHUH I! LV H[SUHVVHG LQ WHUPV RI WKH JHQHUDO FRPSOH[ DPSOLWXGH IXQFWLRQ % DV W! B BLO fFRVK NK ]f BLXW 9 D % FRVK NK FFf ZKHUH FF UHSUHVHQWV WKH FRPSOH[ FRQMXJDWH WHUP $OVR Ucn LV H[SUHVVHG DV Un L%HLXW FFf ZKHUH WKH SULPH LV XVHG WR GHQRWH ZDYH LQGXFHG TXDQWLWLHV )RU FRQYHQLHQFH (T f RI 6HFWLRQ LV UHSURGXFHG EHORZ 6DS S/K XnD8nSG] D >BK "G] Kr Af \ IDnf $Q H[SUHVVLRQ IRU S LV REWDLQHG LQ 6HFWLRQ DQG LV UHSURGXFHG EHORZ 2O/: G] SZn\ 6XEVWLWXWLQJ (T %f LQWR (T %f JLYHV 6DS 3 XDXnSG] DS f ff O R %Of %f %f %f %f UL f Fr Ur GXnZn ? A SJUf ]fG] A ,S fÂ§/fÂ§G]f G] n n B r n r r n S>Zn< G] \2. \ IDnf m2 / B[ 3 %f
PAGE 153
7KH WHUPV RI (T %f DUH QRZ HYDOXDWHG WHUP E\ WHUP 7HUP B LJ G% FRVK N>K ]f BLXW LJ G%n FRVK NK ]f LXfW XnDXnS fÂ§ B 8f ? D G[D FRVK NK J FRVK NK [f D G[' FRVK NK %f % G%A G%AG%B B G%BG%BHLXLW B G%n G%r D FRVK NK G[D G[S G[D GLS G[D GLS G[D GLS fÂ§UfÂ§U SJ FRVK NK ]f S8r8H $VLQK NK SJ FRVK NK ]f IFVLQK NK G% G%n G%n G% G[D GLS G[D GLS Â‘ G%BG@U; GLS f %f %f Â SXnDXnSG] r fÂ‘ R SJ f G% N VLQK NK AG[4 G[S f SJ r G% LIFÂQK NK ?G[D G[S f? \G[D G[S N ? VL FRVK NK ]f G] f KII A VLQK NK IMf N>K IMff NK ? VLQK NK 7KH VHFRQG WHUP BK SJ^Y ]fG] SJ _nU?] KA a 33 7KH WKLUG WHUP I ,I rII G% ZRXLYALR aU rÂ‘ BOXf n\YsV 7 r\ ZW I f A [I f r r & f FRVK NK ]f BLXfW LJ G%n FRVK NK ]f FRVK NK D G[c FRVK NK LJN SVLQK NK ]f BOXOW A LIIr VLQK N>K ]f M FRVK NK J VLQK $K ]f f }}n / A VLQKIFK G[I n [I Z L B LJN VLQK N^K ]f BLXW LJN VLQK N>K ]f :$ 9 R FRVK NK D FRVK NK G% G%n f fÂ§% fÂ§% ,/n6: %,2f %OOf %f %f %f G[c B BGB G[ G] G% G[I G%n I %A%M B "B/ eeE fÂ§V IfÂ§ G[I [I G[I >fIF VLQK IFLf IMf VLQK NK UfB UI! VLQK NK ]f VLQK NK FRVK N>K IMf FRVKIFK ]f %f
PAGE 154
BSJ G ?G% % G%rAM Ur FRVK N>K Ucf FRVK NKf ]f G[c >G[I G[ Kf ?IFVLQK$K Ic 3, G[ 3 G BBQ W [ } Ucf N VLQK N^K IMff fÂ§ G% %r G%r E? &E0A Af / %LQKN^K IMf ? G[ G[ G[ ? VLQK NK IMfn A IF VLQK N^K IMf U nNKFRHKNK VLQKIFK G% G%r Â‘W >G[ G[ %? IF VLQK IFK $ %?QKNKf G% G%r G[ct%fGA% ([SDQGLQJ WKH GHULYDWLYH WHUP LQ (T %f G[ G% G%r f fÂ§%r fÂ§% G[ G[ G% G%r G% G%r U%r % G[A G[A G[ G[ I 5HFRJQL]LQJ WKDW I LV D IUHH LQGH[ VR WKDW 9r G% G[Â 6 9Â% IF% G%r G[" 9M>%r fÂ§N%r %f %f ZKLFK UHVXOWV IURP XVH RI WKH +HOPKROW] HTXDWLRQ ZKLFK JRYHUQV GLIIUDFWLRQ LQ ZDWHU RI FRQVWDQW GHSWK DQG AiU _Yf%_ %f VR WKDW WKH H[SDQGHG GHULYDWLYH WHUP %f EHFRPHV G%f G%r G%G%r G[ % G[ % G[I G[I _9r%_ IF_%_ %f DQG WKH UHVXOW LV O AaÂ=f 7 _9f6_ ) @ %f 7KH DXWKRU ZLVKHV WR H[SUHVV VRPH UHVHUYDWLRQV FRQFHUQLQJ WKH YDOLGLW\ RI (T %f 7KH VKDOORZ ZDWHU DV\PSWRWH IRU WKLV H[SUHVVLRQ LV ]HUR VLQFH IRU VPDOO NKIMf FRWK WK IMf NKILf L IFKAfFRWKIF}AfO ,Q GHHSZDWHU KRZHYHU FRWK$KAf f VR WKDW NK IMf FRWKÂL IMf NK ,I WKH GHULYDWLYH WHUPV DUH QRQ]HUR WKH 6]]
PAGE 155
WHUP ZLOO EH SURSRUWLRQDO WR WKH GHSWK 7KLV ZRXOG PHDQ WKDW GHHS ZDWHU QRQVKRDOLQJ DQG QRQGLVVLSDWLYH ZDYHV LQ YDU\LQJ GHSWKV ZRXOG SURGXFH JUDGLHQWV LQ WKH UDGLDWLRQ VWUHVVHV )RU SODQH ZDYHV WKLV LV QRW D SUREOHP VLQFH WKH FRPSOH[ DPSOLWXGH WHUPV DUH HDVLO\ VKRZQ WR EH ]HUR 7KLV LV VHHQ E\ VXEVWLWXWLRQ RI % DHN[ DQG %r DHaWN[ ZKHUH D _%_ 7KH IRXUWK WHUP LV VWUDLJKW IRUZDUG LJN SVLQK NK ]f BLXW LJN VLQK NK ]f f FRVK NK fÂ§%n 7 FRVK NK B }nJIF 9 R Â‘r! ]f B %A B %HA D FRVK NK n 7 B JN VLQK NK ]f Q JN VLQK N^K ]f @ a D FRVK N>K UMf a VLQKNK %% %f %f %f SZnfG] SJN?%? Â r fIWS VLQK NK ]f VLQKÂK a 9r VLQKIFL! 7KH ILIWK WHUP QHHGV QR HYDOXDWLRQ DQG LV VHHQ WR FDQFHO H[DFWO\ WKH VHFRQG WHUP 7KH VL[WK DQG ODVW WHUP RI (T %f LV Ucnf DQG LV HYDOXDWHG HDVLO\ Ucn %9ff UMnf %H: %%r %rHZWf 6, : a%%n ZKLFK FDQFHOV SDUW RI WKH UHVXOW IURP WKH IRXUWK WHUP $GGLQJ WKH DERYH UHVXOWV IRU DOO WKH WHUPV RI (T %f \LHOGV 3 %f %f %f %f F 3Âa%G%r? A NK ? ?LD GLS \ $ 9 VLQKNKf ADIf _fÂ§NK VLQK NK G[D GLS _9IFe_ N?%?fANKFRWK NK f@ %f
PAGE 156
$33(1',; & /$*5$1*,$1 '(5,9$7,21 2) 7+( *29(51,1* :$9( (48$7,21 7KH JRYHUQLQJ HTXDWLRQ IRU ZDYH PRWLRQ FDQ EH GHULYHG WKURXJK XVH RI D YDULDWLRQDO SULQFLSOH ZKHUH D /DJUDJLDQ LV WKH IXQFWLRQ WKDW LV YDULHG /XNH f VKRZHG WKDW D /DJUDQJLDQ IRU ZDYHV LV JLYHQ E\ / fÂ§ K r= G] &Of .LUE\ f XVHG D WLPH DYHUDJHG /DJUDQJLDQ WR GHULYH D ZDYH HTXDWLRQ )RU WKH IROORZLQJ GHULYDWLRQ HTXDWLRQ &Of ZLOO EH XVHG 7KH WKLUG WHUP LQ (T &Of LV GHWHUPLQHG VWUDLJKW IRUZDUG ZKLOH WKH ILUVW WZR WHUPV QHHG WR EH H[SDQGHG DERXW IM WKH PHDQ ZDWHU OLQH LQ D 7D\ORU VHULHV !W :fD G] >Uc Uff!W _U IO IMf &f :H DVVXPH WKDW WM DQG I! FDQ EH H[SUHVVHG DV rf 9 r9L I! W!R HIW!L &f ZKHUH F LV DQ RUGHULQJ SDUDPHWHU DQG LV WKH GHSWK GHSHQGHQF\ DV GHILQHG LQ (T f )RU WKH VDNH RI VLPSOLFLW\ ZH ZLOO DVVXPH WKDW WKH JUDGLHQW RI IR UHSUHVHQWV WKH VWHDG\ KRUL]RQWDO FXUUHQWV 7KLV PHDQV WKDW IR LV QRW D IXQFWLRQ RI WLPH DQG WKDW : ZLOO QRW EH LQFOXGHG DV LW ZDV LQ WKH GHULYDWLRQ LQ &KDSWHU 7KLV LV EHFDXVH WKH : WHUP LV QRW H[SDQGHG DERXW ] IM DV LW ZDV LQ WKH *UHHQfV IXQFWLRQ GHULYDWLRQ DQG LV QRW QHHGHG LQ RUGHU WR REWDLQ WKH 9IF f 8fMe WHUP ZKLFK DV ZLOO EH VHHQ LV REWDLQHG PRUH GLUHFWO\ LQ WKH SUHVHQW GHULYDWLRQ ,QFOXGLQJ : KHUH ZLOO RQO\ SURGXFH WHUPV WKDW DUH RI KLJKHU RUGHU LQ WKH PHDQ FXUUHQW
PAGE 157
,QWURGXFLQJ &f LQWR &f \LHOGV ÂJYL JK / A IG]AK Wcf H9.I!R9KW!L > I G] H9KIRW!L ^ 9G] > I0G]? K 8Kf .F f \Kf IK >9f@G]` FUML ^ fÂ§ H9NI!RW!LI _m\ Wn 7KH /DJUDQJLDQ / LV DVVXPHG WR EH RI WKH IRUP &f ,W WKHQ IROORZV WKDW / fÂ§ /T I/L H=m I" / ÂR mGL8 eeerf DQG &f &f Â JI!P fÂ§ÂL fÂ§YAR9A ^ U Yr G] U I]G]f P 9ARf Ff IO 8IFR frf I >I9KI r@ 9LAKW!RW!LI[ _Ff r fIWR 7R REWDLQ WKH GHVLUHG OLQHDU JRYHUQLQJ HTXDWLRQ LV YDULHG ZLWK UHVSHFW WR I!L DQG WKHQ LV YDULHG ZLWK UHVSHFW WR Uc? LQ RUGHU WR HOLPLQDWH Uc? 7KH YDULDWLRQDO SULQFLSOH KROGV WKDW WKH /DJUDQJLDQ LV FRQVWDQW XQGHU VPDOO YDULDWLRQV RI WKH XQNQRZQ SDUDPHWHUV ,Q PDWKHPDWLFDO WHUPV DQG '/ '! '/ 'Uf &f FLRf 6LQFH WKH RUGHULQJ SDUDPHWHU f LV DUELWUDU\ WKH DERYH KROGV IRU DOO /f } f f f ,Q WKH Â WHUP I!L LV H[SUHVVHG LQ WHUPV RI LWV GHULYDWLYHV 7KLV UHTXLUHV LQWHJUDWLRQ E\ SDUWV WR
PAGE 158
REWDLQ WKH YDULDWLRQ ZLWK UHVSHFW WR I!? 8SRQ FDUU\LQJ WKLV RXW LW LV IRXQG WKDW 3s G/LB G G8 B G8 [ 'K GID GWnGWIXf KOG9A[f$SSO\LQJ &OOf WR &f DQG UHFRJQL]LQJ WKDW WKH YDULDWLRQ GI!? LV DUELWUDU\ \LHOGV DIWHU FDUU\LQJ RXW WKH LQWHJUDWLRQV LQ Â && mB ,UA&7 YrfÂ§0 nf QX 9IF9r0Lf &f ZKLFK LV UHZULWWHQ DV aMAa AK^&&J9KI!Lf I!LHU N&&Wf &f 9DU\LQJ /L ZLWK UHVSHFW WR LML \LHOGV r W! KI!L &f ZKLFK FDQ EH ZULWWHQ DV r 9L SUU J 'W &f ZKLFK XSRQ VXEVWLWXWLRQ LQWR &f \LHOGV WKH JRYHUQLQJ HTXDWLRQ AU 9A&&S9A[f D rr&&Ifr ZKLFK XSRQ PDNLQJ WKH VXEVWLWXWLRQ WKDW 9AR EHFRPHV Yr r fAU a 9DFF9IFrf U IFfFFfA R FLf 'W &f
PAGE 159
$33(1',; 2%7$,1,1* (48$7,21 f ,Q WKLV $SSHQGL[ ZH VKRZ WKH DOJHEUD LQYROYHG LQ REWDLQLQJ HTXDWLRQ f IURP HTXDn WLRQ f XVLQJ WKH DVVXPSWLRQ WKDW WKH GLVSHUVLRQ UHODWLRQVKLS LV ZULWWHQ DV X fÂ§ D N FRV 28 :H VWDUW E\ XVLQJ WKH IROORZLQJ IRUP RI c! ^J L,\f &rn r &2 L G]f DQG WKHQ PDNH D SKDVH VKLIW E\ XVLQJ WKH VXEVWLWXWLRQ FR}I Â[LFR} G]f VR WKDW LQ WKH ILQDO IRUP B LJ $n^"n \f H} Â &2 Â G]f 'Of 'f 'f 'f DQG DOO \ GHSHQGHQF\ LV VROHO\ LQ WKH FRPSOH[ DPSOLWXGH $n )RU UHIHUHQFH HTXDWLRQ f LV UHSURGXFHG EHORZ 3!]f] SW!YfY NSc! >Z D LnZ9IF f 8f@ LDZ! 8!]f]^Y!YfY^XY![fY^XYLYf] LMMYK! R Gf :ULWLQJ RXW DOO WKH FRPSRQHQWV RI HTXDWLRQ 'f DQG UHDUUDQJLQJ WHUPV JLYHV ^S Xf[ ^S Xf![[ >S 9f!Y@Y NS! RM Rfa! RM8[! LX9Y! LFUZM! 89Â[f\ >89L!Yf[ LX8L![ LX9A!\ 'f 7KH GHULYDWLYHV RI I! GHILQHG E\ HTXDWLRQ DUH DV IROORZV !] fÂ§LJ >Af LN FRV A HrINHRWG]f
PAGE 160
O!][ fÂ§ LIFFR A WEFRVfD fÂ§ N FRV fÂ§ HrA IF FRr rG[f D D ^![fY fÂ§ n, ^Df HrArFRLUf_ LN FRV LN FRV f\ AFrfFRLGrf,, + ,, ^!]fY 7KH DERYH IRUPV RI WKH GHULYDWLYHV UHFRJQL]H WKDW WKH SKDVH IXQFWLRQ Hf NFRWHG[f A GHULYDWLYHV VLQFH N N[ \f +RZHYHU WKH \ GHULYDWLYHV RI WKH SKDVH IXQFWLRQ LQWHJUDO DUH QRW HYDOXDWHG VLQFH WKH SKDVH VKLIW HTXDWLRQ 'f ZLOO HOLPLQDWH DQ\ \ GHSHQGDQFH IURP WKH SKDVH IXQFWLRQ DQG SODFH DOO \ GHSHQGDQFH LQ WKH FRPSOH[ DPSOLWXGH 7KH f ?r ;; WHUP LV QHJOHFWHG IROORZLQJ WKH DVVXPSWLRQ H[SODLQHG LQ VHFWLRQ 6XEVWLWXWLQJ WKH DERYH GHULYDWLYHV LQWR HTXDWLRQ 'f JLYHV U 1 3a8f; f HLILFRrWG[f LN FRV S fÂ§ 8f[fÂ§HnINFROG]f ?U [ R LNFRVS8f f HI LNFRVf[S 8fHANFRWG]n! N FRV >S 8f Â£ÂIN f}mrrf S \A IÂ£HLINHRrWG[ff ? R ,9 9} \ NSHLILFRWLG]f Z DfHfIWFRHWG[f LX8[HAIWFRWG[A 2 2 A Â‘ V LX9YM0 rff G]f LZ$HAI r FRf A ^89fY Hf r FRf G]f 89 G[f fÂ§LN FRV 8Y9fÂ§Hf N FRr G[A fÂ§LN FRV 89Y fÂ§rFRrG]A LNFRV fY89HAI NFRW GO! D D LNFR%89 LNFRV f\89HLNFRW G]!! I $ r1 >89f; AHrNFRLLOff LX!XAHrHRUI[f RMNFRH8AINFRWG]A WX9 'f
PAGE 161
'URSSLQJ WHUPV WKDW FRQWDLQ SURGXFWV RI GHULYDWLYHV RI M DQG GHULYDWLYHV RI WKH FXUUHQW DV ZHOO DV SURGXFWV RI WKH FXUUHQW DQG VHFRQG GHULYDWLYHV RI e LH WHUPV WKDW DUH GHVLJQDWHG E\ DQ RYHUEUDFHf HTXDWLRQ 'f LV UHDUUDQJHG DV ,} >IFFRVS fÂ§ 8f X(M fÂ§ } >IFFRVS fÂ§ (f XO@ A }:L >aIF FRV %S 8f NS X D XIFFRVOM A fÂ§ LN FRV fY89 fÂ§ >9\ Z N FRV 8f IF FRV 8f\9 N FRV (9@ A ` Hrn N &2L >3DHAIF&2Off >.ZLIFFRV(f@ Jf ,W LV HDVLO\ VKRZQ WKDW NFRV^S 8f MM8 R^&fFRV 8f %\ ZULWLQJ WKH GLVSHUVLRQ UHODWLRQVKLS DV X! N FRV 28 LW LV DOVR HDVLO\ VKRZQ WKDW RU D R!N FRV 8 fÂ§ N FRV %8\ )URP HTXDWLRQ LW LV DSSDUHQW WKDW IRU D FRQVWDQW R N FRV f\ D Y 7KXV HTXDWLRQ 'f EHFRPHV ^m >U& FRV 8f@[ e LU&W FRV% 8f A >IFSO FRV f@ LZ$ W9\$ HLINFRWG]n! L9D AHLINFDG]A _S ZKHUH D LNFRV fY89e WHUP KDV EHHQ GURSSHG $W WKLV SRLQW D SKDVH VKLIW LV LQWURGXFHG Y 'f ',2f 'f 'f 'f A B $nHrI LFR}I G[I N FRm G]f 'f
PAGE 162
ZKHUH N DQG DUH D UHIHUHQFH ZDYH QXPEHU DQG ZDYH DQJOH )RU FRPSXWDWLRQDO SXUSRVHV N ZLOO EH DQ DYHUDJHG ZDYH QXPEHU DFURVV D JULG URZ DQG ZLOO EH WDNHQ DV WKH ZDYH DQJOH LQ WKH DEVHQFH RI FXUUHQWV XVLQJ 6QHOOfV ODZ )XUWKHU LW ZLOO EH DVVXPHG VROHO\ IRU FRPSXWDWLRQDO SXUSRVHV WKDW LV HTXDO WR 7KLV JLYHV }>D&WFRV 8f@[ M WII&L FRV 8f MM &W FRV 8f FRV N Nf$ mZ$ L9Y$ }9 fÂ§f LDY9 fÂ§ 'f ?R Y R ([SDQGLQJ WKH GHULYDWLYHV RI e DQG ZLWK D IHZ PLQRU PDQLSXODWLRQV RI WKH WHUPV WKH JRYHUQLQJ HTXDWLRQ f LV REWDLQHG &M&RV &2$[ nÂ&I&RV &2FRVN Nf$ ? ?&}FRV% 8 ?9 $Y N&WO FRV f$ ,&&J ] Z $ 'f
PAGE 163
%,%/,2*5$3+< $UWKXU 56 f f5HIUDFWLRQ RI VKDOORZ ZDWHU ZDYHV WKH FRPELQHG HIIHFWV RI FXUUHQWV DQG XQGHUZDWHU WRSRJUDSK\f 7UDQV $PHU *HRSK\VLFDO 8QLRQ f %DWHPDQ + f 3DUWLDO 'LIIHUHQWLDO (TXDWLRQV RI 0DWKHPDWLFDO 3K\VLFV 'RYHU 3XEn OLFDWLRQV 1HZ
PAGE 164
'DOO\ :5 f f$ QXPHULFDO PRGHO IRU EHDFK SURILOH HYROXWLRQf 06 7KHVLV 8QLn YHUVLW\ RI 'HODZDUH 1HZDUN 'DOO\ :5 f f:DYH WUDQVIRUPDWLRQ LQ WKH VXUI URQHf 7HFKQLFDO 5HSRUW 1R 'HSW RI &RDVWDO DQG 2FHDQRJUDSKLF (QJLQHHULQJ 8QLYHUVLW\ RI )ORULGD *DLQHVYLOOH 'DOO\ :5 'HDQ 5* DQG 'DOU\PSOH 5$ f f $ PRGHO IRU EUHDNHU GHFD\ RQ EHDFKHVf 3URF rA &RDVWDO (QJ &RQI +RXVWRQ 'HDQ 5* DQG 'DOU\PSOH 5$ f :DWHU :DYH 0HFKDQLFV IRU 6FLHQWLVWV DQG (QJLQHHUV 3UHQWLFH+DOO (QJOHZRRG &OLIInV 1HZ HUVH\ (EHUVROH %$ DQG 'DOU\PSOH 5$ f f$ QXPHULFDO PRGHO IRU QHDUVKRUH FLUFXODWLRQ LQFOXGLQJ FRQYHFWLYH DFFHOHUDWLRQ DQG ODWHUDO PL[LQJf 2FHDQ (QJLQHHULQJ 5HSRUW 1R 'HSW RI &LYLO (QJ 8QLYHUVLW\ RI 'HODZDUH 1HZDUN 'HODZDUH (EHUVROH %$ DQG 'DOU\PSOH 5$ f f1XPHULFDO PRGHOLQJ RI QHDUVKRUH FLUFXODn WLRQf 3URF A &RDVWHG (QJ &RQI 6\GQH\ *DUUHWW &5 f f'LVFXVVLRQ WKH DGLDEDWLF LQYDULDQW IRU ZDYH SURSDJDWLRQ LQ D QRQXQLIRUP PRYLQJ PHGLXPf 3URF 5R\ 6RF $ *RXUOD\ 05 f f:DYH VHWXS DQG ZDYH JHQHUDWHG FXUUHQWV LQ WKH OHH RI D EUHDNn ZDWHU RU KHDGODQGf 3URF rA &RDVWDO (QJ &RQI &RSHQKDJHQ +DOHV /= f f(URVLRQ FRQWURO RI VFRXU GXULQJ FRQVWUXFWLRQ UHSRUW ([SHULPHQWDO PHDVXUHPHQWV RI UHIUDFWLRQ GLIIUDFWLRQ DQG FXUUHQW SDWWHUQV QHDU MHWWLHVf 86 $UP\ (QJLQHHU :DWHUZD\V ([SHULPHQW 6WDWLRQ 9LFNHVEXUJ 0LVVLVVLSSL +DQVHQ % DQG 6YHQGVHQ ,$ f f5HJXODU ZDYHV LQ VKRDOLQJ ZDWHU ([SHULPHQn WDO GDWDf 6HULHV 3DSHU ,QVW RI +\GURG\QDPLFV DQG +\GUDXOLF (QJLQHHULQJ 7HFKQLFDO 8QLYHUVLW\ RI 'HQPDUN /\QJE\ +DUULVRQ : DQG :LOVRQ :6 f f'HULYDWLRQ RI QXPHULFDO FDOFXODWLRQ RI ZDYH UHIUDFWLRQf 7HFKQLFDO PHPRUDQGXP 1R 86 $UP\ &RDVWDO (QJLQHHULQJ 5Hn VHDUFK &HQWHU )W %HOYRLU +RULNDZD DQG .XR &7 f f$ VWXG\ RI ZDYH WUDQVIRUPDWLRQ LQVLGH WKH VXUI ]RQHf 3URF A &RDVWDO (QJ &RQI 7RN\R RKQVRQ : f f7KH UHIUDFWLRQ RI VXUIDFH ZDYHV E\ FXUUHQWVf 7UDQV $P *HRSK\V 8QLRQ RKQVRQ : 2f%ULHQ 03 DQG ,VDDFV ' f f*UDSKLFDO FRQVWUXFWLRQ RI ZDYH UHIUDFWLRQ GLDJUDPVf 7HFK 5HSRUW 1R 86 +\GURJUDSKLF 2IILFH :DVKLQJWRQ '& RQVVRQ ,* f f%RRLMfV FXUUUHQWZDYH HTXDWLRQ DQG WKH UD\ DSSUR[LPDWLRQf 3URJ 5HSRUW SS ,QVW +\GURG\Q DQG +\GUDXOLF (QJ 7HFKQLFDO 8QLYHUVLW\ RI 'HQPDUN /\QJE\
PAGE 165
.HOOHU % f f6XUIDFH ZDYHV RQ ZDWHU RI QRQXQLIRUP GHSWKf )OXLG 0HFK .LUE\ 7 f f 3URSDJDWLRQ RI ZHDNO\QRQOLQHDU VXUIDFH ZDWHU ZDYHV LQ UHJLRQV ZLWK YDU\LQJ GHSWK DQG FXUUHQWf 215 7HFK 5HSRUW 5HVHDUFK 5HS &( 'HSW RI &LYLO (QJ 8QLYHUVLW\ RI 'HODZDUH 1HZDUN .LUE\ 7 f f$ QRWH RQ OLQHDU VXUIDFH ZDYH FXUUHQW LQWHUDFWLRQ RYHU VORZO\ YDU\LQJ WRSRJUDSK\f *HRSK\J 5HV &f .LUE\ 7 f f $ PRGHO IRU JUDGXDO UHIOHFWLRQ RI ZHDNO\ WZRGLPHQVLRQDO ZDYHV LQ D VORZO\ YDU\LQJ GRPDLQ ZLWK FXUUHQWV JRYHUQLQJ HTXDWLRQVf 7HFK 5HSRUW 1R 'HSW RI &RDVWHG DQG 2FHDQRJUDSKLF (QJLQHHULQJ 8QLYHUVLW\ RI )ORULGD *DLQHVYLOOH .LUE\ 7 DQG 'DOU\PSOH 5$ f f1XPHULFDO PRGHOLQJ RI WKH QHDUVKRUH UHJLRQf 215 7HFK 5HSRUW 5HVHDUFK 5HSRUW &( 'HSW RI &LYLO (QJ 8QLYHUVLW\ RI 'HODZDUH 1HZDUN .LUE\ 7 DQG 'DOU\PSOH 5$ f f0RGHOLQJ ZDYHV LQ VXUI]RQHV DQG DURXQG LVn ODQGVf :DWHUZD\ 3RUW &RDVWDO DQG 2FHDQ (QJLQHHULQJ /H 0KDXW % DQG .RK 5&< f f2Q WKH EUHDNLQJ RI ZDYHV DUULYLQJ DW DQ DQJOH WR WKH VKRUHf +\GUDXOLF 5HV f /LX 3/) DQG 0HL && f f(IIHFW RI D EUHDNZDWHU RQ QHDUVKRUH FXUUHQWV GXH WR EUHDNLQJ ZDYHVf 7HFKQLFDO 0HPRUDQGXP 1R &RDVWDO (QJLQHHULQJ 5HVHDUFK &HQWHU )RUW %HOYRLU 9D /RQJXHW+LJJLQV 06 Df f/RQJVKRUH FXUUHQWV JHQHUDWHG E\ REOLTXHO\ LQFLGHQW VHD ZDYHV ,f *HRSK\V 5HV f /RQJXHW+LJJLQV 06 Ef f/RQJVKRUH FXUUHQWV JHQHUDWHG E\ REOLTXHO\ LQFLGHQW VHD ZDYHV ,,f *HRSK\V 5HV f /RQJXHW+LJJLQV 06 DQG 6WHZDUW 5: f f5DGLDWLRQ VWUHVV DQG PDVV WUDQVSRUW LQ JUDYLW\ ZDYHV ZLWK DSSOLFDWLRQ WR fVXUI EHDWVff )OXLG 0HFK /RQJXHW+LJJLQV 06 DQG 6WHZDUW 5: f f$ QRWH RQ ZDYH VHWXSf 0DU 5HV f /RQJXHW+LJJLQV 06 DQG 6WHZDUW 5: f f5DGLDWLRQ VWUHVVHV LQ ZDWHU ZDYHV D SK\VLFDO GLVFXVVLRQ ZLWK DSSOLFDWLRQVf 'HHS6HD 5HVHDUFK /XNH & f f$ YDULDWLRQDO SULQFLSOH IRU D IOXLG ZLWK D IUHH VXUIDFHf )OXLG 0HFK f 0HL && f f$ QRWH RQ WKH DYHUDJHG PRPHQWXP EDODQFH LQ WZRGLPHQVLRQDO ZDWHU ZDYHVf 0DU 5HV f 0HL && f 7KH $SSOLHG '\QDPLFV RI 2FHDQ 6XUIDFH :DYHV :LOH\,QWHUVFLHQFH 1HZ
PAGE 166
0LFKH 5 f f0RXYHPHQW RQGXODWRLUHV GH OD PHU HQ SURIRQGHXU FRQVWDDQWWWH RX GFURLVVDQWHf $QQ GHV 3RQW% HW &KDXVVVV 0LOHV : f f6XUIDFH ZDYH VFDWWHULQJ PDWUL[ IRU D VKHOI )OXLG 0HFK 0LOHV : f f2Q VXUIDFHZDYH GLIIUDFWLRQ E\ D WUHQFKf )OXLG 0HFK 0XQN :+ DQG $UWKHU 56 f f:DYH LQWHQVLW\ DORQJ D UHIUDFWHG UD\f LQ *UDYLW\ :DYHV 1%6 &LUF 1HZPDQ 1 f f'LIIUDFWLRQ RI ZDWHU ZDYHVf $SSOLHG 0HFK 5HYLHZ f 1RGD ( 6RQX &5XSHUW 9& DQG &ROOLQV , f f1HDUVKRUH FLUFXODWLRQ XQGHU VHD EUHH]H FRQGLWLRQV DQG ZDYHFXUUHQW LQWHUDFWLRQV LQ WKH VXUI ]RQHf 7HWUD 7HFK 5HSRUW 7& 2f%ULHQ 03 f f:DYH UHIUDFWLRQ DW /RQJ %HDVFK DQG 6DQWD %DUEDUD &DOLIRUQLDf %HDFK (URVLRQ %RDUG %XOOHWLQ f 3HQQH\ :* DQG 3ULFH $7 f f'LIIUDFWLRQ RI VHD ZDYHV E\ EUHDNZDWHUVf 7HFKQLn FDO +LVWRU\ 5HSRUW 1R 'LUHFWRUDWH 0LVH :HDSRQV 'HYHORSPHQW :DVKLQJWRQ '& 3HQQH\ :* DQG 3ULFH $7 f f7KH GLIIUDFWLRQ WKHRU\ RI VHD ZDYHV DQG WKH VKHOWHU DIIRUGHG E\ EUHDNZDWHUVf 3KLO 7UDQV 5R\ 6RF $ 3HUOLQ 0 DQG 'HDQ 5* f f$ QXPHULFDO PRGHO WR VLPXODWH VHGLPHQW WUDQVSRUW LQ WKH YLFLQLW\ RI FRDVWDO VWUXFWXUHVf 0LVH 5HSRUW 1R 86 $UP\ &RDVWDO (QJLQHHULQJ 5HVHDUFK &HQWHU )W %HOYRLU 9LUJLQLD 3KLOOLSV 20 f 7KH '\QDPLFV RI WKH 8SSHU 2FHDQ &DPEULGJH 8QLYHUVLW\ 3UHVV 3LHUVRQ :f f 7KH LQWHUSUHWDWLRQ RI FURVVHG RUWKRJRQDOV LQ WKH ZDYH UHIUDFWLRQ SKHQRPHQDf 7HFK 0HP 1R %HDFK (URVLRQ %RDUG :DVKLQJWRQ '& 3XWQDP $ DQG $UWKXU 56 f f'LIIUDFWLRQ RI ZDWHU ZDYHV E\ EUHDNZDWHUVf 7UDQV $P *HRSK\V 8QLRQ 5DGGHU $& f f2Q WKH SDUDEROLF HTXDWLRQ IRU ZDWHU ZDYH SURSDJDWLRQf )OXLG 0HFK f 6HHOLJ :1 f f7ZRGLPHQVLRQDO WHVWV RI ZDYH WUDQVPLVVLRQ DQG UHIOHFWLRQ FKDUn DFWHULVWLFV RI ODERUDWRU\ EUHDNZDWHUVf 7HFKQLFDO 5HSRUW 1R 86 $UP\ &RDVWDO (QJLQHHULQJ 5HVHDUFK &HQWHU )W %HOYRLU 9LUJLQLD 6KHQJ <3 DQG %XWOHU +/ f f0RGHOLQJ RI FRDVWDO FXUUHQWV DQG VHGLPHQW WUDQVn SRUWf 3URF A &RDVWDO (QJ &RQI &DSH 7RZQ 6KRUH 3URWHFWLRQ 0DQXDO f 86 $UP\ &RDVWDO (QJLQHHULQJ 5HVHDUFK &HQWHU )W %HOYRLU 9LUJLQLD
PAGE 167
6PLWK f f2Q VXUIDFH ZDYHV FURVVLQJ D VWHS ZLWK KRUL]RQWDO VKHDUf )OXLG 0HFK 6YHQGVHQ ,$ f f:DYH KHLJKWV DQG VHWXS LQ D VXUI ]RQHf &RDVWDO (QJ f 7DSSHUW )' f f7KH SDUDEROLF DSSUR[LPDWLRQ PHWKRGf ,Q :DYH 3URSDJDWLRQ DQG 8QGHUZDWHU $FRXVWLFV HGLWHG E\ % .HOOHU DQG 6 3DSDGDNLVf &KDSWHU 6SULQJHU 9HUODJ %HUOLQ 9DQ 'RUQ : f f:LQG VWUHVV RQ DQ DUWLILFLDO SRQGf 0DU 5HV f 9HPXODNRQGD 65 f f(URVLRQ FRQWURO RI VFRXU GXULQJ FRQVWUXFWLRQ UHSRUW &XUn UHQW D ZDYHLQGXFHG FXUUHQW PRGHOf 86 $UP\ (QJLQHHU :DWHUZD\V ([SHULn PHQW 6WDWLRQ 9LFNHVEXUJ 0LVVLVVLSSL 9HPXODNRQGD 65 6ZDLQ $ +RXVWRQ 5 )DUUDU 3' &KRX /: (EHUVROH %$ f f&RDVWDO DQG LQOHW SURFHVVHV QXPHULFDO PRGHOLQJ V\VWHP IRU 2UHJRQ ,Qn OHW 1RUWK &DUROLQDf 7HFK 5HSRUW &(5& 86 $UP\ (QJLQHHU :DWHUZD\V ([SHULPHQW 6WDWLRQ 9LFNHVEXUJ 0LVVLVVLSSL 9HQJD\LO 3 DQG .LUE\ 7 f f6KRDOLQJ DQG UHIOHFWLRQ RI QRQOLQHDU VKDOORZ ZDWHU ZDYHVf 7HFK 5HSRUW 1R &RDVWDO DQG 2FHDQRJUDSKLF (QJLQHHULQJ 'HSDUWn PHQW 8QLYHUVLW\ RI )ORULGD *DLQHVYLOOH 9LVVHU 3f f7KH SURSHU ORQJVKRUH FXUUHQW LQ D ZDYH EDVLQf 5HSRUW 1R 'HSW RI &LYLO (QJLQHHULQJ 'HOIW 8QLYHUVLW\ 'HOIW :KLWKDP *% f /LQHDU DQG 1RQOLQHDU :DYHV :LOH\ 1HZ
PAGE 168
%,2*5$3+,&$/ 6.(7&+ +DUOH\ :LQHU ZDV ERUQ LQ 5RDQRNH 9LUJLQLD RQ 0D\ DQG UDLVHG LQ &XPEHUODQG 0DU\ODQG 7KHUH KH DWWHQGHG SXEOLF VFKRROV URDPHG WKH PRXQWDLQV LQ KLV EDFN \DUG DQG H[FHOOHG LQ GRZQKLOO VNLLQJ $IWHU KLJK VFKRRO KH DWWHQGHG 'DUWPRXWK &ROOHJH ZKHUH KLV ILUVW DWWHPSW DW REWDLQLQJ D FROOHJH GHJUHH HQGHG D IHZ \HDUV VKRUW RI FRPSOHWLRQ 7KLV ZDV IROORZHG E\ VHYHUDO \HDUV RI EHLQJ fRQ WKH URDGf LQ WKH %RKHPLDQ WUDGLWLRQ RI DFN .HURXDF DORQJ ZLWK PDQ\ \HDUV RI SURWHVWLQJ WKH 9LHWQDP ZDU ,Q WKH DXWKRU GLG KLV SDUW WR HQG WKH ZDU E\ EHLQJ LQFDUFHUDWHG LQ WKH )HGHUDO &RUUHFWLRQDO ,QVWLWXWLRQ DW 'DQEXU\ &RQQHFWLFXW IRU UHVLVWLQJ WKH GUDIW 7KH DXWKRUf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f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
PAGE 169
, FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ LJ &KDLUPDQ c 3URIHVVRU RI $HURVSDFH (QJLQHHULQJ 0HFKDQLFV DQG (QJLQHHULQJ 6FLHQFH FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ I A DPHV 7 .LUE\ &RFKDLUUKÂ£Q $VVRFLDWH 3URIHVVRU RI &RDVWDO DQG 2FHDQRJUDSKLF (QJLQHHULQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DULG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH P VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ WIn9" %HQW $ &KULVWHQVHQ 3URIHVVRU RI &LYLO (QJLQHHULQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWR RI 3KLORVRSK\ 5REHUW 'HDQ *UDGXDWH 5HVHDUFK 3URIHVVRU RI &RDVWDO DQG 2FHDQRJUDSKLF (QJLQHHULQJ
PAGE 170
, FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV FI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 3URIHVVRU RI $HURVSDFH (QJLQHHULQJ 0HFKDQLFV DQG (QJLQHHULQJ 6FLHQFH 7KLV GLVVHUWDWLRQ ZDV VXEPLWWHG WR WKH *UDGXDWH )DFXOW\ RI WKH &ROOHJH RI (QJLQHHULQJ DQG WR WKH *UDGXDWH 6FKRRO DQG ZDV DFFHSWHG DV SDUWLDO IXOILOOPHQW RI WKH UHTXLUHPHQWV IRU WKH GHJUHH FI 'RFWRU RI 3KLORVRSK\ $XJXVW 6 'HDQ *UDGXDWH 6FKRRO
PAGE 171
180(5,&$/ 02'(/,1* 2) :$9(,1'8&(' &855(176 86,1* $ 3$5$%2/,& :$9( (48$7,21 %\ +$5/(< 67$1)25' :,1(5 $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( 6&+22/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$ ^81,9(56,7< 2) )/25,'$ /,%5$5,(6
xml version 1.0 encoding UTF8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchemainstance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ELJXI714V_8LM0Q3 INGEST_TIME 20111028T20:17:05Z PACKAGE AA00004971_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
NUMERICAL MODELING OF WAVEINDUCED CURRENTS USING A PARABOLIC
WAVE EQUATION
By
HARLEY STANFORD WINER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988
{UNIVERSITY OF FLORIDA LIBRARIES
NUMERICAL MODELING OF WAVEINDUCED CURRENTS USING A PARABOLIC
WAVE EQUATION
By
HARLEY STANFORD WINER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988
[UNIVERSITY OF FLORIDA LIBRARIES
ACKNOWLEDGEMENTS
The writing of a dissertation is a long process involving interaction with and support
from many people. The author would like to thank all those who in one way or another
offered encouragement, support, advice and helped in the problem solving process. The
contributions of those who are not mentioned by name are in no way diminished by the lack
of individual acknowledgement.
First, the author wishes to express deep appreciation to the members of his committee:
the chairman, Dr. Hsiang Wang, who enticed him to come to Florida and enroll in the
doctoral program, and the cochairman, Dr. James Kirby, who shared his knowledge of wave
equations and numerical modeling, along with Dr. Robert Dean, Dr. Bent Christensen and
Dr. Joseph Hammack. Each of these men has been both a teacher and a friend. The author
wishes to also thank Dr. Peter Sheng who at one time served on the supervisory committee
but due to a prior commitment was unable to be at the defense and officially be on the
committee.
The author would like to thank the staff of the Coastal and Oceanographic Engineering
Laboratory at the University of Florida and the staff of the department of Coastal and
Oceanographic Engineering who helped and aided the author in innumerable ways. Thanks
are also extended to Ms. Lillean Pieter who helped prepare the figures and Dr. Claudio
Neves, a fellow student, who manipulated the I^TgX software bo as to produce a document
in thesis mode.
A special acknowledgement is given to the authorâ€™s parents who encouraged and pushed
him to complete this doctorate. The author would also like to thank his teachers, friends
and fellow students who offered encouragement and interesting discussions.
Lastly the author thanks his soulmate and partner, Esther DeJong, who managed to
be patient with each extension of the final date. Baby cakes, itâ€™s PAinally Done. I love you.
m
TABLE OF CONTENTS
ACKNOWLEDGEMENTS Ãœ
LIST OF FIGURES vi
ABSTRACT ix
CHAPTERS
1 INTRODUCTION 1
1.1 Literature 4
1.2 General Description of the Model 8
2 CIRCULATION MODEL 16
2.1 Governing Equations 16
2.2 Radiation Stresses 24
3 WAVE MODEL 31
3.1 Governing Wave Equation 31
3.2 A Parabolic Approximation 36
3.3 Energy Dissipation Due to Wave Breaking 45
4 FINITE DIFFERENCE SCHEMES AND BOUNDARY CONDITIONS 49
4.1 Numerical Solution of the Governing Equations 49
4.2 Finite Differencing of the Parabolic Wave Equation 52
4.3 Boundary Conditions in the Circulation Model 56
4.4 Boundary Conditions in the Wave Model 59
5 RESULTS 63
5.1 Wave Setup 64
5.2 Longshore Currents 68
iv
70
5.3 Waves On a RipChannel . . .
5.4 ShorePerpendicular Breakwater 76
5.5 Shore Parallel Emerged Breakwater 88
5.6 Comparison with Experimental Results of Gourlay 92
6 CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER STUDY ... 103
6.1 Improved Wave Formulations 104
6.2 Including Wave Reflection 105
6.3 Extending the Model to Include Sediment Transport 105
APPENDICES
A DERIVATION OF THE DEPTH AVERAGED EQUATIONS 107
B DERIVATION OF THE RADIATION STRESS TERMS 113
C LAG RANG LAN DERIVATION OF THE GOVERNING WAVE EQUATION . . 117
D OBTAINING EQUATION (3.79) 120
BIBLIOGRAPHY 124
BIOGRAPHICAL SKETCH 129
v
LIST OF FIGURES
1.1 Computational domain of the model 8
1.2 Flow chart of computer program 11
1.3 Flow chart of circulation model 13
1.4 Flow chart of wave model 14
2.1 Longshore current profiles for different friction formulations for an 11
second, 1.028 meter wave at 17.28 degrees to a 1:25 plane beach 22
2.2 Experimental results for wave setup and setdown. Reprinted from
Bowen, Inman, and Simmons (1968), Journal of Geophysical Research,
vol. 73(8) page 2573, copyright by the American Geophysical Union. . . 30
4.1 Definition sketch of grid nomenclature and coordinate system 56
4.2 Startup function according to Eq. 4.43 with Ci=30 and C2=3 60
4.3 Waves passing through a lateral boundary unaffected by the boundary. 61
5.1 Comparison of the numerical solution of Vemulakonda et al. (1985) for
setup with experimental data of Bowen (1968). Source: Vemulakonda
et al. (1985) 65
5.2 Comparison of numerical results from the model with experimental reÂ¬
sults of Hansen and Svendsen (1979). Wave setup and wave heights for
a 2 second wave on a beach slope of 1:34 and incident wave height of 7
centimeters 67
5.3 Nondimensional longshore velocities vresus the nondimensional disÂ¬
tance offshore from the analytical solution of LonguetHiggins. Reprinted
from LonguetHiggins (1970b), Journal of Geophysical Research, vol.
75(33) page 6793, copyright by the American Geophysical Union 68
5.4 Comparison of experimental results of Visser (1982) and numerical result
for longshore currents due to a 1.1 second wave on a slope of 1:20 for
the deep water conditions H0 = .065m and 0O = 20 degrees 69
5.5 Top: Depth contours, in meters, using Eq. 5.7 with the coordinates as
shown. Bottom: Depth contours as used in the present model. Grid
spacing is 5 meters 71
vi
5.6 Velocity vectors from the results of Ebersole and Dalrymple for a wave
angle of 30 degrees. Scale for velocity vectors is 1 inch = 2.87 m/sec.
Source: Ebersole and Dalrymple (1979) 72
5.7 Velocity vectors obtained from the present model using the same input
conditions as for Ebersole and Dalrymple 73
5.8 Mean water surface elevation contours showing that the water surface is
lower above the trough. Values indicated are in millimeters 74
5.9 Depth contours in meters for the case of a bar with a gap 75
5.10 Velocity vectors for the case of waves over a bar with a gap. Dashed
lines indicate depth contours. Grid spacing equal to 10 meters 75
5.11 Depth contours for the case of normally incident waves upon a shoal.
Grid spacing is .05 meters. Contour interval is .02 meters 77
5.12 Velocity vectors for the case of waves breaking upon a shoal 78
5.13 Maximum velocity across the top of a shoal versus incident wave height
for waves over a shoal on a plane beach. Incident wave height less than
9.1 centimeters are nonbreaking waves. Incident wave heights greater
than 9.1 centimeters produced waves breaking over the shoal 79
5.14 Position of the groin in relation to the grid rows and columns 80
5.15 Results for the numerical model of Liu and Mei (1975). Deepwater
wave height equal to 1 meter, deepwater wave angle equal to 45 degrees.
Source: Liu and Mei (1975) 83
5.16 Results from the present numerical model using the same conditions as
were used by Liu and Mei. Offshore boundary condition of constant waÂ¬
ter surface level which precludes tangential flow at the offshore boundary. 84
5.17 Results from the present numerical model using the same conditions as
were used by Liu and Mei. Offshore boundary condition of no onoffshore
flow. Grid spacing equal to 20 meters 85
5.18 Experimental setup for the tests conducted by Hales (1980). Source:
Hales (1980) 86
5.19 Experimental measurements of the middepth averaged velocities obtain
by Hales. Measurements at one foot intervals. Source: Hales (1980). . . 87
5.20 Ripcurrent velocities adjacent to the down wave side of the jetty. ComÂ¬
parison of experimental and numerical results 88
5.21 Velocity vectors of the currents obtained numerically for the same conÂ¬
ditions as for the Hales experiment. Grid spacing equal to one foot. . . 89
5.22 Definition sketch of grids for the case of a shore parallel breakwater. . . 90
vu
5.23 Wave crests in the lee of a semiinfinite breakwater in a region of constant
depth 91
5.24 Velocity vectors obtained from the model for a 100 meter long breakwater
100 meters offshore for a 1 meter, 6 second wave for increasing angles of
incidence 93
5.25 Stream function contour lines obtained from the Liu model for a 10
second 1 meter wave normally approaching a 700 meter long breakwater
350 meters offshore on a 1 on 50 slope. Source: Liu and Mei (1975). . . 94
5.26 Flow lines obtained from the present model for a 10 second 1 meter wave
normally approaching a 700 meter long breakwater 350 meters offshore
on a 1 on 50 slope 95
5.27 Velocity vectors obtained from the present model for a 10 second 1 meter
wave normally approaching a 700 meter long breakwater 350 meters
offshore on a 1 on 50 slope 95
5.28 Setup contour lines obtained from the present model for a 10 second
1 meter wave normally approaching a 700 meter long breakwater 350
meters offshore on a 1 on 50 slope 96
5.29 Physical layout for the experiment of Gourlay. Source: Gourlay (1974),
Proceedings 14*^ Coastal Engineering Conference, copyright by the AmerÂ¬
ican Society of Civil Engineers, reprinted with permission 97
5.30 Computational domain for the experiment of Gourlay. Grid spacing
equal to .1 meter 98
5.31 Comparison of experimental and numerical results for the experiment of
Gourlay: Wave Setup Contours. Top figure is from Gourlay (1974) ProÂ¬
ceedings 14^ Coastal Engineering Conference, copyright by the AmerÂ¬
ican Society of Civil Engineers, reprinted with permission. The bottom
figure is the computer results 100
5.32 Results from the experiment of Gourlay: Contours of current velociÂ¬
ties and streamlines. Source: Gourlay (1974), Proceedings 14^ Coastal
Engineering Conference, copyright by the American Society of Civil EnÂ¬
gineers, reprinted with permission 101
5.33 Results from the numerical model: Streamlines of the flow 101
5.34 Results from the numerical model: Contour lines of the current velociÂ¬
ties 102
viii
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
NUMERICAL MODELING OF WAVEINDUCED CURRENTS USING A PARABOLIC
WAVE EQUATION
By
HARLEY STANFORD WINER
August 1988
Chairman: Dr. Hsiang Wang
Major Department: Aerospace Engineering, Mechanics and Engineering Science
Prediction of waves and currents in the nearshore region in the presence of proposed
breakwaters is needed in the design process in order to optimize the placement of the
structures. A numerical model is developed which solves for the wave field, the depth
averaged mean currents and the mean water surface elevation within a given computaÂ¬
tional domain with the presence of breakwaters. The model iterates between a wave model
and a circulation model. The wave model uses a parabolic approximation to a combined
refractiondiffraction mild slope equation which includes currents. The circulation model
uses an alternating direction implicit method solution to the depth averaged equations of
momentum using the radiation stress terms obtained through solution of the wave model as
the forcing terms. Each of the terms in the depth averaged equations of momentum involve
assumptions that are discussed in the report. The model is constructed in modular form so
that improvements can easily be incorporated into the model.
The model was employed for several test cases such as rip channels, bar gaps, and
both shore perpendicular and shore parallel breakwaters. Good agreement was obtained
when the model was used to simulate a laboratory experiment of Michael R. Gourlay
(â€œWave setup and wave generated currents in the lee of a breakwater or headland,â€
Proc. 14*^ Coastal Eng. Conf., Copenhagen, 1974, pp. 19761995).
IX
Suggestions are made for improving the model and for including sediment transport in
X
the model.
CHAPTER 1
INTRODUCTION
As waves propagate they are transformed by variations in currents and depths, along
with interaction with obstructions. The major processes of wave transformation include
shoaling, reflection, refraction, diffraction and dissipation. Shoaling is the change of wave
energy, propagation speed and wave height due to changes in the water depth and/or current
field. Refraction is the change in the direction of wave propagation due to the spatial
gradient of the bottom contours and/or current field. Diffraction is the phenomenon of
wave energy being transmitted laterally along a wave crest. Dissipation is the process of
energy loss due to a multitude of possible causes such as bottom percolation, damping
due to viscous effects, turbulence, and wave breaking. As waves approach a shoreline and
break, momentum exchanges can produce currents and mean water level variations. These
currents can be in the form of longshore currents, rip currents, or circulation cells in the lee
of obstructions. They in turn modulate the incident wave field. The process can be quite
complicated depending upon the complexity of the bathymetry, the ambient current field,
and the presence of obstructions.
Coastal engineers and planners have many reasons for wanting to obtain accurate preÂ¬
dictions of waves and currents and water elevations in the coastal rone. For instance, preÂ¬
dictions of extreme conditions are needed for site planning and erosion mitigation studies.
Harbor authorities need wave predictions for structure design and for operational decisions.
This dissertation develops a numerical model for obtaining predictions of waves, currents
and mean surface elevations in a given domain with the offshore wave conditions and the
bathymetry as the input. The model uses a linear wave equation and depth integrated
equations of momentum and mass conservation. Comparison with experimental results
1
2
indicates that the model produces useful and valid predictions of the waves and currents in
the nearshore region in the presence of obstructions such as groins and breakwaters. The
predictions obtained through use of the computer model herein reported are not intended
to be the sole resource for the engineer. A good engineer or planner will seek answers from
multiple sources and cross check them, recognizing the limitations of each method used to
obtain the answer.
Predictions of waves and currents can be obtained for different cases analytically,
through experiments or by numerical solutions. Analytical solutions are obtained only
for simple cases involving idealized beaches using simplified assumptions which are not alÂ¬
ways satisfied in nature. LonguetHiggins and Stewart (1963) give an analytic solution for
wave setup for steady state conditions on a plane beach with straight and parallel conÂ¬
tours, assuming linear waves approaching the beach normally, and spilling breakers where
the wave height shoreward of the breaker line is a constant times the total depth of the
water column. LonguetHiggins (1970a) gives a cross shore profile of the longshore current
for a plane beach with straight and parallel contours, for linear waves approaching the beach
at an angle, spilling breakers, and linearized bottom friction, while neglecting the influence
wave setup has upon the total depth of the water column. In a companion paper, Longuet
Higgins (1970b) extended the solution to include lateral diffusion of the current momentum.
There are other analytical solutions for waves and currents but all are limited to very simple
cases.
The literature is replete with laboratory investigations of waves and currents. Physical
models have advantages and disadvantages. The advantages of experimentation are that
results can be obtained for complex situations where there is uncertainty about the exact
form of the governing equations or about the precise form of any of the component terms
of the equation. Any reasonable amount of detail can be included. Visualization of the
physical processes provides an opportunity to identify problems which otherwise might not
be evident.
3
However there are limitations to the use of laboratory experiments. There is the problem
of scale effects which means that all scaling requirements cannot be satisfied simultaneously;
thus, all features may not be interpreted properly to the prototype. For example, the
time scale dictated by Froude modelling laws in which gravitational forces dominate is
different from that obtained from Reynolds modelling laws in which viscous forces are most
important. Another disadvantage of physical models is the cost of labor to construct and
run large scale experiments. Also, because of the cost of construction, it is not always very
easy to change details in the model. Thus, it is not always possible to isolate one parameter
for study. For example, an investigation of the influence of the size of the gap between
segmented breakwaters might require many reconstructions of the rubble mound structure.
The large size of a model means that it is not easily stored once the initial experiment
is completed. The options are to either dismantle the model upon completion of testing
or to incur some maintenance costs. The significance of this is that the model cannot
always be easily retested if, for example, additional input information becomes available or
man or nature makes a substantial alteration of the prototype after the model has been
dismantled. Also the results of a model for one prototype are not always transferable to
other prototype situations. Often the available space for the physical model will require
the use of a distorted model in which the vertical scale is different from the horizontal
scale. This presents the additional problem that for a distorted model, when modelling
a situation where both refraction and diffraction are equally important, the laws of scale
modelling dictate different time scales for refraction and diffraction. For refraction the
time ratio is Tr â€” \fLy, where the subscript r is for the ratio of the value in the model
to the value in the prototype and T and Lv are time and vertical length respectively. For
diffraction Tr = ~n1" in shallow water and Tr = \/Ljjt in deep water, where Ljjr represents
y/Lvr
the horizontal length scale.
Numerical models do not have many of the disadvantages found with physical models.
Any mechanism for which the governing equations are known can be modeled numerically.
4
There are no limits in terms of scale effects. Numerical models, once developed, can be
adapted easily to many different prototype situations. There is no problem in the storage or
duplication of the model. Today computer facilities are far more available than laboratory
facilities. Computer models easily allow for sensitivity testing of the importance of an
individual variable and any specific feature of the model can be monitored.
Numerical models also have their disadvantages. In fact some of their advantages
are also their disadvantages. For example, being able to monitor any feature also means
that the computer will deliver only the information that is explicitly specified. Thus an
investigation of one aspect will not indicate problems with other aspects. Though any
mechanism for which the governing equation is known can be modeled, the model is only as
good as the assumptions involved in the equations. For example, the forces exerted on the
water column by the complex interaction between the fluid motion and the bottom are not
totally understood and are often greatly simplified. Another example is the use of depth
integrated twodimensional models of fluid flows when the reality is threedimensional. Both
of the above are incorporated in the present study. Knowing the correct equation does not
always make for successful computer modelling since the complexity of the equations may
make for difficulties (for example solving for the complete threedimensional flow field using
the NavierStokes equations). Computer time and cost for large programs can often be
significant.
1.1 Literature
This section is a brief history and literature review of the attempts of the coastal
engineering profession to obtain predictions of waves and currents in the nearshore region
given the offshore wave conditions.
The extension of Snellâ€™s law which governs optical wave refraction to the analogous
problem of water waves over straight and parallel contours has been known for some time.
Johnson, Oâ€™Brien, and Isaacs (1948) reported graphical methods for constructing wave
refraction diagrams. Two methods were reported, one producing wave fronts and the second
5
producing wave rays. Pierson (1951) showed that the method using wave fronts introduced
more error than the wave ray method. These methods proceeded upon the assumption that
the contours are locally straight. The wave rays obtained were important in that the wave
height was determined to vary inversely as the square root of the spacing between the wave
rays. Using these ray construction techniques Oâ€™Brien (1950) was able to demonstrate that
the April 1930 destruction of the Long Beach Harbor breakwater on a seemingly calm day
most probably resulted from the focusing of long wave energy from a South Pacific storm.
Munk and Arthur (1952) showed a method to obtain the wave height directly from the
curvature of the ray and the rate of change of the depth contours. The wave ray tracing
techniques did not include the effects of currents and had the disadvantage (from a numerical
modeling point of view) that wave heights and directions were obtained at positions along
a ray and not at prespecified grid positions. Noda et al. (1974) solved this problem by
devising a numerical scheme that solved a wave energy equation and a statement for the
irrotationality of the wave number to obtain wave height and direction at grid points. This
scheme also included the effects of current refraction but neglected diffraction.
Numerical models of wave induced currents were developed by Noda et al. (1974) in
which time dependent and advective acceleration terms were ignored. Birkemeier and DaiÂ¬
ry mple (1976) wrote a timemarching explicit model which neglected lateral mixing and
advective acceleration. Ebersole and Dalrymple (1979) added lateral mixing and the nonÂ¬
linear advective acceleration terms. Vemulakonda (1984) wrote an implicit model which
included lateral mixing and advective acceleration terms. All of the above used the numerÂ¬
ical scheme of Noda et al. (1974) to obtain wave heights and wave angles. Each of these
models neglected diffraction.
Diffraction of water waves represents some very difficult mathematical problems. HisÂ¬
torically most solutions were based upon the assumption of a flat bottom, since to quote
Newman (1972),
The restriction to constant depth h is important.... Problems involving wave
propagation in a fluid of variable depth are of great interest, and the resulting
6
diffraction effects are among the most important of all ... but these problems
are less tractable than those involving a fluid of constant depth, (page 2)
Penney and Price (1944) showed that Sommerfeldâ€™s solution of the optical diffraction probÂ¬
lem as described by Bateman (1944) is also a solution of the constant depth water wave
diffraction problem. Penney and Price (1952) solved the constant depth diffraction problem
for breakwaters assuming a semiinfinitely long, infinitesimally thin barrier. The solution
is in terms of Fresnel integrals which need to be evaluated numerically. The Shore ProtecÂ¬
tion Manual (SPM,1977) produced many graphical solutions based upon Penney and Price
(1952) that give the wave height and direction in the lee of breakwaters.
The problem of combined refraction and diffraction according to the SPM could be
handled by assuming a constant depth for a few wavelengths so as to obtain the diffraction
effects and then continue with the refraction effects. Perlin and Dean (1983) used essentially
the same procedure in a numerical model which calculated the longshore thrust in terms of
the breaking wave height and angle.
Liu and Mei (1975) solved for the wave field around shore parallel and shore perpendicuÂ¬
lar breakwaters by using a curvilinear coordinate system where one coordinate corresponded
to the wave ray and the other to a curve of constant phase. Their solution was for â€œan amÂ¬
plitude modulation factor D(Â£,f) which accounts for diffractionâ€ (page 72), where Â£ and f
are the two coordinates. An equation for .D(Â£,f) was obtained and a boundary value probÂ¬
lem was established by matching conditions in the incident zone and the shadow zones. For
constant depth the solution for Â£>(Â£,f) is in terms of Fresnel integrals. Upon obtaining the
wave field the radiations stresses were obtained and used to force a circulation model that
solved for the stream function and mean water surface level while ignoring the nonlinear
advective acceleration terms and the lateral diffusion of current momentum. Although Liu
and Mei produced results for the wave field, currents, and setup resulting from the diffracÂ¬
tion of waves by breakwaters, their study is not a wavecurrent interaction model since
there are no current terms in the wave equations and no feedback of current terms from the
circulation model.
7
Berkhoff (1972) derived an equation governing combined refraction and diffraction with
the assumption of a mild slope. Radder (1979) applied the splitting method of Tappert
(1977) to obtain a parabolic equation governing the propagation of the forward propagatÂ¬
ing component of the wave field. Booij (1981) extended the work of Berkhoff to include
the effects of currents and also obtained a parabolic approximation. Kirby (1984) made
corrections to the equation of Booij.
Parabolic wave equations have recently been used in waveinduced circulation models.
Liu and Mei (1975) solved for the wave field by matching conditions between the shadow
zones and the illuminated zone in the lee of a structure. The wave information was then
used to force a circulation model which formulated the governing depth integrated time
averaged momentum equations in terms of a stream function. However this model did not
have current feedback in the wave equation. Yan (1987) and the present study both include
a current feedback for the parabolic wave equation and solve directly for the mean currents.
Prior to the use of parabolic wave equations to obtain the wave forcing in circulation
models, refraction wave models using a numerical scheme derived by Noda et al. (1974)
were used. Birkemeier and Dairymple (1976) produced an explicit numerical solution of
the depth averaged equations of momentum and continuity using the wave model of Noda
et al. (1974). Their model did not include the nonlinear advective acceleration terms
or any mixing. Ebersole and Dalrymple (1980) included the advective acceleration terms
and lateral mixing in an explicit model which also used the refraction scheme of Noda
et al. Vemulakonda (1984) Produced an implicit model which also was based upon the
wave refraction scheme of Noda et al. Yan (1987) used a parabolic approximation to the
mild slope equation which accounts for combined refraction and diffraction. The present
model Improves upon Yanâ€™s model in numerical efficiency, by introducing structures and
most importantly by obtaining the radiation stress terms directly from the complex wave
amplitude.
8
Figure 1.1: Computational domain of the model.
1.2 General Description of the Model
This section is a description of the numerical model that has been constructed to calcuÂ¬
late the wave and current field within a given domain. This domain usually is bordered by
a shoreline along one boundary and an offshore region along the opposite boundary. The
other boundaries are sufficiently far from the region of interest (i.e. a structure or shoal,
etc.) so as to not affect the region of interest. This will be discussed further in the section
on boundary conditions. The computational domain for the purposes of this investigation
consists of a rectilinear coordinate system. Figure 1.1 shows the computational domain.
The model consists of two parts, a circulation model and a wave model. The circulation
model computes the mean water surface level and the depth averaged mean currents using
the radiation stresses from the waves as the forcing terms and includes bottom friction,
advective acceleration terms and the lateral diffusion of current momentum. The wave
model determines the radiation stress terms by first solving for the complex amplitude as a
9
function of the total depth and the depth averaged mean currents, and then computing the
radiation stresses from the complex amplitudes. The model starts with initial conditions
of zero mean currents and zero mean water surface elevation. The offshore incident wave
height is initially zero and then is smoothly increased up to its given value using a hyperbolic
tangent curve. This is done so as to minimize any transients due to a shock startup.
The model iterates between the wave model and the circulation model until steady state
convergence is obtained. Each of the two parts of the model will be described separately.
The circulation model uses an alternating direction implicit (ADI) method of solution.
Three equations are solved for three unknowns. The unknowns are fj the mean water surface
level, and U and V, the two components of the depth averaged mean current. The equations
are the continuity equation and the x and ydirection momentum equations. Each iteration
of the circulation model solves for the three unknowns on the next time level. In other words
each iteration advances the values of fj, U and V by a time increment, At. For the purposes
of this investigation the model solves for steady state conditions using the time marching
character of the circulation model. The model can be used to solve for time dependent
wave fields, but this would require the substitution of a time dependent wave equation in
the wave model. The ADI method first solves, grid row by grid row, for U at the next time
level and an intermediate value of fj, by solving the x direction equation of motion and the
continuity equation using an implicit numerical scheme. This solution is in terms of the
known values of 17,V,and fj at the present time level. Once all the grid rows are solved in
the x direction, the y equation of motion and the continuity equation are solved, again by
an implicit scheme, grid column by grid column, for V and fj at the next time level in terms
of U at the new time level, fj at the intermediate level and V at the present level. For each
of the implicit schemes in the two directions, the governing equations yield a tridiagonal
matrix. The tridiagonal matrix and the method of solution will be discussed later.
The wave model also yields a tridiagonal matrix equation. The governing equation is
a parabolic equation for the complex amplitude A. The complex amplitude contains phase
10
information meaning among other things that y variations of the wave field are contained
within the complex amplitude. The term â€œparabolic equationâ€ indicates that there are
first order derivatives in the x direction and second order derivatives in the y direction.
The method of solution is to solve for the amplitude on the 1+1 grid row in terms of the
known amplitude values on the I grid row. The solution scheme starts on the first grid row
using the prescribed offshore incident wave and then advances grid row by grid row to the
last grid row. To advance to the next grid row the amplitude values on the 1+1 grid row
are implicitly formulated using a CrankNicolson scheme. This will be further discussed in
section 4.2 which describes the finitedifferencing scheme for the parabolic wave equation.
Once the complex amplitudes are known the radiation stress terms are computed in terms
of the absolute value of the amplitude and the gradients of the amplitude. This is described
in section 2.2.
The circulation model and the wave model are the heart of the program; however there
is much more. In all there are some 27 subroutines in the program. These subroutines cover
everything from input and output to convergence checks and flooding of the shoreline. The
main program is compartmentalized into many subroutines so as to facilitate any changes
or future upgrading. For example, the bottom shear stress, a term of importance in the
equations of motion, currently uses a crude approximation. If a better approximation is
developed that is also computationally efficient, it will be an easy task to replace the present
subroutine that computes the bottom shear with the newer subroutine. There are many
places where the model can be improved with better assumptions and better approximaÂ¬
tions. These will be discussed in detail in the final chapter outlining recommendations for
further study. The modular construction of the model also allows for ease in substituting
one wave model for another.
Figure 1.2 shows a flow chart of the computer program. The subroutine INPUT is called
to initialize the values of the unknowns and to input the water depth, the offshore wave
height and direction and the position of any structures. The three subroutines WAVE,
11
Figure 1.2: Flow chart of computer program.
12
SHEAR, and LATMDC compute the radiation stresses, the bottom shear stress, and the
lateral mixing coefficients. These terms are the forcing terms in the equations of motion
and are used in the subroutine CIRC, which is the circulation model, to solve for fj, U and V.
There is a decision tree which directs the program to the three subroutines WAVE, SHEAR,
and LATMIX at selected intervals. This is purely arbitrary and is done in the interest of
computer speed. It is assumed that the forcing terms do not change at a great rate and can
thus be updated periodically. This updating is more frequent during the startup time while
the offshore incident wave height is being increased. The CHECK subroutine is to check the
model for convergence. Since steady state conditions are assumed, this subroutine indicates
convergence when the absolute value of the percentage difference between the updated values
of U,V, and fj and the previous values is less than an arbitrarily small specified value. The
UPDATE subroutine simply updates the values of the unknowns for the next iteration. The
FLOOD subroutine allows for the addition or elimination of grid rows at the shoreline to
allow for beach flooding due to setup.
Figure 1.3 is a flow chart of the circulation model. The subroutines XCOEF and
YCOEF determine the coefficients of the tridiagonal matrix equation for the general cases
in the x and y directions. The subroutine TRIDA solves the tridiagonal matrix equation
using the doubleelimination scheme described by Carnahan, Luther, and Wilkes (1969).
Subroutines XCOJ1 and XCOJN are adaptations of XCOEF to determine the coefficients
of the tridiagonal matrix for the special case along the lateral boundary where J=1 or
J=N. Likewise YCOIM is an adaptation of YCOEF for the I=M grid row which borders
the shoreline. These adaptations are necessary because boundary conditions necessitate
different numerical formulation of some of the derivative terms. Similarly with the presence
of an emerged impermeable barrier bordering the shore side of the JJJ grid row, special
subroutines YCJJJ and YCJJJ1 are used on the grid rows I=JJJ and I=JJJ+1 so as to
properly include the boundary conditions for these grid rows, and for shore perpendicular
groins subroutines XCJGR and XCJGR1 incorporate the boundary conditions for a groin
13
Figure 1.3: Flow chart of circulation model.
14
Figure 1.4: Flow chart of wave model.
located between the J = JGR and the J = JGR+1 grid columns.
Figure 1.4 is a flow chart of the wave model. Subroutine WAVNUM determines the wave
number k, the group velocity Ct and the intrinsic frequency a at each of the grid centers.
These terms are used in the governing parabolic equation. Subroutine AMPCO determines
the coefficients of the tridiagonal matrix and CTRIDA solves the matrix equation for the
complex amplitude on the next grid row. CTRIDA is essentially the same as TRIDA except
that it has been modified to handle complex numbers. The subroutine WW is called by the
AMPCO subroutine in order to determine the dissipation coefficient following the work of
Dally (1980,1987).
Chapter 2 describes the circulation model in greater detail. Individual sections will
cover the governing equations and the formulation of each of the component terms. ChapÂ¬
ter 3 describes the wave model, the derivation of the linear wave equation, the parabolic
15
approximation, the manner of including wave energy dissipation due to wave breaking and
the method of obtaining the radiation stress terms directly from the complex amplitude.
The fourth chapter gives details of the numerical methods used to obtain solution of the
governing equations and the boundary conditions employed. The fifth chapter presents the
results of applications of the model and comparisons with other studies both numerical and
experimental, as well as some analytical solutions to some idealized cases. The final chapter
summarizes the study and makes recommendations for future research.
CHAPTER 2
CIRCULATION MODEL
2.1 Governing Equations
The governing equations for the circulation model are the depth integrated time averÂ¬
aged horizontal equations of momentum
dU TTdU â€ždU dfj 1 1
aT + ual + + 9al+ jDrâ€œ ~ jDr
+
J_ (dstx asxv\ 1 Â¿>r,
pD \ dx dy ) p dy
dV TrdV T rdV dfj 1 1
aT + 0 ai + v a7 + 3ai + Jd'1" ~ Jdt
dx
+
_i_ f^xv 1 = Q
pD \ dx dy ) p dx
(2.1)
(2.2)
and the continuity equation
w + Ã©'Â£'D)+Â¿
(2.3)
where
U = x component of the mean current
V = y component of the mean current
fj = mean water surface elevation
p = water density
D = the total water depth = h0 + fj
h0 = local stillwater depth
rj = the lateral shear stress due to turbulent mixing
Tix = x component of the bottom shear stress
16
17
rky = y component of the bottom shear stress
t,x = x component of the surface shear stress
t,v = y component of the surface shear stress
and  *^xy and Svv are the radiation stress components which arise from the excess
momentum flux due to waves. It should be noted that U and V include the wave induced
mass transport.
These equations are obtained by integrating the local x and y momentum equations and
the continuity equation over the depth of the water column and then time averaging the
results. This is demonstrated in Appendix A, which follows closely the work of Ebersole and
Dalrymple (1979). For greater details of the derivation the reader is referred to Ebersole
and Dalrymple (1979) or Phillips (1969).
These are the same equations that have been used by other modelers such as Birke
meier and Dalrymple (1976), Ebersole and Dalrymple (1979), Vemulakonda (1984), and
Yan (1987). Birkemeier and Dalrymple used an explicit numerical scheme and neglected
the advective acceleration terms and the lateral mixing. Ebersole and Dalrymple added the
advective acceleration and lateral mixing terms but still used an explicit numerical scheme.
Vemulakonda introduced an implicit numerical formulation for these equations. Each of
these investigators used a refraction wave model based upon the work of Noda et al. (1974)
which solves for the wave height and wave angle at given grid points. Yan (1987) used a
combined refractiondiffraction wave equation to determine the complex wave amplitude
(magnitude and phase) and then solved for the wave angle in terms of the slope of the waÂ¬
ter surface. Each of these studies obtained the radiation stress components in terms of the
wave height and wave angle. The present study differs from the previous ones in that the
radiation stress terms are obtained directly from the complex amplitude and its gradients.
This will be discussed in section 2.2.
The use of Eqs. (2.12.3) involves many assumptions. The primary assumption is
that the flow field can be represented in two dimensions using depth and time averaged
18
equations. This assumption implies that vertical fluctuations of the flow which are zero in
the time average can be ignored. Many details of the flow may be missed by this assumption
of purely horizontal flow. Of course reducing the problem to a twodimensional flow has
the distinct advantage of reducing an essentially intractable problem to a simpler tractable
form.
Some of the terms in Eqs. (2.12.3) also involve certain assumptions. These terms
are the bottom friction, the surface wind stress and the lateral mixing. These are all
phenomenon that are not completely understood so that assumptions are made concerning
the complex forces that are represented. These assumptions usually involve the use of
empirical coefficients. In this study the wind forces are ignored so the assumptions involved
with the surface stress will not be discussed in detail here. It would be a relatively simple
task, though, to add a wind surface stress to the model, using a quadratic shear stress as
formulated by Birkemeier and Dairymple (1976) and also used by Ebersole and Dedrymple
(1979) following the work of Van Dorn (1953) in determining the value of the empirical
coefficient.
The interaction of an oscillatory fluid flow with a real bottom is very complex. Many
physical phenomena may be at work. With sand bottoms there can be percolation into the
sand and frictional effects resulting in wave energy dissipation. With mud bottoms where
the mud is essentially a dense fluid layer there can be damping and internal waves along
the mudwater interface. Wave energy losses due to bottom effects are not included in the
model but can easily be incorporated into the wave dissipation subroutine. Resuspension
and settling of particles also constitutes a complex phenomenon with density changes and
momentum transfers. All of these effects are glossed over in the model by assuming a rigid
impermeable bottom. As with all of the models discussed above the bottom friction term is
represented in the quadratic form. Noda et al. (1974), Birkemeier and Dalrymple (1976) and
Vemulakonda (1984) all used the weak current formulation as proposed by LonguetHiggins
(1970a) in which
19
Tbz = pF uorl U
(2.4)
Thv = pF uori V
(2.5)
where F is a drag coefficient (of the order of .01) and uorj, is the time average over a wave
period of the absolute value of the wave orbital velocity at the bottom. From linear wave
theory it is found that
O IT
Uort = T sinh kh (2,6)
where H is the wave height and T is the wave period, k is the wavenumber, and h = D is
the total water depth. Ebersole and Dalrymple used a more complete quadratic formulation
Tb = pF Ã¼*ÃÃ¼t
(2.7)
where the overbar represents the time average over a wave period and ut is comprised of both
the mean current and the wave orbital velocities. The total velocity vector tT* is expressed
as
ut = (U + Â¿eos#)Â» + (V + Ã¼sin0),7 (2.8)
so that its magnitude is given by
u1 = VU2 + V2 + u2 + 2UÃš cos 9 + 2V Ã¼ sin 6 (2.9)
The wave orbital velocity Ã¼ is expressed as
Ã¼ â€” um cos at
(2.10)
where um is the maximum wave orbital velocity at the bottom which is found to be
xH
â€”
T sinh kh
The x and y components of the time averaged bottom friction are thus
Thx = ^ JQ ({/ + Â¿m cos icos at) â€¢ utd(<7t)
Hv = ^ Jo (Y + Â¿m sin icos at) â€¢ u{d(ot)
(2.11)
(2.12)
(2.13)
where tij is written as
20
tTj = \Ju2 + V2 + titeos2 at + 2Uum cos at cos 9 + 2VÃºm cos at sin 0 (214)
The integrals in Eqs. (2.12) and (2.13) are then evaluated by a 16 point summation using
Simpsonâ€™s Rule.
In the present study several different formulations are used and compared. The first is
the formulation given by Longuet Higgins (1970a) and used by several previous investigaÂ¬
tors. The second is a modification of the LonguetHiggins formulation with partial inclusion
of current strength in the friction formula.
T* = pF \uorb\ U + pF\U\U
(2.15)
rbv = pF\uorb\V + PF\U\V
(2.16)
It is thought that this formulation has some value since in the lee of structures where there
is relatively weak wave activity, there may still be relatively strong currents generated by
the gradients of the setup outside of the sheltered region. The third formulation is similar
to that of Ebersole and Dalrymple in that the full quadratic expression is used. However the
wave orbital velocities at the bottom are expressed as gradients of the complex amplitude
and the resulting integral is determined using an eight point Gaussian integration. It is
found that an eight point Gaussian integration of a cosine squared integration over an
interval of 2x is accurate to within 1 percent.
The quadratic formulation of the bottom friction is used
Tb = pF uitT( (2.17)
where tT* is composed of both the mean current and the wave orbital velocities. The wave
orbital velocities are expressed as the real part of the gradient of the complex velocity
potential $ which is expressed as
a> _ â– â€¢,A(JÂ»V) cÂ°sh k[h0 + z) iut
9 a cosh kh
a
(2.18)
21
where A is the complex amplitude, u/ is the absolute frequency, a is the intrinsic or relative
frequency following the current, and h0t h and k are as previously explained. Equations
governing $ and a parabolic approximation solution for A are presented in the next chapter.
The total velocity vector tT< is expressed as
ut =
= Â£,+s(ff)]r+[v+*(!;)]
where SR indicates the real part, so that its magnitude is given by
Â«i  f2+v!+(Â£)+(Â£)+[Â«(Â£)]â€™+[* (Â£)F
The x and y components of the time averaged bottom friction are thus
â€œ 'rjÂ£[a + *{i)]''51*
Tty  >4 Jo \v+*(%)}â– w
(2.19)
(2.20)
(2.21)
(2.22)
As in Ebersole and Dalrymple, the assumption is made that only the gradients of A are
retained and the gradients of the depth and spatially varying quantities such as k and a
are ignored. This is essentially the assumption of a flat bottom and very slowly varying
currents.
Figure 2.1 plots the magnitude of the longshore current as a function of the distance
offshore for the various friction formulations discussed above. These velocity profiles were
obtained numerically for an eleven second wave with a deep water wave height of 1.028
meters (1 meter at offshore grid of the computational domain, using a AX of 10 meters
and 30 grid rows) and a deep water wave angle of 17.28 degrees (10 degrees at the model
boundary) approaching a 1:25 plane beach. A friction factor, F, of .01 was used for each case.
Using a friction formulation that ignores the contribution of the waves results in a very weak
friction and thus a large longshore current. Using the Longuet Higgins formulation gives a
friction slightly less than the more proper time averaged formulation, while a combination of
the LonguetHiggins formulation and the strong current model (no wave contribution) yields
a slightly stronger friction. For practical numerical considerations the latter formulation is
VELOCITY (M/SEC)
22
in
Figure 2.1: Longshore current profiles for different friction formulations for an 11 second,
1.028 meter wave at 17.28 degrees to a 1:25 plane beach.
most often used since it requires less computation time and more importantly the Gaussian
integration involving the reed part of the gradient of complex amplitudes at times produces
stability problems.
The lateral mixing term in Eqs. (2.1) and (2.2) also involve several simplifying assumpÂ¬
tions. The first assumption is found in the derivation of the two depth integrated equations
of motion where it is assumed that the turbulent shear stress term pu'v1 is independent
of depth. The other assumption is that the process of momentum transfer due to turbulent
fluctuations can be represented by a product of a mixing length parameter and derivatives
of the mean current. This is what is done in the present study and also in previous studies.
Ebersole and Dalrymple (1979), Vemulakonda (1984), and Yan (1987) all used the formuÂ¬
lation given by LonguetHiggins (1970b) in which he developed an analytical expression for
the cross shore distribution of the longshore current velocities using a linearized friction
formulation as described above, linear waves at an angle to the normal of a plane beach,
23
and a spilling breaker assumption. LonguetHiggins wrote
/ dU dV\ ..
" = aÂ¿ + t*al) (2'23)
in which ex and ev Eire mixing length coefficients.
LonguetHiggins reasoned that the mixing length parameter ex must tend to zero as
the shoreline was approached. He assumed that ex is proportional to the distance from the
shoreline, X, multiplied by a velocity scale which he chose as the shallow water celerity
y/gh. Thus ex is written as
ex = NX \/gh (224)
where N is a dimensionless constant for which LonguetHiggins chose the limits as
0 < N < 0.016
(2.25)
Obviously, for physical reasons there needs to be some limit on the scale of these eddies.
Ebersole and Dalrymple arbitrarily chose the value of ex near the breaker line to be the
maximum value and for ex to be constant offshore from the breaker line. The same limit
upon ex is used in this study. The coefficient ev is assumed to be a constant and it is
arbitrarily assigned the maximum value of ex.
In the present study the further assumption is made in computing that the cross
derivatives are negligible (a common assumption in fluid mechanics). Therefore, the lateral
mixing terms become
(*Â»f)
(2.26)
/ sv\
\Â£xdx)
(2.27)
in the x direction and
in the y direction. This means that the lateral mixing term is represented by a purely lateral
diffusion of momentum. That the lateral mixing term is a lateral diffusion term has some
significance in that the diffusion term is expressed explicitly and thus there results a time
step limitation to an implicit model. This will be discussed in section 4.1 dealing with the
finite differencing of the equations governing the circulation model.
24
2.2 Radiation Stresses
In a series of papers in the early 1960s LonguetHiggins and Stewart (1962, 1963, 1964)
introduced and developed the mathematics of the radiation stresses. These stresses which
can be conceptually thought of as the â€œflux of excess momentumâ€ can be obtained through
integration of the momentum equations over the water column as is shown in Appendix A.
Using linear wave theory the radiation stress components are found to be
Szz = E n(cos2 6 + 1) â€” ^
At
Svv = E n(sin2 0 + 1) â€” ^
At m
Szv = En cos 0 sin 6
(2.28)
(2.29)
(2.30)
where E is the wave energy equal to pgH2, n is the ratio of the group velocity to the wave
celerity and 6 is the angle the wave rays make with the x axis.
If the wave height and wave angle are known at each grid point, the radiation stresses
and their gradients are readily determined for use in the governing equations. To determine
the wave height and angle at each grid point Birkemeier and Dalrymple (1976), Ebersole
and Dalrymple (1979) and Vemulakonda (1984) each used the following scheme which is
based upon the irrotationality of the wave number and an equation for the conservation of
wave energy. The irrotationality of the wave number determines changes in the wave angle
due to refraction while the energy equation determines changes in the wave height due to
shoaling and refraction. This scheme was first introduced by Noda et al. (1974).
The irrotationality of the wave number
V x Ã = 0 (2.31)
where
k = \k\ cos Oi + fc sin Oj
(2.32)
is expanded in Cartesian coordinates to obtain
ae dO
cos 6â€”I sin 0â€” + sin 0
ox ay
fc dx
â€” cos 9
J_d\k\
1*1 dy
(2.33)
25
For waves on a current the dispersion relation is
oj = a + k
Ãœ = tanh \k\h + fc cos 9 + \k\sin 9
(2.34)
Equation (2.34) is differentiated to eliminate and from Eq. (2.33). Using a
forward difference derivative in x and a backward difference derivative in y, 9{j is obtained
as a function of cosÃ, sin 0, U,V,k,h and their derivatives as well as 6 at the adjoining grids.
The sin 6 and cos 0 terms are expressed in terms of the surrounding four grids using a 2nd
order Taylor expansion. The solution is an iterative process alternating solution of the 6
equation with solution of the dispersion relationship (2.34) for k = Jt which is accomplished
by a NewtonRaphson method until acceptable convergence is obtained for k and 9.
Next an energy equation is solved for the wave height. A conservation of energy equation
is given by Phillips (1969 ed. Eq. 3.6.21) as
_+_{m + CJ> + Sj,._i = 0 (2.35)
Assuming steady state conditions and expanding in Cartesian coordinates gives
(V + C,cos t)1 + (V + C, sin Â«)i+ A(U + C, cos S)
dU
dU
+ aZ.(y + cs sin 9) + gzx~qÂ¿ + Tvz^[ + fÂ¡
dy
dy
'XV
dV _ dV
dx Â°vv dy
0 (2.36)
where
a
XX
avv =
'TV
â€” Tyl â€”
n(cos2 9 + 1) â€” i
2
n(sin2 9 + 1) â€” 
2
n cos 9 sin 9
Defining
Qi j = \{^U + C>cose) + ^V + Co8in + 9 }
(2.37)
where
dU dU dV _ dV
zxdÂ¿ + Tvx~dÃ + Tzv~dÂ¿(Tvv'dÂ¿.
(2.38)
a =
26
and noting that U,V,Cly and a are known quantities; taking forward difference derivatives
in x and backward difference derivatives in y and solving for the wave height, Hij, at the
*,j grid yields
Hij =
(V+C,Â»in g)g,,y_l _ (U+C, cot e)Hi+1j
AY ~ AX
V+C4Â»ini U+C, coei r,
ÃY" Â¿X
(2.39)
AY Af LJjj
The computation is a row by row iteration proceeding from large i (offshore in the
coordinate system of the above cited authors) to small i. During the computation of each
Hij the breaking height is also calculated and if Hij exceeds this value, then the breaking
wave height is used. Several breaking height indices are possible. Miche (1944) gave as the
theoretical limiting wave steepness the condition
^r = .142tanh(fcj,hÂ¡,)
Lb
(2.40)
where the b subscript indicates breaking conditions. Le MÃ©hautÃ© and Koh (1967) indicated
from experimental data that the limiting steepness is
â€” = .12tanh(fcÂ¡,hÃ,)
Lb
(2.41)
or
J5T* = ^.12tanh(Jfe*fc*) (2.42)
Kb
This is the breaking height criterion used by Noda et al. (1974). Perhaps the simplest
breaking height criterion is the spilling breaker assumption
Hh = .78ht (2.43)
which is used by Vemulakonda (1984) and Vemulakonda et al. (1985).
The scheme described above of determining the wave angle 6 from the irrotationality
of the wave number (2.33) and the dispersion relation (2.34) and the wave height from
the energy equation (2.35) has its limitations. Principally this scheme does not take into
account the effects of diffraction. This limitation is surmounted by use of a parabolic wave
equation. However this presents another problem in that the wave angle is not determined
27
in a parabolic equation solution. Only the complex amplitude, which contains the magÂ¬
nitude and the phase of the wave, is obtained. Yanâ€™s (1987) solution to this problem is
to obtain 0 directly from the complex amplitude by assuming that the slope of the water
surface resulting from the waves indicates the direction of wave propagation. This value
for 0 and twice the absolute value of the complex amplitude for the wave height are then
used in Eqs. (2.282.30) to obtain the radiation stresses. Hie value thus obtained for 0 is
then also used in the dispersion relationship. For plane waves this works fine; however, in
diffraction zones where shortcrested wave are encountered problems may arise, since the
instantaneous slope of the wave at a particular location can be other than in the direction
of wave propagation.
In this report the radiation stresses are obtained directly from the complex amplitude
and its gradients using equations developed by Mei (1972) and repeated in his book (1982).
The basis of the derivation is to use the definition of the velocity potential to substitute for
the wave induced velocities in the definition for the radiation stresses.
For an inviscid incompressible fluid the instantaneous equations of motion are
+ = â€” Vp â€” pgez (2.44)
Vf = 0 (2.45)
where q represents the velocity components, p is the pressure and ex is the z or vertical unit
vector. Denoting the horizontal components of q as ua,a = 1,2 and the vertical component
of q as w, the boundary conditions are
= 0 (2.46)
= tv (2.47)
on z = rj, where rÂ¡ is the instantaneous free surface, and the bottom boundary condition
dh
u, â€” = Â«, (2.48)
q is assumed to be composed of a time averaged part q and a fluctuating part q'\ i.e.
dr)
drÂ¡
dt +Up'dxp
f= 9t*,y ,z) + q,(x,y,z,t)
(2.49)
28
Integrating a horizontal component of (2.44) vertically over the depth of the water column
from z = h to z = rÂ¡, using the boundary conditions (2.462.48) and then taking time
averages (the same procedure as in appendix A) the following definition of the radiation
stresses is obtained:
ro
Sap = p u'au'pdz + Safi
J â€” h
[i
Pdz~ y(Â»i + *0J
h Â¿
+ yW)1
(2.50)
where the overbar is used to indicate the time averaged values and 6ap is the Kronicker
delta function
tap
for a = P
for a ^ f3
(2.51)
Mei (1972) was able to express Eq. (2.50) in terms of the general complex amplitude
function B by relating it to the velocity potential and the free surface height by
_ ig cosh k[h0 + z) iut
a cosh kh
t) = B e~iut
(2.52)
(2.53)
An expression for p is obtained by integrating the vertical component of Eq. (2.44)
dw dw
fai+inâ€˜f3rt+,m
dw
dz
(2.54)
Integrating from any depth z to the surface, using Leibnitzâ€™s rule for interchanging the
order of differentiation and integration, employing the two surface boundary conditions,
Eqs. (2.46) and (2.47) and the continuity Eq. (2.45), Eq. (2.54) yields:
d fi d H
= pgin ~ z) +pâ€”J wdz + pâ€”J upwdzpw\ (2.55)
Taking time averages gives
rfl du'w'
p{x,y,*) = P9{r) ~ z) + Pjt dx dz ~ p(w'm)2 (2.56)
Substitution of Eqs. (2.52),(2.53) and (2.56) into (2.50) and making the assumption that
horizontal derivatives of k, h and a can be neglected (this is essentially a mild slope asÂ¬
sumption), the following is obtained after straightforward but tedious calculations which
29
are shown in Appendix B:
c P9 L (dBdB*) 1 f 2kh \
4  ydxa dxp J k2 V sinh2fch/
Â°ap
IBI
2 kh
sinh 2 kh
+
dxa dip
(VhB2  *2Â£2)Â¿(2*hcoth2*h  1)] } (2.57)
where B* denotes the complex conjugate of B. In obtaining Eq. (2.57) use was made of the
flat bottom Helmholtz equation
Vj;Â£ + Jfc2fi = 0 (2.58)
The general complex amplitude function B is related to the complex amplitude A which
is solved for in the wave model by
B = Ae^flcolidz^
(2.59)
There are some problems with Eq. (2.57). One may question the assumption of a
flat bottom; however the inclusion of the slope and gradients of the wave number and the
intrinsic frequency leads to intractable mathematics. The use of the localized flat bottom
is also implicit in the derivation of LonguetHiggins and Stewart to obtain Eqs. (2.28 2.30)
since in carrying out the integrations involved it is assumed that horizontal derivatives of
the velocity potential are limited to derivatives of the phase function. For plane waves
using a spilling breaker assumption, results using Eq. (2.57) differ little from those using
Eqs. (2.282.30); however in diffraction zones Eq. (2.57) is superior for reasons mentioned
previously.
The major problem with the use of Eq. (2.57) is also a major problem associated with
the use of LonguetHigginsâ€™s formulation. This problem is that in both formulations the
initiation of wave breaking implies the dissipation of energy and thus the forcing of setÂ¬
up and longshore currents starting at the breaker line. The physical reality is somewhat
different in that in the initial stages of breaking, energy is not immediately dissipated but
rather is transformed from organized wave energy to turbulent energy and to a surface roller
(Svendsen 1984). There is a space lag between the initiation of breaking and the dissipation
30
30
23
20
1.3
l 0
03
0
03
6
4
2
>
0
â€¢2
4
eat.
Figure 2.2: Experimental results for wave setup and setdown. Reprinted from Bowen, InÂ¬
man, and Simmons (1968), Journal of Geophysical Research, vol. 73(8) page 2573, copyright
by the American Geophysical Union.
of energy. This can be seen in figure 2.2 which gives the results for the experiments of Bowen
et al. (1968). Analytical and numerical solutions for the mean water surface elevation will
give a setup starting at the break point while the solid fine in the lower graph shows
that setup measured in the experiment begins shoreward of the break point. This will be
discussed further in Chapter 5.
CHAPTER 3
WAVE MODEL
3.1 Governing Wave Equation
Severed methods have been used to derive the governing equation for the propagation
of waves over varying topography and varying currents. Among the methods used are perÂ¬
turbation techniques, multiple scales, a Lagrangian and the use of Greenâ€™s second identity.
With the assumption of a mild slope and weak current conditions each of these methods
should yield the same equation. For illustrative purposes the method of using Greenâ€™s secÂ¬
ond identity will be used here to derive a linear equation. A Lagrangian derivation of the
same equation is presented in Appendix C. The linear equation is used since second order
waves exhibit singularities as the water depth goes to zero.
A velocity potential t is assumed such that the water particle velocities are given by
V$, where V is the threedimensional gradient operator
d  d * d 
(31)
and *,j, and k are the unit vectors in the x,y, and z directions, respectively. The velocities
are a superposition of steady currents and wave induced motion. The velocity potential is
given by
t = to + eft (3.2)
where to is defined such that its gradient yields U,V, and W which are the steady current
terms, and ft is the wave component of the velocity potential. An undefined scaling e
is used so as to separate the wave and steady current part of the velocity potential. The
wave part of the velocity potential is composed of two parts: /, which contains the depth
dependency and which is a function of the horizontal coordinates and time. The depth
31
32
dependency / is given by linear theory as
j _ cosh k(h0 + z)
cosh kh
(3.3)
where h0 is the depth of the bottom referenced to the still water line, and h is the total
water depth and k is the wave number vector. It should be noted that / is also a function
of the horizontal coordinates since k and h0 vary with horizontal position.
For incompressible, irrotational flow the governing equation for the velocity potential
is the Laplace equation
VV = 0 h0
where r? is the instantaneous water surface elevation.
Using a horizontal gradient operator V/, defined by
_ d Â» d Â«
Vkfc,+^J
the Laplace equation is written as
+ zz â€” 0
h0 < z < T]
(3.5)
(3.6)
The governing equation is complemented by the following boundary conditions. On the free
surface the kinematic free surface boundary condition is given by
â€¢ VhV = 4>x
z = rÂ¡
The dynamic free surface boundary condition with zero atmospheric pressure
1,
4>t + ^(VÂ¿)2 + gz = 0
z = rÂ¡
(3.7)
(3.8)
and the bottom boundary condition
 Vh â€¢ Vfcho = 4>t
z = h0
(3.9)
The derivation technique is to write Greenâ€™s second identity for and / as defined by
Eqs. (3.2) and (3.3) respectively.
rfi rfi
Ã fzzdz  Ã dz = [ft  fz]lt
W â€” ft0 J ~ ItQ
(3.10)
33
Then Eq. (3.6) is used to substitute for zz in the first term of Eq. (3.10) and the boundary
conditions are used to substitute for z at the mean free surface and at the bottom in the
third term of Greenâ€™s identity. Equation (3.2) is used and terms of order e only are retained
in order to obtain the governing equation for .
In order to obtain z at fj it is assumed that
tj = tj + eq.
(3.11)
The kinematic free surface boundary condition, Eq. (3.7), is then expanded in a Taylor
series about z = fj,
_ d
+ e"aÂ¡
dn
 z
= 0
J*=fJ
(3.12)
solving for z using Eqs. 3.2 and 3.11 while retaining only terms of order e yields
dfj
4>* Uâ€”
A total horizontal derivative
dt
+ U â€¢ Vhfj + Vh â€¢ Vfcfj  fjWz
*=tf
(3.13)
D 9 ft yn
Di = ai + uâ€™Vh
(3.14)
is introduced so that the first two terms on the right hand side of Eq. (3.13) can be written
as the total horizontal derivative of fj. The final manipulation of Eq. (3.13) is to recognize
that â€” ^ at the surface is the horizontal divergence of the horizontal current (= V>, â€¢ Ãœ).
Thus
* fl= + fjVh â€¢ Ãœ
(3.15)
rj is eliminated from Eq. (3.15) by expanding the dynamic free surface boundary condition,
Eq. (3.8), in a Taylor series about z = fj.
4>t + + gz
*=T)
= 9*1 +
+cÃ±
M=fÂ¡
Â¿Â« + Â¿(V*)2
, 1 3
9 + *t+ 2^(V^)
= 0
J*=
(3.16)
Using Eq. 3.2 to substitute for <)> the order t terms yield a solution for rj
(3.17)
34
Substituting into Eq. (3.15) gives
The first term of the Greenâ€™s identity using Eq. (3.6) becomes
r f.,dz=r fvi+dz=r w^p^dz (3.19)
* â€œ ho * â€ ho * ho
Now
Vh = U + efVh4> + e4>'Vhf (3.20)
and
^h â€¢ = Vfc â€¢ Ãœ + eVfc â€¢ (/ V/,^) + â€¢ V/,/ + z
Retaining only terms of order e and dropping the last term since it is second order in the
slope the integral is equal to
r /Vfc.(/V^)dz r fVh4>Vhf dz (3.22)
J ~ho * â€œ/lo
These two terms are handled using Leibnitzâ€™s rule of integration.
vaT /(/vâ€žÂ¿)dz = r vhf(fvj)dz+r fvh(fvht)dz
* ho * /lo * hp
+^hVf2 U + _aâ€ž V/,<Â£ (3.23)
Using this relationship and recognizing that / 
depth of /2 is CCt/g, where C is the wave celerity, Ca is the wave group velocity and g is
the gravitational constant, the first term is equal to
 V* â€¢ (â€”9Vj) + Vhfj + Vhh0f21â€”aâ€ž
9
where the z has been dropped.
The second term of Eq. (3.10) is evaluated quite easily.
dz â€” z
(3.24)
(3.25)
35
The first integral above is dropped since there are only steady current terms. The remaining
term, dropping the e, is
= ~r f*f"dz = k74>r fid.â€”*â„¢* (3.26)
Jh0 Jho 9
The third term in Eq. (3.10) is
H. \H = Hu In Hu Uo (3.27)
The first term on the right hand side is equal to t q since / 1^= 1 and has already been
determined. The second term on the right hand side of Eq. (3.27) is easily determined using
the bottom boundary condition Eq. (3.9)
 Hm fcâ€ž=/Vfc^*Vh/i0 _fco (3.28)
Using Eq. (3.20) this becomes
 H. Uâ€ž= fU â€¢ Vkh0 _fco +e/2 _fco Vhj> â– Vhh0 + efivhf â– Vhh0 _fco (3.29)
The first term is dropped since it not order e and the third term is clearly higher order in
the slope. Thus, again dropping the t
Mk=/!lVkfVA (3.30)
This term cancels the bottom boundary term resulting from the Leibnitz integration above.
The fourth term of Eq. (3.10) is also easily determined.
â€œ 4>fu A0= ~^81DcMh Â°kh ^ *"fco= ~k tanh kh â€ Y* ^ (331)
Evaluating at the mean water surface and retaining only terms of order e yields
Mu th=~+ (332)
Collecting terms from (3.24),(3.26),(3.18),(3.30) and(3.32) yields the following hnear
governing equation for .
+ (V* â€¢ Ã¼)Â£Â£  Vfc(CCfVfc* + (a2  k'CCM = 0 (3.33)
This equation was first derived by Kirby (1984).
36
3.2 A Parabolic Approximation
Equation (3.33) is a hyperbolic wave equation. It can be altered to an elliptic equation
by assuming that any time dependency is solely in the wave phase. Before doing this an
energy dissipation term DD is introduced as follows. It is assumed that Eq. (3.33) holds
for the case of no energy dissipation; then for the case with dissipation the left hand side is
balanced by any energy that is dissipated. The governing equation is then written as
+ (V* â€¢  Vk(CCfV*Â¿) + (
In the derivation of the governing hyperbolic equation using a Lagrangian as presented in
Appendix C the DD term can be viewed as a virtual work term.
Following Kirby (1983) the DD term is assumed to be of the form
Dd>
DD= WD( (3'35)
where w is an undefined coefficient indicating the strength of the dissipation. Eventually
the w term will be related to the energy dissipation due to wave breaking following the work
of Dally, Dean and Dalrymple (1984).
Using Eq. (3.35) in Eq. (3.34) and making the assumption that the only time dependency
is in the phase, i.e.
dd>
= iu) (3.36)
reduces the hyperbolic equation (3.34) to the elliptic form
2iuU â– + 0 â€¢ Vh(U â€¢ Vj) + (Vfc â€¢ ff)(Ãœ â– VhÂ¿)  Vâ€ž â€¢ (CCtVhÂ¿)
+{ = Ãºrw<Â£ (3.37)
where only the phase contribution to the horizontal derivative of ^ is retained in obtaining
the term on the right hand side of (3.37).
Before deriving a parabolic approximation to Eq. (3.37), the need for such an approxiÂ¬
mation should be discussed. There are two major computational drawbacks to numerically
37
solving an elliptic equation. The first is that solution is required simultaneously for each
grid. This means that for a large computational domain, say m by n grids, it is necessary
to invert anmxnbymxn matrix. If m and n are sufficiently large, problems may be
encountered in terms of the interned memory storage, and even if the available storage is
not a problem, the number of computations required to invert the matrix may make for
an exceedingly slow solution. In some situations this may be acceptable; however for the
purposes of a wave induced circulation model in which repeated solution for the wave field
is necessary the program would take a very long time to run. The other disadvantage inÂ¬
volved in the numerical solution of an elliptic equation is that boundary conditions must be
specified at all of the boundaries. In many applications the only known boundary condition
is the offshore, or incident wave amplitude. For example, the domain of interest may be the
wave field in the lee of an obstruction where the only down wave condition is that the waves
freely pass through the down wave boundary. Or in the present study in which the wave
amplitude is known to go to zero at the shoreline a problem ensues in that the shoreline is
(for purposes of the circulation model) at the grid edge and not at the grid point.
A parabolic equation eliminates both of these problems. Since the solution proceeds grid
row by grid row where each solution uses the results of the previous grid row, no down wave
information is required. Since only one grid row is solved at a time, the solution requires
only that a tridiagonal matrix equation be solved to obtain values for the grid row. The only
required boundary conditions are the conditions on the first grid row (usually the offshore
boundary) and lateral boundary conditions which are usually a derivative condition that
allows for the wave to pass through the boundary without any reflection at the boundary.
There are several ways to transform the elliptic Eq. (3.37) to a parabolic equation. The
most direct is to make certain assumptions concerning the length scale of variations in the x
direction. At the heart of the parabolic approximation is the assumption that the direction
of wave propagation is essentially along the x axis. For waves propagating at an angle to
38
the x axis the proper form of 4> is
Â¿ = g y) e*(/ * co" * dz+f * ,in * Â«*Â»> (3.38)
and the proper form of the dispersion relationship is
u> = a + kcosOU + fcsin 6 V (3.39)
where ff is the angle of the wave propagation relative to the x axis. The essence of a parabolic
approximation is to assume that the angle that the wave rays make with the x axis is small
so that the ksin0 term can be neglected in Eqs. (3.38) and (3.39) and cos 6 is assumed to
be unity. The manner of doing this varies with different authors. Yan (1987) retains the
complete dispersion relation (Eq. (3.39)) but deletes the y dependency in Eq. (3.38) in the
derivation of his governing wave equation. He solves for 6 by determining the slope of the
instantaneous water surface. For long crested waves this method is satisfactory although
requiring additional computations. However, this method may produce erroneous results in
diffraction zones where the presence of short crested waves will produce a horizontal comÂ¬
ponent of the Blope normal vector that is different from the direction of wave propagation.
Kirbyâ€™s (1983) method is to make the assumption that
Â¿=tgA(?lVle<(ftdz) (340)
o
so that the e,fktin6dv Gf Â¿he phase function is absorbed into the amplitude function,
A.
Using Eq. (3.40) directly in Eq. (3.37) and further assuming that derivatives of Ax are
small compared to derivatives of the phase function (i.e. that Â« ikAz ) will yield a
parabolic equation. For illustration the case of linear diffraction in the absence of currents
and in a region of constant depth is examined. Substitution of Eq. (3.2) into the Laplace
equation (3.6) yields the Helmholtz equation
VÂ£Â¿ + *2Â¿ = 0
(3.41)
This equation is also obtained directly from Eq. (3.37) for the stated assumptions of no
currents, no dissipation, and constant depth. Substitution of (3.40) into the Helmholtz
equation gives
Axx + 2ikAx + AVV = 0 (3.42)
and employing the above mentioned assumption concerning derivatives of Ax yields the
linear parabolic equation
2 ikAx + AVV = 0 (3.43)
Another method of obtaining a parabolic equation is to use a splitting method in which
it is assumed that 4> is composed of forward moving and backward moving components that
are uncoupled.
4> = + Â¡j>~ (3.44)
Uncoupled equations for + and are developed and since the equations are uncoupled
only the equation for + is retained. In the following the work of Kirby (1983) and Booij
(1981) is closely followed.
The second order elliptic equation
d_ /1Â£Â£\
dx Vt dx )
+ = 0
(3.45)
can be split exactly into equations governing forward and backward disturbances $+ an $
which are given by
d$+
dx
d$~
dx
+Â¿7$+
â€”Â»7~
Booij chose the relation
(3.46)
(3.47)
(3.48)
40
where Â£ is an as yet unknown operator on . Substituting into Eq. (3.45) and neglecting
derivatives of ( alone leads to the result
4>zz +
<Â¡>z + = o
(3.50)
The elliptic equation (3.37) is now put into the form of Eq. (3.50). Letting p = CCg
for notational convenience and expanding some of the terms of Eq. (3.37) in their spatial
coordinates gives
[p4>z)z + (p4>y)v + k2p + (u>2  a2 + iw(V/, â€¢ U))4> + iaw4>
~{U2k)z ~ {V2iv)v  (UV]>z)v  (UV4>v)z + 2 tuff Vj = 0 (3.51)
Booij chose to neglect terms involving the square of the mean current components assuming
that they are smaller than terms involving p and then arranged the above equation as
4>zz + P l[Pz + 2iuU)4>x + k2 ^1 + 4> â€” Q (3.52)
where
M4> = (w2  a2 + tw(Vfc â€¢ Ãœ) + Â«0w)<Â¿ + 2iuV4>y + [p4>v)v (3.53)
Kirby points out that this choice by Booij leads to his inability to derive the correct wave
action equation. Kirby instead does not drop any of the squares of the mean current
components until after the splitting is complete. By assuming that the waves are oriented
in the x direction the dispersion relation is written as
a = oj â€” kU
(3.54)
leading to the result
(jj2 â€” (j2 â€” 2ukU  (kU)2
(3.55)
Using this relationship Kirby then arranged Eq. (3.51) as
h, + (p  V')'(p  U*)J, + (1 + â€˜^ ) i = 0
(3.56)
41
where
Ml = {2ukU + iÃ¼j{VhÃœ) + i
+ {{pV2)lv}v + 2iuÃœ Vhi (3.57)
Comparison with (3.50) gives
 *â€™(1+W^y) <3'58)
Y = (359)
or solving for 7 and Â£
â€¢> = (360)
Ã  *Â»<,*)Â»(!+ 35^)* (3.61)
Booij points out that the splitting matrix approach employed by Radder (1979) is equivalent
to the choice
1R = *(1 + 2A:2(p  Z72))
Â£r = ki(pU2)2
Using the results (3.60) and (3.61) in Eq. (3.46) along with the definition (3.49) gives the
parabolic equation
^ (1 +
M
k2(p U2)
By expanding the pseudooperators in the form
M
) *i+ } =
(
1 +
1 +
Ã)Â¡Ã+ â€œ (I +
M
k2(p â€” U2)J \ 4k2{pU2)
M \f .. . 3 M
iYi+ * (i+Â¡Â¡
)*+
k2(pâ€”U2)) \ 4 k2(p â€” U2)j
the result of both expansions may be summarized in the general equation
(3.63)
(3.64)
fz{kib (â€˜ + kâ€™frlu*)) Â¿+} = Vâ€™)1â€™ 0 + Â»toU>)) ** (3'65)
42
where
â€¢{\
Radder
i BÃ³oij
(3.66)
and where P2 = Pi + j for both approximations. Kirby chose the approximation with
Pi = 0 since for his derivation which was for weakly nonlinear waves it should correspond
to the information contained in the nonlinear Schrodinger equation. With Pi = 0 Eq. (3.65)
becomes
1
k(p  U2)fc +  { fc(p  u2)}x 4>+ = ik\p  U2)j>+ + 'M4>+
(3.67)
where M+ is defined by Eq. (3.57). An equation for the complex amplitude A is then
obtained by making the substitution
(3.68)
which after dropping squares of the components of the mean current results in
(C, + V)A, +  >v \ (7) t A
4[cc<(p)J/iA=0 (3'69)
Kirby then further modified the equation by a shift to a reference phase function using the
relation
A = A'ei(lxfkdx) (3.70)
where k can be taken as an averaged wave number, resulting in
(C, + U)A'X + i{k  k)(CÂ¡ + U)A' + \ ^
The shift to a reference phase function is intended to incorporate the phase information
in the complex amplitude so that the instantaneous water surface can be easily recovered
in the form
q(x,y) = 3i{AV/Â£d1}
(3.72)
43
where 9? designates the real part. This is an important consideration especially when obÂ¬
taining the radiation stresses.
It is important that in obtaining Eq. (3.69) that squares of the products of the current
components be retained until the last step. The reason for this is seen in that terms such
as kUV and kU2 are combined with terms of ojV and uU using the dispersion relationship
to give terms of aU and aV.
Equation (3.71) has been successfully used by Kirby in the form given and with modifiÂ¬
cations. The modifications are the addition of a nonlinear term which makes the equation
applicable for weakly nonlinear waves. However for present purposes Eq. (3.71) presents a
certain problem which will shortly be corrected. The problem is that for waves propagating
at an angle to the x axis it is not possible to obtain a correct form of the conservation
of wave action. The case of no current on a beach with straight and parallel contours is
given as an illustration. The equation for the conservation of wave action reduces to the
conservation of wave energy flux. In mathematical terms this is
^Cscos0a2J =0 (3.73)
where a is the wave amplitude. For waves approaching the beach in a direction normal to
the beach Eq. (3.71) reduces to
+Â¿(C,M' = 0 (3.74)
Multiplying Eq. (3.74) by the complex conjugate of Aand add to it the complex conjugate
equation of (3.74) multiplied by A', yields
[C,>l2]s = 0 (3.75)
which is Eq. (3.73) for the special case of cos# = 1. However it is not possible to obtain
Eq. (3.73) from Eq. (3.71) for the general case of waves at angle 6 on a beach with straight
and parallel contours.
The importance of this inability to give the correct form of Eq. (3.73) is that if energy
flux outside the breaker line is not conserved then gradients of the Sxv radiation stress
44
will generate currents offshore of the breaker line in the circulation model. Although
such currents in general will be of a small magnitude, they can cause some numerical
computational problems. The problem though is easily corrected by using as the dispersion
relationship
u) = a + k cos 9 U (3.76)
and the substitution
_*.A(x,y) j(f kcottdx)
(3.77)
Rather than using a splitting method, the procedure will be to go directly from Eq. (3.51)
using Eq. (3.77) and writing
u>2 â€” a1 = 2uik cos 9U â€” (k cos 6 U)"1
(3.78)
in place of Eq. (3.55). Dropping the (j)zx term and all terms containing products of the
current components that cannot be reduced as explained above results in the following
parabolic equation.
Cg cos 6 + U'
(Cgcos 9 + U) Ax + 
At
A + V Av +
5(7),
cos29)A i CCg(Â£)â€žj +^A = 0
Now adding a phase shift
icoiÃdzâ€”f tcostdz)
so that the free surface can be recovered by
rj(i,y) = AV'(/tco,fdz)
(3.79)
(3.80)
(3.81)
where 6 is the wave angle obtained by Snellâ€™s law in the absence of any obstructions or y
variation of the depth or current. The following governing equation is thus obtained.
CgCosO + U'
(Ct cos 6 + U)A'X  t(k cos 0  k coe 0)(Ct cos 9 + U)A' + â€”
At
+VK + \ (Â£) A'   coS! H)A'  j [cc.(^)
A'
+ ^A' = 0 (3.82)
45
Details of the algebra and assumptions involved in obtaining Eq. (3.79) are given in ApÂ¬
pendix D. For computational purposes it will be assumed that cos# = cos#.
From Eq. (3.82) the correct form of the conservation of wave action can be obtained.
For the case of no y variation of the wave number Eq. (3.82) reduces to
(CtcoB6 + U)A'x + 
Cg COS # + U
A' + VA[
1+5Â©/
â– i*C,(lcos!9)A'Ã[cC((^)â€ž] +A = 0
(3.83)
Multiplying Eq. (3.83) by the complex conjugate of A and adding to it A times the complex
conjugate equation of (3.83) yields
(C, cos# + U)
+
\V\A\2]
a
X
a
= 0
(3.84)
which is a closer approximation to the exact expression which should have a (C, sin 0 + V)
term instead of V in the y derivative term.
3.3 Energy Dissipation Due to Wave Breaking
Kirby (1983) found a method of relating the dissipation coefficient w to the work of
Dally (1980) who proposed that the decay of energy flux in the surf zone was proportional
to the amount of excess energy flux. The excess energy flux is the difference between the
actual flux and a stable energy flux. In mathematical terms this relationship is expressed
as
(ECt)x = j[ECg(ECt).] (3.85)
where
EC,
E
Câ€ž
K
h
the energy flux
wave energy = ^pgH2
the rate of energy propagation which is â€”
ok
constant
the water depth
46
(ECt)t = the â€˜stableâ€™ energy flux.
H = the local wave height
The stable energy flux is obtained by assuming it to be a function of the height of a wave
propagating over a flat bottom after breaking ceases, and that this height is a constant
times the water depth, (y) = 7. The constants K and 7 were determined by analysis of
experimental data obtained from Horikawa and Kuo (1966).
Kirby related Dailyâ€™s work to the dissipation term w by deriving a conservation equation
for the wave action Garrett (1967) and Bretherton and Garrett (1968) showed that for
waves propagating in a moving and variable medium, the quantity that is conserved is the
wave action (7) > which is the wave energy divided by the intrinsic frequency. Kirby derived
a wave action conservation equation from the linear hyperbolic wave Eq. (3.34) using the
definition of W as given by Eq. (3.35):
^ + (Vfc â€¢ u)^  Vk(CC9Vkj) + (
Kirby follows the approach of Jonsson (1981) assuming that * can be represented as
t* = Reie (3.87)
where R and 0 are reeil. From linear theory it is obvious that
R â€” â€” A. (3.88)
G
To lowest order the derivatives of Â© are approximated as
0< = w (3.89)
Vfc0 = k (3.90)
Substituting into (3.86) gives the complex equation
n2p DR
bÂ¡t + (VÂ» â€¢  vÂ»(cc,vkR)  + 2Â»Â«,)
;{vA (0o)R+2{o0) â– VhR + Vâ€ž â€¢ (kCC,)R + 2kCC, â– VkÃ¼l
â€”lowif) = 0 (3.91)
47
Since R and 0 are real, the reed and imaginary parts of (3.91) must individually equal zero.
Taking the imaginary part and multiplying by R gives
V* â€¢ (kCCg)R2  awR2 =
(3.92)
which using the substitution kCCa = aCt is written as
{(rR2)t + Vj, â€¢ (aR2[U + Cg)) = awR2
Using (3.88) R2 may be replaced by
Ri = = 2p
a2 a \a J
which yields the wave action conservation equation
Assuming a time steady wave field and neglecting currents, (3.95) becomes
(3.93)
(3.94)
(3.95)
(ECg)x = w E
(3.96)
Comparing (3.96) with (3.85) and noting that [Cg), = Cg, w is written as
â€˜if)
(3.97)
where H is the local wave height given by 2A.
In obtaining Eq. (3.97) Kirby neglected the effects of the currents. However, the same
expression (3.97) is obtained when currents are included in the derivation. Dally (1987)
extended his work to investigate wave breaking on currents and proposed the following
equation:
+ U) (f )]i = ~Yo [EC' ~ [ECl)t] (3 98)
For the case of a steady state wave field propagating in the x direction the conservation of
wave action Eq. (3.95) reduces to
<^> (!)],=Â»(!)
(3.99)
48
Setting the right hand sides of Eqs. (3.98) and (3.99) equal to each other results in Eq. (3.97).
At first glance it may appear unusual that the same expression for the dissipation
coefficient results for the case of currents as for the case of no currents. This can be
explained in part by the fact that Dailyâ€™s work was with waves where breaking was depth
limited. For the case of current limited breaking (for example a stopping current), another
expression is necessary. In this study Eq. (3.97) will be used, since in general the current
field will be moderately weak. Such an assumption of weak currents was invoked in the
derivation of the governing parabolic equation.
CHAPTER 4
FINITE DIFFERENCE SCHEMES AND BOUNDARY CONDITIONS
4.1 Numerical Solution of the Governing Equations
The finite differencing of the governing equations in the circulation model is derived
by a matrix analysis as suggested by Sheng and Butler (1982). A rational analysis of the
complete Eqs. (2.12.3) including all the nonlinear terms is not possible. However guidance
can be obtained through an analysis of the linearized equations. For this purpose the linear
equations for the special case of a flat bottom and without any of the forcing terms will be
examined.
fÃÃ± an f)v
(4.1)
(4.2)
(4.3)
Let W be a vector of fj,U and V
dfj au dV n
7T + Dâ€” + Dâ€” = 0
dt dx dy
au dfj
ar + sai = 0
dL + gdA = 0
at ydy
w =
n
u
V
(4.4)
Using Eq. (4.4), the governing equations are written as
Wt + AWz + BWV = 0
where A and B are the following matrices
(4.5)
â€™ 0
D
0 â€˜
' 0
0
D '
A =
9
0
0
and
B =
0
0
0
0
0
0
. 9
0
0
(4.6)
To find an implicit finite difference scheme for the governing equations, Eq. (4.5) is
finite differenced as
jyn+i _ jyn
AT
+ AW?+1 + BWf+1 = 0
(4.7)
49
50
This equation is solved for Wn+1 in terms of Wn
(1 + 2AZ + 2A y)Wn+l = Wn
(4.8)
where
1 AT
= 2AAX{
and
Aâ€ž = B
1 â€žAT
2 AY
(4.9)
4Aj;AvWn+i is added to the left hand side of the equation and 2AyWn is added and subÂ¬
tracted to the right hand side of Eq. (4.8) so that the equation can be factored to get
(1 + 2A.)(1 + 2Xv)Wn+1 = (1  2Av)Wn + 2XvWn (4.10)
An intermediate value W * is introduced so that the above equation can be written as
two equations
(l + 2AI)W* = (l2Aâ€ž)Wn
(4.11)
(1 + 2Xv)Wn+1 = W* + 2X vWn
(4.12)
Note that by eliminating W* from the two Eqs. (4.11) and (4.12), Eq. (4.8) is recovered
with the addition of an error of 4A2AyWn+l which has been introduced.
Writing out the component equations of Eq. (4.11) gives
AT AT
V* = yn9^vfjn
(4.13)
(4.14)
(4.15)
(4.16)
While writing out the component equations of Eq. (4.12) gives
Vn+1 + = fj* + D^6yVn
un+1 = u*
yn+1 , c Â¿Â¡n+1 _ yÂ» . c =n
v +9AY6yV ~v +gAY6^
(4.17)
(4.18)
(4.19)
51
Using Eq. (4.15) to eliminate V* in Eq. (4.19) and (4.18) to eliminate U* in Eq. (4.14) the
following are obtained. For the x direction
AT AT
Â«* + daxSiU = â€œ days,vâ€™'
AT
Un+1 + gâ€”6xfÂ¡* = Un
and for the y direction
AT AT
r+1 + v'"+1 = Â«â€¢ + D^e,vn
AT
yn+1 + g^8vfjn+l = Vn
(4.20)
(4.21)
(4.22)
(4.23)
The above equations are numerically stable for any combination of AT, AX, and AY.
However as a circulation model they are incomplete in that they do not contain any of the
forcing terms that drive the circulation. In particular in this model wave forcing represented
by gradients in the radiation stresses and bottom friction are of concern. Bottom friction is
especially important because at steady state conditions gradients in the radiation stresses
must be balanced by either gradients in the water surface elevation or by bottom friction
forces. In the absence of friction forces the model could yield continuously accelerating
currents. The method of adding the forcing terms is somewhat arbitrary in that the rational
derivation of the finite differencing scheme of Sheng and Butler cannot be expanded to
include these essentially non linear terms. The addition of these terms to the model also
limits the latitude of the ratio of jn a manner that is not analyzable. That both the
friction and the radiation stresses are nonlinear should be recognized. The nonlinearity
of the velocity friction term is quite evident since it is represented by the quadratic term
pF\U\U. The radiation stress terms, which represent excess momentum flux induced by
waves, are functions of the wave height squared, based on linear wave theory. In shallow
water the wave height H is a function of the total water depth, h0 + fj so that H2 is a
function of fj2.
All the forcing terms and the nonlinear acceleration terms are added at the present or
time level with the exception of the bottom friction term. The quadratic friction term
52
is finite differenced as pF\Un\Un+1 and the friction term as suggested by LonguetHiggins is
differenced as pF uofj Un+1, where uorj is in terms of values at the n time level. However,
when using the integral quadratic expression, it is evaluated entirely in terms of the present
or known values.
The nonlinear advective acceleration terms are added at the nth time level using central
difference derivatives. In the x direction momentum equation these terms are
TT + Uiâ€”lJ ^ AT/" I V i TA i T/ T + l â€” l
U,J 2AX h ^y
(4.24)
and in the y direction equation they are
jWj + u<u + fuÂ» + 'â€” (4'25)
In the model of Vemulakonda (1984) which is based upon the WIFM (Waterways ExÂ¬
periment Station Implicit Flooding Model) (Butler 1978) model, a three time level leapfrog
scheme is used. This gives the advantage that all of the forcing terms (and the nonlinear
acceleration terms) are added at the time centered level. However this has some disadÂ¬
vantages in that additional computer memory and computational time are required and
leapfrog schemes can exhibit instabilities. In the present study a two time level scheme
is used and the wave forcing terms as well as the nonlinear acceleration terms are added
at the present time level. Since, for all of the applications in this report, the model is run
until steady state conditions are obtained, it is reasoned that there is no great necessity to
properly time center all of the terms. If the use of the model is extended to time dependent
applications, this may need to be corrected.
4.2 Finite Differencing of the Parabolic Wave Equation
Dividing Eq. (3.82) by (CycosÃ + U) yields
f C, cos B+U j
A!z + i(k cos Ã“ â€” k cos 6) A1 + ^ ^
;A' +
+
2 (C,cosÃ+ 17) (C,cosÃ + 17) v
, i kCa(l â€” cos2 Ã) A,
v 2 (Ct cosÃ + U)
(Y.) LA
2(Cycosi + i7) \a )
2(Cg cosÃ + U)
CC,
4A
+
w
2(CjCosÃ + U)
A' = 0
(4.26)
53
The following notation is used to describe the finite differencing of Eq. (4.26): Fj
represents the value of F[i,j) where F is any term and i and j designate the grid row
and grid column. Because of the staggered grid system used for the circulation model as
previously explained, in which the velocity components are specified at the grid edges, the
velocity terms U and V are averaged across the grid so as to produce a grid centered term
for use in the wave model. Thus 17) in the wave model is actually \[Uj + Â£^â€™+1) and similarly
for V and other grid designations.
A CrankNicholson finite differencing scheme is used for Eq. (4.26). All terms not
involving an x derivative are an average of the value of the term on the * grid row and on
the i + l grid row. Terms having y derivatives are centered atÂ», j on the i row and Ã +1, j
(X'+1â€”X'.)
on the Â« + 1 grid row. The one term with an x derivative is written as â€”* ^ 1 . Thus all
terms are centered at Â» + ,j. The finite differencing of Eq. (4.26) is written as
1
+ 2
Â£lzA
AX
+ ' [(jfcâ€™+1 cos 0,+1  fcâ€˜+1 cos 9i+1)A$+1 + (Â£â€™ cos 0*  k) cos 0â€˜)A']
1 f [(C7g cos g + C/)/o]Â«>1 ~ [(Cg cos g + f7)/cx]Â«
h 2AX \ \[Ca cos 0 + U)/(r]fl + \[Ca cos 0 + U)/o
Vcr
(Ce cos 6 + U)
r
 A;ii
2AY
l}(4+â€˜+4)
f Vo V a;+i  a;.,]
\[Cacos6 + U)J. 2AY
1
+ 2
2(Cycos0 + U)
[Ca COS0 + U)
<+' r(V/Â»);tt  (V/a))t\
i
+
2(CÂ¡ cos 0 + 17)
)!(
2AY
w*)Ui  PH1
2AY
4
t
4
kCg{l cosJ0)l*+1 A<+1 +
(Cg cos 0 + U)
kCg{ 1 â€” cos2 0)
â– (#),
[Cg cos 0 + U)
â€¢+l
P
Â«'+!
V]
(Â«),
+
4+1
V)
[Cg cos 0 + i/);+1 (C, cos 0 + Â£7)â€˜
/
1 Ã w*+1 A,+1 w* A\ )
+ 4 { (Cycos0 + Â£0*+Â» + [Cg cos0 + U)) J Â° ^'27)
54
where p = CCg for notational convenience and the y derivative of p ^ j is written as
â€˜ (rUi + pj) [(Â£)â€™+, ~ (Â£))]  (Pj + p;~i) [(f)â€™  ($)'â€ž
2(AY)2
(4.28)
In order to evaluate w)+1 by the method to be explained in section 3.3 an intermediate
value of A)+1 is obtained by the explicit formulation
A,+1 â€” A*
3 Ã + i(k* cos O'  k'j cos Ol)A'j
AX
1 Ã [(Ct cos 0 + V)/<,])+'  \{CÂ¡ cos 0 + U)/a]
1 [icâ€ž
' AX 1 [(C, cos 0 + U)/afAi + [(Cy cos e + )/a]
(
!H
(Cg cos 0 + U)/a
j+i
2AY
+
(  V
(YMUi 
A* *
kCt(l  cos2 0)
\2[Cg cos 0 + U) J .
2AY
1 2
(Cpcosi + i/)
AÃ
n
V)
T +
2 (Cg cos 0 + U);. _r 2(Cg cos 0 + U)) Â°
When Eq. (4.27) is written out using Eq. (4.28) and w*+1 obtained through use of
(4.29) there results for any particular values of i and j an equation for the three unknown
values A)+\, A)+1, and Alt.\. The equation is put into the form
ai aYA + hi 4+1 + Cj Aâ€™ti = dj (4.30)
where the coefficients aÂ¡, bj, and c; are known quantities involving the properties of the
environment (depth, current, celerity, etc.) and the dj involves the environmental properties
and the known values of the amplitude on the Â» grid row. When Eq. (4.30) is written for
all the j values, j equal to 1 through N, for a specified i value, the ensemble of N equations
takes the form of a tridiagonal matrix equation. For illustrative purposes let N be equal to
four. The resulting matrix equation is
r^i
aj ci
Aâ€™+1
â€˜ d\ â€˜
a2 &2 c2
4+1
d2
03 b$ Cj
A*3+1
dz
Â£
*
*
i
4+1
. d* .
L 4+1J
(4.31)
55
Equation (4.31) contains four simultaneous equations with six unknowns. Two boundÂ¬
ary conditions should be specified. At each boundary a functional relationship is established
between the grid inside the domain and the grid outside of the domain. For the wave propÂ¬
agation model there are two types of boundary conditions that are readily available for use
in a computer model. These are a totally reflective boundary condition which would be
used to model a laboratory experiment in a wave basin, and nonreflecting noninterfering
boundary which will allow waves to freely pass through the boundary. For the reflective
boundary, the mathematical statement of the boundary condition is ^ = 0. At the lower
boundary this is finite differenced as = 0 which gives the functional relationship
A\ = A0. (4.32)
At the upper boundary the functional relationship for a reflective boundary is
An+i = An.
(4.33)
For the nonreflective boundary the mathematical statement is = â€”imA where m is
the longshore component of the wavenumber. Since the lateral boundaries are sufficiently
far from any diffraction effects, only refraction and shoaling effects are present. Thus
using Snellâ€™s Law m = ksin# = kosin0o. The non reflective boundary condition is finite
differenced at the lower boundary as
â€”Â£y~~ =Y[Al + AÂ°] (434)
which gives
Aq =
fj.ÃºnAY)
1 +
tmAYA
2 )
Ai
(4.35)
Similarly at the upper boundary the functional relationship used to eliminate Ajv+i is
(l+ imAY)
AN+1 = (! ,mAY) An (436)
The use of either pair of boundary conditions, (4.32) and (4.33) or (4.35) and (4.36) to
eliminate Aq and A;v+i will then yield an N by N tridiagonal matrix equation comprising N
56
Figure 4.1: Definition sketch of grid nomenclature and coordinate system.
equations for N unknowns. Returning to the illustrative example where N is equal to four,
the matrix Eq. (4.31) becomes
6i + 7\d\ ci
02 f>2 C%
O3 63
CS
' A\+1 '
â€˜ di '
A'2+1
Â¿2
4+1
Â¿3
[ 4+1J
. d< .
(4.37)
where 7\ and 7n represent the functional relationships as described above on the j= 1 and
jâ€”N boundaries.
Equation (4.37) is solved easily for the unknown AJ+1 using a double sweep method.
In the computer program the subroutine CTRIDA uses an algorithm of Carnahan, Luther,
and Wilkes (1969) to efficiently accomplish the double sweep solution of eq. (4.37).
4.3 Boundary Conditions in the Circulation Model
As shown in figure 4.3 a staggered grid system is used so that velocities are expressed
at the grid edge and fj and the wave properties are defined at the grid center. So that the
wave model will have full grids it is desirable to place the boundaries of the model at the
grid edges on all sides, with the exception of the offshore grid row where the wave conditions
are an input or boundary condition for the wave model. Figure 1.1 can also be referred to.
The offshore boundary condition that is used in the circulation model is that the mean
57
water surface level is not changed by any of the circulation within the model, mathematically
ETA(1,J) = 0 J = 1, N
(4.38)
This condition necessitates that the offshore boundary be sufficiently far from any setup or
setdown. Other possible offshore conditions are a noflow condition or a radiation condition.
The no flow condition would conserve water within the computational domain so that the
volume of water involved in set up would be compensated by an equal volume of setdown
offshore. In an open ocean environment the open ocean can be conceived of as an infinite
reservoir. However in modeling laboratory experiments a no flow condition as the offshore
boundary condition would be the most appropriate. A radiation boundary condition as
proposed by Butler and Sheng (1984) imposes the condition
dr) dr)
â€” + C â€” = 0
dt dx
(4.39)
This condition allows long waves to freely propagate out of the computational domain
without any reflection. This tends to eliminate any seiche within the computational domain
which is induced by the startup transients. This condition was applied at one time in the
present model but it did not preserve the offshore zero setup condition and was changed in
favor of the more rigid zero water surface elevation condition. If the model is adapted for
use with time varying wave trains (an adaptation that the model can easily accommodate)
then the radiation condition will become imperative.
At the shoreline the boundary condition is that there is no flow into or out of the beach.
Since the model allows for flooding of the beach the index of the shoreline changes. The
model allows for differential flooding along the shoreline, for example, due to differences of
wave intensity in the lee of structures. For each grid column in the y direction the last grid
is the IWET(J) grid, where IWET(J) is allowed to change with changes of the mean water
level. The numerical statement of the no flow boundary condition at the shoreline is
t/(IWET(J)+l,J) = 0
J=1,N
(4.40)
58
This boundary condition is used in other models, though not all models allow for differential
flooding of the shoreline. Yan (1987), whose model was used to investigate waves and
currents around a tidal or fluvial discharge, used the no flow condition away from the
discharge and defined Â£/(IWET(J)+l,J) as equal to the discharge per grid divided by the
water depth at the grids where discharge occurred.
Other conditions are also imposed at the shoreline boundary. These are that the wave
height and the longshore velocity are both zero. These conditions are imposed at the
shoreward grid edge of the IWET(J) grids even though the wave height and longshore
velocities are specified a half grid space form the shoreline. This is done by writing x
derivatives of the radiation stress terms and of the V velocities that need to be positioned
at the center of the IWET(J) grid in terms of three points. The three points are at the
IWET(J)1 and IWET(J) grids and the zero value at the shoreline at the edge of the
IWET(J) grid.
For the lateral boundary conditions several alternatives are available. Birkemeier and
Dalrymple (1976) and Ebersole and Dairymple (1979) both used a periodic or repeating
beach. This meant that computations were performed between the second and grid,
and values outside of the computational domain are determined according to
Q(I,1) = Q{I,N) (4.41)
Q{I,N + 1) = Q{I,2) for all I (4.42)
for any quantity Q.
At one point this was used in the present model until the problem of what should have
been transient edge waves generated by wave obstructions during the startup were found
to be trapped in the model. The periodic boundary condition allowed for disturbances
propagating out of the domain at one lateral boundary to enter the domain at the other
boundary. For straight and parallel contours, the periodic beach condition worked well in
the present model but with the presence of breakwaters another condition had to be used.
For the lateral boundary conditions there are two choices, either closed or open bound
59
aries. Closed boundaries are easily applied by the condition of no flow across the boundary.
Such a choice of boundary conditions would accurately duplicate the physical conditions exÂ¬
isting in a laboratory basin (or for prototype conditions within a harbor). Open boundaries
are obtained by specifying derivative conditions. In the present study an open boundary
is obtained by specifying that there are no gradients of the depth and the wave properÂ¬
ties across the lateral boundaries. This necessitates that the lateral boundaries are placed
sufficiently far from any obstructions or any y gradients caused by the obstructions. An
alternative open boundary condition would be that there is no gradient of the flow. This
would, however, put the position of the boundary condition at the center of the bordering
grid rather than at the edge of the grid and could possibly introduce error at the boundary.
4.4 Boundary Conditions in the Wave Model
In the wave model only lateral boundary conditions are needed for computational purÂ¬
poses. The offshore condition is merely the specified wave amplitude and wave angle. No
condition is needed at the shoreline since the parabolic wave model marches grid row by
grid grow and needs only the lateral conditions.
The offshore condition is that each time the wave model is called the wave amplitude
on the first grid row is specified as
A(1,J) = C(t) ~~A0(J) J=1,N (4.43)
where A0(J) is defined in the input subroutine according to
A0(J) = eâ€˜*osin 0oJaY (444)
where k0 is the wave number along the first grid row and 90 is the angle the wave ray
makes with the x axis. This offshore condition necessitates that the water depth be constant
along the offshore grid row. This is another reason for choosing the offshore condition in
the circulation model to be a constant zero. With the presence of wave obstructions and
current deflectors such as jetties it is possible that with other offshore conditions the mean
water surface elevation could vary along the offshore grid row. The use of one condition in
60
o
Figure 4.2: Startup function according to Eq. 4.43 with Ci=30 and C2â€”3.
the wave model which is contrary to the condition in the circulation model can introduce
instabilities in the overall model. The interaction of the two models is more sensitive than
either of the models individually.
The C(t) term in equation 4.43 is a startup function which smoothly goes from zero
to unity so as to bring the incident wave height from zero to H0 and minimize any shock
loading of the model. Mathematically it is given by
C(t) = .5
t
1 + tanh(â€” C2)
w
(4.45)
where C\ and C2 are arbitrary constants to adjust the slope and offset of the curve. FigÂ¬
ure 4.2 shows equation 4.45 for C\ and C2 equal to 30 and 3 respectively.
The lateral boundary conditions are either open or reflective. The reflective condition
is expressed as
(4.46)
while the open condition is expressed as
= â– 'â€¢"A (4.47)
where m is the longshore component of the wavenumber which in the absence of diffraction
effects remains constant
m = k0 sin 60 = k sin 9
(4.48)
61
Direction of
Wave Propagation
x
Figure 4.3: Waves passing through a lateral boundary unaffected by the boundary.
according to Snellâ€™s law. Equation 4.46 or 4.47 are used to eliminate the amplitude outside
of the domain from the tridiagonal matrix equation. To illustrate equation 4.47 is finite
differenced at the lower lateral boundary as
A(I,1)  A(I,0) _ . A(I,1) + A(I,0)
AY im 2
Solving for A(I,0)
i _ imAY
A(I,0) = tVa(I,1) (4.50)
1 + im^
Thus A(I,0) is eliminated in terms of A(I,1). similarly at the upper boundary A(I,N+1) is
eliminated in terms of A(I,N). The same procedure is used with equation 4.46 which yields
the simple relationships
A(I,0) = A(I,1) (4.51)
and
A(I,N+1) = A(I,N)
(4.52)
62
It is important that the boundary conditions 4.46 and 4.47 be used as prescribed
above. The parabolic equation gives N equations containing the N+2 unknowns A(I+1,J)
for J=0,N+1. The two boundary conditions are thus used to reduce the number of unÂ¬
knowns to a number equal to the number of equations. If equation 4.47 is written in terms
of A(I+1,1) and A(I+1,2) at the lower boundary and in terms of A(I+1,N1) and A(I+1,N)
at the upper boundary and the governing wave equation is written for J=2,N1, then errors
may be introduced at the boundaries. This is especially true for the case of a reflective
boundary. Figure 4.3 shows waves freely passing through the lateral boundaries without
any noticeable reflection for the case of a plane wave with a period of 10 seconds propagating
over constant depth at an angle of 10 degrees to the x axis.
CHAPTER 5
RESULTS
The numerical model combining the iterative interaction of the circulation model and
the wave model as described in the previous chapters was tested for several cases in which
either analytical or experimental results are available. The model was also run for cases
where previous numerical results are available.
The two cases involving analytical solutions are for setup induced by normally incident
waves on a plane beach and longshore currents generated by incident plane waves at an angle
on a plane beach. The first section of this chapter covers setup, comparing the results
with the analytical solutions of LonguetHiggins and Stewart (1964) and the experimental
results of Bowen et al. (1968) and Hansen and Svendsen (1979). The second section covers
longshore currents and compares results with the analytical solution of LonguetHiggins
(1970a,b) and experimental results of Visser (1982).
The model was run for the case of a ripchannel on a plane beach. Comparisons with the
numerical results of Birkemeier and Dairymple (1976) and Ebersole and Dalrymple (1979)
are given in the third section. A rip current was also generated for the case of normal waves
breaking on a shore parallel bar with a gap. No comparisons are available for this case.
The model was run for the cases of shore perpendicular and shore parallel breakwaters
represented in the model as infinitesimally thin impermeable barriers. For the shore perÂ¬
pendicular barrier, the results are compared with the experimental results obtained at the
Waterways Experiment Station as reported by Hales (1980). For the shore parallel barrier
the results are compared with the numerical results of Liu and Mei (1975) who did not
include a wave current feedback mechanism.
In the final section the model was used to simulate the laboratory experiments of
63
64
Gourlay (1974) in which the shoreline in the lee of a breakwater was made to curve and
meet the breakwater as occurs with tombolo formation. The results of the model were in
close agreement with Gourlayâ€™s experimental values.
5.1 Wave Setup
LonguetHiggins and Stewart (1964) gave a simple but elegant derivation for the wave
induced setup on a plane beach assuming linear shallow water theory. For the steady state
case of normally incident waves on a plane beach without any y variations and without any
currents the depth integrated equations of motions reduce to
gdi + Â±as~=0
(5.1)
dx pD dx
In shallow water for normally incident waves the radiation stress term from linear wave
theory, Eq. (2.28), becomes
Sâ€ž = E(n+l) = ^=^Wff2
(5.2)
A spilling breaker assumption is made so that the wave height if is a constant ratio k to
the water depth D = h0 + fj,
H = k (h0 + fj) (5.3)
Substituting equations (5.2) and (5.3) into Eq. (5.1) gives
5rj _ _ 3 dK2(h0 + fj)1 _ 3 2 fdhQ dfj\
dx 16(h0 I fj) dx 8* V dx + dx)
Solving for gives the result
dfj _ /c2 dh0
dx 1 + /c2 dx
that the slope of the water surface inside the breaker line is a constant times the slope of
the beach.
Previous models are able to produce the results of the analytical expression but to
date none, including the present, have matched the experimental results. In figure 5.1 it
is seen that the slope of the mean water surface produced by the wave setup is nearly
65
DISTANCE FROM STILLWATER LINE
IN BEACH, x CMS
3.0
2.5
2.0
1.5
CMS
1.0
IS?
0.5
0
0.5
Figure 5.1: Comparison of the numerical solution of Vemulakonda et al. (1985) for setup
with experimental data of Bowen (1968). Source: Vemulakonda et al. (1985)
a constant. For the numerical results the setup starts at the breaking point, whereas in
the experimental results the water surface elevation is nearly constant for some distance
shoreward of the breaking point and then there is a change in the slope of the water surface
elevation. The numerical results follow the analytical solution. This is an area that needs
further investigation. The reason for the difference between the analytical and experimental
results is that the analytical formulation is predicated upon the radiation stress term being
a function of the wave height, so that as the wave height is reduced after the breaking point
so is the radiation stress term. As pointed out by Svendsen (1984) the radiation stress term
remains nearly constant after breaking for a distance roughly equal to five or six times the
breaking depth. This distance is termed by Svendsen as the outer or transition region in
which there is a rapid change of wave shape. In the inner region the wave shape according
to Svendsen remains nearly constant and is characterized by a surface roller or bore. By
incorporating the surface roller Svendsen is able to obtain close agreement for wave height
and setup to the experimental results of Hansen and Svendsen (1979). However this close
66
agreement is just for the inner region. For the outer region there is no viable solution other
than that the radiation stress term and the water surface elevation remain nearly constant.
Further there is no analytical separation of the outer and inner regions other than the a
po8teri definition of where there is a sharp change in the slope of the water surface elevation.
This definition is not useful in a model that seeks to determine the water surface elevations.
Svendsenâ€™s work shows promise and should be pursued; it is not however at this point
viable for inclusion in a numerical circulation model. For one it needs to be expanded for the
case of waves at a angle (this need will be discussed further in the next section). Another
limitation is that Svendsenâ€™s model is for monotonically decreasing depths, which is a severe
restriction (for example excluding barred profiles).
Svendsen conceptually addresses the question concerning the spacial difference between
the initiation of the setup in the analytical and experimental results. The analytical model
assumes that once breaking commences energy dissipation also begins. Svendsen states that
the organized potential energy of the wave prior to breaking is transformed into forward
momentum in the outer region. In keeping with the definition of the radiation stresses as
the flux of excess momentum, i.e. momentum that would not be present in the absence of
the wave, it is easy to conceptualize that the radiation stress term should remain nearly
constant just after breaking. The problem is in assuming that the wave form in the breaking
zone is characterized the same as outside the breaking zone. The author, at present, offers
no solution other than to recognize the existence of the problem.
In the present study results for the setup on a plane beach with normally incident
waves differ somewhat slightly from the analytical solution as given by LonguetHiggins and
Stewart. This is because the present model does not use the spilling breaker assumption.
Instead wave energy is dissipated according to a model suggested by Dally (1980) which is
explained in section 3.3. Thus the slope of the mean water surface is not a constant times
the bottom slope. A result from the experiments of Hansen and Svendsen (1979) show this
clearly in figure 5.2. The figure shows a slight curvature in the wave height after breaking
67
mm
Figure 5.2: Comparison of numerical results from the model with experimental results of
Hansen and Svendsen (1979). Wave setup and wave heights for a 2 second wave on a beach
slope of 1:34 and incident wave height of 7 centimeters.
and in the slope of the setup. Although the problem of where the setup begins is still
present, the curvature of the mean water surface is the same for both cases. It is also quite
clear that the diminution of the wave height is not constant as is assumed by the spilling
breaker assumption.
In figure 5.2 the breaking wave height in the experiment is much greater than in the
model. This is because linear wave theory cannot adequately predict wave shoaling. Ven
gayil and Kirby (1986) were able to duplicate the shoaling recorded in the experiments of
Hansen and Svendsen (1979) by numerically solving coupled equations for the many harÂ¬
monic components. Their results showed that in the shoaling process prior to wave breaking,
wave energy was transmitted to the higher frequency components. Their work however is
not adaptable for use in the current model since it is not known how to dissipate energy
68
Figure 5.3: Nondimensional longshore velocities vresus the nondimensional distance offÂ¬
shore from the analytical solution of LonguetHiggins. Reprinted from LonguetHiggins
(1970b), Journal of Geophysical Research, vol. 75(33) page 6793, copyright by the AmeriÂ¬
can Geophysical Union.
from the various frequency components once breaking is initiated.
5.2 Longshore Currents
LonguetHiggins (1970b) gives an analytical solution for the longshore current generated
by obliquely incident waves on a plane beach. The solution involves a spilling breaker
assumption and assumptions concerning the form of the bottom friction and the lateral
mixing. For steady state conditions the y equation of motion reduces to a balance between
the gradient of the Sxy term and the bottom friction with the lateral mixing serving to
distribute the current. Details of the solution can be found in the cited paper.
Figure 5.3 gives the results of the analytical solution of LonguetHiggins. In the figure
the ordinate V is the nondimensional velocity v/v0, where t; is the actual velocity and v0
is the velocity at the breaking point obtained for the case of no lateral mixing.
57t. , .isinfl .
v0 =Â¿(fffis)i (5.6)
6 C
The abscissa X is the nondimensional distance z/zj, where zj is the distance from the
69
Wave SetUp Line Breaker Line
Figure 5.4: Comparison of experimental results of Visser (1982) and numerical result for
longshore currents due to a 1.1 second wave on a slope of 1:20 for the deep water conditions
H0 â€” .065m and 60 = 20 degrees.
shore to the breaker line.
Previous models have reproduced the results of the Longuet Higgins solution by incorÂ¬
porating all of the same assumptions (i.e. spilling breakers and bottom friction). Since the
present model does not contain the same assumptions comparison will not be made with the
analytical results of LonguetHiggins. Instead comparison is made with the experimental
results of Visser (1982). For both the analytical and the numerical results it should be noted
that the same inadequacy is present as was discussed in the previous section. In both cases
it is implied that wave energy dissipation commences as soon as the wave begins breaking.
Thus both analytical and numerical solutions initiate a longshore current starting at the
breaker line. It is reasonable to expect that the same holds for the y component of the
onshore flux of excess momentum as for the x component. That is, that the Sxv term also
remains nearly constant over a certain unspecified distance as the organized momentum flux
of the non breaking wave becomes turbulent momentum upon breaking. This is illustrated
in figure 5.4 which shows a comparison of the experimental results of Visser (1982) and the
model results for the same test conditions. The maximum current is nearly the same in
both the model and the experiment but the maximum is further offshore in the model. This
can be attributed to the initiation of a longshore current in the model at the breaker line.
70
5.3 Waves On a RipChannel
Both Birkemeier and Dalrymple (1976) and Ebersole and Dalrymple (1979) examined
the case of waves incident upon a shore with the presence of a ripchannel. The bottom
topography was given according to a formula first used by Noda et al. (1974)
h(x,y) = mx jl + Aes^)1/S sin10 ^(y  ztan^) (5.7)
where ft is the angle the ripchannel makes with the x axis which is, in their coordinate
system, directed offshore from the still water line, m is the beach slope, A is an amplitude
of the bottom variation and A is the length of the repeating beach. The origin of the
coordinates is at the intersection of the still water line and the trough as is shown in
figure 5.5 which gives the depth contours for the case where the slope is .025 and ft is equal
to 30 degrees.
Due to the restriction of the present model that there be no y gradients at the lateral
boundaries the above formulation is modified to be
h(x ) = f mx{nAes(^)1/S8in10f(yztan^)}  J(y  xtanÂ£) < x
V 1 [mi  j(y â€” x tan ft)\ > x V â€™ '
Also shown in figure 5.5 are the contours obtained through use of the modified Eq. 5.8
for ft equal to 30 degrees and a .025 slope. The two formulas differ only in that the repeating
rip channels in the Noda formulation are removed. For application in the present model the
presence of the trough in the farthest offshore grid row is removed so that the depth in the
offshore grid row remains constant in compliance with the requirements of the boundary
conditions as discussed in sections 4.3 and 4.4.
Figure 5.6 shows the results obtained by Ebersole and Dalrymple for the case of a wave
angle of 30 degrees a wave height of 1.0 meters and a period of 4 seconds. Figure 5.7 gives
the results for the present model. Even though the scales of the two figures are different it
is evident that there is agreement between the two. The length scale of the eddy is about
30 meters for both cases. In the caption to the figure of his results, Ebersole indicated that
the lateral mixing coefficient may have been large. The present author does not believe
71
Figure 5.5: Top: Depth contours, in meters, using Eq. 5.7 with the coordinates as shown.
Bottom: Depth contours as used in the present model. Grid spacing is 5 meters.
72
Figure 5.6: Velocity vectors from the results of Ebersole and Dalrymple for a wave angle
of 30 degrees. Scale for velocity vectors is 1 inch = 2.87 m/sec. Source: Ebersole and
Dalrymple (1979).
73
Figure 5.7: Velocity vectors obtained from the present model using the same input condiÂ¬
tions as for Ebersole and Dalrymple.
74
Figure 5.8: Mean water surface elevation contours showing that the water surface is lower
above the trough. Values indicated are in millimeters.
this to be the case. In the present study the lateral mixing coefficient was reduced to as
low a value as 2 meter2/second without substantially changing the magnitude of the eddy
adjacent to the rip trough. However Ebersole did find a significant difference between the
case of no lateral mixing and with lateral mixing. Results for the present model for the
case of no lateral mixing are not available since the model sometimes exhibits an instability
without the stabilizing influence of the lateral mixing terms. Figure 5.8 shows the mean
water surface elevation contours obtained from the model which show that the water mean
surface is lower above the trough.
The agreement between the results of the Ebersole model and the present model for
the case of waves incident upon a ripchannel should not be surprising since diffraction in
this case is small compared to current and depth refraction.
Another test result is the case of normally incident waves breaking on a bar in which
75
Figure 5.9: Depth contours in meters for the case of a bar with a gap.
Figure 5.10: Velocity vectors for the case of waves over a bar with a gap. Dashed lines
indicate depth contours. Grid spacing equal to 10 meters.
76
there is a gap in the bar. Figure 5.9 gives the contour lines for this test case. The topography
was created by adding an inverted parabola on to a pÃeme beach with a slope of 1:40. The gap
was created by suppressing the bar with a symmetric smooth tangent function. Figure 5.10
shows the velocity vectors resulting from a 6 second wave with an incident wave height of 1
meter. The contours of the bottom topography are represented by dashed lines. The driving
force of the circulation is the return flow through the gap in the bar of the water that is
carried over the bar by the breaking process. These results may not be totally correct given
the arguments of section 5.1. If there were no gap in the bar the waves breaking on the bar
would push water shoreward until sufficient setup was produced. However with a gap in
the bar lateral flow is allowed and setup cannot be maintained.
Kirby (conversation with the author) has remarked that he has observed the initiation
of a very strong current in the direction of the waves produced by waves breaking on a sand
mound in a small wave basin. This was tested in the model using normally incident waves
on a slope of .02 with a shoal described by the following equation which is the shoal used
in the Delft experiments of Berkhoff et al. (1982).
Ã.45 i < â€”5.82m
.45 â€” .02(5.82 + x) (Ã“j)1 + (Ã)1 > > 0 (5.9)
.45  .02(5.82 + x)  .s/l  (*%)â€™  (*)! + .3 (rfg)1 + (I)1 < 1.0
Figure 5.11 shows the depth contours for this test case. Figure 5.12 shows the velocity
vectors obtained for the case of a barely breaking wave. For the case of a non breaking
wave the velocity vectors plotted to the same scale are not discernable. Figure 5.13 plots
the max velocity obtained in the shoal region versus the incident wave height. At a wave
height of about 10 centimeters there is a sharp change in the slope of the curve indicating
increased velocities once breaking commences.
5.4 ShorePerpendicular Breakwater
The model is tested here for the case of waves interacting with a shoreperpendicular
breakwater. The breakwater is represented as an infinitesimally thin impermeable barrier
located in between the J = Jbrk and J = Jbrk+i grid columns and extending from
77
Figure 5.11: Depth contours for the case of normally incident waves upon a shoal. Grid
spacing is .05 meters. Contour interval is .02 meters.
78
Figure 5.12: Velocity vectors for the case of waves breaking upon a shoal.
79
Figure 5.13: Maximum velocity across the top of a shoal versus incident wave height for
waves over a shoal on a plane beach. Incident wave height less than 9.1 centimeters are
nonbreaking waves. Incident wave heights greater than 9.1 centimeters produced waves
breaking over the shoal.
80
Figure 5.14: Position of the groin in relation to the grid rows and columns.
the shore boundary to the I = Ibrk grid row. Figure 5.4 illustrates the position of the
breakwater. Mathematically, this thin impermeable barrier behaves as an internal solid
boundary that prevents any activity on one side of the barrier from being transmitted to
the other side. This requires modifications to both wave and current models.
In the wave model an internal boundary condition is imposed so that at the breakwaÂ¬
ter face wave energy is totally reflected. This is accomplished by imposing the reflection
condition
f* = o
dy
(5.10)
on both sides of the breakwater from grid row I = I brk to the shoreline, or
A[I,Jbrk) = A[I,Jbrk+\) for all I > Ibrk (511)
When the wave equation is expressed in finite difference for J = Jbrk, A[I, Jbrk+i)i which
is on the other side of the barrier, is eliminated by virtue Eq. 5.11. Likewise when the wave
equation is applied to the other side of the barrier A[I, Jbrk) is also eliminated. Further
81
reflective conditions are imposed at the barrier by assuming that the derivatives of any
property perpendicular to the breakwater has a zero value at the breakwater.
In the circulation model the interned boundary condition of impermeability at the barÂ¬
rier is imposed by setting the velocity component normal to the barrier equal to zero.
V(I,Jbrk+i) = 0 for all I > Ibrk (5.12)
To insure that currentinduced momentum is not diffused across the barrier, subroutines
were introduced in the computation for the x sweep for the two grid columns adjacent to
the barrier.
Two sample cases were run for comparison purposes. The first case is a breakwater
extending 400 meters from the shoreline on a beach with a composite slope of 1 on 10 for
the first one hundred meters and a slope of 1 on 100 beyond that point. This is a geometry
selected by Liu and Mei (1975). A deep water wave height of 1 meter and a deep water
wave angle of 45 degrees was used by Liu and Mei. Using linear shoaling and refraction
this gives a wave height of .84 meters and a wave angle of 31.15 degrees as the wave input
conditions at the offshore grid row 800 meters from shore in a depth of 16.95 meters.
The solution method of Liu and Mei was to solve for the wave field in the absence of
currents and mean water level changes and then use the wave field as a constant forcing in the
solution of the circulation. By ignoring the lateral mixing and advective acceleration terms
and assuming steady state conditions, and by expressing the two horizontal components of
the depth integrated flow as derivatives of a stream function ip, the two horizontal equations
of motion were reduced to one Poisson equation
VV = /(*,y) (5.13)
where the / term contains the forcing due to water level gradients, friction and wave forcing.
This equation represents a boundary value problem with two unknowns ip and fj. Upon
specifying boundary values an initial guess for the water surface fj is made and then Eq. 5.13
is solved for ip. With this new solution for ip the value of rj is updated by solving steady
82
state equations of motion. This process is repeated until convergence.
Figures 5.15 and 5.16 show the results for Liu and Mei and the present study respecÂ¬
tively. Figure 5.15 shows the contours of the stream function whereas figure 5.16 shows
flow lines which give the direction of the flow but not the intensity of the flow. The present
model shows a separation eddy at the tip of the breakwater which is absent in Liuâ€™s results.
Both models show a series of rip cells in the up wave side of the jetty which result from
the interaction of the incident and reflected waves. There are not as many rip cells in the
present model as the weaker ones farther upstream are buried by the stronger longshore
current. Neither the Liu model nor the present one show a rip current at the downwave side
of the jetty. In figure 5.16 the flow lines in the offshore region are directed in the onshore
 offshore direction which is different from the stream function contour lines in figure 5.15.
This is because of a difference in the offshore boundary condition in the two models. In
the Liu model the offshore condition is a constant xp which means that the flow must be
tangential to the offshore boundary. In the example used in the present model the offshore
condition is a constant water surface (fj = 0) which forces the tangential velocity to be zero
at the offshore boundary. An attempt to change the offshore condition to a zero normal
flow condition did not accurately conserve water within the computational domain, even
though convergence was achieved. Figure 5.17 shows the flow lines resulting from using no
onshore flow at the offshore boundary.
The second case is to simulate the experiments conducted at the Waterways Experiment
Station and reported by Hales (1980). A jetty perpendicular to the shoreline extended 15
feet out from the still water line on a plane beach with a slope of 1 on 20. The plane beach
extended beyond the jetty tip until a depth of one foot was reached and further offshore
the laboratory basin was a constant depth of one foot. Waves were sent at an angle by
use of a moveable wave paddle and wave guides which were set up to follow the wave rays
at the ends of the wave paddle. Figure 5.18 shows the experimental setup for these tests.
These wave guides in essence made the experiments to be in a closed basin and greatly
83
Figure 5.15: Results for the numerical model of Liu and Mei (1975). Deepwater wave height
equal to 1 meter, deepwater wave angle equal to 45 degrees. Source: Liu and Mei (1975).
100 200 300 400
_j 1 1 1
Scale (meters)
84
Figure 5.16: Results from the present numerical model using the same conditions as were
used by Liu and Mei. Offshore boundary condition of constant water surface level which
precludes tangential flow at the offshore boundary.
85
Figure 5.17: Results from the present numerical model using the same conditions as were
used by Liu and Mei. Offshore boundary condition of no onoffshore flow. Grid spacing
equal to 20 meters.
86
Figure 5.18: Experimented setup for the tests conducted by Hales (1980). Source: Hales
(1980).
87
WOVE PERIOD = 1.0 SECONDS
WOVE HEIGHT NEQR GENERATOR = 0.139 PEFT
Figure 5.19: Experimental measurements of the middepth averaged velocities obtain by
Hales. Measurements at one foot intervals. Source: Hides (1980).
affected some of the circulation patterns away from the jetty. Figure 5.19 shows the average
middepth velocities recorded on the down wave side of the jetty. Note the strong offshore
current next to the boundary. Since the boundaries are less than two jetty lengths from the
jetty, one must ask whether the circulation near the jetty was also affected. The numerical
simulation of the Hales experiments was performed using open lateral boundaries, since
to fully represent the experimental conditions in which the lateral boundaries are neither
straight (i.e. curving wave rays) nor perpendicular to the shoreline boundary would require
mapping techniques that are not yet included in the capability of the model.
Figure 5.20 shows a comparison of the experimental and numerical results obtained for
the magnitude of the rip current adjacent to the down wave side of the jetty for the case of
a 1 second wave with an angle of incidence of 30 degrees and a one foot depth wave height
88
Figure 5.20: Ripcurrent velocities adjacent to the down wave side of the jetty. Comparison
of experimental and numerical results.
of .139 ft. One would expect the depth averaged velocity to be somewhere between the
middepth velocity and the bottom velocity. The model appears to slightly underpredict
the rip current. In figure 5.19 a circulation cell in the lee of the groin is barely discernable.
In figure 5.21 the circulation cell is very obvious.
5.5 Shore Parallel Emerged Breakwater
In order to solve for the wave and current field resulting from the interaction of waves
and an emerged breakwater several boundary conditions at the breakwater are incorporated
into the model in addition to the boundary conditions already discussed in sections 4.3
and 4.4. The breakwater is represented as being infinitesimally thin and impermeable.
Figure 5.22 shows the position of the breakwater on the shoreward side of the I = Ibrk grid
row. The impermeability condition implies that neither waves nor currents pass through
the breakwater. The no current condition is easily imposed in the circulation model by
making Cf(lBjyr+i, J) equal to zero for the entire length of the breakwater. In addition the
circulation model is modified so that momentum diffusion and derivatives do not take place
89
Figure 5.21: Velocity vectors of the currents obtained numerically for the same conditions
as for the Hales experiment. Grid spacing equal to one foot.
90
> X " "
y
y
Figure 5.22: Definition sketch of grids for the case of a shore parallel breakwater,
across the barrier.
In the wave model an adaptation is made so as to prescribe the condition that the
wave height is zero on the lee side of the breakwater. This condition on the lee side of the
breakwater is physically not realistic since common observation reveals that waves actually
propagate along the lee side of a breakwater. The parabolic model however marches waves
in but one direction. This assumption however will give good results several grids away
from the breakwater. Figure 5.23 shows the crests of the diffracted waves in the lee of a
semiinfinite breakwater for a region of constant depth. This figure shows that the zero
wave height condition on the lee side of the breakwater does not adequately represent the
waves along the breakwater or allow wave energy to properly propagate in the lee of the
breakwater; however, important information can be obtained through use of the marching
parabolic wave equation method. To set the wave height to zero on the lee side of the
breakwater the ACOEF subroutine is modified so as to advance the wave one half delta x
from the center of the I = BRK grid row to the position of the breakwater at the edge of
91
Figure 5.23: Wave crests in the lee of a semiinfinite breakwater in a region of constant
depth.
the grid row. The amplitude is then set to zero at the breakwater and remains the same
elsewhere. The wave is then advanced another one half delta x so as to be back at the grid
centers and then is marched grid row by grid row to the shoreline. With the exception of
the two one half delta x steps the wave model is otherwise unchanged.
A test case was arbitrarily chosen of a breakwater 100 meters long 100 meters from
the shoreline on a plane beach with a 1 on 40 slope. The wave conditions again were
arbitrarily chosen to be a six second period and a wave height of 1 meter on the offshore
grid row. Figure 5.24 shows the velocity vectors obtained for the case of normal incidence
and for the cases of increasing angle of incidence. For normal incidence two circulation cells
and a well defined rip current in the lee of the breakwater result. The rip current is fed
by the longshore currents which result from the difference in the setup in the lee of the
breakwater and the region exposed to direct waves. As the angle of incidence increases a
longshore current develops and the circulation pattern becomes skewed. The rip current
is increasingly skewed in the direction of the longshore current with increasing strength of
92
the longshore current and the upstream circulation cell becomes increasingly weaker. The
other cell remains but is shifted down current.
The emerged breakwater program was also run for the case of a shore parallel breakwater
700 meters in length, 350 meters offshore on a 1 on 50 slope. This is a case tested by Liu
and Mei and represents the condition for the Santa Monica offshore breakwater. The wave
conditions were a 10 second wave of normal incidence with a deep water wave height of 1
meter. Figure 5.25 shows the symmetrical stream function contour lines obtained from the
Liu model for half of the region in the lee of the breakwater. Figure 5.26 shows the flow lines
obtained from the present model and figure 5.27 shows the velocity vectors for this case.
Figure 5.28 shows the contour lines of the mean water surface or setup. There is obviously
greater setup in the exposed region than in the lee of the structure but no well defined rip
current is seen in figure 5.27 or in the results of Liu shown in figure 5.25. Both models show
a strong longshore current at the shoreline directly behind the end of the breakwater, but
this current does not extend to the center of the breakwater since there is a distance over
300 meters in which the intense longshore current can be diffused and directed offshore.
5.6 Comparison with Experimental Results of Gourlay
Gourlay (1974) reported results of laboratory experiments in which currents were genÂ¬
erated by the diffracted waves in the lee of a breakwater. His experiment was designed to
demonstrate that longshore variation of the wave height could generate longshore currents.
The experimental wave basin was thus constructed so that the shoreline was everywhere parÂ¬
allel with the diffracted wave crests. Figure 5.29 shows the laboratory set up for Gourlayâ€™s
experiments. A 1 on 10 concrete beach is parallel to the incoming wave crests in the exÂ¬
posed zone and in the shadow zone the slope is also 1 to 10 but the beach is curved with
a constant radius centered on the breakwater tip. Gourlayâ€™s reasoning in choosing this
arrangement was that â€œsince the diffracted wave crests have theoretical curves which can
be approximated by a circle the influence of waves breaking at an angle to the beach was
almost completely eliminated.â€ (page 1978)
93
5o
Figure 5.24: Velocity vectors obtained from the model for a 100 meter long breakwater 100
meters offshore for a 1 meter, 6 second wave for increasing angles of incidence.
94
Figure 5.25: Stream function contour lines obtained from the Liu model for a 10 second 1
meter wave normally approaching a 700 meter long breakwater 350 meters offshore on a 1
on 50 slope. Source: Liu and Mei (1975).
95
Figure 5.26: Flow lines obtained from the present model for a 10 second 1 meter wave
normally approaching a 700 meter long breakwater 350 meters offshore on a 1 on 50 slope.
Figure 5.27: Velocity vectors obtained from the present model for a 10 second 1 meter wave
normally approaching a 700 meter long breakwater 350 meters offshore on a 1 on 50 slope.
96
Figure 5.28: Setup contour lines obtained from the present model for a 10 second 1 meter
wave normally approaching a 700 meter long breakwater 350 meters offshore on a 1 on 50
slope.
An experiment of Gourlay was numerically duplicated by making several adaptations
to the overall scheme. Figure 5.30 shows the numerical computational grids used in the proÂ¬
gram. Rather than model the entire wave basin it was decided for computational economy
to assume the wave paddle location to be situated 3 meters offshore of the breakwater and
that the wave paddle produced waves only in half of the domain seaward of the breakwater.
The offshore boundary condition in the circulation model was a noflow condition through
the wave paddle. This condition may slightly influence the results obtained for the mean
water levels since the computer model will not have as great an off shore reservoir from
which the wave set up must come as in the physical experiment. An adaptation to the
flooding routine was made to allow for variable flooding. This meant that the last grid
of each column could vary independently of the neighboring grid columns. This change
necessitated the incorporation of many decision statements within the circulation model so
that there would be no flow between wet grids and dry grids unless the water elevation was
high enough to initiate flooding. At present flooding is allowed only in the x or onoff shore
direction.
97
Figure 5.29: Physical layout for the experiment of Gourlay. Source: Gourlay (1974), ProÂ¬
ceedings 14^ Coastal Engineering Conference, copyright by the American Society of Civil
Engineers, reprinted with permission.
98
Figure 5.30: Computation sil domain for the experiment of Gourlay. Grid spacing equal to
.1 meter.
99
In the wave model part of the program, the problem of wet and dry grids is handled
using a technique proposed by Kirby and Dalrymple (1986) by treating the dry grids as
though they have a very small (say one millimeter) depth of water. This technique allows for
keeping a uniform computational domain for the wave model while avoiding any numerical
problems such as dividing by zero depths.
Gourlayâ€™s experiment used a wave height of 9.1 centimeters with a wave period of 1.5
seconds. The depth in the offshore constant depth region was 20 centimeters. For the
numerical computation a grid size of .1 meters was used. This created a 60 by 63 grid
starting domain which grew to a 64 by 63 grid domain once set up was established.
The numerical results show close agreement with the experiment results. Figure 5.31
shows the experimental results for the mean water surface contours resulting from the wave
setup and the corresponding results obtained from the numerical model. The agreement
appears to be quite good. There is however an area of slight differences where the shoreline
connects to the breakwater. The computer model fails to show any setup in contrast to the
experimental results which show a slight setup in this region. An explanation for this is
that the parabolic approximation used in the wave model does not propagate waves along
the breakwater as occurs in the physical experiment. This in turn may also help explain
the absence of the small stagnation eddy shown in figure 5.32.
Figure 5.32 shows the experimental results for the circulation. The figure shows both
the contours for the magnitude of the velocity and some streamlines indicating the direction
of the flow. Figure 5.33 shows the stream lines obtained from the numerical values while
figure 5.34 shows the contour lines for the magnitude of the velocities obtained numerically.
Again good overall agreement is in evidence. The rip current in the exposed area in parÂ¬
ticular is very well shown in the numerical results. The differences are that, as mentioned
earlier, there is no stagnation eddy shown in figure 5.33 and the center of the primary eddy
is slightly farther behind the breakwater in the physical experiment than in the model reÂ¬
sults. Similarly as seen in figure 5.34 the maximum currents obtained in the model are not
100
1.1 i 1 / /JJ ///>/;?/ 7771
Figure 5.31: Comparison of experimental and numerical results for the experiment of
Gourlay: Wave Setup Contours. Top figure is from Gourlay (1974) Proceedings 14^
Coastal Engineering Conference, copyright by the American Society of Civil Engineers,
reprinted with permission. The bottom figure is the computer results.
101
Figure 5.32: Results from the experiment of Gourlay: Contours of current velocities and
streamlines. Source: Gourlay (1974), Proceedings 14^ Coastal Engineering Conference,
copyright by the American Society of Civil Engineers, reprinted with permission.
Figure 5.33: Results from the numerical model: Streamlines of the flow.
102
Figure 5.34: Results from the numerical model: Contour lines of the current velocities.
as far behind the breakwater as they were in the experiment. Both of these spacial disparÂ¬
ities between numerical and physical results may result from the failure of the parabolic
approximation to adequately propagate wave energy fully into the lee of the breakwater.
In comparing the velocities obtained in the model to the experimentally measured values it
should be noted that Gourlayâ€™s results are for surface velocities whereas the model results
are for depthaveraged currents. This may be a factor in explaining that the maximum in
figure 5.32 for the experimental results is larger than the maximum in figure 5.34 for the
numerical results.
The differences in the experimental and numerical results which were mentioned above
should be kept in perspective. Overall the agreement is good both qualitatively and quanÂ¬
titatively.
CHAPTER 6
CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER STUDY
A numerical solution for the circulation and waves in coastal regions has been presented.
The solution method uses an alternating direction implicit method of solving the depth
integrated equations of momentum and continuity. Wave forcing terms are obtained through
a parabolic approximation to the linear mild slope equation for wave propagation over
varying depths and currents. The numerical solution involves an iteration between the
parabolic wave solution and the ADI circulation solution.
The model was run for several test cases. These include cases with varying topograÂ¬
phy, such as rip channels and bar gaps. The model was also tested with shore parallel and
shore perpendicular breakwaters represented as infinitesimally thin impermeable barriers.
Comparisons were made with analytical and experimental results. These comparisons were
favorable. In particular comparison of the results from the model in simulating the experiÂ¬
ments of Gourlay (1974) show the predictive capability of the model. The model was also
run for an idealized shore parallel breakwater varying the aspect ratio of the length of the
breakwater to the distance offshore of the breakwater to determine the relative importance
of this parameter.
The primary conclusion of this study is that it is possible to get good predictive reÂ¬
sults using a numerical model that incorporates some simplifications of the physical reality.
In the present study the threedimensional flows that are encountered in nature are repÂ¬
resented by twodimensional depth averaged velocities. Simplifying assumptions are also
made concerning bottom friction, momentum transfer or mixing, wave breaking and wave
energy dissipation. For each of these topics it is possible to argue that nature is greatly difÂ¬
ferent and more complex than the assumptions embodied in this report. Yet the results (in
103
104
particular the comparison to the laboratory experiments of Gourlay) confirm the predictive
capability of the numerical model. The simplifications mentioned above make this possiÂ¬
ble, for there must be a mathematical statement of the physical process before numerical
modelling of the process can even be attempted.
In making recommendations for further study we should distinguish between short range
objectives and the possibilities that may become open to numerical modelers by advances
in the computer hardware field. It is not difficult to envision within a few decades or less,
super computers with thousands of parallel processors. Such advances will make many of the
present computational methods obsolete. For example the parabolic approximation is made
solely to put the wave equation in a solvable form with respect to the capabilities of present
day computers. With sufficient parallel processors it is conceivable that it may be practical
to solve the full elliptic equation explicitly. We shall not dwell further on speculation about
future generations of computers. In the near term, the numerical model presented in this
report can be refined and the applications can be extended. The model is constructed in
modular form so that any new formulation for any of the component processes can easily
be incorporated into the model. The model is thus open for continual improvement.
6.1 Improved Wave Formulations
The model can be improved by better wave modeling and by a more correct formulation
of the radiation stress terms once breaking has commenced. The formulation used in this
model and previous models for the radiation stresses initiates setup and longshore currents
at the breaker fine whereas as explained in section 5.1 experiments show otherwise.
An improved wave model will assign a nonzero value of the amplitude in the lee of
a shore parallel breakwater. Research should be done to establish a method of doing this
so as to enable more wave energy to propagate into the lee of the structure. Dalrymple
and Kirby (1988) have found a method to determine a nonzero wave height in the lee of a
breakwater. However their solution is in the form of a Fourier integral which is not suited
to the complex amplitude form required in the present model.
105
6.2 Including Wave Reflection
The author has made unsuccessful attempts to include wave reflection in the model.
Kirby (1986) used coupled equations governing the interaction of a forward and a backward
(reflected) propagating wave that are coupled through a slope interaction. He was able to
obtain the reflection and transmission coefficients over submerged breakwaters that agree
well with the experiments of Seelig (1980). Once the arrays of the complex amplitudes of
the two waves is determined, the radiation stress terms can be determined from equations
that have been derived following the same procedure as used by Mei (1972) but defining
as being comprised of two components. In fact this procedure can also be used to obtain
the radiation stress terms for any number of components. The only limit being the amount
of computation required as the number of components increases.
Smith (1987) following the work of Miles (1967) for wave reflection and transmission
over a step derived equations for the complex reflection and transmission from a submerged
plane barrier. An attempt was made using this formulation to obtain the circulation and
wave pattern from a submerged breakwater represented as a plane impermeable barrier.
However problems of numerical stability were encountered. It is suspected that the sharp
gradients between the region with the presence of the reflected wave and regions of its
absence adjacent to the barrier produced unrealistic gradients in the radiation stress terms.
An attempt was made to include the reflected wave for the case of an emerged plane barrier
with the same unstable result. This is a topic for further investigation.
6.3 Extending the Model to Include Sediment Transport
The model as it currently stands can be extended to include sediment transport. The
wave model can be slightly amended to retain wave energy dissipation as an array. If a
formulation of sediment entrainment can be made in terms of the wave energy dissipation
and the bottom shear stress, everything else is in place to accomplish the sediment transport
model. A simple numerical mechanism would be to assume steady state conditions for
certain time intervals and solve for the wave and current field and the sediment concentration
106
as a function of known variables and then solve a separate model for sediment balance and
then update the depths and proceed iterating between the sediment balance model and the
present wavecurrent model marching in time. Since it has already been demonstrated that
the wave current model gives adequate results in the lee of a shore parallel breakwater, a
good test for such a sediment transport model wold be to see if it could yield a tombolo
formation in the lee of a shore parallel breakwater.
APPENDIX A
DERIVATION OF THE DEPTH AVERAGED EQUATIONS
In this appendix a derivation of the governing equations for the circulation model is
presented. This appendix closely follows the work of Ebersole and Dalrymple (1979).
Before proceeding with the derivations of the depth integrated time averaged equations
of motion and the continuity equation the Leibnitz Rule for taking the derivative of an
integral is stated since it is invoked throughout the derivation.
Ty C /(I'V) = C dJ^Tix + my)'y)dJÂ£  '^y)â€™y)dJJT
The Leibnitz Rule can also be used in reverse to take derivatives outside of an integral.
The kinematic boundary conditions are also reproduced since the use of the Leibnitz Rule
will yield bottom and surface terms that will be eliminated through use of these boundary
conditions. On the free surface the boundary condition is
dr) dr)
dr)
Wr' = Tt+'xaiJrv"Ty
(A.2)
and at the bottom the boundary condition
(A.3)
In the circulation model it is assumed that the still water depth does not vary with time so
the first term of Eq. (A.3) is dropped (actually for time varying depth essentially the same
equations are obtained).
The continuity equation (or the conservation of mass equation as it is abo called) in
three dimensions is
, d(/>u) d(pv) d(pw)
dt dx dy dz
(A.4)
107
108
Integrating over the depth from z â€” â€”h0 to z = rÂ¡ and using the Leibnitz Rule to take
derivatives outside the integrad, the continuity equation becomes
d f* . dr] d [i
atLpiâ€™p'Tt+TzL'n,i'
h0
f \ dr> r \ dh0 d n
Si Vâ€”â–
/ s
(A.5)
The terms with an overbrace are p times the free surface kinematic boundary condition
and the terms with an underbrace are p times the bottom boundary condition. Eliminating
both of these from Eq. (A.5) gives
d n . dr* . d rn J
/ Pdz+*~ pudz + â€” pvdz 0
dt Jh0 ox Jâ€”k0 dvJh0
(A.6)
Equation (A.6) is now time averaged. Time averaging means averaging over a wave
period. The time average of a term is denoted by an overbar. Thus
^ 1 fT
â– H
Fdt
(A.7)
where F represents any term and T is the wave period. Letting u and v be comprised of a
mean flow, a wave induced flow, and an arbitrary or random turbulence component
u = Ã¼ + Ãº + u'
V = V + V + v1
(A.8)
(A.9)
The turbulent fluctuations are assumed to be of a high frequency relative to the wave
frequency so that their time averages over the wave period are zero. By substituting the
above expressions for u and v into Eq. (A.6) and taking time averages the continuity equation
is
^ {p{K + â€™?)}+ ^ iPÃ¼iho + Â»?)} + JÂ¿f_k PÃºdz+ (A.10)
A {Â«Â«(*.+ fl)} + = o
(AH)
109
The time averages of the vertically integrated wave induced velocities Ã¼ and v are not zero.
These quantities are known as the wave induced mass flux.
Defining
U = u + u
and
V = v + 5
(A.12)
where
u =
1 Jn
r / pudz and
o + fj)Jk0
p{h0 + *?)
the time averaged depth integrated continuity equation is
%Ã+kuQlo+t,)+â€˜Â§Â¡Â¡viho+tl)
1 ri
v â€” ~Ti rr / pi> dz
p{h0 + fi)Jho
(A.13)
(A.14)
The assumption that the density is constant in both time and space has used in obtaining
both Eq. (A.11) and Eq. (A.14). The assumption of a constant bottom has also been used
in obtaining (A.14). If the bottom is not constant in time a term would be included in
Eq. (A.14).
The x and y momentum equations are handled in the same manner as the continuity
equation above. The x and y direction momentum equations are
du du2 duv duw 1 dp 1 ,drxx drvx drxx.
Tt + + + IT = } Ãš + + if + sT>
dv t duv t dv2 t dvw _ 1 dp 1 fdrxv drvv drZ]/
dt dx dy dz
= zir + 1A
+
}
(A.15)
(A.16)
p dy p dx dy dz
The same procedure is followed for both the x and y equations. Only Eq. (A.15) will be
dealt with here. Integrating the left hand side over the depth and applying the Leibnitz
Rule to take the derivative outside the integral gives
it Tip\ d /*'J , dr] dh0
(LHS> = ai/Â»." ~aT +
if +
dxjh0 ndx Â° dx
dy J
, , x dr) . . dh0
uvdz  (uv)â€ž â€”  (uv).fco â€” +
h0 dy dy
(UW)l ~ (UW)ho
(A.17)
110
The terms in Eq. (A.17) that are designated by an overbrace are tif, times the kinematic
free surface condition and those designated by an underbrace are u_j,0 times the bottom
boundary condition. These are both dropped leaving
, â€ž d r dr, dr
(LHS) = * U UdZ+d~x U UdZ+d~y U UV dZ ^
Substituting the definitions given for u and v in Eqn. (A.8) and (A.9) and recognizing that
the turbulent fluctuating components are of a high frequency relative to the wave frequency
so that the time averages of integrals involving the product of a turbulent component and
a wave induced component are zero gives
d~r d~r d ~r d~r
LHS = â€” / udz+ â€” Ãºdz + â€” u?dz+â€” tfdz (A.19)
dtJko dt]K0 dxJho dxjh0 â€™
d ~r~T~Z d ~r d ~r d ~r
+ Â£ {u')2dz + â€” / 2UÃš dz + â€” uv dz + â€” / uiidz (A.20)
dxJho dxjh0 dyJh0 dyJho
d ~r d ~r d ~r
+ X Ã¼vdz+â€” Ãœi)dz + â€” u'v'dz (A.21)
oyjho oyJh0 dy JK0
using the definitions of Eqs. (A.12) and (A.13) this becomes
+aiuvv,Â°+Â«)+TyS'h' \w^)Lrdâ€˜Lriz
(A.22)
where the two terms designated by underbraces result from being added to complete the
U2 and UV terms.
The right hand side after integrating over the depth assuming a constant density and
assuming an inviscid fluid such that no horizontal viscous stresses exist, i.e. rtx and Tyx are
zero, is
RHS =
if r ^
p Jho dx p Jâ€”h0 dz
(A.23)
Using the Leibnitz Rule and averaging over a wave period the right hand side becomes
1 d r . 1 dn 1 dhn 1 1
RHS = â€”â€” / pdz+ + PKozâ€” + Tzz~~ rnir (A.24)
pdxjh0 p dx p dx p â€ p v '
Ill
TZXq and rMX_ko are the time averaged surface and bottom shear stresses. Assuming that
the free surface pressure pn is zero (the dynamic free surface boundary condition) the right
hand side is
___ 1 d 7* 7 1 dh~0 11 ,
RHS = â€”â€” / pdz + p.hâ€”Â± + TÂ¿TiÂ¿ (A.25)
pdxjh0 p dx p p â€™
The mean pressure at the bottom can be expressed as the sum of the wave induced or
dynamic pressure and the hydrostatic pressure
Pfco = PdVnko + P9[h0 + r>)
(A.26)
so that the product Ph0fa can be manipulated to read
dh0 1 dh0 13.,. _.2, . .dfj
'â€™*â€¢17 = + 2Ã¡Ã' l(hÂ°+
(A.27)
Terms from both the right and left hand sides are combined as the radiation stresses.
Defining
sâ€ž = f'jp+,07) iz  \P9(hâ€ž + Â«)*  rM'Y
~r> i ~Ã± r>
S,^j ^vdz  ^ ,0 i, j^dz
Sn = /_ JÂ» + ,07) dz  1,,(*. + Â«)>  ^7) {/I P5 ^
(A.28)
(A.29)
(A. 30)
and making the assumptions that the product of the dynamic pressure at the bottom and
the bottom slope is negligible, that the Jif?h0(u')2dz term can be ignored and that the
lateral friction tj caused by momentum fluxes due to turbulent fluctuations is independent
of depth and that rj is defined by rÂ¡ = â€”pu'v', the x direction equation of momentum is
written as
3 Â»r2,l , ^ , d rnr/L . . ldSxx . 1 dS,
+ 
XV
5^ + Â« + arU '<*â€¢ + *Â» + TyUV^ + + , 8x ' , 8â€ž
+,(Â». +fl5+(ÃAÂ±Â«Ã^irâ€ž + In, = o
OX P Oy p p
(A.31)
By subtracting the depth integrated continuity equation and dividing by the total depth,
the final form of the x momentum equation is
8U dU_ dU_ 1
dt + dx dy + p{h0 + fj)
dST
dS.
xv
dx
dy .
(A.32)
112
dfj 1 drÂ¡
+ 9TZ + T1
: +
â€”rn* = 0
(A.33)
dx pdy p{h0 +p{h0 + rÂ¡)
The same procedure gives as the depth integrated time averaged y momentum equation
dV TTdV rrdV 1 dSt
1 U h V 1
dt dx dy p(h0 + fj) dx
xy
dS,
vv
dy .
drÂ¡ 1 dri 1
+ 9dyJr pdx p[h0 + f}) T'v
p(K + v)
rr Hv = 0
(A.34)
(A.35)
APPENDIX B
DERIVATION OF THE RADIATION STRESS TERMS
The basis of the derivation is to use the definition of the velocity potential to substitute
for the wave induced velocities in the definition for the radiation stresses.
d
, _ d
w =
zp dz
where is expressed in terms of the general complex amplitude function B as
_ 1 (_ig â€žcoshk(h0 + z) _iut
2 V a
B cosh kh
+ cc)
where c.c. represents the complex conjugate term. Also rÂ¡' is expressed as
r,' = l(Beiut + c.c)
where the prime is used to denote wave induced quantities.
For convenience Eq. (2.50) of Section 2.2 is reproduced below.
S<*p = P u'au'pdz + 6ap ^ pdz  y {h0 + Â»j)2 + y(r?')2
An expression for p is obtained in Section 2.2 and is reproduced below
d IL .W1
dz  p{w'y
Substituting Eq. (B.5) into Eq. (B.4) gives
Sap = P I Ku'pdz + SaP
* â€ /l o
(B.l)
(B.2)
(B.3)
(B.4)
(B.5)
ri â€ž r* ( r* du'w' \
J ^ P9Ã1 ~ Z)dz + J ^ ( pj ^â€”dz\dz
'  .. ' _ Â° ' * * '
/
p[w'Y dz  Y{K + r))2 + ^(Â»?')2
Â«O L
(B.6)
113
114
The terms of Eq. (B.6) are now evaluated term by term.
Term 1:
, _ 1 / ig dB cosh k[h0 + z) _iut ig dB' cosh k(h0 + z) iu)t
u'au'p â€”
, = 1 (_
Uâ€œ 2 \ a dxa cosh kh
ff2 cosh2 k(h0 + 2)
a dxD
cosh kh
(B.7)
55 51T dB^dB_ _ dB_dB_e2iut _ dB' dB* Xut
4a2 cosh2 kh dxa dxp dxa dip dxa dip
dxa dip
â€”râ€”r pg cosh2 k(h0 + z)
pUÂ°Ut = 2Vsinh 2kh
pg cosh2 k(h0 + 2)
Jfcsinh 2 kh
dB dB' dB' dB
+
dxa dip dxa dip
â€˜ (dB_W\
dip )
(B.8)
(B.9)
Ã pu'auâ€™pdz =
* â€œ* o
P9
â€¢1
( dB 55* y
k sinh 2kh
^5xa dip )
P9
*1
( dB 55* \1
ksmh2kh
\dxa dxp )\
Â£Â»(****:) i(i+
4 dip J k2 \ si
/ cosh2 k(h0 + 2) dz
â€¢ hp
^ (sinh 2k(h0 + fj) + 2k[h + fj))
2kh \
sinh 2khJ
The second term:
J_h pg{v  z)dz = pg ^r)z 
hÂ«
... â€ž(* + *)*
P9^~
The third term:
f If *ff dB LUÃ.U ".yo T â€œJ _lu,f , â€¢ ># I.UOU T *J
â€˜ ~ 2 { a dxt 6 + " â€œ 6
cosh k(h0 + 2) _iult ig dB' cosh k(h0 + z)
cosh kh a dxÂ¡ cosh kh
igk psinh k(h0 + z) _lult ^ iff* D, sinh k[h0 + 2) j
cosh kh
g sinh 2A:(h0 + 2)
â€¢ Â»it' â€” L. ^
4sinh2fch 5x, ' dx(
tv
, _ 1 / Â«ffk sinhk[h0 + 2) lult iff* .sinh fc(h0 + z)
2 V <7 cosh kh a cosh kh 6 J
dB 55* '
â€”5 + â€”5
u^ur =
(BIO)
(B.ll)
(B.12)
(B. 13)
(B.14)
/;
dxÂ¡
_ 9_d_
4 dx
dz =
dB
dxf
dB' J r
_ ?_L ÃÂ£Â£b â€”sÃ f_
4 5xf 5xf 5xf J [2)fc sinh 2fc(/ic + ij) 21: sinh 2k[h0 + tj)_
r,} sinh 2k{h0 + 2)
sinh 2Â¿h
cosh 2k[h0 + Â»;)
cosh2fc(h0 + 2)
(B.15)
115
_pg d \dB B, P* / cosh 2k[h0 + rÂ¡) cosh 2k(hâ€ž + z)
4 dxÂ¡ [dxf dx( J Jhâ€ž \2fcsinh2A:(h0 + fj
PI
4
?lJL
4 dx
P9 d
___n .
, t , * j  â€”w , Â» rÂ¡) 2k sinh 2k{h0 + fj))
â€” â€”a* + d^B] ( cosh2fe(/lo + ^) (h , â€”\ sinh2fc(/i0 + fj) \
dx( dx( dx( J \ 21; sinh 2k(h0 + fj)' 0 ^ 4fc2 sinh 2k{h0 + fj) J
r 1 '2fchcosh 2fc/i sinh2A:/i
as . as* J /
:f [dx( + axf BJ (
41:2 sinh 2fch 4k2B\nh2kh)
dB
dB'
T^&B' + iÂ¿;B\^2khcoth2khv
Expanding the derivative term in Eq. (B.16)
dx(
dB dB* â€ž
â€”Â£* + â€”B
ax
(
dx.
d2B a2B* dBdB*
rB* + ^iB + 2
ax2
ax2
axf ax
f
Recognizing that f is a free index so that = V*
a2B
ax2
5Ã7
= VÂ¿B = Jfc2B
d2B'
dx?
= V\B* = â€”k2B'
(B.16)
(B.17)
which results from use of the Helmholtz equation which governs diffraction in water of
constant depth and
2^Â§!r=2vâ€˜B! (B18)
so that the expanded derivative term (B.17) becomes
a2B . a2B* dBdB*
ax2 B + ax2 B + 2dx( dx(
= 2V*B2  2Jfc2B
(B.19)
and the result is
/l ("/," lÂ£fÃZ) = T (Vâ€˜S1 " 2F  1] (B.20)
The author wishes to express some reservations concerning the validity of Eq. (B.20).
The shallow water asymptote for this expression is zero since for small k(h+fj), coth 2Jt(h +
*1) ~ tk[h+n) > and 2fc(h+r?)coth2fc(/i+fj)l 0. In deepwater, however, coth2Jfc(h+Â»j) â€œ 1
so that 2k(h + fj) coth2k(h + fj)  1 =Â£ 2kh. If the derivative terms are nonzero, the Szz
116
term will be proportional to the depth. This would mean that deep water nonshoaling and
nondissipative waves in varying depths would produce gradients in the radiation stresses.
For plane waves this is not a problem, since the complex amplitude terms are easily shown
to be zero. This is seen by substitution of B = ae,kx and B* = ae~tkx where a = B.
The fourth term is straight forward
igk psinh k(h0 + z) _iut igk sinh k(h0 + z) ,â€¢
cosh kh
+ â€”B'
cosh kh
#_ i ( Â»'gfc
2 V o
, , = â– *(> +z) _ B.^, _
4 a2 cosh2 kh ' >
77T2 _ 1 g2k2 sinh2 k(h + z) n gk sinh2 k{h0 + z)
( } "4 a2 cosh2 k[h + rj) ~ sinh2Jkh BB
(B.21)
(B.22)
(B.23)
 Ã p(w')2dz = pgk\B\2 Ã
Jâ€”h0 "â€”h0
2 sinh2 k(ha + z)
sinh2Â¿h ~ 4 1 1 V* sinh 2kh,
The fifth term needs no evaluation and is seen to cancel exactly the second term.
The sixth and last term of Eq. (B.6) is rÂ¡')2 and is evaluated easily
rÂ¡' = 1 (Â£eâ€˜wi + B*eiut)
(r/)2 = 1 (BeW + 2 BB* + B*e2,wt)
SI
W=2^*= 2
fOT=fl*2
which cancels part of the result from the fourth term.
Adding the above results for all the terms of Eq. (B.6) yields
P91
(B.25)
(B.26)
(B.27)
(B.28)
c P9Ã~(9BdB*\ 1 ^ , 2kh \
4 I \5za dip y A:2 V sinh2kh)
+
^af)
I 2â€”2kh
1 1 sinh 2kh
dxa dip
(V*Â£2  k2\B\2)^(2kh coth 2kh  1)] 
(B.29)
APPENDIX C
LAGRANGIAN DERIVATION OF THE GOVERNING WAVE EQUATION
The governing equation for wave motion can be derived through use of a variational
principle where a Lagragian is the function that is varied. Luke (1967) showed that a
Lagrangian for waves is given by
L â€” Jh [^â€˜ + 3Z
dz
(C.l)
Kirby (1984) used a time averaged Lagrangian to derive a wave equation. For the following
derivation, equation (C.l) will be used. The third term in Eq. (C.l) is determined straight
forward, while the first two terms need to be expanded about fj the mean water line in a
Taylor series.
L=fâ€™4+i:
4>t +
(W)J
dz + {rÂ¡  rÂ¡)4>t â€ž +(f7  fj) ^
(C.2)
We assume that rj and can be expressed as
*) = V + cÂ»?i
= o + efi (C.3)
where e is an ordering parameter and / is the depth dependency as defined in Eq. (3.3). For
the sake of simplicity we will assume that the gradient of fo represents the steady horizontal
currents. This means that fo is not a function of time and that W will not be included as
it was in the derivation in Chapter 3. This is because the W term is not expanded about
z = fÂ¡ as it was in the Greenâ€™s function derivation and is not needed in order to obtain the
(Vfc â€¢ U)jÂ£ term which as will be seen is obtained more directly in the present derivation.
Including W here will only produce terms that are of higher order in the mean current.
117
118
Introducing (C.3) into (C.2) yields
Â¿gvi gh20
L = ?Y + Wrh + ^YÂ±9Â£ + Â«t>itJ'h fdz+^(h0 + fj) +
cVfc^oVfc<^i [ f dz + eV/,^0^1 ( Ã Vnfdz+f ftdz\ +
Jh0 Uhâ€ž JKc )
(y
2
Iâ€™ [/Vâ€ž/+//,]dz} +
e2rÂ¡iit + crti ( â€”hÂ£Â°^ + cVhfoVh4>i + eV^^o^i/* *j +0(e'
! 2
The Lagrangian L is assumed to be of the form
>)}
(C.4)
It then follows that
L â€” Lq + fLi + e* L2 + fSZ>s + 1
Â¿0iÂ£iS+M(*.+,)
and
(C.5)
(C.6)
Â¿1 = gf>m + â€”Â¿1, + â€”v^oV^! { r v*/ dz + r fzdz) + m (V^o)2 (c.7)
0 fl Ufco â€¢/*â€ž J 2
ti'Vhti Ã [f^hf + / /*] oifx #J (c.8)
* â€œ/lo
To obtain the desired linear governing equation is varied with respect to i and then Li
is varied with respect to rji in order to eliminate rÂ¡\. The variational principle holds that the
Lagrangian is constant under small variations of the unknown parameters. In mathematical
terms
and
DL
D
DL
Dr)!
= 0
= 0
(C.9)
(cio)
Since the ordering parameter e is arbitrary the above holds for all L,â€¢ Â» = 0,1,2, â€¢ â€¢ â€¢. In the
Li term \ is expressed in terms of its derivatives. This requires integration by parts to
119
obtain the variation with respect to \. Upon carrying this out it is found that
PÂ± = dLi_ d dU _ dU x
D
Applying (C.ll) to (C.8) and recognizing that the variation d\ is arbitrary yields after
carrying out the integrations in Â¿2
CC â€” k2CC
 V*(â€”*v*^) + M ')  nu  v*(v*^o9i) = o (C.12)
9 9
which is rewritten as
+ ^h{CCgVhi)  i(
Varying Li with respect to iji yields
0*71 + iÂ« + Vh\ = 0
(C.14)
which can be written as
lD!
Vl = pr
g Dt
(C.15)
which upon substitution into (C.13) yields the governing equation
^ + (V^o)^  Vfc(CC,VÂ»*) + (a2  k'CC'fa = 0
which upon making the substitution that 0 = V/,^o becomes
+ (v* * " v*(cc,v*^) + (
(c.i6)
Dt
(C.17)
APPENDIX D
OBTAINING EQUATION (3.79)
In this Appendix we show the algebra involved in obtaining equation (3.79) from equaÂ¬
tion (3.51) using the assumption that the dispersion relationship is written as
u â€” a ( k cos OU
We start by using the following form of <Â¡>
 ig eÂ«'(/ * CO. i dz)
and then make a phase shift by using the substitution
Ã coÂ»f dxJ k cos 6 dx)
so that in the final form
_ ig A'(X>V)e*(f Icoiidz)
(D.l)
(D.2)
(D.3)
(D.4)
and all y dependency is solely in the complex amplitude A'. For reference equation (3.51)
is reproduced below.
(P4>z)z + (pv)v + k2p(Â¡> + [w2  a2 + i'w(Vfc â€¢ U)j I i(7w<Â£
{U24>z)z{V24>V)V{UV4>X)V{UVÃV)X + 2Ã = o (d.5)
Writing out all the components of equation (D.5) and rearranging terms gives
{p ~ U2)xx + (p  U2)4>zx + [(p  V2)4>v] v + k2p4> + (a;2  o2)*
+iojUx4> + iuVv4> + icrw(j>  (UVx)v  [UVv)x + 2iuUx + 2iuV^>v = 0 (D.6)
The derivatives of defined by equation D.2 are as follows
4>x = â€”ig [(^) + ik COS 9^ câ€˜(/ k cot 0dz)
120
121
ZZ â€”
+ 2tfccos0^ + i(Jbcosi)a
â€” k2 cos2 9 â€”1 fc co* *dx)
a a J
{4>z)v â€”
"'I
{(a) e*(^*co,#
+i(k cos 9)y e<(/*cÂ°Â«#rf*)j
II
'^e.(/fcco.idx)J
H
II
G>x)v
The above forms of the derivatives recognize that the phase function eâ€˜(/icoÂ»edz) y
derivatives since k = k(x, y). However the y derivatives of the phase function integral are
not evaluated since the phase shift (equation D.3) will eliminate any y dependance from
the phase function and place all y dependance in the complex amplitude. The ( â€”)
v Â° / XX
term is neglected following the assumption explained in section 3.2. Substituting the above
derivatives into equation (D.6) gives
(pU2)x () e*'(/tco,#dl) + ikcos9(p  U*)gei
\
+2tftcos 9(p â€” U2) () e'(f kcotedx) + i(kcos 6)z(p â€” U2')â€”ei(fkco*9dx)
k2 cos2 9 [p  U2) â€”eâ€™(fkco,tdz) +
(p fÃ¡eHf tcosedz)\
V Jo
l\ J V* /rj
+k2pei(fkcotidx) + (u2  a2)eâ€™(ftcoetdx) + iuUxe^f kcoâ€˜9 dx^
o o a
/ A S
+twVy~e^f *â€œ*6dx) + iwAe^f kcoâ€˜9^  {UV)v eâ€™(/ *coâ€˜9dz>
2 UV
! 6 dx)
tJfccos 9UUV â€”e'(f kco,e dl)
â€”ik cos 9UVv â€”*co*9dz^  i{kcos 9)vUVkcot 9 dl>
a a
2ikcos9UV (Ã¡ei(fkâ€œÂ»edx)'j â€” i(kcos 0)vUVâ€”kcot9dz)
r ^ S
~(UV)z (^eâ€˜(/4co,Ã
2ojkcoe9U^fkcot9dx^ + 2iuV (Â¿(f kcc"9dx)'j
= 0(D.7)
122
Dropping terms that contain products of derivatives of j and derivatives of the current as
well as products of the current and second derivatives of Â£ (i.e. terms that are designated
by an overbrace) equation (D.7) is rearranged as
IÂ» [fccosi(p â€” U2) + u/E/j â€” + 2Â» [fccosÃ(p â€” E/2) + u/l/] ^ +Â»Wi4
+ [~fc2 cos2 0(p  U2) + k2p + u/2  a2  2uk cos OU^ ^ â€” i(k cos 0)vUV â€”
+t [Vv(w  kcosOU)  [k cos 6U)yV  k cos 0UV] j j e^f k COi â€˜ dx>>
+ [P(ae,^fcCO,i<,l)) + 2s â€” Jbcos17)] ^Ã¡eHftcctdx)J = ^D 8)
It is easily shown that
kcos0{p  U2) +
By writing the dispersion relationship as
<7 = u> kcosOU
it is also easily shown that
ui â€” a = 2o>k cos 6U â€” {k cos 6 Uy
From equation D.2 it is apparent that for a constant o/,
(fccos0)â€ž = ov
Thus equation (D.8), becomes
{â€¢ [a{Ct cosÃ + U)]x j + 2ia(Ct cosÃ + E7) ^
+ [fc2p(l  cos2 Ã)] J + iwA + IVvA + Â«7vV^} eâ€˜(/*â€œâ€¢Â«<**)
+2iVa ^ei(fkcot8dx)'j + p tco. tdx)^
where a i(k cos 0)vUVÂ£ term has been dropped.
At this point a phase shift is introduced
v.
= 0
(D.9)
(DIO)
(DU)
(D.12)
(D.13)
^ _ Â¿IgUf icoÂ»f dxf k coÂ« 6 dz)
(D.14)
123
where k and 0 are a reference wave number and wave angle. For computational purposes k
will be an averaged wave number across a grid row and 6 will be taken as the wave angle in
the absence of currents using Snellâ€™s law. Further it will be assumed solely for computational
purposes that 0 is equal to 8. This gives
Â»[
2(Cg cos 0 + U) cos 0(k  k)A
+Â«wA + iVvA + 2Â»V<7 (â€”) +iavVâ€” = 0 (D.15)
\o Jv o
Expanding the derivatives of Â£ and with a few minor manipulations of the terms the
governing equation (3.79) is obtained.
(C, cos 0 + Â¿*+ Â«'(
+V +
kCt(l  cos2 0)A   ICCg
(i).l
z
w
+ A = 0(D.16)
BIBLIOGRAPHY
Arthur, R.S. (1950), â€œRefraction of shallow water waves: the combined effects of currents
and underwater topography,â€ Trans. Amer. Geophysical Union, 31(4): 549552.
Bateman, H. (1944), Partial Differential Equations of Mathematical Physics, Dover PubÂ¬
lications, New York.
Berkhoff, J.C.W. (1972), â€œComputation of combined refractiondiffraction,â€ Proc. 13*^
Coastal Eng. Conf., Vancouver, 471490.
Berkhoff, J.C.W. (1976), â€œMathematical models for simple harmonic linear water waves:
wave diffraction and refraction,â€ Publ. no. 163, Delft Hydraulics Laboratory,
Delft.
Birkemeier, W.A., and Dalrymple, R.A. (1976), â€œNumerical models for the prediction of
wave setup and nearshore circulation,â€ Ocean Engineering Report No. 3, Dept,
of Civil Eng., University of Delaware, Newark, Delaware.
Booij, N. (1981), â€œGravity waves on water with nonuniform depth and current,â€ Report
no. 811, Dept, of Civil Eng., Delft Univ. of Technology, Delft.
Bowen, A.J. (1969), â€œThe generation of longshore currents on a plane beach,â€ J. Mar
Res.. 27(2): 206215
Bowen, A.J., Inman, D.L., and Simmons, V.P. (1968), â€œWave â€˜set downâ€™ and â€˜setupâ€™,â€
J. Geophys. Res., 73(8): 25692577.
Bretherton, F.P., and Garrett, C.J.R. (1968), â€œWavetrains in inhomogeneous moving
media,â€ Proc. Roy. Soc., A 302: 529554.
Butler, H.L. (1978), â€œ Coastal flood simulation in stretched coordinates,â€ Proc. 16^
Coastal Eng. Conf., Hamburg, 10301048.
Butler, H.L. (1980), â€œEvolution of a numerical model for simulating long period wave
behavior in oceanestuarine systems,â€ Estuarine and Wetlands Processes with
Emphasis on Modeling, Marine Science Series, vol.ll, Plenum, New York.
Butler, H.L., and Sheng, Y.P. (1984), â€œBoundary conditions for limited area modeling,â€
Proc. 19fck Coastal Eng. Conf., Houston, 858870.
Carnahan, B., Luther, H.A., and Wilkes, J.O. (1969), Applied Numerical Methods, WiÂ¬
ley, New York.
124
125
Dally, W.R. (1980), â€œA numerical model for beach profile evolution,â€ M.S. Thesis, UniÂ¬
versity of Delaware, Newark.
Dally, W.R. (1987), â€œWave transformation in the surf rone,â€ Technical Report No.
69, Dept, of Coastal and Oceanographic Engineering, University of Florida,
Gainesville.
Dally, W.R., Dean, R.G. and Dalrymple, R.A. (1984), â€œ A model for breaker decay on
beaches,â€ Proc. 19*^ Coastal Eng. Conf., Houston, 8298.
Dean, R.G., and Dalrymple, R.A. (1983), Water Wave Mechanics for Scientists and
Engineers, PrenticeHall, Englewood Cliff's, New Jersey.
Ebersole, B.A., and Dalrymple, R.A. (1979), â€œA numerical model for nearshore circulation
including convective acceleration and lateral mixing,â€ Ocean Engineering Report
No. 21, Dept, of Civil Eng., University of Delaware, Newark, Delaware.
Ebersole, B.A., and Dalrymple, R.A. (1980), â€œNumerical modeling of nearshore circulaÂ¬
tion,â€ Proc. 17^ Coasted Eng. Conf., Sydney, 27102725.
Garrett, C.J.R. (1967), â€œDiscussion: the adiabatic invariant for wave propagation in a
nonuniform moving medium,â€ Proc. Roy. Soc., A 299: 2627.
Gourlay, M.R. (1974), â€œWave setup and wave generated currents in the lee of a breakÂ¬
water or headland,â€ Proc. 14*^ Coastal Eng. Conf., Copenhagen, 19761995.
Hales, L.Z. (1980), â€œErosion control of scour during construction; report 3, Experimental
measurements of refraction diffraction, and current patterns near jetties,â€ U.S.
Army Engineer Waterways Experiment Station, Vickesburg, Mississippi.
Hansen, J.B., and Svendsen, I.A. (1979), â€œRegular waves in shoaling water. ExperimenÂ¬
tal data,â€ Series Paper 21, Inst, of Hydrodynamics and Hydraulic Engineering,
Technical University of Denmark, Lyngby.
Harrison, W., and Wilson, W.S. (1964), â€œDerivation of numerical calculation of wave
refraction,â€ Technical memorandum No. 6, U.S. Army Coastal Engineering ReÂ¬
search Center, Ft. Belvoir.
Horikawa, K., and Kuo. CT. (1966), â€œA study of wave transformation inside the surf
zone,â€ Proc. 10^ Coastal Eng. Conf., Tokyo, 217233.
Johnson, J.W. (1947), â€œThe refraction of surface waves by currents,â€ Trans. Am.
Geophys. Union, 28: 867874.
Johnson, J.W., Oâ€™Brien, M.P., and Isaacs, J.D. (1948), â€œGraphical construction of wave
refraction diagrams,â€ Tech. Report No. 2, U.S. Hydrographic Office, Washington,
D.C.
Jonsson, I.G. (1981), â€œBooijâ€™s currrentwave equation and the ray approximation,â€ Prog.
Report 54, pp 720, Inst. Hydrodyn. and Hydraulic Eng., Technical University of
Denmark, Lyngby.
126
Keller, J.B. (1958), â€œSurface waves on water of nonuniform depth,â€ J. Fluid Mech.. 4:
607614.
Kirby, J.T. (1983), â€œ Propagation of weaklynonlinear surface water waves in regions
with varying depth and current,â€ ONR Tech. Report 14, Research Rep. CE8337
Dept, of Civil Eng., University of Delaware, Newark.
Kirby, J.T. (1984), â€œA note on linear surface wave current interaction over slowly varying
topography,â€ J. Geophys. Res., 89(C1): 745747.
Kirby, J.T. (1986), â€œ A model for gradual reflection of weakly twodimensional waves
in a slowly varying domain with currents: governing equations,â€ Tech. Report
No. 60, Dept, of Coasted and Oceanographic Engineering, University of Florida,
Gainesville.
Kirby, J.T., and Dalrymple, R.A. (1982), â€œNumerical modeling of the nearshore region,â€
ONR Tech. Report 11, Research Report CE8224, Dept, of Civil Eng., University
of Delaware, Newark.
Kirby, J.T., and Dalrymple, R.A. (1986), â€œModeling waves in surfzones and around isÂ¬
lands,â€ J. Waterway, Port, Coastal and Ocean Engineering, 112: 7893.
Le MÃ©hautÃ©, B., and Koh, R.C.Y. (1967), â€œOn the breaking of waves arriving at an angle
to the shore,â€ J. Hydraulic Res., 5(1): 6788.
Liu, P.LF., and Mei, C.C. (1975), â€œEffect of a breakwater on nearshore currents due to
breaking waves,â€ Technical Memorandum No. 57, Coastal Engineering Research
Center, Fort Belvoir, Va.
LonguetHiggins, M.S. (1970a), â€œLongshore currents generated by obliquely incident sea
waves, I,â€ J. Geophys. Res., 75(33): 67786789.
LonguetHiggins, M.S. (1970b), â€œLongshore currents generated by obliquely incident sea
waves, II,â€ J. Geophys. Res., 75(33): 67906801.
LonguetHiggins, M.S., and Stewart, R.W. (1962), â€œRadiation stress and mass transport
in gravity waves, with application to â€˜surf beatsâ€™,â€ J. Fluid Mech.. 13: 481504.
LonguetHiggins, M.S., and Stewart, R.W. (1963), â€œA note on wave setup,â€ J. Mar. Res..
21(1): 410.
LonguetHiggins, M.S., and Stewart, R.W. (1964), â€œRadiation stresses in water waves; a
physical discussion, with applications,â€ DeepSea Research, 11: 529562.
Luke, J.C. (1967), â€œA variational principle for a fluid with a free surface,â€ J. Fluid Mech..
27(2): 395397.
Mei, C.C. (1972), â€œA note on the averaged momentum balance in twodimensional water
waves,â€ J. Mar. Res.. 31(2): 97104.
Mei, C.C. (1982), The Applied Dynamics of Ocean Surface Waves, WileyInterscience,
New York.
127
Miche, R. (1944), â€œMouvement ondulatoires de la mer en profondeur constaanttte ou
dÃ©croissante,â€ Ann, des PontB et ChaussÃ©ss.
Miles, J.W. (1967), â€œSurface wave scattering matrix for a shelf," J. Fluid Mech.. 28:
755767.
Miles, J.W. (1982), â€œOn surfacewave diffraction by a trench,â€ J. Fluid Mech.. 115: 315
325.
Munk, W.H., and Arther, R.S. (1952), â€œWave intensity along a refracted ray,â€ in
Gravity Waves, NBS Circ. 521: 95108.
Newman, J.N. (1972), â€œDiffraction of water waves,â€ Applied Mech. Review, 25(1): 17.
Noda, E., Sonu, C.J., Rupert, V.C., and Collins, J.I. (1974), â€œNearshore circulation under
sea breeze conditions and wavecurrent interactions in the surf zone,â€ Tetra Tech
Report TC1494.
Oâ€™Brien, M.P. (1950), â€œWave refraction at Long Beasch and Santa Barbara, California,â€
Beach Erosion Board Bulletin. 4(1): 112.
Penney, W.G., and Price, A.T. (1944), â€œDiffraction of sea waves by breakwaters,â€ TechniÂ¬
cal History Report No. 26, Directorate Mise. Weapons Development, Washington,
D.C.
Penney, W.G., and Price, A.T. (1952), â€œThe diffraction theory of sea waves and the
shelter afforded by breakwaters,â€ Phil. Trans. Roy. Soc., A 244: 236253.
Perlin, M., and Dean, R.G. (1983), â€œA numerical model to simulate sediment transport
in the vicinity of coastal structures,â€ Mise. Report. No. 8310, U.S. Army Coastal
Engineering Research Center, Ft. Belvoir, Virginia.
Phillips, O.M. (1969), The Dynamics of the Upper Ocean, Cambridge University Press.
Pierson, W.J. (1951), â€œ The interpretation of crossed orthogonals in the wave refraction
phenomena,â€ Tech. Mem. No. 21, Beach Erosion Board, Washington, D.C.
Putnam, J.A., and Arthur, R.S. (1948), â€œDiffraction of water waves by breakwaters,â€
Trans. Am. Geophys. Union, 29: 481490.
Radder, A.C. (1979), â€œOn the parabolic equation for water wave propagation,â€ J. Fluid
Mech.. 95(1): 159176.
Seelig, W.N. (1980), â€œTwodimensional tests of wave transmission and reflection charÂ¬
acteristics of laboratory breakwaters,â€ Technical Report No. 801, U.S. Army
Coastal Engineering Research Center, Ft. Belvoir, Virginia.
Sheng, Y.P., and Butler, H.L. (1982), â€œModeling of coastal currents and sediment transÂ¬
port,â€ Proc. 18^ Coastal Eng. Conf., Cape Town, 11271148.
Shore Protection Manual (1977), U.S. Army Coastal Engineering Research Center, Ft.
Belvoir, Virginia.
128
Smith, J. (1987), â€œOn surface waves crossing a step with horizontal shear,â€ J. Fluid
Mech.. 175: 395412.
Svendsen, I.A. (1984), â€œWave heights and setup in a surf zone,â€ Coastal Eng. 8(4):
303329.
Tappert, F.D. (1977), â€œThe parabolic approximation method,â€ In Wave Propagation and
Underwater Acoustics (edited by J.B. Keller and J.S. Papadakis) Chapter 5,
Springer Verlag, Berlin.
Van Dorn, W. (1953), â€œWind stress on an artificial pond,â€ J. Mar. Res.. 12(3): 249276.
Vemulakonda, S.R. (1984), â€œErosion control of scour during construction; report 7, CurÂ¬
rent  a waveinduced current model,â€ U.S. Army Engineer Waterways ExperiÂ¬
ment Station, Vickesburg, Mississippi.
Vemulakonda, S.R., Swain, A., Houston, J.R., Farrar, P.D., Chou, L.W., Ebersole, B.A.
(1985), â€œCoastal and inlet processes numerical modeling system for Oregon InÂ¬
let, North Carolina,â€ Tech. Report CERC856, U.S. Army Engineer Waterways
Experiment Station, Vickesburg, Mississippi.
Vengayil, P., and Kirby, J.T. (1986), â€œShoaling and reflection of nonlinear shallow water
waves,â€ Tech. Report No. 62, Coastal and Oceanographic Engineering DepartÂ¬
ment, University of Florida, Gainesville.
Visser, P.J. (1982), â€œThe proper longshore current in a wave basin,â€ Report No. 821,
Dept, of Civil Engineering, Delft University, Delft.
Whitham, G.B. (1974), Linear and Nonlinear Waves. Wiley, New York.
Yan, Y. (1987), â€œNumerical modeling of current and wave interactions on an inletbeach
system,â€ Technical Report No. 73, Dept, of Coastal and Oceanographic EngineerÂ¬
ing, University of Florida, Gainesville.
Yue, D.KP. (1980), â€œNumerical study of Stokes wave diffraction at grazing incidence,â€
Sc. D. dissertation, Dept, of Civil Eng., M.I.T., Cambridge.
BIOGRAPHICAL SKETCH
Harley Winer was born in Roanoke, Virginia on May 30,1944 and raised in Cumberland,
Maryland. There he attended public schools, roamed the mountains in his back yard and
excelled in downhill skiing. After high school, he attended Dartmouth College where his
first attempt at obtaining a college degree ended a few years short of completion. This was
followed by several years of being â€œon the roadâ€ in the Bohemian tradition of Jack Kerouac
along with many years of protesting the Vietnam war. In 1972 the author did his part
to end the war by being incarcerated in the Federal Correctional Institution at Danbury,
Connecticut for resisting the draft.
The authorâ€™s second attempt to obtain a college degree was at Lincoln University, a
small black liberal arts college located in southeastern Pennsylvania. Here he was introduced
to the political struggles of black Americans and the cultures of African people. He was
awarded a Bachelor of Arts Cum Laude in Philosophy in 1973. This degree qualified the
author for employment successively as an assembly line worker, a longshoreman, a taxi
driver and as a carpenter.
In 1978 the author enrolled as an undergraduate engineering student at the University
of Delaware. He made straight Aâ€™s and received a Bachelor of Science in Civil Engineering
in 1981. Awarded a Davis Fellowship, he continued his engineering studies as a graduate
student at the University of Delaware. He earned an M.S. in Civil Engineering in 1985 a
few years after he had moved to Florida to pursue a doctorate in coastal engineering.
In 1982 the author gave up downhill skiing and learned scuba diving upon moving to
Florida. In addition to being a certified scuba diver, the author has a private pilot license
with both land and sea ratings and is an accomplished flutist.
129
I certify that I have read this study and that in my opinion it conforms to acceptÂ¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
ig, Chairman / Â¡
Professor of Aerospace Engineering,
Mechanics and Engineering Science
I certify that 1 have read this study and that in my opinion it conforms to acceptÂ¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
f
v..
James T. Kirby, CochairrhÃ¡n
Associate Professor of Coastal and
Oceanographic Engineering
I certify that I have read this study arid that in my opinion it conforms to acceptÂ¬
able standards of scholarly presentation and is fully adequate, m scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Ben* A. Christensen
Professor of Civil Engineering
I certify that I have read this study and that in my opinion it conforms to acceptÂ¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Docto, of Philosophy.
Robert G Dean
Graduate Research Professor of Coastal
and Oceanographic Engineering
I certify that I have read this study and that in my opinion it conforms to acceptÂ¬
able standards cf scholarly presentation and is fully adequate, in scope and quality as a
dissertation for the degree of Doctor of Philosophy.
Professor of Aerospace Engineering,
Mechanics and Engineering Science
This dissertation was submit ted to the Graduate Faculty of the College of Engineering and
to the Graduate School and was accepted as partial fulfillment of the requirements for the
degree cf Doctor of Philosophy.
August 198S
Dean, Graduate School
UNIVERSITY OF FLORIDA
3 1262 08556 8029
104
particular the comparison to the laboratory experiments of Gourlay) confirm the predictive
capability of the numerical model. The simplifications mentioned above make this possiÂ¬
ble, for there must be a mathematical statement of the physical process before numerical
modelling of the process can even be attempted.
In making recommendations for further study we should distinguish between short range
objectives and the possibilities that may become open to numerical modelers by advances
in the computer hardware field. It is not difficult to envision within a few decades or less,
super computers with thousands of parallel processors. Such advances will make many of the
present computational methods obsolete. For example the parabolic approximation is made
solely to put the wave equation in a solvable form with respect to the capabilities of present
day computers. With sufficient parallel processors it is conceivable that it may be practical
to solve the full elliptic equation explicitly. We shall not dwell further on speculation about
future generations of computers. In the near term, the numerical model presented in this
report can be refined and the applications can be extended. The model is constructed in
modular form so that any new formulation for any of the component processes can easily
be incorporated into the model. The model is thus open for continual improvement.
6.1 Improved Wave Formulations
The model can be improved by better wave modeling and by a more correct formulation
of the radiation stress terms once breaking has commenced. The formulation used in this
model and previous models for the radiation stresses initiates setup and longshore currents
at the breaker fine whereas as explained in section 5.1 experiments show otherwise.
An improved wave model will assign a nonzero value of the amplitude in the lee of
a shore parallel breakwater. Research should be done to establish a method of doing this
so as to enable more wave energy to propagate into the lee of the structure. Dalrymple
and Kirby (1988) have found a method to determine a nonzero wave height in the lee of a
breakwater. However their solution is in the form of a Fourier integral which is not suited
to the complex amplitude form required in the present model.
105
6.2 Including Wave Reflection
The author has made unsuccessful attempts to include wave reflection in the model.
Kirby (1986) used coupled equations governing the interaction of a forward and a backward
(reflected) propagating wave that are coupled through a slope interaction. He was able to
obtain the reflection and transmission coefficients over submerged breakwaters that agree
well with the experiments of Seelig (1980). Once the arrays of the complex amplitudes of
the two waves is determined, the radiation stress terms can be determined from equations
that have been derived following the same procedure as used by Mei (1972) but defining
as being comprised of two components. In fact this procedure can also be used to obtain
the radiation stress terms for any number of components. The only limit being the amount
of computation required as the number of components increases.
Smith (1987) following the work of Miles (1967) for wave reflection and transmission
over a step derived equations for the complex reflection and transmission from a submerged
plane barrier. An attempt was made using this formulation to obtain the circulation and
wave pattern from a submerged breakwater represented as a plane impermeable barrier.
However problems of numerical stability were encountered. It is suspected that the sharp
gradients between the region with the presence of the reflected wave and regions of its
absence adjacent to the barrier produced unrealistic gradients in the radiation stress terms.
An attempt was made to include the reflected wave for the case of an emerged plane barrier
with the same unstable result. This is a topic for further investigation.
6.3 Extending the Model to Include Sediment Transport
The model as it currently stands can be extended to include sediment transport. The
wave model can be slightly amended to retain wave energy dissipation as an array. If a
formulation of sediment entrainment can be made in terms of the wave energy dissipation
and the bottom shear stress, everything else is in place to accomplish the sediment transport
model. A simple numerical mechanism would be to assume steady state conditions for
certain time intervals and solve for the wave and current field and the sediment concentration
106
as a function of known variables and then solve a separate model for sediment balance and
then update the depths and proceed iterating between the sediment balance model and the
present wavecurrent model marching in time. Since it has already been demonstrated that
the wave current model gives adequate results in the lee of a shore parallel breakwater, a
good test for such a sediment transport model wold be to see if it could yield a tombolo
formation in the lee of a shore parallel breakwater.
APPENDIX A
DERIVATION OF THE DEPTH AVERAGED EQUATIONS
In this appendix a derivation of the governing equations for the circulation model is
presented. This appendix closely follows the work of Ebersole and Dalrymple (1979).
Before proceeding with the derivations of the depth integrated time averaged equations
of motion and the continuity equation the Leibnitz Rule for taking the derivative of an
integral is stated since it is invoked throughout the derivation.
Ty C /(I'V) = C dJ^Tix + my)'y)dJÂ£  '^y)â€™y)dJJT
The Leibnitz Rule can also be used in reverse to take derivatives outside of an integral.
The kinematic boundary conditions are also reproduced since the use of the Leibnitz Rule
will yield bottom and surface terms that will be eliminated through use of these boundary
conditions. On the free surface the boundary condition is
dr) dr)
dr)
Wr' = Tt+'xaiJrv"Ty
(A.2)
and at the bottom the boundary condition
(A.3)
In the circulation model it is assumed that the still water depth does not vary with time so
the first term of Eq. (A.3) is dropped (actually for time varying depth essentially the same
equations are obtained).
The continuity equation (or the conservation of mass equation as it is abo called) in
three dimensions is
, d(/>u) d(pv) d(pw)
dt dx dy dz
(A.4)
107
108
Integrating over the depth from z â€” â€”h0 to z = rÂ¡ and using the Leibnitz Rule to take
derivatives outside the integrad, the continuity equation becomes
d f* . dr] d [i
atLpiâ€™p'Tt+TzL'n,i'
h0
f \ dr> r \ dh0 d n
Si Vâ€”â–
/ s
(A.5)
The terms with an overbrace are p times the free surface kinematic boundary condition
and the terms with an underbrace are p times the bottom boundary condition. Eliminating
both of these from Eq. (A.5) gives
d n . dr* . d rn J
/ Pdz+*~ pudz + â€” pvdz 0
dt Jh0 ox Jâ€”k0 dvJh0
(A.6)
Equation (A.6) is now time averaged. Time averaging means averaging over a wave
period. The time average of a term is denoted by an overbar. Thus
^ 1 fT
â– H
Fdt
(A.7)
where F represents any term and T is the wave period. Letting u and v be comprised of a
mean flow, a wave induced flow, and an arbitrary or random turbulence component
u = Ã¼ + Ãº + u'
V = V + V + v1
(A.8)
(A.9)
The turbulent fluctuations are assumed to be of a high frequency relative to the wave
frequency so that their time averages over the wave period are zero. By substituting the
above expressions for u and v into Eq. (A.6) and taking time averages the continuity equation
is
^ {p{K + â€™?)}+ ^ iPÃ¼iho + Â»?)} + JÂ¿f_k PÃºdz+ (A.10)
A {Â«Â«(*.+ fl)} + = o
(AH)
109
The time averages of the vertically integrated wave induced velocities Ã¼ and v are not zero.
These quantities are known as the wave induced mass flux.
Defining
U = u + u
and
V = v + 5
(A.12)
where
u =
1 Jn
r / pudz and
o + fj)Jk0
p{h0 + *?)
the time averaged depth integrated continuity equation is
%Ã+kuQlo+t,)+â€˜Â§Â¡Â¡viho+tl)
1 ri
v â€” ~Ti rr / pi> dz
p{h0 + fi)Jho
(A.13)
(A.14)
The assumption that the density is constant in both time and space has used in obtaining
both Eq. (A.11) and Eq. (A.14). The assumption of a constant bottom has also been used
in obtaining (A.14). If the bottom is not constant in time a term would be included in
Eq. (A.14).
The x and y momentum equations are handled in the same manner as the continuity
equation above. The x and y direction momentum equations are
du du2 duv duw 1 dp 1 ,drxx drvx drxx.
Tt + + + IT = } Ãš + + if + sT>
dv t duv t dv2 t dvw _ 1 dp 1 fdrxv drvv drZ]/
dt dx dy dz
= zir + 1A
+
}
(A.15)
(A.16)
p dy p dx dy dz
The same procedure is followed for both the x and y equations. Only Eq. (A.15) will be
dealt with here. Integrating the left hand side over the depth and applying the Leibnitz
Rule to take the derivative outside the integral gives
it Tip\ d /*'J , dr] dh0
(LHS> = ai/Â»." ~aT +
if +
dxjh0 ndx Â° dx
dy J
, , x dr) . . dh0
uvdz  (uv)â€ž â€”  (uv).fco â€” +
h0 dy dy
(UW)l ~ (UW)ho
(A.17)
110
The terms in Eq. (A.17) that are designated by an overbrace are tif, times the kinematic
free surface condition and those designated by an underbrace are u_j,0 times the bottom
boundary condition. These are both dropped leaving
, â€ž d r dr, dr
(LHS) = * U UdZ+d~x U UdZ+d~y U UV dZ ^
Substituting the definitions given for u and v in Eqn. (A.8) and (A.9) and recognizing that
the turbulent fluctuating components are of a high frequency relative to the wave frequency
so that the time averages of integrals involving the product of a turbulent component and
a wave induced component are zero gives
d~r d~r d ~r d~r
LHS = â€” / udz+ â€” Ãºdz + â€” u?dz+â€” tfdz (A.19)
dtJko dt]K0 dxJho dxjh0 â€™
d ~r~T~Z d ~r d ~r d ~r
+ Â£ {u')2dz + â€” / 2UÃš dz + â€” uv dz + â€” / uiidz (A.20)
dxJho dxjh0 dyJh0 dyJho
d ~r d ~r d ~r
+ X Ã¼vdz+â€” Ãœi)dz + â€” u'v'dz (A.21)
oyjho oyJh0 dy JK0
using the definitions of Eqs. (A.12) and (A.13) this becomes
+aiuvv,Â°+Â«)+TyS'h' \w^)Lrdâ€˜Lriz
(A.22)
where the two terms designated by underbraces result from being added to complete the
U2 and UV terms.
The right hand side after integrating over the depth assuming a constant density and
assuming an inviscid fluid such that no horizontal viscous stresses exist, i.e. rtx and Tyx are
zero, is
RHS =
if r ^
p Jho dx p Jâ€”h0 dz
(A.23)
Using the Leibnitz Rule and averaging over a wave period the right hand side becomes
1 d r . 1 dn 1 dhn 1 1
RHS = â€”â€” / pdz+ + PKozâ€” + Tzz~~ rnir (A.24)
pdxjh0 p dx p dx p â€ p v '
Ill
TZXq and rMX_ko are the time averaged surface and bottom shear stresses. Assuming that
the free surface pressure pn is zero (the dynamic free surface boundary condition) the right
hand side is
___ 1 d 7* 7 1 dh~0 11 ,
RHS = â€”â€” / pdz + p.hâ€”Â± + TÂ¿TiÂ¿ (A.25)
pdxjh0 p dx p p â€™
The mean pressure at the bottom can be expressed as the sum of the wave induced or
dynamic pressure and the hydrostatic pressure
Pfco = PdVnko + P9[h0 + r>)
(A.26)
so that the product Ph0fa can be manipulated to read
dh0 1 dh0 13.,. _.2, . .dfj
'â€™*â€¢17 = + 2Ã¡Ã' l(hÂ°+
(A.27)
Terms from both the right and left hand sides are combined as the radiation stresses.
Defining
sâ€ž = f'jp+,07) iz  \P9(hâ€ž + Â«)*  rM'Y
~r> i ~Ã± r>
S,^j ^vdz  ^ ,0 i, j^dz
Sn = /_ JÂ» + ,07) dz  1,,(*. + Â«)>  ^7) {/I P5 ^
(A.28)
(A.29)
(A. 30)
and making the assumptions that the product of the dynamic pressure at the bottom and
the bottom slope is negligible, that the Jif?h0(u')2dz term can be ignored and that the
lateral friction tj caused by momentum fluxes due to turbulent fluctuations is independent
of depth and that rj is defined by rÂ¡ = â€”pu'v', the x direction equation of momentum is
written as
3 Â»r2,l , ^ , d rnr/L . . ldSxx . 1 dS,
+ 
XV
5^ + Â« + arU '<*â€¢ + *Â» + TyUV^ + + , 8x ' , 8â€ž
+,(Â». +fl5+(ÃAÂ±Â«Ã^irâ€ž + In, = o
OX P Oy p p
(A.31)
By subtracting the depth integrated continuity equation and dividing by the total depth,
the final form of the x momentum equation is
8U dU_ dU_ 1
dt + dx dy + p{h0 + fj)
dST
dS.
xv
dx
dy .
(A.32)
112
dfj 1 drÂ¡
+ 9TZ + T1
: +
â€”rn* = 0
(A.33)
dx pdy p{h0 +p{h0 + rÂ¡)
The same procedure gives as the depth integrated time averaged y momentum equation
dV TTdV rrdV 1 dSt
1 U h V 1
dt dx dy p(h0 + fj) dx
xy
dS,
vv
dy .
drÂ¡ 1 dri 1
+ 9dyJr pdx p[h0 + f}) T'v
p(K + v)
rr Hv = 0
(A.34)
(A.35)
APPENDIX B
DERIVATION OF THE RADIATION STRESS TERMS
The basis of the derivation is to use the definition of the velocity potential to substitute
for the wave induced velocities in the definition for the radiation stresses.
d
, _ d
w =
zp dz
where is expressed in terms of the general complex amplitude function B as
_ 1 (_ig â€žcoshk(h0 + z) _iut
2 V a
B cosh kh
+ cc)
where c.c. represents the complex conjugate term. Also rÂ¡' is expressed as
r,' = l(Beiut + c.c)
where the prime is used to denote wave induced quantities.
For convenience Eq. (2.50) of Section 2.2 is reproduced below.
S<*p = P u'au'pdz + 6ap ^ pdz  y {h0 + Â»j)2 + y(r?')2
An expression for p is obtained in Section 2.2 and is reproduced below
d IL .W1
dz  p{w'y
Substituting Eq. (B.5) into Eq. (B.4) gives
Sap = P I Ku'pdz + SaP
* â€ /l o
(B.l)
(B.2)
(B.3)
(B.4)
(B.5)
ri â€ž r* ( r* du'w' \
J ^ P9Ã1 ~ Z)dz + J ^ ( pj ^â€”dz\dz
'  .. ' _ Â° ' * * '
/
p[w'Y dz  Y{K + r))2 + ^(Â»?')2
Â«O L
(B.6)
113
114
The terms of Eq. (B.6) are now evaluated term by term.
Term 1:
, _ 1 / ig dB cosh k[h0 + z) _iut ig dB' cosh k(h0 + z) iu)t
u'au'p â€”
, = 1 (_
Uâ€œ 2 \ a dxa cosh kh
ff2 cosh2 k(h0 + 2)
a dxD
cosh kh
(B.7)
55 51T dB^dB_ _ dB_dB_e2iut _ dB' dB* Xut
4a2 cosh2 kh dxa dxp dxa dip dxa dip
dxa dip
â€”râ€”r pg cosh2 k(h0 + z)
pUÂ°Ut = 2Vsinh 2kh
pg cosh2 k(h0 + 2)
Jfcsinh 2 kh
dB dB' dB' dB
+
dxa dip dxa dip
â€˜ (dB_W\
dip )
(B.8)
(B.9)
Ã pu'auâ€™pdz =
* â€œ* o
P9
â€¢1
( dB 55* y
k sinh 2kh
^5xa dip )
P9
*1
( dB 55* \1
ksmh2kh
\dxa dxp )\
Â£Â»(****:) i(i+
4 dip J k2 \ si
/ cosh2 k(h0 + 2) dz
â€¢ hp
^ (sinh 2k(h0 + fj) + 2k[h + fj))
2kh \
sinh 2khJ
The second term:
J_h pg{v  z)dz = pg ^r)z 
hÂ«
... â€ž(* + *)*
P9^~
The third term:
f If *ff dB LUÃ.U ".yo T â€œJ _lu,f , â€¢ ># I.UOU T *J
â€˜ ~ 2 { a dxt 6 + " â€œ 6
cosh k(h0 + 2) _iult ig dB' cosh k(h0 + z)
cosh kh a dxÂ¡ cosh kh
igk psinh k(h0 + z) _lult ^ iff* D, sinh k[h0 + 2) j
cosh kh
g sinh 2A:(h0 + 2)
â€¢ Â»it' â€” L. ^
4sinh2fch 5x, ' dx(
tv
, _ 1 / Â«ffk sinhk[h0 + 2) lult iff* .sinh fc(h0 + z)
2 V <7 cosh kh a cosh kh 6 J
dB 55* '
â€”5 + â€”5
u^ur =
(BIO)
(B.ll)
(B.12)
(B. 13)
(B.14)
/;
dxÂ¡
_ 9_d_
4 dx
dz =
dB
dxf
dB' J r
_ ?_L ÃÂ£Â£b â€”sÃ f_
4 5xf 5xf 5xf J [2)fc sinh 2fc(/ic + ij) 21: sinh 2k[h0 + tj)_
r,} sinh 2k{h0 + 2)
sinh 2Â¿h
cosh 2k[h0 + Â»;)
cosh2fc(h0 + 2)
(B.15)
115
_pg d \dB B, P* / cosh 2k[h0 + rÂ¡) cosh 2k(hâ€ž + z)
4 dxÂ¡ [dxf dx( J Jhâ€ž \2fcsinh2A:(h0 + fj
PI
4
?lJL
4 dx
P9 d
___n .
, t , * j  â€”w , Â» rÂ¡) 2k sinh 2k{h0 + fj))
â€” â€”a* + d^B] ( cosh2fe(/lo + ^) (h , â€”\ sinh2fc(/i0 + fj) \
dx( dx( dx( J \ 21; sinh 2k(h0 + fj)' 0 ^ 4fc2 sinh 2k{h0 + fj) J
r 1 '2fchcosh 2fc/i sinh2A:/i
as . as* J /
:f [dx( + axf BJ (
41:2 sinh 2fch 4k2B\nh2kh)
dB
dB'
T^&B' + iÂ¿;B\^2khcoth2khv
Expanding the derivative term in Eq. (B.16)
dx(
dB dB* â€ž
â€”Â£* + â€”B
ax
(
dx.
d2B a2B* dBdB*
rB* + ^iB + 2
ax2
ax2
axf ax
f
Recognizing that f is a free index so that = V*
a2B
ax2
5Ã7
= VÂ¿B = Jfc2B
d2B'
dx?
= V\B* = â€”k2B'
(B.16)
(B.17)
which results from use of the Helmholtz equation which governs diffraction in water of
constant depth and
2^Â§!r=2vâ€˜B! (B18)
so that the expanded derivative term (B.17) becomes
a2B . a2B* dBdB*
ax2 B + ax2 B + 2dx( dx(
= 2V*B2  2Jfc2B
(B.19)
and the result is
/l ("/," lÂ£fÃZ) = T (Vâ€˜S1 " 2F  1] (B.20)
The author wishes to express some reservations concerning the validity of Eq. (B.20).
The shallow water asymptote for this expression is zero since for small k(h+fj), coth 2Jt(h +
*1) ~ tk[h+n) > and 2fc(h+r?)coth2fc(/i+fj)l 0. In deepwater, however, coth2Jfc(h+Â»j) â€œ 1
so that 2k(h + fj) coth2k(h + fj)  1 =Â£ 2kh. If the derivative terms are nonzero, the Szz
116
term will be proportional to the depth. This would mean that deep water nonshoaling and
nondissipative waves in varying depths would produce gradients in the radiation stresses.
For plane waves this is not a problem, since the complex amplitude terms are easily shown
to be zero. This is seen by substitution of B = ae,kx and B* = ae~tkx where a = B.
The fourth term is straight forward
igk psinh k(h0 + z) _iut igk sinh k(h0 + z) ,â€¢
cosh kh
+ â€”B'
cosh kh
#_ i ( Â»'gfc
2 V o
, , = â– *(> +z) _ B.^, _
4 a2 cosh2 kh ' >
77T2 _ 1 g2k2 sinh2 k(h + z) n gk sinh2 k{h0 + z)
( } "4 a2 cosh2 k[h + rj) ~ sinh2Jkh BB
(B.21)
(B.22)
(B.23)
 Ã p(w')2dz = pgk\B\2 Ã
Jâ€”h0 "â€”h0
2 sinh2 k(ha + z)
sinh2Â¿h ~ 4 1 1 V* sinh 2kh,
The fifth term needs no evaluation and is seen to cancel exactly the second term.
The sixth and last term of Eq. (B.6) is rÂ¡')2 and is evaluated easily
rÂ¡' = 1 (Â£eâ€˜wi + B*eiut)
(r/)2 = 1 (BeW + 2 BB* + B*e2,wt)
SI
W=2^*= 2
fOT=fl*2
which cancels part of the result from the fourth term.
Adding the above results for all the terms of Eq. (B.6) yields
P91
(B.25)
(B.26)
(B.27)
(B.28)
c P9Ã~(9BdB*\ 1 ^ , 2kh \
4 I \5za dip y A:2 V sinh2kh)
+
^af)
I 2â€”2kh
1 1 sinh 2kh
dxa dip
(V*Â£2  k2\B\2)^(2kh coth 2kh  1)] 
(B.29)
APPENDIX C
LAGRANGIAN DERIVATION OF THE GOVERNING WAVE EQUATION
The governing equation for wave motion can be derived through use of a variational
principle where a Lagragian is the function that is varied. Luke (1967) showed that a
Lagrangian for waves is given by
L â€” Jh [^â€˜ + 3Z
dz
(C.l)
Kirby (1984) used a time averaged Lagrangian to derive a wave equation. For the following
derivation, equation (C.l) will be used. The third term in Eq. (C.l) is determined straight
forward, while the first two terms need to be expanded about fj the mean water line in a
Taylor series.
L=fâ€™4+i:
4>t +
(W)J
dz + {rÂ¡  rÂ¡)4>t â€ž +(f7  fj) ^
(C.2)
We assume that rj and can be expressed as
*) = V + cÂ»?i
= o + efi (C.3)
where e is an ordering parameter and / is the depth dependency as defined in Eq. (3.3). For
the sake of simplicity we will assume that the gradient of fo represents the steady horizontal
currents. This means that fo is not a function of time and that W will not be included as
it was in the derivation in Chapter 3. This is because the W term is not expanded about
z = fÂ¡ as it was in the Greenâ€™s function derivation and is not needed in order to obtain the
(Vfc â€¢ U)jÂ£ term which as will be seen is obtained more directly in the present derivation.
Including W here will only produce terms that are of higher order in the mean current.
117
118
Introducing (C.3) into (C.2) yields
Â¿gvi gh20
L = ?Y + Wrh + ^YÂ±9Â£ + Â«t>itJ'h fdz+^(h0 + fj) +
cVfc^oVfc<^i [ f dz + eV/,^0^1 ( Ã Vnfdz+f ftdz\ +
Jh0 Uhâ€ž JKc )
(y
2
Iâ€™ [/Vâ€ž/+//,]dz} +
e2rÂ¡iit + crti ( â€”hÂ£Â°^ + cVhfoVh4>i + eV^^o^i/* *j +0(e'
! 2
The Lagrangian L is assumed to be of the form
>)}
(C.4)
It then follows that
L â€” Lq + fLi + e* L2 + fSZ>s + 1
Â¿0iÂ£iS+M(*.+,)
and
(C.5)
(C.6)
Â¿1 = gf>m + â€”Â¿1, + â€”v^oV^! { r v*/ dz + r fzdz) + m (V^o)2 (c.7)
0 fl Ufco â€¢/*â€ž J 2
ti'Vhti Ã [f^hf + / /*] oifx #J (c.8)
* â€œ/lo
To obtain the desired linear governing equation is varied with respect to i and then Li
is varied with respect to rji in order to eliminate rÂ¡\. The variational principle holds that the
Lagrangian is constant under small variations of the unknown parameters. In mathematical
terms
and
DL
D
DL
Dr)!
= 0
= 0
(C.9)
(cio)
Since the ordering parameter e is arbitrary the above holds for all L,â€¢ Â» = 0,1,2, â€¢ â€¢ â€¢. In the
Li term \ is expressed in terms of its derivatives. This requires integration by parts to
119
obtain the variation with respect to \. Upon carrying this out it is found that
PÂ± = dLi_ d dU _ dU x
D
Applying (C.ll) to (C.8) and recognizing that the variation d\ is arbitrary yields after
carrying out the integrations in Â¿2
CC â€” k2CC
 V*(â€”*v*^) + M ')  nu  v*(v*^o9i) = o (C.12)
9 9
which is rewritten as
+ ^h{CCgVhi)  i(
Varying Li with respect to iji yields
0*71 + iÂ« + Vh\ = 0
(C.14)
which can be written as
lD!
Vl = pr
g Dt
(C.15)
which upon substitution into (C.13) yields the governing equation
^ + (V^o)^  Vfc(CC,VÂ»*) + (a2  k'CC'fa = 0
which upon making the substitution that 0 = V/,^o becomes
+ (v* * " v*(cc,v*^) + (
(c.i6)
Dt
(C.17)
APPENDIX D
OBTAINING EQUATION (3.79)
In this Appendix we show the algebra involved in obtaining equation (3.79) from equaÂ¬
tion (3.51) using the assumption that the dispersion relationship is written as
u â€” a ( k cos OU
We start by using the following form of <Â¡>
 ig eÂ«'(/ * CO. i dz)
and then make a phase shift by using the substitution
Ã coÂ»f dxJ k cos 6 dx)
so that in the final form
_ ig A'(X>V)e*(f Icoiidz)
(D.l)
(D.2)
(D.3)
(D.4)
and all y dependency is solely in the complex amplitude A'. For reference equation (3.51)
is reproduced below.
(P4>z)z + (pv)v + k2p(Â¡> + [w2  a2 + i'w(Vfc â€¢ U)j I i(7w<Â£
{U24>z)z{V24>V)V{UV4>X)V{UVÃV)X + 2Ã = o (d.5)
Writing out all the components of equation (D.5) and rearranging terms gives
{p ~ U2)xx + (p  U2)4>zx + [(p  V2)4>v] v + k2p4> + (a;2  o2)*
+iojUx4> + iuVv4> + icrw(j>  (UVx)v  [UVv)x + 2iuUx + 2iuV^>v = 0 (D.6)
The derivatives of defined by equation D.2 are as follows
4>x = â€”ig [(^) + ik COS 9^ câ€˜(/ k cot 0dz)
120
121
ZZ â€”
+ 2tfccos0^ + i(Jbcosi)a
â€” k2 cos2 9 â€”1 fc co* *dx)
a a J
{4>z)v â€”
"'I
{(a) e*(^*co,#
+i(k cos 9)y e<(/*cÂ°Â«#rf*)j
II
'^e.(/fcco.idx)J
H
II
G>x)v
The above forms of the derivatives recognize that the phase function eâ€˜(/icoÂ»edz) y
derivatives since k = k(x, y). However the y derivatives of the phase function integral are
not evaluated since the phase shift (equation D.3) will eliminate any y dependance from
the phase function and place all y dependance in the complex amplitude. The ( â€”)
v Â° / XX
term is neglected following the assumption explained in section 3.2. Substituting the above
derivatives into equation (D.6) gives
(pU2)x () e*'(/tco,#dl) + ikcos9(p  U*)gei
\
+2tftcos 9(p â€” U2) () e'(f kcotedx) + i(kcos 6)z(p â€” U2')â€”ei(fkco*9dx)
k2 cos2 9 [p  U2) â€”eâ€™(fkco,tdz) +
(p fÃ¡eHf tcosedz)\
V Jo
l\ J V* /rj
+k2pei(fkcotidx) + (u2  a2)eâ€™(ftcoetdx) + iuUxe^f kcoâ€˜9 dx^
o o a
/ A S
+twVy~e^f *â€œ*6dx) + iwAe^f kcoâ€˜9^  {UV)v eâ€™(/ *coâ€˜9dz>
2 UV
! 6 dx)
tJfccos 9UUV â€”e'(f kco,e dl)
â€”ik cos 9UVv â€”*co*9dz^  i{kcos 9)vUVkcot 9 dl>
a a
2ikcos9UV (Ã¡ei(fkâ€œÂ»edx)'j â€” i(kcos 0)vUVâ€”kcot9dz)
r ^ S
~(UV)z (^eâ€˜(/4co,Ã
2ojkcoe9U^fkcot9dx^ + 2iuV (Â¿(f kcc"9dx)'j
= 0(D.7)
122
Dropping terms that contain products of derivatives of j and derivatives of the current as
well as products of the current and second derivatives of Â£ (i.e. terms that are designated
by an overbrace) equation (D.7) is rearranged as
IÂ» [fccosi(p â€” U2) + u/E/j â€” + 2Â» [fccosÃ(p â€” E/2) + u/l/] ^ +Â»Wi4
+ [~fc2 cos2 0(p  U2) + k2p + u/2  a2  2uk cos OU^ ^ â€” i(k cos 0)vUV â€”
+t [Vv(w  kcosOU)  [k cos 6U)yV  k cos 0UV] j j e^f k COi â€˜ dx>>
+ [P(ae,^fcCO,i<,l)) + 2s â€” Jbcos17)] ^Ã¡eHftcctdx)J = ^D 8)
It is easily shown that
kcos0{p  U2) +
By writing the dispersion relationship as
<7 = u> kcosOU
it is also easily shown that
ui â€” a = 2o>k cos 6U â€” {k cos 6 Uy
From equation D.2 it is apparent that for a constant o/,
(fccos0)â€ž = ov
Thus equation (D.8), becomes
{â€¢ [a{Ct cosÃ + U)]x j + 2ia(Ct cosÃ + E7) ^
+ [fc2p(l  cos2 Ã)] J + iwA + IVvA + Â«7vV^} eâ€˜(/*â€œâ€¢Â«<**)
+2iVa ^ei(fkcot8dx)'j + p tco. tdx)^
where a i(k cos 0)vUVÂ£ term has been dropped.
At this point a phase shift is introduced
v.
= 0
(D.9)
(DIO)
(DU)
(D.12)
(D.13)
^ _ Â¿IgUf icoÂ»f dxf k coÂ« 6 dz)
(D.14)
123
where k and 0 are a reference wave number and wave angle. For computational purposes k
will be an averaged wave number across a grid row and 6 will be taken as the wave angle in
the absence of currents using Snellâ€™s law. Further it will be assumed solely for computational
purposes that 0 is equal to 8. This gives
Â»[
2(Cg cos 0 + U) cos 0(k  k)A
+Â«wA + iVvA + 2Â»V<7 (â€”) +iavVâ€” = 0 (D.15)
\o Jv o
Expanding the derivatives of Â£ and with a few minor manipulations of the terms the
governing equation (3.79) is obtained.
(C, cos 0 + Â¿*+ Â«'(
+V +
kCt(l  cos2 0)A   ICCg
(i).l
z
w
+ A = 0(D.16)
BIBLIOGRAPHY
Arthur, R.S. (1950), â€œRefraction of shallow water waves: the combined effects of currents
and underwater topography,â€ Trans. Amer. Geophysical Union, 31(4): 549552.
Bateman, H. (1944), Partial Differential Equations of Mathematical Physics, Dover PubÂ¬
lications, New York.
Berkhoff, J.C.W. (1972), â€œComputation of combined refractiondiffraction,â€ Proc. 13*^
Coastal Eng. Conf., Vancouver, 471490.
Berkhoff, J.C.W. (1976), â€œMathematical models for simple harmonic linear water waves:
wave diffraction and refraction,â€ Publ. no. 163, Delft Hydraulics Laboratory,
Delft.
Birkemeier, W.A., and Dalrymple, R.A. (1976), â€œNumerical models for the prediction of
wave setup and nearshore circulation,â€ Ocean Engineering Report No. 3, Dept,
of Civil Eng., University of Delaware, Newark, Delaware.
Booij, N. (1981), â€œGravity waves on water with nonuniform depth and current,â€ Report
no. 811, Dept, of Civil Eng., Delft Univ. of Technology, Delft.
Bowen, A.J. (1969), â€œThe generation of longshore currents on a plane beach,â€ J. Mar
Res.. 27(2): 206215
Bowen, A.J., Inman, D.L., and Simmons, V.P. (1968), â€œWave â€˜set downâ€™ and â€˜setupâ€™,â€
J. Geophys. Res., 73(8): 25692577.
Bretherton, F.P., and Garrett, C.J.R. (1968), â€œWavetrains in inhomogeneous moving
media,â€ Proc. Roy. Soc., A 302: 529554.
Butler, H.L. (1978), â€œ Coastal flood simulation in stretched coordinates,â€ Proc. 16^
Coastal Eng. Conf., Hamburg, 10301048.
Butler, H.L. (1980), â€œEvolution of a numerical model for simulating long period wave
behavior in oceanestuarine systems,â€ Estuarine and Wetlands Processes with
Emphasis on Modeling, Marine Science Series, vol.ll, Plenum, New York.
Butler, H.L., and Sheng, Y.P. (1984), â€œBoundary conditions for limited area modeling,â€
Proc. 19fck Coastal Eng. Conf., Houston, 858870.
Carnahan, B., Luther, H.A., and Wilkes, J.O. (1969), Applied Numerical Methods, WiÂ¬
ley, New York.
124
125
Dally, W.R. (1980), â€œA numerical model for beach profile evolution,â€ M.S. Thesis, UniÂ¬
versity of Delaware, Newark.
Dally, W.R. (1987), â€œWave transformation in the surf rone,â€ Technical Report No.
69, Dept, of Coastal and Oceanographic Engineering, University of Florida,
Gainesville.
Dally, W.R., Dean, R.G. and Dalrymple, R.A. (1984), â€œ A model for breaker decay on
beaches,â€ Proc. 19*^ Coastal Eng. Conf., Houston, 8298.
Dean, R.G., and Dalrymple, R.A. (1983), Water Wave Mechanics for Scientists and
Engineers, PrenticeHall, Englewood Cliff's, New Jersey.
Ebersole, B.A., and Dalrymple, R.A. (1979), â€œA numerical model for nearshore circulation
including convective acceleration and lateral mixing,â€ Ocean Engineering Report
No. 21, Dept, of Civil Eng., University of Delaware, Newark, Delaware.
Ebersole, B.A., and Dalrymple, R.A. (1980), â€œNumerical modeling of nearshore circulaÂ¬
tion,â€ Proc. 17^ Coasted Eng. Conf., Sydney, 27102725.
Garrett, C.J.R. (1967), â€œDiscussion: the adiabatic invariant for wave propagation in a
nonuniform moving medium,â€ Proc. Roy. Soc., A 299: 2627.
Gourlay, M.R. (1974), â€œWave setup and wave generated currents in the lee of a breakÂ¬
water or headland,â€ Proc. 14*^ Coastal Eng. Conf., Copenhagen, 19761995.
Hales, L.Z. (1980), â€œErosion control of scour during construction; report 3, Experimental
measurements of refraction diffraction, and current patterns near jetties,â€ U.S.
Army Engineer Waterways Experiment Station, Vickesburg, Mississippi.
Hansen, J.B., and Svendsen, I.A. (1979), â€œRegular waves in shoaling water. ExperimenÂ¬
tal data,â€ Series Paper 21, Inst, of Hydrodynamics and Hydraulic Engineering,
Technical University of Denmark, Lyngby.
Harrison, W., and Wilson, W.S. (1964), â€œDerivation of numerical calculation of wave
refraction,â€ Technical memorandum No. 6, U.S. Army Coastal Engineering ReÂ¬
search Center, Ft. Belvoir.
Horikawa, K., and Kuo. CT. (1966), â€œA study of wave transformation inside the surf
zone,â€ Proc. 10^ Coastal Eng. Conf., Tokyo, 217233.
Johnson, J.W. (1947), â€œThe refraction of surface waves by currents,â€ Trans. Am.
Geophys. Union, 28: 867874.
Johnson, J.W., Oâ€™Brien, M.P., and Isaacs, J.D. (1948), â€œGraphical construction of wave
refraction diagrams,â€ Tech. Report No. 2, U.S. Hydrographic Office, Washington,
D.C.
Jonsson, I.G. (1981), â€œBooijâ€™s currrentwave equation and the ray approximation,â€ Prog.
Report 54, pp 720, Inst. Hydrodyn. and Hydraulic Eng., Technical University of
Denmark, Lyngby.
126
Keller, J.B. (1958), â€œSurface waves on water of nonuniform depth,â€ J. Fluid Mech.. 4:
607614.
Kirby, J.T. (1983), â€œ Propagation of weaklynonlinear surface water waves in regions
with varying depth and current,â€ ONR Tech. Report 14, Research Rep. CE8337
Dept, of Civil Eng., University of Delaware, Newark.
Kirby, J.T. (1984), â€œA note on linear surface wave current interaction over slowly varying
topography,â€ J. Geophys. Res., 89(C1): 745747.
Kirby, J.T. (1986), â€œ A model for gradual reflection of weakly twodimensional waves
in a slowly varying domain with currents: governing equations,â€ Tech. Report
No. 60, Dept, of Coasted and Oceanographic Engineering, University of Florida,
Gainesville.
Kirby, J.T., and Dalrymple, R.A. (1982), â€œNumerical modeling of the nearshore region,â€
ONR Tech. Report 11, Research Report CE8224, Dept, of Civil Eng., University
of Delaware, Newark.
Kirby, J.T., and Dalrymple, R.A. (1986), â€œModeling waves in surfzones and around isÂ¬
lands,â€ J. Waterway, Port, Coastal and Ocean Engineering, 112: 7893.
Le MÃ©hautÃ©, B., and Koh, R.C.Y. (1967), â€œOn the breaking of waves arriving at an angle
to the shore,â€ J. Hydraulic Res., 5(1): 6788.
Liu, P.LF., and Mei, C.C. (1975), â€œEffect of a breakwater on nearshore currents due to
breaking waves,â€ Technical Memorandum No. 57, Coastal Engineering Research
Center, Fort Belvoir, Va.
LonguetHiggins, M.S. (1970a), â€œLongshore currents generated by obliquely incident sea
waves, I,â€ J. Geophys. Res., 75(33): 67786789.
LonguetHiggins, M.S. (1970b), â€œLongshore currents generated by obliquely incident sea
waves, II,â€ J. Geophys. Res., 75(33): 67906801.
LonguetHiggins, M.S., and Stewart, R.W. (1962), â€œRadiation stress and mass transport
in gravity waves, with application to â€˜surf beatsâ€™,â€ J. Fluid Mech.. 13: 481504.
LonguetHiggins, M.S., and Stewart, R.W. (1963), â€œA note on wave setup,â€ J. Mar. Res..
21(1): 410.
LonguetHiggins, M.S., and Stewart, R.W. (1964), â€œRadiation stresses in water waves; a
physical discussion, with applications,â€ DeepSea Research, 11: 529562.
Luke, J.C. (1967), â€œA variational principle for a fluid with a free surface,â€ J. Fluid Mech..
27(2): 395397.
Mei, C.C. (1972), â€œA note on the averaged momentum balance in twodimensional water
waves,â€ J. Mar. Res.. 31(2): 97104.
Mei, C.C. (1982), The Applied Dynamics of Ocean Surface Waves, WileyInterscience,
New York.
127
Miche, R. (1944), â€œMouvement ondulatoires de la mer en profondeur constaanttte ou
dÃ©croissante,â€ Ann, des PontB et ChaussÃ©ss.
Miles, J.W. (1967), â€œSurface wave scattering matrix for a shelf," J. Fluid Mech.. 28:
755767.
Miles, J.W. (1982), â€œOn surfacewave diffraction by a trench,â€ J. Fluid Mech.. 115: 315
325.
Munk, W.H., and Arther, R.S. (1952), â€œWave intensity along a refracted ray,â€ in
Gravity Waves, NBS Circ. 521: 95108.
Newman, J.N. (1972), â€œDiffraction of water waves,â€ Applied Mech. Review, 25(1): 17.
Noda, E., Sonu, C.J., Rupert, V.C., and Collins, J.I. (1974), â€œNearshore circulation under
sea breeze conditions and wavecurrent interactions in the surf zone,â€ Tetra Tech
Report TC1494.
Oâ€™Brien, M.P. (1950), â€œWave refraction at Long Beasch and Santa Barbara, California,â€
Beach Erosion Board Bulletin. 4(1): 112.
Penney, W.G., and Price, A.T. (1944), â€œDiffraction of sea waves by breakwaters,â€ TechniÂ¬
cal History Report No. 26, Directorate Mise. Weapons Development, Washington,
D.C.
Penney, W.G., and Price, A.T. (1952), â€œThe diffraction theory of sea waves and the
shelter afforded by breakwaters,â€ Phil. Trans. Roy. Soc., A 244: 236253.
Perlin, M., and Dean, R.G. (1983), â€œA numerical model to simulate sediment transport
in the vicinity of coastal structures,â€ Mise. Report. No. 8310, U.S. Army Coastal
Engineering Research Center, Ft. Belvoir, Virginia.
Phillips, O.M. (1969), The Dynamics of the Upper Ocean, Cambridge University Press.
Pierson, W.J. (1951), â€œ The interpretation of crossed orthogonals in the wave refraction
phenomena,â€ Tech. Mem. No. 21, Beach Erosion Board, Washington, D.C.
Putnam, J.A., and Arthur, R.S. (1948), â€œDiffraction of water waves by breakwaters,â€
Trans. Am. Geophys. Union, 29: 481490.
Radder, A.C. (1979), â€œOn the parabolic equation for water wave propagation,â€ J. Fluid
Mech.. 95(1): 159176.
Seelig, W.N. (1980), â€œTwodimensional tests of wave transmission and reflection charÂ¬
acteristics of laboratory breakwaters,â€ Technical Report No. 801, U.S. Army
Coastal Engineering Research Center, Ft. Belvoir, Virginia.
Sheng, Y.P., and Butler, H.L. (1982), â€œModeling of coastal currents and sediment transÂ¬
port,â€ Proc. 18^ Coastal Eng. Conf., Cape Town, 11271148.
Shore Protection Manual (1977), U.S. Army Coastal Engineering Research Center, Ft.
Belvoir, Virginia.
128
Smith, J. (1987), â€œOn surface waves crossing a step with horizontal shear,â€ J. Fluid
Mech.. 175: 395412.
Svendsen, I.A. (1984), â€œWave heights and setup in a surf zone,â€ Coastal Eng. 8(4):
303329.
Tappert, F.D. (1977), â€œThe parabolic approximation method,â€ In Wave Propagation and
Underwater Acoustics (edited by J.B. Keller and J.S. Papadakis) Chapter 5,
Springer Verlag, Berlin.
Van Dorn, W. (1953), â€œWind stress on an artificial pond,â€ J. Mar. Res.. 12(3): 249276.
Vemulakonda, S.R. (1984), â€œErosion control of scour during construction; report 7, CurÂ¬
rent  a waveinduced current model,â€ U.S. Army Engineer Waterways ExperiÂ¬
ment Station, Vickesburg, Mississippi.
Vemulakonda, S.R., Swain, A., Houston, J.R., Farrar, P.D., Chou, L.W., Ebersole, B.A.
(1985), â€œCoastal and inlet processes numerical modeling system for Oregon InÂ¬
let, North Carolina,â€ Tech. Report CERC856, U.S. Army Engineer Waterways
Experiment Station, Vickesburg, Mississippi.
Vengayil, P., and Kirby, J.T. (1986), â€œShoaling and reflection of nonlinear shallow water
waves,â€ Tech. Report No. 62, Coastal and Oceanographic Engineering DepartÂ¬
ment, University of Florida, Gainesville.
Visser, P.J. (1982), â€œThe proper longshore current in a wave basin,â€ Report No. 821,
Dept, of Civil Engineering, Delft University, Delft.
Whitham, G.B. (1974), Linear and Nonlinear Waves. Wiley, New York.
Yan, Y. (1987), â€œNumerical modeling of current and wave interactions on an inletbeach
system,â€ Technical Report No. 73, Dept, of Coastal and Oceanographic EngineerÂ¬
ing, University of Florida, Gainesville.
Yue, D.KP. (1980), â€œNumerical study of Stokes wave diffraction at grazing incidence,â€
Sc. D. dissertation, Dept, of Civil Eng., M.I.T., Cambridge.
BIOGRAPHICAL SKETCH
Harley Winer was born in Roanoke, Virginia on May 30,1944 and raised in Cumberland,
Maryland. There he attended public schools, roamed the mountains in his back yard and
excelled in downhill skiing. After high school, he attended Dartmouth College where his
first attempt at obtaining a college degree ended a few years short of completion. This was
followed by several years of being â€œon the roadâ€ in the Bohemian tradition of Jack Kerouac
along with many years of protesting the Vietnam war. In 1972 the author did his part
to end the war by being incarcerated in the Federal Correctional Institution at Danbury,
Connecticut for resisting the draft.
The authorâ€™s second attempt to obtain a college degree was at Lincoln University, a
small black liberal arts college located in southeastern Pennsylvania. Here he was introduced
to the political struggles of black Americans and the cultures of African people. He was
awarded a Bachelor of Arts Cum Laude in Philosophy in 1973. This degree qualified the
author for employment successively as an assembly line worker, a longshoreman, a taxi
driver and as a carpenter.
In 1978 the author enrolled as an undergraduate engineering student at the University
of Delaware. He made straight Aâ€™s and received a Bachelor of Science in Civil Engineering
in 1981. Awarded a Davis Fellowship, he continued his engineering studies as a graduate
student at the University of Delaware. He earned an M.S. in Civil Engineering in 1985 a
few years after he had moved to Florida to pursue a doctorate in coastal engineering.
In 1982 the author gave up downhill skiing and learned scuba diving upon moving to
Florida. In addition to being a certified scuba diver, the author has a private pilot license
with both land and sea ratings and is an accomplished flutist.
129
I certify that I have read this study and that in my opinion it conforms to acceptÂ¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
ig, Chairman / Â¡
Professor of Aerospace Engineering,
Mechanics and Engineering Science
I certify that 1 have read this study and that in my opinion it conforms to acceptÂ¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
f
v..
James T. Kirby, CochairrhÃ¡n
Associate Professor of Coastal and
Oceanographic Engineering
I certify that I have read this study arid that in my opinion it conforms to acceptÂ¬
able standards of scholarly presentation and is fully adequate, m scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Ben* A. Christensen
Professor of Civil Engineering
I certify that I have read this study and that in my opinion it conforms to acceptÂ¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Docto, of Philosophy.
Robert G Dean
Graduate Research Professor of Coastal
and Oceanographic Engineering
I certify that I have read this study and that in my opinion it conforms to acceptÂ¬
able standards cf scholarly presentation and is fully adequate, in scope and quality as a
dissertation for the degree of Doctor of Philosophy.
Professor of Aerospace Engineering,
Mechanics and Engineering Science
This dissertation was submit ted to the Graduate Faculty of the College of Engineering and
to the Graduate School and was accepted as partial fulfillment of the requirements for the
degree cf Doctor of Philosophy.
August 198S
Dean, Graduate School
NUMERICAL MODELING OF WAVEINDUCED CURRENTS USING A PARABOLIC
WAVE EQUATION
By
HARLEY STANFORD WINER
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988
{UNIVERSITY OF FLORIDA LIBRARIES

