Energy Balance Model for Breaking Solitary Wave Runup

Ying Li1 and Fredric Raichlen2

Abstract: The runup of breaking solitary waves on a plane beach has been investigated. Experimental measurements demonstrate the
effect of the impact of the jet from a plunging breaking solitary wave and the postbreaking bore formed on the resultant runup. An
empirical method was developed based on energy conservation principles to provide an estimate of the runup of breaking solitary waves
on a plane slope. Energy dissipation associated with wave breaking was estimated using the results from a numerical model developed by
Li. The results from this highly simplified energy conservation model agree reasonably well with experiments, and this model appears to
be useful in predicting breaking solitary wave runup on a plane beach.
DOI: 10.1061/共ASCE兲0733-950X共2003兲129:2共47兲
CE Database keywords: Wave runup; Breaking waves; Solitary waves; Energy dissipation.

Introduction propagating up a 1:15 uniform slope in the laboratory. 共Fig. 2 was

The ocean waves generated by seismic events 共tsunamis兲 usually constructed from still photographs taken at different locations
have long wavelengths and small wave heights. As they propagate along a plane beach during 10 different experiments with the
shoreward, due to the offshore bathymetry, the wave heights can same wave generated for each photograph.兲 The complexity of the
increase significantly, leading to wave breaking near the shoreline problem and its theoretical intractability are obvious in Fig. 2.
共Yeh 1991; Shuto 1997兲. Shuto 共1997兲, from an analysis of obser- Thus, most of the investigations for breaking wave runup are
vations from the 1933 Showa Great Tsunami, indicated that for experimental and numerical; the latter are based on various ap-
waves with heights greater than about 4 m plunging breaking proximations.
waves were observed in the nearshore region for over 25% of the Solitary waves or combinations of negative and positive
cases analyzed. The broken waves can run up at the shoreline, solitary-like waves 共‘‘N’’ waves兲 are often used to simulate the
inundating coastal regions and causing large property damage and runup and shoreward inundation of tsunamis 共Synolakis 1986,
loss of life. Therefore, it is important when determining seismic 1987; Zelt 1991; Zelt and Raichlen 1991; Tadepalli and Synolakis
sea wave mitigation efforts to understand and to be able to predict 1996; Li 2000; Li and Raichlen 2002兲. It is interesting to note that
the runup associated with breaking waves. Due to the mathemati- the 1983 Nihankai-Chibu earthquake tsunami on the north Akita
cal difficulties caused by the complexities of the fluid motion coast of Japan produced graphic evidence supporting the use of a
associated with wave breaking, a fully theoretical approach is not solitary wave representation of a tsunami at a number of coastal
possible. locations as well as evidence of nearshore wave breaking 共Shuto
An example of a storm wave breaking at a beach in Hawaii is 1985兲. These observations are further supported by remarkable
presented in Fig. 1. This photograph shows the evolution of the photographs of the tsunami at several locations showing both
breaking wave/runup process starting 共from the left兲 with an in- nearshore plunging breaking waves and shoreward propagating
cipient breaking wave and proceeding to the right with the plung- bores 共Saito et al. 1984兲. Such waves can model many of the
ing jet from the breaking wave leading to the jet impingement on characteristics of tsunamis, and, using a simple plane beach, im-
the front face of the wave with the resultant large splashup and
portant characteristics of the runup process can be studied in the
finally the runup on the beach.
laboratory. Indeed, the results for the simple two-dimensional
For solitary waves the process is quite similar; however, the
case of a solitary wave propagating in a constant depth and im-
interaction of the rundown with the runup process, which would
pinging on a plane sloping beach can yield results that are useful
be present in the waves of Fig. 1, is eliminated. Fig. 2 shows the
breaking and the resulting ‘‘splashup’’ process of a solitary wave for three-dimensional numerical modeling of coastal sites.
The maximum runup, defined as the vertical distance above
1 the still water level that the wave reaches on a slope is an impor-
Principle Software Engineer, Oracle Corporation, M/S 3op12,
500 Oracle Pkwy., Redwood Shores, CA 94065. E-mail:
tant parameter needed to define tsunami mitigation procedures.
[email protected] Early experiments of maximum runup of solitary waves, non-
2
Professor Emeritus, Civil Engineering and Mechanical Engineering, breaking and breaking, reported by Hall and Watts 共1953兲 have
California Institute of Technology, Pasadena, CA 91125. E-mail: been used in the past to verify analytical results and the accuracy
[email protected] of numerical models. Hall and Watts 共1953兲 measured the maxi-
Note. Discussion open until August 1, 2003. Separate discussions mum runup of solitary waves on five different slopes. The waves
must be submitted for individual papers. To extend the closing date by were generated by what is now considered to be an overly sim-
one month, a written request must be filed with the ASCE Managing
plified method—i.e., by pushing a vertical plate without using a
Editor. The manuscript for this paper was submitted for review and pos-
sible publication on March 15, 2001; approved on July 10, 2002. This programmed plate trajectory. The shape of the incident solitary
paper is part of the Journal of Waterway, Port, Coastal, and Ocean wave was not described in those experiments. However, it is be-
Engineering, Vol. 129, No. 2, March 1, 2003. ©ASCE, ISSN 0733- lieved that the leading face of the wave up to the crest has a shape
950X/2003/2-47–59/$18.00. that is reasonably similar to a solitary wave.

JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003 / 47
 Fig. 1. Plunging breaking wave with accompanying splashup 共Kaanapali Beach, Maui, Hawaii兲

Battjes 共1974兲 used dimensional analysis and physical reason- 1:19.85 slope that were originally propagating in a constant depth
ing to analyze the characteristics of the runup of periodic waves h o , and the following expressions for the runup were obtained:
on plane slopes. He showed that the breaking criterion, breaker
Nonbreaking: R/h o ⫽2.831冑cot ␤ 共 H/h o 兲 5/4 (5)
type, breaker height-to-depth ratio, and the maximum runup were
approximately governed by one parameter referred to as the ‘‘surf Breaking: R/h o ⫽1.109共 H/h o 兲 0.582 (6)
similarity’’ parameter, ␨, first used by Iribarren and Nogales
共1949兲 The nonbreaking formula was obtained from his approximate
nonlinear theory and confirmed by experimental data 关see also Li
␨⫽tan ␤/ 共 H/L o 兲 1/2 (1) and Raichlen 共2001兲兴. Eq. 共6兲 for breaking waves was obtained
where L o ⫽deep water wavelength of the incident periodic wave; empirically for a beach slope of 1:19.85. Synolakis 共1986兲 also
and H⫽wave height at the toe of the slope inclined at angle ␤ to measured the time history and the spatial wave shape for breaking
the horizontal. Battjes 共1974兲 has proposed that the runup of pe- solitary waves during the runup process.
riodic breaking waves when normalized by the incident wave In the study presented here, the runup process of breaking
height can be expressed as solitary waves was investigated from the viewpoint of the conser-
vation of energy. The runup was inferred from the balance be-
R/H⫽␨ (2) tween the incident wave energy and the energy dissipated and
Kobayashi and Karjadi 共1994兲 extended the surf similarity pa- reflected during the process and the potential energy of the runup
rameter proposed by Battjes 共1974兲 for solitary wave runup and tongue at the time of maximum runup. Thus, the model developed
fitted the breaking data of Synolakis 共1986兲 and the numerical is referred to as an ‘‘energy balance model’’ 共EBM兲. The runup
data from their own model 共which includes an ad hoc dissipation predicted by this model will be compared to experiments con-
term兲, and proposed an empirical expression for the maximum ducted during this investigation as well as the results of others.
runup normalized by the incident wave height as This method provides a means of predicting breaking solitary
wave runup on plane hydrodynamically smooth slopes for a wide
R/H⫽2.955␨ 0.395 (3) range of conditions; i.e., slopes and incident relative wave
where the wavelength of the solitary wave L o in Eq. 共1兲 was heights.
defined by them as
L o ⫽gT 2 /2␲ (4) Experimental Equipment
where g⫽acceleration due to gravity; and T⫽representative time A schematic of the solitary wave runup experiments is presented
period of a solitary wave 共Goring 1978兲. 关The description of the in Fig. 3 along with the notation used in this study. Waves were
selection of T is given in detail by Kobayashi and Karjadi generated using a programmable vertical bulkhead wave genera-
共1994兲.兴 tor. The nonlinear method to define the wave-plate trajectory de-
The experiments of Synolakis 共1986兲 and Zelt 共1991兲 with veloped by Goring 共1978兲 was used to generate first-order solitary
nonbreaking and breaking solitary waves are of importance in waves. 关Interested readers can find a detailed discussion of the
confirming analytical and numerical models of the runup process wave generating system and the laboratory setup used in Li
due to the precision of their experimental techniques compared to 共2000兲.兴
those of other earlier studies. Synolakis 共1986兲 measured the The wave tank used for these experiments 共36.6 m long, 0.4 m
maximum runup of nonbreaking and breaking solitary waves on a wide, and 0.61 m deep兲 and denoted herein as CIT 共California

48 / JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003
 Fig. 2. Solitary wave splashup on 1:15 slope for incident wave; height H/h o ⫽0.40

Institute of Technology兲 had a 13.7 m long hydrodynamically tory as used in the CIT tank. In addition to these data, data were
smooth slope 共beach兲 installed at one end of the tank. The slope used from a study by Synolakis 共1986兲 using the CIT tank and the
was composed of five sections of precision aluminum plates, and same wave generation procedure as in this study, but with a dif-
the angle of the beach to the horizontal was adjustable. The de- ferent slope also constructed of smooth aluminum plates. A third
viation of the slope from a plane surface was less than ⫾1 mm. set of experimental data were used to compare to the EBM. These
As shown in Fig. 3, the origin of the coordinate system was were the results of the Hall and Watts 共1953兲 study conducted in
chosen at the initial shoreline position, with the positive x-axis a concrete tank 25.9 m long, 4.27 m wide, and 1.22 m deep, with
directed offshore toward the wave generator and the vertical a simple impulsive movement of a vertical plate generating the
y-axis directed upward, measured from the still water level. The wave.
water level in the wave tank was measured by a point gauge
located at the toe of the slope with an accuracy of ⫾0.1 mm. Water depths for data used in this study from the experiments
Experiments were conducted with this arrangement for a slope at CIT and CERC for a beach slope of cot ␤⫽15 ranged from a
of one vertical to 15 horizontal (cot ␤⫽15). In addition to experi- minimum of 24.98 to a maximum of 76.20 cm. The data used
ments conducted in this apparatus, experimental results were ob- from Synolakis 共1986兲 for cot ␤⫽19.85 corresponded to water
tained in a wave tank located at the Coastal Engineering Research depths greater than 19.5 cm. The data from Hall and Watts 共1953兲
Center 共CERC兲 of the U.S. Army Corps of Engineers Waterways used herein for a beach slope of cot ␤⫽5.67 were for depths
Experiment Station, Vicksburg, Miss. The CERC experiments greater than about 15 cm.
were conducted in a wave tank 45.7 m long, 0.9 m wide, and 0.9 In the CIT experiments for cot ␤⫽15, the relative incident
m deep with a 1:15 slope at one end using the same plate trajec- wave height, H/h o , was determined using a resistance wave

JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003 / 49
 Fig. 3. Definition sketch for breaking solitary wave runup

gauge located at half of a characteristic length, L/2, of the inci- within less than ⫾1%. 共To obtain the same time origin for each
dent solitary wave from the toe of the slope, where L/h o experiment, a second wave gauge located at a fixed position on
⫽8.52冑H/h o and h o is the constant water depth. The wave gauge the slope was used.兲
was calibrated before and after each experiment. The wave shape In addition to the use of a high-speed video camera, the maxi-
near the breaking point and the maximum runup position were mum runup was determined by visual observations and markers
determined using a high-speed video camera 共Redlake Camera, that indicated the tip of the runup tongue. For the high-speed
Model HR5OO兲 operating at 125 frames/s and a shutter speed of videos, the camera was mounted above the slope with the axis of
1/2,000 s with a resolution of 480⫻420 pixels. Due to limitations the lens oriented perpendicular to the slope, with attention given
of data storage and the maximum area of camera coverage with- to the intersection between the tip of the runup tongue and the dry
out losing spatial accuracy, each experiment was repeated twice, slope. A scale attached to the slope permitted runup measure-
with the camera moved after the first experiment. In this way, the ments. In the case of the simple observations at the time of maxi-
complete wave profile during the runup process was obtained. mum runup, the elevation of the markers was determined using a
This technique was considered satisfactory, since it has been point gauge. A comparison of these methods is presented in Fig.
found from past experiments that the wave generator can repro- 4. It can be seen that the data using the visual observation tech-
duce incident wave amplitudes from one experiment to the next to niques agree well with those from the high-speed videos. As a

Fig. 4. Comparison of maximum runup obtained from high-speed video and visual observations for slope of 1:15

50 / JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003
 Fig. 5. Typical plunging breaking solitary wave showing initiation of splashup

consequence, most of the maximum runup measurements for that this splashup is caused by the translation of the impact point
cot ␤⫽15 were obtained using the simpler visual method. of the jet up the slope and its interaction with the front face of the
The experiments of Synolakis 共1986兲 for cot ␤⫽19.85 used a wave. Shortly after the incident jet impacts the free surface the
resistance wave gauge to define the incident wave and an array of originally smooth wave surface around the impact point becomes
capacitance wave gauges to obtain the profile of the runup tongue rough, as seen in Fig. 2共d兲, suggesting that a short wave is gen-
and the maximum runup. In the earlier experiments by Hall and erated propagating in a direction opposite that of the incident
Watts 共1953兲, both the wave height and the maximum runup el- wave. The ‘‘rough’’ region caused by this wave grows as the
evation were determined visually. splashup process continues. As the incident wave moves shore-
ward, the shape of the splashup 共reflected兲 jet changes and curves
back toward the incident wave. Finally, the incident jet disappears
Results and Discussion of Results and the reflected jet collapses 关Figs. 2共i and j兲 and 6共b兲兴.
The splashup process described perhaps may be used to ex-
Breaking Wave Characteristics and Splashup plain the origin of the counterrotating vortices proposed by Skjel-
breia 共1987兲 in the postbreaking region. This is illustrated in Fig.
Several types of breaking waves are commonly described in the
7, showing schematically the incident jet, the reflected jet, the
literature—spilling, surging, collapsing, and plunging breaking.
reverse flow under the impingement point, and the water motion
For the slopes, waves, and depths used in this study, most of the
as if water were in the aerated regions beneath the jets 共the direc-
waves broke as plunging breakers. A photograph of a breaking
tion of the flow is represented by the arrows兲. The following three
solitary wave for an incident relative wave height of H/h o
possible vortices are illustrated: 共1兲 a clockwise vortex formed by
⫽0.30 is shown in Fig. 5 just after the plunging jet reaches the
the incident jet and the reverse flow under the jet; 共2兲 the clock-
free surface. This picture clearly shows the disturbance that is
initiated at the free surface by the breaking process. Two different wise vortex formed by the reflected jet as it strikes the bed of the
processes appear to occur in the postbreaking region at the point slope; and 共3兲 the vortex that would be formed between them if
of impact of the plunging jet, which might affect runup. water were in the space. This latter vortex rotates counterclock-
In the first case, if the incident solitary wave height is small wise, establishing the pattern of counterrotating vortices proposed
relative to the offshore depth, the wave breaks nearer the original by Skjelbreia 共1987兲.
shoreline than do larger waves. Thus, the impact point of the jet The reason for the difference between wave breaking without
generated by smaller plunging breaking waves may be on the dry and with splashup 关Fig. 6共a兲 compared to Fig. 6共b兲兴 is puzzling,
slope rather than on the water surface. It is interesting that for this but perhaps it can be explained by considering two simple cases
case, the incident jet simply collapses after breaking, producing of a moving water jet impinging on 共1兲 a dry inclined plate; or 共2兲
the runup tongue that then propagates upslope. A photograph of the surface of a quiescent pool with a small depth. Because the
this type of plunging breaking wave-slope interaction is shown in scale of the jet and its Reynolds number are relatively large, it is
Fig. 6共a兲. reasonable to neglect viscous effects and the effects of surface
An example of the second postbreaking process is shown in tension relative to the kinematics and dynamics of the problem.
Figs. 2 and 6共b兲, where the plunging jet impacted the forward face Considered first is the case of a moving jet impacting a dry
of the wave with a shoreward directed jet generated at the impact sloping surface. If one assumes zero vorticity associated with the
point. This jet impact initiated the splashup/runup process. In this jet, potential theory can be used to describe the jet-plate interac-
case, the jet is reflected at an angle with the bottom that is greater tion 共Milne-Thomson 1968兲. The impingement point, considered
than the corresponding angle of the incident jet. It is postulated as the center of the jet, is a stagnation point that separates the flow

JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003 / 51
 Fig. 6. Two cases of plunging breaking solitary waves on 1:15 slope: 共a兲 without splashup, H/h o ⫽0.10; 共b兲 with splashup, H/h o ⫽0.40

running up the slope from that running down the slope. Therefore, its jet advance shoreward with a relatively large velocity 共ap-
the runup tongue is simply composed of the water associated with proximately the wave celerity in the constant depth region兲. To an
the portion of the jet running up the slope. observer moving with the jet, the downslope force that acts on the
When the plunging jet impacts in a small depth, the jet inter- jet as it impacts the quiescent region is relatively large and the
acts with the downstream still water surface before reaching the water associated with it can be deflected upward violently. It is
bottom. This process is complex, and only a tentative explanation postulated that this action results in the observed angle of the
is proposed here. When the translating jet impacts the water sur- reflected jet relative to the bottom being much larger than the
face, it will push a ‘‘wedge shape’’ portion of water that was corresponding angle of the incident jet. The transition between the
originally still 共zero water particle velocity兲 forward up the slope two types of solitary wave breaking is shown in Table 1, where
to form the runup tongue. At the same time, the momentum ex- the corresponding ratio of wave height to depth in the constant
change between the jet and the water upslope redirects the jet depth portion of the tank is presented for the three beach slopes
upward and away from the slope; i.e., it prevents it from running that were used in this study.
along the slope as in the dry slope case. The breaking wave and

Table 1. Breaking Conditions and Transition from Breaking without

Splashup to Breaking with Splashup
Slope (H/h o ) breakinga (H/h o ) transitionb
1:5.67 0.119 0.27
1:15 0.04 0.14
1:19.85 0.03 0.17
a
Fig. 7. Cartoon showing proposed vortices generated by plunging Initiation of breaking.
b
Transition between breaking without splashup and breaking with
breaking wave
splashup.

52 / JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003
Energy Conservation Considerations
In the following discussion, the incident wave energy is specified
and a general energy balance equation relating to the runup pro-
cess is presented. It should be noted that both postbreaking pro-
cesses described earlier can be treated by this method. This equa-
tion assumes that the potential energy at the time of maximum
runup is a function of the maximum runup and the volume of the
incident solitary wave. The empirical expression that is presented
for energy dissipation during wave breaking is based on the bore
representation of postbreaking conditions mentioned earlier and
proposed by Li 共2000兲; see also Li and Raichlen 共2002兲. It will be
used in this model to account for the energy loss. Using the pro-
posed energy dissipation model and energy conservation consid-
erations, the prediction of the maximum runup is presented and
compared to experiments.

Energy Balance Equation

Once the solitary wave is produced, either by the wave generator
in experimental conditions or by a tectonic event or landslide in Fig. 8. Energy balance model
nature, the wave with height H propagates toward the shoreline
and runs up to a maximum elevation R. After reaching the posi-
tion of maximum runup, the runup tongue retreats and produces
the major reflected wave propagating offshore. 共There is also a
minor reflected wave that is produced during the runup process c⫽ 冑g 共 h o ⫹H 兲 (9)
due to wave-slope interaction; this will be discussed shortly.兲 One may refer to Fig. 3 for the definition of these parameters.
Consider the control volume, ⌫, shown by the dotted line in The total volume per unit width (V), the kinetic energy per
Fig. 3. The control volume is selected so that all of the wave unit width (E K ), and the potential energy per unit width (E P ) of
motion is included in the domain, ⌫, at time t o before the wave a solitary wave with wave height H are obtained by integrating
reaches the slope, and at time t 1 , which is the time of maximum these and related expressions from ⫺⬁⭐x⭐⬁

冕
runup. The wave motion and velocity at the seaward boundary of
⬁
⌫ are so small that both the mass and the momentum influx at that V⫽ ␩dx⫽ 冑共 16/3兲 h 3o H (10)
boundary are neglected for t o ⭐t⭐t 1 . The left boundary of ⌫ ⫺⬁

冕
includes the initial shoreline and the maximum runup tongue on
⬁
the slope. Therefore, the runup and rundown processes are con- E K⫽ ␳ 共 ␩⫹h o 兲共 u 2 /2兲 dx⫽ 共 4␳g/3) 兲 H 3/2h 3/2
o
fined to the chosen control volume. This guarantees that there is ⫺⬁
no mass and momentum flux out of the control volume ⌫ for time
t o ⭐t⭐t 1 . ⫹ 共 4␳g/15) 兲 H 5/2h 1/2
o (11)

冕
The energy balance model is developed by considering the ⬁
conservation of energy inside the control volume, ⌫. Since there E p⫽ 共 ␳g/2兲 ␩ 2 dx⫽ 共 4␳g/3) 兲 H 3/2h 3/2 (12)
o
is no mass or momentum flux at either the left or the right bound- ⫺⬁
ary during the interval t o ⭐t⭐t 1 , the work done by the wave at where ␳⫽fluid density. The potential energy 关Eq. 共12兲兴 is the
the left and right boundaries of the control volume is zero. The potential energy in the fluid column due only to the solitary wave.
energy transformation for t o ⭐t⭐t 1 is investigated using the It can be shown that the kinetic energy is at most about 8%
EBM. greater than the potential energy for relative wave heights less
The energy terms involved in the runup process at the time of than about 0.5. Therefore, to simplify the discussion, the follow-
maximum runup, t⫽t 1 , are shown diagrammatically in Fig. 8. ing equation for the kinetic and potential energies is used:
This figure shows that the sum of the potential energy at the time
of maximum runup, E P , plus the energy dissipated, E D , and the E K ⬇E P ⫽ 共 4␳g/3) 兲 H 3/2h 3/2
o (13)
energy reflected from the slope, E R , must equal the incident wave
As the solitary wave propagates toward the plane beach and
energy, E I . 共At the time of maximum runup the kinetic energy in
runs up the slope, a portion of the kinetic energy must be trans-
the runup tongue must essentially be zero.兲 The relative magni-
formed into potential energy; i.e., the kinetic energy decreases
tude of each energy term is represented approximately by the
and the potential energy slowly increases. At the position of maxi-
width of each arrow; these terms will be discussed individually
mum runup, the potential energy reaches a maximum, while the
presently.
kinetic energy is a minimum and has a value close to zero 关this
Initially, all of the energy 共the kinetic energy and potential
process has been described by Li 共2000兲兴. After the maximum
energy兲, E I , inside the control volume ⌫ is contained in the inci-
runup is reached the rundown process begins, accompanied by a
dent solitary wave. To define E I , the following form of the soli-
decrease in potential energy and a corresponding increase in ki-
tary wave, the water particle velocity, and the wave celerity up to
netic energy.
the first-order accuracy are taken as:
During this process, a portion of the energy will be reflected
␩ 共 x,t 兲 ⫽H sech2 关 冑3H/4h 30 共 x⫺ct 兲兴 (7) from the slope and a portion of the energy is dissipated. The
energy dissipation, E D , is considered first. It is caused by several
u 共 x,t 兲 ⫽c␩/ 共 ␩⫹h o 兲 (8) mechanisms—friction at the free surface, i.e., between the air and

JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003 / 53
Fig. 9. Experimental water surface–time history at distance L/2 offshore of toe of slope; H/h o ⫽0.28 for slope of 1:15. The solid line is the water
surface–time history after eliminating the runup tongue; the dashed line is the water surface–time history without eliminating the runup tongue.

the free surface; friction at the bottom of the wave tank; and, most The energy balance for the runup process at an arbitrary loca-
important, wave breaking. This is expressed as tion on the slope shoreward of breaking can be expressed as
E D ⫽E FF ⫹E FB ⫹E B (14) E I ⫽E K ⫹E P ⫹E B ⫹E R (15)
where E D ⫽total dissipated energy; and E FF ⫽energy dissipated
where E K and E P ⫽kinetic and potential energies, respectively,
by friction at the air-water interface. The energy dissipated by
present in domain ⌫ during the runup process; and E I ⫽total in-
bottom friction is denoted as E FB , and E B is the energy dissipa-
cident wave energy as given by
tion associated with wave breaking, see also Fig. 8. For this prob-
lem, the energy dissipated at the free surface is assumed small E I ⫽E K ⫹E P ⬇ 共 8/3) 兲 ␳g 共 Hh o 兲 3/2 (16)
compared to the incident wave energy; thus, E FF is neglected.
It can be shown that the energy dissipation rate due to bottom The energy dissipation due to wave breaking⫽E B , and E R
friction in a constant depth tank is proportional to the attenuation ⫽energy associated with reflection.
in wave height dH/dx. Keulegan 共1948兲 analyzed the rate of The reflected wave associated with runup mentioned in the
energy loss associated with the laminar boundary layer beneath a literature usually refers to the reflected wave due to both reflec-
solitary wave propagating over a smooth surface. Naheer 共1978兲 tion from the slope before the wave reaches its maximum runup
also investigated the energy dissipation and viscous damping of and the wave that propagates offshore generated by the rundown
solitary waves propagating in a constant depth over a smooth and process. In this study, the reflected wave will refer only to the
a rough bottom, and the viscous dissipation and attenuation were former; i.e., for the time t o ⭐t⭐t 1 .
found to be small compared to the energy associated with the Water surface time histories are presented in Fig. 9 at a loca-
incident wave. tion that corresponds to one-half of the characteristic length of a
Therefore, for breaking solitary wave runup on a relatively solitary wave offshore of the toe of the slope (L/2h o
steep slope, the major energy dissipation is due to wave breaking. ⫽4.26冑H/h o ) for two cases—共1兲 where the runup tongue is ex-
This is especially true in experiments where the wave tank bottom perimentally eliminated; and 共2兲 where it is not. The portion of
and the slope surface are smooth. 共For rough slopes and other the water surface time history between Line A and Line B in Fig.
conditions, e.g., small slopes leading to relatively large upslope 9 shown as a solid line includes the reflected wave generated
propagation distances, bottom dissipation probably cannot be ne- before the time of maximum runup, t 1 , is reached. The wave that
glected, and corresponding empirical formulas must be used in is the main reflected wave and is generated by the rundown pro-
the energy balance model to represent this energy loss.兲 For sim- cess when the runup tongue is not eliminated is shown in Fig. 9 as
plicity in the following discussion, the dissipation caused by bot- the dashed line extending from Line B to the end of the recording;
tom friction will be neglected. Thus, in Eq. 共14兲 the energy dis- i.e., the dashed line for t * ⬎80, where t * ⫽t 冑g/h o .
sipation is considered solely due to wave breaking; i.e., E D A series of experiments were conducted to define E R ; i.e., the
⬇E B . energy between A and B in Fig. 9 共i.e., the solid line in the figure兲.

54 / JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003
 Fig. 10. Experimental arrangement used to measure wave reflection

Fig. 10 shows the experimental arrangement used to measure the would normally comprise the runup tongue overtops the slope and
wave energy reflected from the slope during the interval t o ⭐t is captured in the collection box. In this way the wave generated
⭐t 1 . Instead of the relatively long slope used for experiments to by the rundown process is eliminated, and only the reflected wave
determine wave runup, the slope used in the reflection measure- due to the wave-slope interaction before the runup process is
ment only extends from the wave tank bottom to the initial shore- measured.
line position 共the still water surface兲. A collection box with its Fig. 11 shows the maximum height of this ‘‘slope’’ reflected
initial water level lower than that in the wave tank is located wave measured at the toe of the slope as a function of incident
shoreward of the end of the slope. Therefore, when the incident wave height. It is seen that even for H/h o ⫽0.45 the relative
solitary wave runs up the slope, the portion of the wave that height of the maximum reflected wave, H r /h o , before the run-

Fig. 11. Variation of maximum reflected wave height as function of incident wave height after eliminating runup tongue for slope of 1:15

JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003 / 55
 Fig. 12. Variation of energy dissipated during breaking with beach slope and relative wave height

down process begins is less than 0.03. For such small waves, the weighted essentially nonoscillatory 共WENO兲 shock capturing
water surface time history of the reflected wave can be converted scheme employed in gas dynamics. Wave breaking and post-
into a spatial representation using linear theory. The reflected breaking propagation are handled automatically by this approach
wave energy, E R , including both potential and kinetic energy can and ad hoc terms are not required. A computational domain map-
be computed by integrating spatially over the wave to obtain the ping technique was used to model the shoreline motion.
potential energy. Assuming linear theory for the relatively small To define the energy dissipation associated with wave break-
reflected wave, the kinetic energy is assumed to be equal to the ing, the total energy in the domain ⌫ was determined as a function
potential energy. From these experiments and analysis, the energy of time for a range of relative wave heights, H/h o , and slopes,
in the reflected wave was found to be negligible compared to the
cot ␤, employing the numerical method developed by Li 共2000兲 to
incident wave energy.
determine the potential and kinetic energies. Considering the as-
The kinetic energy, E K , associated with the runup tongue on
sumptions mentioned earlier, the difference between the total en-
the slope at the time of maximum runup is also very small com-
pared to the incident wave energy 共Li 2000兲. Grilli and Svendsen ergy in the domain at t⫽t o and at t⫽t 1 gives the energy dissi-
共1989兲 calculated the nonbreaking runup on a slope using the pated. This is expressed as E B /E I and is shown as the ordinate in
boundary element method, and found that the kinetic energy in Fig. 12 plotted as a function of H/h o for given beach slopes. The
the runup tongue at the time of the maximum runup is almost data obtained by Li 共2000兲 from the numerical method briefly
equal to zero. Therefore, the kinetic energy term in the energy described earlier are indicated by the symbols. An empirical equa-
conservation equation, Eq. 共15兲, at the time of maximum runup, is tion was fitted to these data, and is described in the following:
neglected; i.e., E K ⬇0. In the following discussion, the remaining
terms in Eq. 共15兲, E B and E P , are treated. E B /E I ⫽C 关 A ln共 cot ␤ 兲 ⫹B 兴 (17)

where
Potential Energy
As seen in Eq. 共15兲, the energy dissipated by wave breaking must A⫽⫺0.470共 H/h o 兲 ⫹0.534 (18a)
be estimated to define the potential energy at the time of maxi-
mum runup. Li 共2000兲 developed an estimate of this from a nu- B⫽2.165共 H/h o 兲 ⫺1.154 (18b)
merical study of the energy transformation associated with the
runup process. This is summarized as follows. C⫽0.190关 ln共 H/h o 兲兴 ⫹0.969 (18c)
The general characteristics of the breaking process from the
time of breaking through the duration of the postbreaking propa- The fit of Eq. 共17兲 to the data is reasonable for the range of H/h o
gation are assumed to be reasonably represented as a propagating shown, with differences at most of the order of 10%.
bore. A numerical model using a bore structure to treat this pro- With the energy dissipation defined 关Eq. 共17兲兴, Eq. 共15兲 is used
cess of wave breaking and shoreward propagation was developed to predict the runup, modified as
by Li 共2000兲; it is also presented in Li and Raichlen 共2002兲. The
nonlinear shallow water wave equations were solved using the E P /E I ⫽1⫺E B /E I (19)

56 / JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003
 Fig. 13. Variation of shape factor with relative wave height and beach slope from numerical model of Li 共2000兲

Thus, the task is to define the relative runup, R/h o , from the triangular in cross section, ␣ would be equal to 0.33. From Eq.
normalized potential energy, E P /E I , as determined from Eq. 共21兲, it is seen that ␣ may be a function of both the relative
共19兲. incident wave height and the angle of the plane beach to the
Dimensional analysis is used as a guide to define the maxi- horizontal. To determine the dependence of the shape factor on
mum runup for a breaking solitary wave from the potential energy H/h o and ␤, the variation of ␣ with these parameters was inves-
at the time of maximum runup. The dependent variables for the tigated using the numerical model developed by Li 共2000兲 for
runup process are grouped into three categories, as follows: 共1兲 three slopes (cot ␤⫽15, 19.85, and 50兲 and for 0.025⬍H/h o
the incident wave parameters—the wave height H, the water ⬍0.5. The results are presented in Fig. 13 for the time of maxi-
depth in the constant depth region h o , and the total volume of the mum runup. It is seen that for H/h o ⬎0.05 共all breaking solitary
wave per unit width V; 共2兲 the angle of the slope, ␤, and the waves for these slopes兲 the shape factor is essentially independent
maximum runup of the solitary wave on the slope, R; and 共3兲 of both the relative incident wave height and the beach slope, and
physical constants—the acceleration due to gravity g and the den- it is approximately equal to 0.12. We will return to the consider-
sity of the water, ␳. In functional form, this relation is ation of this parameter in the next section.
f 共 E P ,H,h o ,V,cot ␤,R,␳,g 兲 ⫽0 (20)
Prediction of Maximum Runup from Energy Considerations
Choosing the variables h o , ␳, and g as the independent physical The variation of the maximum relative runup, R/h o , with the
variables, we obtain the following nondimensional form of Eq. relative incident wave height and beach slope as obtained for
共20兲: three slopes (cot ␤⫽5.67, 15, and 19.85兲 is presented in Fig. 14
E P /␳gh 3o ⫽F 共 R/h o ,V/h 2o ,H/h o ,cot ␤ 兲 (21) using experimental data from Hall and Watts 共1953兲, Li 共2000兲,
and Synolakis 共1986兲, respectively. These data show, as expected,
The relation among the nondimensional variables in Eq. 共21兲 can that for breaking waves the maximum runup decreases as the
only be obtained from numerical simulations or experiments. The beach slope decreases for a given incident wave height, and for a
following expression is presented to relate the potential energy at given slope, and for a given slope generally there is an increase in
the time of maximum runup to the characteristics of the runup relative runup with increasing incident wave height. This varia-
tongue from the nondimensional variables in Eq. 共21兲: tion can be predicted from energy conservation considerations
E P /␳gh 3o ⫽␣ 共 R/h o 兲共 V/h 2o 兲 (22) from Eqs. 共19兲, 共17兲, 共18兲, 共16兲, and 共22兲 as the following:
R/h o ⫽ 兵 关 1⫺ 共 E B /E I 兲兴 /1.5␣ 其 共 H/h o 兲 (23)
The choice of the arrangement of terms in Eq. 共22兲 is essentially
the definition of potential energy—the product of the weight of with all quantities known a priori except for the shape factor, ␣. It
the displaced volume (␳gV), where at the time of maximum was found empirically that the best fit of Eq. 共23兲 to the experi-
runup the volume in the runup tongue must be equal to the vol- mental data was achieved for ␣⫽0.18. This is larger than the
ume in the incident wave, and the distance from the still water shape factor of 0.12 obtained numerically. This is probably ex-
level to the center of mass (␣R). In effect, the quantity ␣ in Eq. plained by the fact that although the numerical method described
共22兲 describes the shape of the runup tongue and is referred to by Li 共2000兲 predicts gross characteristics such as the maximum
herein as the shape factor. For example, if the runup tongue were runup, due to the ‘‘thinness’’ of the tip of the runup tongue one

JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003 / 57
 Fig. 14. Variation of maximum runup with relative wave height and slope; EBM compared to experiments

would not expect the shape of the runup tongue obtained experi- Acknowledgments
mentally to be accurately predicted by that theory. It should be
noted that Eqs. 共3兲 and 共6兲 agree well with the data for the slope This investigation was supported by the National Science Foun-
with cot ␤⫽19.85, as would be expected, since the expressions dation 共NSF兲 through Grant CMS-9523414. The writers appreci-
were obtained empirically from the data of Synolakis 共1986兲. ate the interest in and attention given to this study by Dr. Clifford
However, Eq. 共3兲 from Kobayashi and Karjadi 共1994兲 does not Astill of the NSF, and the comments made by one of the anony-
agree with the data for the other two slopes (cot ␤⫽5.67 and 15兲 mous reviewers.
or the expressions determined from the EBM. This might be ex-
pected because of the strong correlation of breaking wave dissi-
pation with beach slope not included in the development of Koba- References
yashi and Karjadi 共1994兲.
Battjes, J. A. 共1974兲. ‘‘Surf similarity.’’ Proc., 14th Conf. on Coastal
Engineering, Vol. 1, ASCE, New York, 466 – 480.
Conclusions Goring, D. G. 共1978兲. ‘‘Tsunami: The propagation of long waves onto a
shelf.’’ Rep. No. KH-R-38, W. M. Keck Laboratory of Hydraulics and
Water Resources, California Institute of Technology, Pasadena, Calif.
The following major conclusions were drawn, based on the ex- Grilli, S. T., and Svendsen, I. A. 共1989兲. ‘‘Computation of non-linear
perimental results: wave kinematics during propagation and runup on a shelf.’’ Proc.,
1. There is a dramatic difference between the splashup of the NATO Advanced Research Workshop on Water Wave Kinematics,
breaking wave jet, depending on whether the jet impinges on North Atlantic Treaty Organization, Brussels, Belgium, 387– 413.
the still water surface, the forward face of the wave, or on a Hall, J. V., and Watts, J. W. 共1953兲. ‘‘Laboratory investigation of the
solid surface. However, the runup does not appear to be af- vertical rise of solitary waves on impermeable slopes.’’ Technical
fected by this, and it changes continuously with relative Memorandum No. 33, Beach Erosion Board, U.S. Waterways Experi-
wave height. ment Station, U.S. Army Corps of Engineers, Vicksburg, Miss.
Iribarren, C. R., and Nogales, C. 共1949兲. ‘‘Protection des ports.’’ XVIIth
2. Reflected wave energy associated with the wave-slope inter-
Int. Navigation Congress, Section II, Communication, 31– 80.
action, i.e., not the wave generated by rundown, appears to
Keulegan, G. H. 共1948兲. ‘‘Gradual damping of solitary waves.’’ J. Res.
be negligible compared to the energy of the incident wave. Natl. Bur. Stand., 40, 487–501.
3. Neglecting the bottom dissipation associated with wave Kobayashi, N., and Karjadi, E. A. 共1994兲. ‘‘Surf-similarity parameter for
propagation on a slope, the runup appears to be reasonably breaking solitary-wave runup.’’ J. Waterw., Port, Coastal, Ocean
well predicted by a simple model based on energy conserva- Eng., 120共6兲, 645– 650.
tion principles for the range of slopes investigated and for Li, Y. 共2000兲. ‘‘Tsunamis: Non-breaking and breaking solitary wave
0⭐H/h o ⭐0.45, i.e., the limits of the experimental data. runup.’’ Rep. No. KH-R-60, W. M. Keck Laboratory of Hydraulics and

58 / JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003
 Water Resources, California Institute of Technology, Pasadena, Calif. Skjelbreia, J. E. 共1987兲. ‘‘Observation of breaking waves on sloping bot-
Li, Y., and Raichlen, F. 共2001兲. ‘‘Solitary wave runup on plane slopes.’’ J. toms by use of laser Doppler velocimetry.’’ Rep. No. KH-R-48, W. M.
Waterw., Port. Coastal, Ocean Eng., 127共1兲, 33– 44. Keck Laboratory of Hydraulics and Water Resources, California In-
Li, Y., and Raichlen, F. 共2002兲. ‘‘Non-breaking and breaking solitary stitute of Technology, Pasadena, Calif.
wave runup.’’ J. Fluid Mech., 456, 295–318. Synolakis, C. E. 共1986兲. ‘‘The runup of long waves.’’ PhD thesis, Cali-
Milne-Thomson, L. M. 共1968兲. Theoretical hydrodynamics, Macmillan, fornia Institute of Technology, Pasadena, Calif.
New York. Synolakis, C. E. 共1987兲. ‘‘The runup of solitary waves.’’ J. Fluid Mech.,
Naheer, E. 共1978兲. ‘‘The damping of solitary waves.’’ International Jour- 185, 523–545.
nal of Hydraulic Research, 16, 235–249. Tadepalli, S., and Synolakis, C. E. 共1996兲. ‘‘Model for the leading waves
Saito, A., Hamada, M., Kosuge, S., Emoto, K., Kitahara, M., and Ke-
of tsunamis.’’ Phys. Rev. Lett., 77, 2141–2145.
muyama, M. 共1984兲. ‘‘A report on the 1983 Nihonkai-Chubu earth-
Yeh, H. H. 共1991兲. ‘‘Tsunami bore runup.’’ Tsunami hazards, E. Bernard,
quake.’’ Rep., Tokai Univ., Shizuoka, Japan.
ed., Kluwer Academic, Boston, 209–220.
Shuto, N. 共1985兲. ‘‘The Nihonkai-Chubu earthquake tsunami in the north
Akita coast.’’ Coast. Eng. Japan, 28, 255–264. Zelt, J. A. 共1991兲. ‘‘The runup of non-breaking and breaking solitary
Shuto, N. 共1997兲. ‘‘A natural warning of tsunami arrival.’’ Perspectives waves.’’ Coastal Eng., 15共3兲, 205–246.
on tsunami hazard reduction, G. Hebenstreit, ed., Kluwer, Dordrecht, Zelt, J. A., and Raichlen, F. 共1991兲. ‘‘Overland flow from solitary waves.’’
The Netherlands, 157–173. J. Waterw., Port, Coastal, Ocean Eng., 117共3兲, 247–263.

JOURNAL OF WATERWAY, PORT, COASTAL AND OCEAN ENGINEERING © ASCE / MARCH/APRIL 2003 / 59