Wave Transmission at Submerged Rubblemound Breakwaters

Stuart R. Seabrook1 and Kevin R. Hall2

Abstract

Submerged rubblemound breakwaters are becoming more popular as a potential

alternative to coastal protection measures where a moderate degree of energy transmission
is acceptable. Such situations include areas where vegetative shore protection is existing
or proposed or in the event that an existing shore protection structure has become
damaged or under designed and a method is needed to reduce the incident wave energy.
Although there have been previous investigations on the performance of submerged
rubblemound breakwaters, there are only a few design equations available to the design
engineer. Those available are based on a limited range of input design variables and as a
result are insufficient in some cases.

Physical model studies were performed at the Queen's University Coastal Engineering
Research Laboratory (QUCERL) in Kingston, Canada to assess the performance of
submerged rubblemound breakwaters under a wide range of design conditions in two-
dimensional (2-D) and three-dimensional (3-D) settings. The tests include a number of
wide crested structures to provide data where previous investigations have not. The
results show that the relative submergence, incident wave height and structure crest width
are the most important design variables.

A number of potential design equations were evaluated by statistical analysis methods.

The proposed design equation fits the 2-D test data well and provides moderate agreement
with the 3-D test results. Although physical testing is suggested for all design applications
due to the complexity of site specific considerations, the proposed equation does provide a
good preliminary design tool for submerged rubblemound breakwaters.

1
Department of Civil Engineering, Queen's University, Kingston, Ont.,K7L 3N6

2
Professor, Department of Civil Engineering, Queen's University, Kingston, Ont. K7L 3N6

2000
 COASTAL ENGINEERING 1998 2001

1.0 Introduction

Submerged rubblemound breakwaters are simply rubblemound structures constructed with

a crest elevation below the local water level. Although they are well suited to situations
where minimal visual intrusion is desired and where it is desirable to maintain a moderate
degree of energy transfer between the shoreline and the offshore region for environmental
reasons, their performance is sensitive to water level changes and it is not practical to
expect transmission coefficients as low as those achievable with conventional surface
piercing structures. However, there are benefits associated with the potentially smaller
material requirements for stable submerged structures and the ability to rehabilitate
existing structures by simply reducing the incident wave conditions with a submerged
breakwater.

Numerous physical and numerical investigations have been performed for various
submerged structure configurations and materials. In general, the physical processes at
submerged rubblemound breakwaters can be defined for three regions in the vicinity of the
structure as indicated in Figure 1. Relevant nomenclature is also indicated.

In Region 1, the incident wave shoals on the rising face of the breakwater, considerable
non-linear wave transformations take place as bound waves are developed (Beji and
Battjes, 1993) and some wave breaking is initiated. A portion of the incident energy is
also reflected from the front breakwater face. Wave breaking continues into Region 2
where significant non-linear interactions occur between the various wave phases.
Harmonic generation occurs as energy is transferred from the fundamental wave frequency
to higher harmonic frequencies (Driscoll, Dalrymple and Grilli, 1993). Some wave energy
is also dissipated on the breakwater crest through friction and air entrainment as well as
within the breakwater structure. In Region 3, the free and bound transmitted waves
dissociate as they travel into the deeper water. This generally results in a broadening
energy spectra as the various wave components travel with their own celerity.

Numerical modeling efforts have met with some success in representing the transformation
of weakly non-linear incident waves (Ohyama and Nadaoka, 1993; Driscoll, Dalrymple
and Grilli, 1993; Beji and Battjes, 1994; Losada, Silva and Losada, 1996). Although all
models are reported to reflect some of the physical modeling data well, none of the
approaches can fully model the breaking and non-linear decomposition process on a
theoretical basis. Therefore, given the uncertainties associated with the present state-of-
the-art in numerical modeling, it may be most appropriate for the design engineer to
consider more general design equations based on physical modeling results.
 2002 COASTAL ENGINEERING 1998

4^~ Region 1- -»*- Region 2 Region 3 ~fr$

Figure 1 : Submerged Breakwater Nomenclature

Relatively few design equations have been developed to date for submerged rubblemound
breakwaters. Those available have been developed by Seelig (1980) for surface piercing
and submerged permeable breakwaters, by Ahrens (1987) for reef-type breakwaters and
by Van der Meer (1991) for low crested and submerged rubblemound structures. Seelig's
equation was developed with very little submerged breakwater data and Ahren's equation
is not directly applicable to conventional submerged rubblemound breakwaters given the
reshaping nature of reef breakwaters. Van der Meer's equation was developed from a
considerable volume of test data from a number of authors but some variables were not
varied to a large degree; a limited variation in crest width was perhaps the most important
shortcoming of this research. As a result, none of the existing equations are sufficient for
application over a wide range of design conditions.

The objectives of the research presented in this paper are therefore as follows:
1. To test a sufficiently wide range of submerged breakwater geometries, ensuring a
broad range of crest widths, under a relatively large range of incident wave
conditions in a 2-D setting to assess the validity of the existing design equations.
2. To extend and modify the existing design equations as necessary to provide some
physical basis for the dimensionless parameters utilized.
3. To perform 3-D testing for a number of conditions similar to those tested in the
2-D apparatus to assess the validity of the proposed equations for application in
more realistic 3-D environments.
 COASTAL ENGINEERING 1998 2003

2.0 Experimental Setup

The majority of the physical tests were 2-D in nature and were performed in a 1 meter
wave flume at QUCERL. These test results provided the data for development of the
proposed design equations. Subsequent 3-D testing was carried out in the wave basin at
QUCERL using a smaller set of test variables. The results of these tests were used in the
evaluation of alternative design equations.

2-D Testing

The testing setup for the 2-D tests is shown in Figure 2. The wave flume is 47.0 m long,
1.2 m deep and 1.0 m wide and is equipped with a flapper type wave generator. A
plywood beach was constructed in the flume, upon which the test breakwaters were
constructed. The beach permits testing of the submerged breakwaters in relatively large
incident waves. The submerged breakwater cross section consisted of a core of relatively
course core material (D50c = 0.017 m) and two layers of primary armour (DSOa = 0.059 m).
A second armour size (D50a = 0.037 m) was used in some tests. The armour size was
determined such that the breakwater remained stable during testing. The stone size
required for the most severe testing condition was determined using the stability equation
of Vidal et al. (1993). The stone size recommended for the "total slope" section was used
for the entire breakwater.

Figure 2 : 2-D Testing Configuration

Incident and transmitted waves were measured by two wave probe arrays, located in front
and rear of the breakwater respectively. The probes were capacitive type water level
gauges, sampling the water surface at 20 Hz. Reflection from the rear wall of the wave
flume was minimized using a 1:10 beach of rubberized hair in front of a porous matrix of
concrete blocks.

In total, approximately 800 tests were performed with irregular waves. A number of tests
were also performed with regular waves in order to confirm the presence of physical
phenomenon observed by previous authors. The testing program involved 13 submerged
breakwater geometries tested under 5 different water levels with a number of incident
2004 COASTAL ENGINEERING 1998

wave characteristics.

All irregular wave spectra tested were Jonswap with a = 0.0081 and y = 3.3. The signals
were generated using the National Research Council of Canada's (NRC), GEDAP wave
generation and analysis package. Given the physical characteristics of the flume and the
mechanical response of the paddle to the input signal, the generated wave spectral
characteristics may vary from the target characteristics by 5% to 10%. A summary of the
irregular wave characteristics tested is provided in Table 1.

Table 1: Irregular Wave Characteristics in 2-D Tests

Wave Set W01 W02 W03 W04 W05 W06 W07 W08 W09 W10 Wll

Hmo (m) * 0.05 0.05 0.05 0.10 0.10 0.10 0.15 0.15 0.15 0.20 0.20

Tp(s)* 1.2 1.5 2.0 1.2 1.5 2.0 1.2 1.5 2.0 1.5 2.0
* Note: Wave characteristics are target values - those measured in flume may vary to some degree.

3-D Testing

The wave basin at QUCERL is approximately 30 m by 25 m by 1.2 m deep and is

equipped with a piston-type wave paddle 10.5 m long. All dimensions for the 3-D tests
were scaled at 0.63 times those of the 2-D tests and the submerged breakwater structure
was constructed on the floor of the basin. A concrete beach was constructed behind the
breakwater and covered with sand to minimize the reflection and provide a qualitative
assessment of beach development. A subset of testing parameters was used in the 3-D
tests, with 3 breakwater geometries and 3 water depths, all tested at incident angles of
90 ° and 60°. The general testing configuration is shown in Figure 3 below.

KEY

AP5 Wa.**6txteWaS
Wcws
Prctes

©VI StomegedSreocwote,
KB
Velocity
PKfcel
Series! 00)
AP8

300 SCO 1C0

!.0m

\p] AP3 AP6

.~Cona0f& teach

••'^m*— 2.5m—'

(a) 3-D Testing at 90° (b) 3-D Testing at 60°

Figure 3: 3-D Testing Configurations

 COASTAL ENGINEERING 1998 2005

The submerged breakwaters used in the 3-D tests were constructed using a relatively fine
core material (D50c = 0.004 m) and the smallest armour tested in the flume (Z)5a,=0.037 m).
The wave signals were generated from Jonswap spectra. Because the breakwater for the
3-D tests was not constructed on an elevated platform, it was not possible to generate
waves as large as those used in the 2-D tests. A summary of the irregular wave
characteristics used in the 3-D tests is provided in Table 4.

Table 2 Irregular Wave Characteristics in 3-D Tests

Wave Set W01 W02 W03 W04 W05 W06 W07 W08 W09 W10 Wll

Hmo (m)* 0.032 0.032 0.032 0.032 0.063 0.063 0.063 0.063 0.095 0.095 0.095

Tp(s)* 0.95 1.19 1.59 1.98 0.95 1.19 1.59 1.98 0.95 1.19 1.59
* Note: Wave characteristics are approximate - those measured in basin may vary to some degree.

Water level probes were placed throughout the area behind the submerged breakwater,
and one probe was located near the paddle to provide an indication of the incident wave
height. Two velocity probes were also moved throughout the area behind the submerged
breakwater to provide a qualitative indication of the general velocity patterns. The
locations of the various probes are shown in Figure 3 above.

Data Sampling and Analysis

Sampling for each test was performed at 20 Hz over a period of 100 waves. Full
reflection analysis was undertaken for the 2-D data to separate the incident and reflected
spectra. The reflection analysis was performed using a least squares analysis of 3 probes
of the 5 probe array (Mansard and Funke, 1987). The incident wave characteristics at
probe arrays 1 and 2 were used to define the transmission coefficient K, such that:

where H, and //, are the incident Hm0 values at probe arrays 2 and 1 respectively.

The significant wave height (//,) at individual probes was used to define the transmission
coefficient for the 3-D tests. The incident wave was defined by Probe 5 and the
transmission coefficient was computed at various locations behind the submerged
breakwater. Although the use of Hs instead ofHm0/ to define K, is not consistent with the
analysis of the 2-D data, it reduces the effect of energy that has leaked into the testing area
or been reflected into the lee of the breakwater, on the characteristic wave height estimate.

The transmission coefficient values were analyzed with respect to the various incident
wave and structure characteristics. This analysis involved a simple graphical trend analysis
of the data, followed by a comprehensive statistical analysis of alternative design equations
relating K, to the most important design variables. The trend analysis was conducted with
 2006 COASTAL ENGINEERING 1998

dimensional and dimensionless variables considered to be important to the transmission

process. The most important dimensionless variables were ascertained by dimensional
analysis of the transmission process.

It is generally accepted that the transmitted wave at a submerged breakwater is a function

of a number of variables.
H, = /(P, g, |i, D50a, n, H„ L, d„ 6, B, h„ h) (2)

where p, g and fi are density, gravitational acceleration and dynamic fluid viscosity
respectively and the other variables are defined in Figure 1. A dimensionless form of this
expression can be developed using p, g and Ht as basic or repeating variables such that:

H, PH,jrg D L ds B K
K. = — = <p ( !—L- 5 n , , 6 , .-*-)
' H, ' ~H, H, '' ~H, H

These basic dimensionless variables were used to develop alternative dimensionless

variables which are more relevant to the physical processes affecting transmission at
submerged breakwaters. Dimensional and statistical analysis of these relevant variables
was undertaken to develop a design equation for transmission at submerged rubblemound
breakwaters.

3.0 Results

The observations made during testing were generally consistent with observations noted
from previous investigations. The incident wave spectra were broadened with a shift of
energy to higher frequencies as the waves passed the breakwater and there was evidence
of harmonic generation and subsequent dispersion as individual waves passed over the
structure. The most obvious process affecting K, was wave breaking. Inspection of the
general trends defined by the 2-D test data indicate that K, is most sensitive to the depth c
submergence d„ the incident wave height H, and the crest width B. To a lesser degree, K
is influenced by the period of the incident wave (Tp), the breakwater armour dimensions
{D50a) and the breakwater slopes (6).

Typical trends observed during the 2-D tests are summarized in Figures 4 through 8. It i
evident that the transmission increases with increased ds, increased H, and increased B. I
is also shown that a small increase in K, is observed with increasing Tp, increasing D50a an
steeper slopes (increasing Tan ff).
 COASTAL ENGINEERING 1998 2007

Submerged Breakwater Transmission Submerged Breakwater Transmission

Ktvsds(B=0.6m,Tp~2.0s) Ktvsds(B=0.6m,Hs- 0.10 m)

1
0.9 0.9 • \

0,8 " 0.! . » ._

0.7 . . .. >.... ; . .,..- • M< - 0.05m 0.7
. I,.-'' .. ? Tp -1.11
0.6 0.5
J0.5
:y • . : : . : |
:

0.4
4""^ "'
i Ks-O.fSm 0.4
Tp~1Jt

0.3 ! • 0.3 Tp-2.Q*

"'.•'.'.

0.2 ;Ki-0.2Om 0.2

0.1 0.1
0
-r -
i
-
0 0.05 0.7 0.15 0.2 0 0.05 0.1 0.15 0.2
d$lm) dt(m)

Figure 4 Figure 5
Effect ofd, and H, on K, Effect ofd. and T on K,

Submerged Breakwater Transmission Submerged Breakwater Transmission

KtvsB(ds=0.05m,Hs- 0,10 m) Kt »s Slope (ds=0.05 m, Hs - 0.10 m)

0.9 0.9
0.5 0.8
0.7 0.7
j ;'p 1.21 ; 'Tp-l.lt
0.6 O.f
*0.5 i >i; r *0.5
0.4 .". .:>. :! " 1.5 s
0.4 c~~..- *
; • r/>-i.s<

0.3 :
• i irp- 2.0 s 0.3 i j rp-2.0*
0.2 . t -.:• 0.2
0.( 0.) :
0.5 ( (.5 02 0.3 0.4 0.5 0.5 0.7
>M Tan ffteta (B=0.6 m)

Figure 6 Figure 7
Effect of B and T on K, Effect of 6 and T„ on K,

Submerged Breakwater Transmission

Kt vs D50 (ds=0.05 m, Hs - 0.10 m)

Tp -1.2 •

S£0.5
Tp-f.St
0.4
0.3 Tp~2.0s
L~y---.r.yt

0.035 0.04 0.045 0.05 0.055 0.05

050 (m) (6=2.5 m)

Figure 8
Effect ofD„, and T„ on K,
 2008 COASTAL ENGINEERING 1998

Three dimensional test data showed similar trends in general but the scatter in the data
was much more evident. The transmission coefficients were generally higher in the 3-D
tests: this is attributed to a number of factors including diffraction of wave energy into the
lee of the breakwater and reflection of wave energy from the testing apparatus. The results
show that the relative submergence is the most influential factor under low submergence
conditions while crest width is important under higher submergence conditions.

The effects of submergence depth, incident wave height and crest width are shown in
Figure 9. This figure is based on the transmission coefficient immediately behind the mid-
point of the breakwater with incident waves perpendicular to the structure.

3-D Testing : Submerged Breakwaters

Effect of ds, Hs and B

». -
B - 0.19 m
'
3 - 0.38 m
•
3 • 0.95 m

0 12 3 4 5
ds/Hs

Figure 9 : Typical 3-D Testing Results

The predicted values of the transmission coefficient for the 2-D test variables were
generated using van der Meer's Equation for submerged and low crested breakwaters and
Ahren's Equation for reef breakwaters, and are shown in Figures 10 and 11.

Test Data vs. Ahrens' Equation Test Data vs. van der Meer's Equation

l i
B-Uffl
o S » 0.3 m
JHr« — B*0.6m
'JIiys_? ssia t* B = 0.6 m
W^X
5 1
A
B«1.Sm
z
/i *.-° r B"1.Sm
„ rrrf * B * 2.5 m

i ir-
„* V ri' " a-2.5 m
ir ^H B - 3.5 m

\A «j sd&#W Ueat. = Ob*.

0 0.2 0.4 0.6 0.8
S • 3.S m

Measured Kt
Measured Kt Mess. • Obs.

Figure 10 : Predicted J^ - Ahrens' Eqn. Figure 11: Predicted K, -Van der Meer's Eqn.
 COASTAL ENGINEERING 1998 2009

The results show that these equations are not suitable to represent transmission for
structures tested in this study, particularly when the crest width is large. The inability of
the existing design equations to predict suitable transmission coefficients for the tested
conditions indicates that there is a need for an improved design equation.

4.0 Development of an Improved Design Equation

Previous investigations have indicated that d/Hm0, is the most important dimensionless
variable affecting transmission. This observation was supported by these tests. Numerous
dimensionless variables considered, including variables discussed in previous authors
works. Only those found to be significant in defining the transmission process are
discussed here.

Dimensionless variables representing wave breaking, overtopping, frictional losses and

internal flow losses were found to be important in defining the transmission process.
These variables are discussed below.

i. The wave breaking process is considered to be represented by the dimensionless

variable d/Hm0i (relative submergence). The relationship between K, and d/Hm0i for
all of the 2-D test data is shown in Figure 12. The effect of relative submergence
is very evident when d/Hm0i is small. This is expected since the majority of waves
are breaking and under breaking wave conditions, the unbroken wave (transmitted)
has been found to be closely related to the water depth (ds). Given the scatter,
there are obviously other factors playing an important role in the process as well.

All 2-D Test Data

Kt vs. ds/HI

1 1 *11 ft
1 . 4 i!"

fr-A --
<H %
- i

X ~JuLkM
^ySr
. r i

—itSJt—
lr"T~^
Is P
.. ...... _
0 12 3 4
ds/HI

Figure 12 : Effect of Relative Submergence

As the relative submergence increases, its influence is reduced substantially. This

is expected as the relative portion of the incident waves which would break is
reduced as the submergence depth increases.
 2010 COASTAL ENGINEERING 1998

ii. The effect of overtopping is expected to be relatively important in defining the

transmission coefficient at submerged breakwaters, especially under wave breaking
conditions. As the relative submergence approaches zero, transmission at the
submerged breakwater was observed to become a function of the potential for
overtopping as well as transmission through the breakwater structure, in particular,
the large armour crest material.

Typically, overtopping rates are a function of the wave steepness and structure
geometry. A dimensionless form of the structure crest width and the local wave
height (Hm0/B) was found to be representative of the overtopping effect.

iii. As the wave passes over the breakwater, some energy is lost to frictional
dissipation on the surface of the structure. The dimensionless variable
representing this process was loosely based on the empirical Darcy-Weisbach
expression for head loss. This requires the assumption that the flow velocity can
be represented by the velocity of a gravity wave and results in the dimensionless
variable dfl^/BD^.

iv. Flow within the breakwater structure will also result in some energy loss as a wave
travels over a submerged breakwater. Given the relatively high porosity of the
armour layer, the effect of the wavelength on the fluid velocity, the effect of
submergence depth on the portion of flow within the armour layer and the effect of
the crest width on the overall drag losses, the dimensionless variable selected to
represent drag losses was Bd/LD50a.

On the basis of these dimensionless variables, statistical fitting was used to develop a
suitable design equation. A number of alternative equations were considered in an effort
to develop a design equation which provided:
• a good statistical fit (R2),
• a normal distribution of residuals,
• predictions which are well bounded (ie. 0.0<K,< 1.0),
• physically relevant variables with a minimum number of fitted parameters

Through trial and error, a finalized form of the design equation was developed such that
these criteria were generally satisfied for the 2-D test data.

-0.65(^)-1.09(-J.) B d d H
1 - (e "' * 0.047(—i-) - 0.067(-i_4) (4)
L D,. B V,n

The proposed equation fits the 2-D test data well, resulting in an R2 value of 0.914. Given
the uncertainty associated with the fitting of statistical parameters, all of the parameters of
the proposed design equation were adjusted to their upper or lower 95% confidence
 COASTAL ENGINEERING 1998 2011

interval values such the change in predicted K, value was maximized. The results, as
shown in Figure 13 show that the prediction is not very sensitive to these changes and as a
result, the prediction is relatively robust in its relation to the physical variables affecting
the transmission phenomenon.

Proposed Kt Equation
Sensitivity to Parameters (95 % C.I.)

m
H- tf^
i?T.:
"' • t\ "A
Est. Param.

fvm- ^J +9S% Param.

SF;-i*i r-
1
,,sua
vmtt
-95% Param.

•1
0 0.2 0.4 0.6 0.8
Observed Kt

Figure 13 : Sensitivity to Parameter Estimates

Although the proposed equation does not fit the 3-D data as well, the prediction is still
relatively good at low to moderate transmission coefficients (Figure 14).

Fit of Equation for 3-D Test Data

Est Kt vs Obs. Kt Behind Breakwater

Proposed Equation

Est.- Obs.

Figure 14 : Observed vs Predicted K, - 3-D Tests

5.0 Conclusions

Based on an extensive set of 2-D and 3-D tests of wave transmission at submerged
breakwaters, a number of conclusions can be drawn.
i. The transmission coefficient at submerged breakwaters is most sensitive to the
 2012 COASTAL ENGINEERING 1998

relative submergence (d/H,).

ii. The relative crest width is another very important factor which has not been
adequately accounted for in previous design equations,
iii. An improved design equation for transmission at submerged breakwaters would
be:
-0.6S(1)-1.09(^) B d d H
£, = 1 - (e ' + 0.047( i-) - 0.067(-i—-)) (4)
L D,„ B D,h

The proposed equation represents the comprehensive set of test data well
(R2=0.914) and is robust in its relation to the physical variables which significantly
affect the transmission process.
The equation is well bounded over the range of test data, which is considered to be
representative of typical design conditions. The equation does, however, become
unbounded when B becomes very large or very small due to the size of the 3rd and
4th terms. Therefore, it is recommended that caution be used when applying the
equation outside of the following variable ranges.

0 < '- s 7.08

£D,„
d H (S)
0 s —:—'- s 2.14
 COASTAL ENGINEERING 1998 2013

References:

Ahrens, J.P. "Characteristics of Reef Breakwaters". CERC Technical Report 87-17,

1987.

Beji, S. and Battjes, J.A. "Experimental Investigation of Wave Propagation Over a Bar",
Coastal Engineering, 19:1-2, 1993, pp. 151-162.

Beji, S. and Battjes, J.A. "Numerical Simulation of Nonlinear Wave Propagation Over a
Bar", Coastal Engineering, 23:1-2, 1994, pp. 1-16.

Driscoll, A.M., Dalrymple, R.A. and Grilli, S.T. "Harmonic Generation and Transmission
Past a Submerged Rectangular Obstacle", 23rd International Conference in Coastal
Engineering, American Society of Civil Engineers, New York, Vol. 1, 1993, pp. 1142-
1160.

Losada, I.J., Silva, R. and Losada, M.A. "3-D Non-Breaking Regular Wave Interaction
with Submerged Breakwaters", Coastal Engineering, 28, 1996, pp. 229-248.

Mansard,E.P.D. and Funke, E.R. "On the Reflection Analysis of Irregular Waves",
Technical Report No. TR-HY-017, National Research Council of Canada, Ottawa, 1987.

Ohyama, T. and Nadaoka, K. "Modelling the Transformation of Nonlinear Waves Passing

over a Submerged Dyke", 23 rd International Conference on Coastal Engineering,
American Society of Civil Engineers , New York, Vol. 1, 1993, pp. 526-539.

Seabrook, S.R. "Investigation of the Performance of Submerged Rubblemound

Breakwaters", M.Sc. Thesis, Queen's University, Kingston, 1997.

Seelig, W.N. "Estimation of Wave Transmission Coefficients for Overtopping of

Impermeable Breakwaters", CERC Coastal Engineering Technical Aid 80-7, 1980.

Van der Meer, J.W. "Stability and Transmission at Low-Crested Structures", Delft
Hydraulics Publication 453, 1991.

Vidal, C, Losanda, M.A., Medina, R., Mansard, E.P.D. and Gomez-Pina, G. "Universal
Analysis for the Stability of Both Low Crested and Submerged Breakwaters", 23rd
International Conference on Coastal Engineering, American Society of Civil Engineers ,
New York, Vol. 2, 1993, pp. 1679-1692.