International Journal of Small Craft Technology

BANK EROSION FROM SMALL CRAFT WAVE WAKE IN SHELTERED WATERWAYS

G J Macfarlane, Australian Maritime College, Australia
G Cox, Kamira Holdings Pty Ltd, Australia
J Bradbury, Tasmanian Department of Primary Industry and Water, Australia
SUMMARY
The wave wake generated by large vessels has received high-profile research and regulatory responses in recent years.
Conversely, recreational vessel wave wake can impact significantly on sheltered waterways, yet the sector receives little
research funding and is often regulated with simplistic criteria. Small craft wave wake, comprising smaller waves and
shorter periods, can be quantified adequately with fewer variables, opening the possibility for over-arching guidelines
that are not site-specific.
This paper analyses past small craft wave wake studies, including bank erosion studies, to determine which measures of
erosion potential are the most descriptive. Past approaches using a single criterion are shown to be generally inadequate
and often misleading. A multi-criteria approach has been adopted to ensure that all the erosive components of highspeed, small craft wave wake are accounted for. A possible implementation method is discussed.
NOMENCLATURE
E
Em
Fr
Frh
g
h
H
Hm
L
n
T
Tm
T1
v
y

1.

Energy per metre wave crest length [Jm-1]

Energy of the maximum wave [Jm-1]
Length Froude number [v(gL)-1/2]
Depth Froude number [v(gh)-1/2]
Acceleration due to gravity [9.81ms-2]
Water depth [m]
Wave height [m]
Maximum wave height [m]
Waterline length [m]
Wave decay exponent
Wave period [s]
Period of the maximum wave [s]
Period of the leading wave [s]
Vessel speed [ms-1]
Lateral distance between vessel sailing line and
measurement point [m]
Constant
Pi
INTRODUCTION

Vessel wave wake (also commonly referred to as wash

or wake wash) has been a prime topic for study over the
past two decades, though it no longer attracts quite the
same attention since industry has gained a general
understanding of the primary issues. Sufficient science
has been developed to allow for the regulation of the
most damaging vessels without actually perfecting the
science.
It is known that wave wake issues can differ
considerably depending upon the size and/or speed of
the vessel(s) and the location(s) in which they operate.
As a result, it may be useful to categorise particular
scenarios into the following three distinct regions, with

2008 Royal Institution of Naval Architects

reference to examples of rivers and harbours in

Australia:
a) Highly Sensitive Regions - This region includes very
sheltered waterways such as rivers with very limited
fetch and/or width. They often have steep, cohesive
banks that are highly susceptible to erosion by
vessel wave wake. Vessel speeds are likely to be
restricted to a small range of sub-critical depth
Froude numbers. Vessel operation at trans-critical
depth Froude numbers should be avoided and
operation at super-critical depth Froude numbers
may be limited to only very small craft (less than
about five metres length). Examples include the
lower Gordon River, upper reaches of the
Parramatta River and sections of the Noosa River.
b) Moderately Sensitive Regions - This region includes
semi-sheltered estuaries such as the lower reaches
of large rivers and harbours or areas where
shorelines have been artificially armoured to
withstand increased wave action. Vessel speeds are
likely to be restricted to a range of sub-critical depth
Froude numbers. The possible exceptions may
include certain small craft and larger wave wakeoptimised craft that could operate at some supercritical depth Froude numbers. In such cases,
multiple criteria may be required to determine
acceptable speeds for each vessel type (this is
discussed in more detail in later sections). Operation
at trans-critical depth Froude numbers should be
limited to acceleration and deceleration between the
sub and super-critical conditions. Examples include
the lower reaches of the Parramatta and Brisbane
Rivers and sheltered areas of Sydney Harbour.
c) Coastal Regions - In these more exposed regions,
wave wake criteria generally only apply to large
high-speed craft operating at trans or super-critical
depth Froude numbers. Minimal problems eventuate
from almost all vessels operating at sub-critical

speeds. Some existing criteria applied to high-speed

vessels are based on acceptable levels from
conventional (i.e., not high-speed) vessels
operating at sub-critical speeds. Often the criteria
are imposed due to adverse safety risks for other
users of the waterway (and shoreline) as a result of
large/long vessel waves generated at high speeds.
Examples include Scandinavian coastal regions and
Marlborough Sounds in New Zealand.
The differences between types of waterways are
discussed further in Section 12.
The limited number of wave wake regions where wave
wake is of concern within Australia (such as those of
the Gordon, Parramatta and Brisbane River ferry
services) have been the subject of individual studies
that have sought vessel-specific solutions, as opposed
to an over-arching methodology that would allow for a
desktop evaluation of any vessel in any waterway,
Macfarlane and Cox (2007).
Australia does have a relatively large recreational
boating population that utilises the limited sheltered
waterways available. This is not dissimilar to the USA,
where the majority of recreational boating is enjoyed on
fresh water lakes and rivers, as well as sheltered coastal
waterways, rather than the open ocean itself. It
therefore makes sense to attempt to develop guidelines
for vessel wave wake that allows for the sustainable use
of sheltered waterways.
In Australia, this has generally come in response to a
perceived erosion event. It can be said with some
certainty that maritime regulatory authorities have been
reactive in their approach to wave wake and erosion. A
partial exception are the Gordon River services which,
operating within a National Park and World Heritage
Area, are regulated by a land management rather than
maritime agency. There the initial response in the early
1990s was reactive, but became proactive with the
implementation of a long-term monitoring and vessel
certification process that is on-going today, Bradbury et
al (1995), Bradbury (2007a).
2.

RELEVANT WAVE WAKE STUDIES

The Australian Maritime College (AMC) has conducted

studies of wave wake and erosion in several
recreational sheltered waterways in South-east
Queensland, Macfarlane and Cox (2004). The
Queensland Government had a pressing desire to find
the causal links between the wave wake of certain
vessels and the erosion they may have caused in the
Noosa, Brisbane, Bremer, Maroochy and Mary Rivers.
One aim was to raise awareness of potential effects of
new classes of vessels and activities such as
wakeboarding before erosion occurs, so that regulatory
bodies are not reliant solely on reactionary measures.

The AMCs studies collected wave wake data from

controlled field experiments on a range of small craft,
but without actually measuring corresponding erosion.
Instead, erosion studies undertaken on the Gordon
River prior to 1995 were re-analysed in an attempt to
derive relationships between measured small craft wave
wake and erosion thresholds. A set of operating criteria
were developed from the re-analysed Gordon River
data.
In a separate study of different bank types, similar
controlled field experiments were conducted on the
Gordon River between 1997 and 2005 by the
Tasmanian Department of Primary Industry and Water,
but in this instance both the wave wake and the
subsequent shoreline disturbance were recorded,
Bradbury (2007b), Macfarlane (2006). An aim of this
paper is to compare the results of these two different
studies and define the areas of correlation.
3.

WAVE WAKE AND BANK EROSION

3.1

INTRODUCTION

At the outset, it must be accepted that there may never

be a rigorous theory that links vessel wave wake and
erosion. This is similarly the case in coastal
engineering, where beach erosion is predicted by a
number of largely empirical and statistical rules
developed over many decades. Those rules may have a
grounding in basic science and engineering, but they
are underpinned by empirical equations and a reality
that can only be represented statistically, with
accompanying error as a consequence. One only has to
review The Coastal Engineering Manual (formerly the
Shore Protection Manual) by The US Army Corps of
Engineers (2002), to see that it is weighted heavily with
model test results and empirical tables.
Similarly, it must also be accepted that most natural
waterways are dynamic environments subject to
erosional and/or depositional processes. Not all erosion
events can be blamed on vessel wave wake. In many
instances local land use practices such as riparian (river
bank) vegetation removal and farming, as well as
waterway issues such as regulation, channelisation,
extractive processes and upstream or downstream
development (of flood protection or harbours, for
instance), can be the root cause of upstream erosion.
Recreational boating often simply becomes the focus of
attention for an otherwise existing and complex
problem.
In contrast to naturally-occurring wave climates, a
vessels wave wake is characterised by short event
duration and a broad spectral spread of wave
parameters that do not lend themselves to the
application of conventional statistical methods.

2008 Royal Institution of Naval Architects

Instead, the principle statistics of concern may well

relate to the extent to which certain wave wake
parameters exceed those of the existing wave climate in
a particular area.
3.2

RECREATIONAL CRAFT

In developing comprehensive, but simple, bank erosion

criteria for recreational craft, a number of factors must
be realised:
1. Recreational boating is not a substantial direct
revenue source for marine regulatory authorities,
so receives limited attention, hence limited
funding.
2. When funding for maritime scientific investigation
is limited and a political solution must be found,
recreational boaters are soft targets. It is often
easier and cheaper to apply a blanket speed limit to
boating activities than to police it.
3. Vessel wave wake complaints are often used to
mask other community concerns such as the noise
generated by high-speed craft and the loss of
amenity. Communities and governments react
strongly to tangible evidence such as bank erosion,
regardless of the cause, whereas noise and loss of
amenity are more subjective, somewhat less
tangible, and therefore less likely to attract
regulation.
4. Shoreline erosion can very often be the result of
land use issues, engineering works, river regulation
or climate change and sea level rise.
5. Regulators, builders and owners of small craft have
scant information relating to vessel parameters
such as displacement, dimensions and hull design.
Often only very simplified parameters must be
relied upon to determine wave wake potential.
6. Every possible combination of bank type, bank
material, riparian vegetation and river bathymetry
cannot be covered, and indeed may not need to be.
7. The influence of environmental variables such as
shallow water must be limited as they introduce an
exceptional number of additional parameters.
Fortuitously, the vast majority of recreational vessels
using sheltered waterways are small, high-speed
monohulls, typically used for water skiing and
recreational fishing. There are often a smaller number
of slow-speed vessels such as professional fishing
vessels, houseboats and workboats, and some
commercial charter and ferry operators. If the vessel
length is sufficiently large and it is engaged in a
commercial operation, case-by-case testing and
approval can be implemented.
4.

wave height, period, water depth and direction of

propagation. For the purpose of comparing vessels or a
single vessel at a range of speeds the last two can often
be held reasonably constant at site appropriate values,
leaving height and period as the key variables.

RELEVANT WAVE WAKE

CHARACTERISTICS

For the purposes of studying small craft wake waves,

four parameters are necessary to adequately describe a

2008 Royal Institution of Naval Architects

It is necessary to identify the waves of

geomorphological interest and focus upon them. Of the
two vessel-generated wave types, transverse and
divergent, it is the divergent systems that dominate in
high-speed vessel wakes. Transverse waves can be
significant when generated by displacement hull forms
or heavy, transom-sterned high-speed craft traversing at
displacement speeds, and especially where the
waterways are very narrow. Transverse wave height
(and therefore energy) decays faster than divergent
wave height with lateral separation from the sailing
line, but this decay becomes bounded by the shoreline.
Being more of a concern with slow vessel speeds,
transverse waves are best controlled by changes to
operating speeds and vessel design. The focus here is
on small recreational craft, and therefore on divergent
waves.
For other applications a degree of vessel induced
erosion may be acceptable but limits may need to be
placed on how much erosion is to be permitted. Earlier
work on the Gordon (von Krusenstierna 1989, Nanson
et al 1994) drew attention to an increase in the rate of
erosion as waves became larger (as opposed to simply
higher) and found that simple measures could explain
much of the erosion. However, allowing some erosion
is more complicated because one must then consider
the cumulative effects of all waves exceeding the
erosion threshold.
In analysing vessel wave wakes, the two parameters of
maximum wave height and the corresponding wave
period for the highest wave (often termed the maximum
wave) have therefore been adopted as the primary
measures. The importance of quantifying wave wakes
with simple measures is critical when assessing small
craft wave wake impacts. If the measures were
complicated, statistically difficult to represent or costly
to collect and collate, regulatory authorities would be
reluctant to pursue a path of boating management
through scientific understanding. Blanket speed limits
might be a typical response but these, which to be
effective must be specified for the worst offender, are
likely to be overly restrictive for other vessel classes.
Our primary measures, height of the maximum wave
and its corresponding period, appear to exhibit certain
predictable relationships at high vessel speeds, which is
essential to development of a simple but sound method
for predicting small craft wave wake. Cox (2000)
demonstrated for high-speed craft travelling at subcritical depth Froude numbers that divergent wave
height is largely a function of length-displacement ratio
and the corresponding period is largely a function of

vessel waterline length. Analysis using the AMCs

wave wake database, Macfarlane and Renilson (2000),
and by others (Warren, 1991) clearly supports this.
Vessel hull form has only a limited bearing on highspeed, deep water wave wake, as demonstrated in
Figures 1 and 2 (data obtained from the AMC wave
wake database). Figure 3, taken from field tests in
Queensland, shows how high-speed wave period
(divided by the square-root of L) collapses to a narrow,
constant band at high speeds.
In Figure 4, the most common wave wake parameters,
such as energy, power and height, show growing values
with increasing vessel speed, peaking at a certain speed
(normally about Fr = 0.5) and then decreasing back to a
lower level. Similarly, wave period also grows with
increasing vessel speed, peaks, but tends to level off
rather than decrease at higher speeds. Regardless of
which wave wake parameter is used as the erosion
indicator, it is clear that there may be two distinct
operating speed ranges slow speed and high speed,
with intermediate transitional speeds to be especially
avoided.
Planing craft in particular are burdened by this
transition hump where resistance and hence wave
wake is high. In some sports, such as wake boarding,
this is viewed by the proponents as beneficial. Many
boaters will explain anecdotally how they believe it is
better to travel at high speeds in sheltered areas and this
reasoning has long been used as a justification for
transiting at speed. The current science would not
support such a generalisation since the waves from
small planing vessels have been demonstrated to be
capable of eroding both muddy and sandy banks (even
at the greatest distance allowed by the sheltered
waterbodies examined).
5.

WAVE MEASUREMENT

If waves are measured too close to the passing vessel

(within one boat length laterally, though this can be
speed-dependent), the measured waves may be subject
to localised interactions. If they are measured too far
from the vessel, the dispersing wake waves may be
substantially affected by the existing wind wave
environment.
During the Noosa and Maroochy River field studies the
reference point for wave measurement was standardised
so that the results were directly comparable. The
nominal lateral distance from the sailing line to the
wave probe was set at 23 metres. This distance had
some relevance to these locations, as they have an
average width of about 100 metres. It was surmised that
most vessels would navigate in the centre half of the
river, so would not normally stray closer to the bank
than one-quarter of the width. The 1997-2005 Gordon
River studies, having been undertaken by different

personnel and for different reasons and on a river

almost twice as wide, had a greater lateral separation of
a nominal 50 metres between measurement point and
sailing line.
To compare the results of trials where different lateral
separations between sailing line and measurement point
are used, wave heights must be corrected according to
their decay with distance. The relationship between the
maximum diverging wave height and lateral distance of
deep water waves varies according to Equation 1
(Macfarlane and Renilson, 1999):
Hm = yn

(Equation 1)

The variable is a vessel-dependent function of speed.

The exponent n has theoretical values of - and -
for divergent and transverse waves respectively when
measured at the points of intersection of these two
wave trains, Sorensen (1973). During field trials the
wave measurement point is most likely not at the point
of intersection, so the exponent values may vary. It was
decided that a - decay exponent was appropriate for
deep water divergent waves, recognising that it is not
necessarily absolute (nor applicable in those relatively
rare cases where transverse waves have greater
geomorphic effect). Analysis of the AMCs wave wake
database shows that the deep water, divergent wave
decay exponent generally varies between a range from
-0.22 to -0.4, where -0.33 is considered a reasonable
engineering approximation (Macfarlane, 2002).
6.

USEFUL WAVE WAKE MEASURES IN

SHELTERED WATERWAYS

Historically, wave height has been used as the primary

comparative measure for vessel wave wake. It is
possibly the simplest parameter to measure and this
fulfils another desirable requirement it is within
public perception where subjective visual observation
must substitute for engineering measurement. Similar
comments were made by Lesleighter (1964) in his
analysis of ski boat wave wake on the Hawkesbury
River, where he found that inflated anecdotal claims of
excessive wave wake height could not be substantiated
by measurement.
In the authors opinion, the historical use of wave
height alone, or indeed any single criterion, cannot
possibly reflect the true erosion potential of a vessels
wave wake. Wave period is a strong indicator of the
potential to move sediment in any shoreline
environment, either through the period-dependent
orbital velocity below the surface of shallow, but
unbroken, waves, or through the gravity driven jets of
plunging breakers. Period, along with height, is
required to calculate both wave energy and wave
power. Sheltered waterways generally see only a wind
wave environment. Wind waves of short fetch (and

2008 Royal Institution of Naval Architects

even waves of longer fetch, such as wind-driven ocean

seas) exhibit a disproportionate growth relationship
between wave height and period, disproportionate in
that wave height grows more rapidly than period but
both have equal weighting in calculating wave energy.
Table 1 shows example hindcast wind waves values for
different wind speeds and fetches.
It is clear that increasing either fetch or wind speed
leads to much faster growth in wave height than wave
period. Consequently it can be argued that sheltered
waterways experience occasional wind wave height
variations of several hundred percent, but with limited
accompanying wave period growth. Sheltered
shorelines in a wind wave environment are often
dynamically stable. Beach areas, if they exist, adjust in
response to the prevailing wave climate and sediment
budget. Other landforms in low wave energy
environments may typically owe their genesis to
processes not associated with waves. When there is a
substantial increase in incident wave period beyond
what such landforms would normally experience the
shoreline may experience erosion. Not only are the
longer period waves more energetic but orbital currents
capable of entraining sediment extend to greater depths.
Where mud flats are present, shoaling long-period wake
waves may form higher breakers more likely to plunge.
Small craft traversing at high speeds in sheltered
waterways can generate wave periods far longer than
those which occur naturally.
The geomorphic impact of wind waves is not evenly
felt throughout river systems and the greatest impacts
occur at the downwind ends of reaches. In contrast,
vessel wave wake impacts are more evenly spread
throughout the waterway, with diverging waves
especially impacting upon shorelines that would not
otherwise be subjected to a significant incident wave
climate. The wave wakes of high-speed craft, in
particular, are dominated by the divergent wave system
and, as the depth Froude number becomes supercritical, all waves propagate obliquely to the sailing
line.
7.

WAVE ENERGY OR POWER?

Wave energy and wave power are both used in coastal

engineering. Assuming a simplified, sinusoidal wave
form, wave energy (per wavelength and unit crest
width) is proportional to H2T2 and wave power (derived
from energy density) is proportional to H2T.
Power is a useful descriptor of wave energy over a
period of time, such as may be found in the statistical
analysis of an incident wave field acting over a long
timeframe. In the case of the wave wake of a passing
vessel, the waves generated are discrete events and so
do not necessarily lend themselves to description on a

2008 Royal Institution of Naval Architects

statistical time basis. It is felt that wave energy may be

a better measure for such discrete events.
The Noosa and Maroochy River studies introduced
another derived parameter wave energy per unit wave
height, or a HT2 relationship when reduced to its
principal variables, Macfarlane and Cox (2004). This
parameter has shown empirically to display the most
promising correlation between incident waves and
erosion. Short period waves of less than 2 seconds
period, such as the maximum waves generated by small
craft and sheltered waterway wind waves themselves,
do not shoal to any appreciable degree before they
break. Moreover, a wave breaks when the water depth
approximately equals the wave height, so its energy is
concentrated into a depth of water equal to the wave
height. Consequently, it is believed that energy per unit
wave height is a measure of the energy content in a
short-period wave just at the time of breaking, and
therefore the energy being dissipated onto the shoreline.
It is agreed that this explanation has not been tested and
serves only as a possible explanation of the empirical
strength of the HT2 relationship with erosion rates.
8.

BANK EROSION STUDIES

In Australia there have been several significant studies

that have attempted to measure bank erosion from
vessel wave wake. The first were academic
collaborations on the Gordon River in the early 1990s
(von Krusenstierna 1990, Nanson et. al., 1994) and the
second is an on-going study on the Gordon River
conducted by Bradbury (refer Bradbury et al, 1995).
There was a desktop study of the Swan River
(Pattiarachi 1990), but its analysis technique was
rudimentary and the results inconclusive.
Von Krusenstierna attempted to measure erosion using
erosion pins set into sandy banks, from which
commercial vessels were subsequently banned. The
waves of a passing vessel were measured and the
resultant erosion was measured. Unfortunately the wave
wake data was analysed and averaged in such a way as
to make it almost unrecognisable. However, it was
demonstrated that there was a threshold of wave wake
values below which the rate of erosion was regarded as
less significant, and such thresholds were evident for all
of the wave wake measures recorded. In the Noosa
River study, this manipulated data was recovered
somewhat and threshold wave energy and period values
were extracted. At the time they appeared reasonable,
but there was no comparative data available.
From the mid 1990s further experiments were
conducted on cohesive muddy banks lining the Gordon
River reaches remaining open to commercial traffic.
Although of cumulative concern, the amount of erosion
per vessel pass was expected to be less than 0.1 mm
and therefore undetectable by measurement of erosion

pins in the field. Instrumental measurement of the

turbidity (degree of suspension of solid material)
resulting from sediment suspension in normally very
clear water was therefore used as a proxy for erosion.
Since the land manager had a pressing need for a
criterion to distinguish appropriately low wake energy
vessels the suggestion from early (and limited) data of
an initial threshold to sediment movement at a wave
height of 75 mm was used to define the maximum
acceptable wave. Subsequent work demonstrated this
to be overly simplistic and that wave period was also an
effective influence. That point was most graphically
demonstrated by the extreme turbidity caused by the
low but long waves generated by small planing craft.
However limiting wave height and period
independently was found overly restrictive in that it
excluded many of the wave wake events that did not
cause any erosion.
Four graphs from one of several sites used in the ongoing study are presented, showing turbidity near the
river bank (measured at two different water depths)
against the maximum wave parameters of height
(Figure 5), period (Figure 6), energy (Figure 7) and
power (Figure 8). A fifth graph, Figure 9, shows
turbidity against the derived parameter of HT2 (energy
per unit wave height). From these graphs, several
salient features become apparent:
All graphs define very definite threshold values
below which turbidity is essentially zero (ie. within
the range of instrumental and background noise).
Wave height is a relatively poor indicator of
erosion potential. One wave height value of
178mm shows zero turbidity, yet the secondhighest recorded turbidity event occurs for another
wave at this same height.
There is close correlation between wave period and
turbidity.
There is similarly close correlation between both
wave energy and power with turbidity.
The derived parameter HT2 exhibits the tightest
grouping of all data points.
9.

OPERATING CRITERIA

9.1

SINGLE OR MULTIPLE CRITERIA?

The historical application of a single operating

criterion, most notably a limit on the height of the
maximum wave, has been demonstrated to be at best
unreliable and at worst incorrect. For example, a single
criterion of wave height was adopted for vessel
operations on Sydney Harbour yet there have been
reports of significant foreshore damage (Kogoy, 1998).
Moreover, attempts to remedy the lop-sided nature of a
single parameter criterion by expanding into a wave
energy or wave power form may be inefficient in
containing all erosive components of vessel wave wake,
so that some vessel types might still be over-restricted

when limits are based upon experimentally determined

thresholds to erosion.
On the other hand, where limits have been determined
by desktop (or otherwise incomplete) studies they have
not always been appropriate for all relevant vessel
types.
For instance, some high-speed vessels,
particularly those claimed to possess wave wake
reducing characteristics (which are more strictly often
only wave height reducing characteristics by way of
high length-displacement ratio) have the potential to
satisfy an apparently reasonable energy criterion but
still cause erosion. Prime examples of this are the
various low-wave wake ferries operating in Sydney
and Brisbane, Australia. Such vessels have been found
capable of generating wave periods considerably in
excess of the existing waterway wave climate (up to 4-5
times longer), but with low accompanying height when
travelling at high speed. The possibility of this is
illustrated graphically in Figure 4, where the most
commonly applied wave wake parameters of energy,
power and wave height reduce gradually at high speeds.
An energy criterion may be reasonably defined for
commercial vessels operating at any speed. However,
if commercial vessels are forced to operate at low speed
then it is possible that smaller, high-speed recreational
craft may meet the energy criterion yet still create an
erosive wave wake, since aspects of geomorphic
response may be linked more to wave period than
height.
The particular sensitivity of sheltered waterways to
incident wave period led to the belief that multiple
criteria were the key to any operating limits. Highspeed, deep water wave period was shown to remain
essentially constant whereas other parameters decreased
steadily as speed increased well into the high-speed
range. The other benefit of including a period-based
criterion was the strong correlation at high speed
between the period of the maximum wave and a
vessels waterline length. This was an important part of
the criterion simplification process necessary for
eventual application in practical situations.
9.2

CRITERIA DEVELOPMENT

From an analysis of the original Gordon River erosion

studies (von Krusenstierna, 1990) two criteria were
derived both for the Noosa and Brisbane Rivers, to be
applied jointly; an energy criterion and a wave period
criterion. The data plot used to derive the energy
criterion could also be used to derive a period criterion.
By then using the relationship between wave period and
waterline length, as discussed in Section 4, the period
criterion simply became an upper limit of waterline
length.
The wave wake data presented in von Krusenstierna
had been collected by a group dominated by

2008 Royal Institution of Naval Architects

geographers. The data was presented in a modified

statistical form, using parameters such as significant
wave height and period, with the raw data being
discarded. Unfortunately this data format is
incompatible with present day wave wake analysis, but
there was sufficient information available to recover
certain values in the form of the maximum wave height,
Hm, and a statistical derivation of the period of the
leading wave, T1.
When HmT12 was plotted against T1 as a log-log graph
(Figure 10), the Gordon River erosion data tended to
clump into three distinct groups low erosion,
moderate erosion and substantial erosion. The derived
values of HmT12 and corresponding T1 values were
converted into energy and waterline length limits as
follows:
T1 was related to vessel static waterline length
using statistical values derived from ship model
wave wake testing, knowing that there was
reasonable predictability in the relationship for the
deep water, high-speed condition. The empirical
relationship used was:
22L
(Equation 2)
T1 =
3g

HmT12 was converted to energy by multiplying by a

fundamental wave height, which was one that
was derived from field work and used for many
years as the operating criterion for the Gordon
River, corrected for the different lateral distance to
the measurement point.

The threshold energy and period values for the upper

bound of the low erosion group became the Noosa
River criteria, the Noosa River being a very sheltered
river with sensitive banks. The lower bound of the
moderate erosion group became the Brisbane River
criteria, the upper Brisbane River being a more
energetic environment.
Using the low erosion grouping does not imply
zero erosion, as there are some events in this
grouping where modest erosion was recorded. Further
discussion of a possible zero threshold is presented in
Section 11.
10.

DERIVED CRITERIA

measured at a point approximately 23 metres abreast of

the sailing line.
10.1(b) Period-Based Waterline Length Criterion
A vessel capable of satisfying the energy criterion at
any speed (in knots) greater than 3.04L (i.e., a length
Froude number >0.5) should also have a static
waterline length (L) less than 5.2 metres. Vessels
longer than 5.2 metres waterline length should be
restricted to those speeds less than 3.04L that satisfy
the energy criterion.
10.2

BRISBANE RIVER

10.2(a) Energy Criterion

The energy per metre of crest length of the maximum
wave is to be less than 180 J/m, i.e.,
1962Hm2Tm2 180 J/m

(Equation 4)

measured at a point approximately 23 metres abreast of

the sailing line.
10.2(b) Period-Based Waterline Length Criterion
A vessel capable of satisfying the energy criterion at
any speed (in knots) greater than 3.04L shall also have
a static waterline length (L) less than 9.0 metres.
Vessels longer than 9.0 metres waterline length shall be
restricted to those speeds less than 3.04L that satisfy
the energy criterion.
It is noted that the energy threshold for the Brisbane
River is three times that of the Noosa River. In coastal
engineering terms, energy states tend to jump in orders
of magnitude, not in incremental percentages. In many
respects the push by designers to improve the wave
wake characteristics of their vessels by a nominal
modest percentage is likely to be somewhat
inconsequential in erosion terms. Generally, a design
either will or will not work small changes to design
parameters like hull spacing, waterline beam, draught,
etc are unlikely to turn an erosive design into an
acceptable one. These double criteria highlight the
design dilemma lengthening the hull to reduce the
displacement-length ratio and hence wave height will
only increase the wave period. This has been the legacy
of the low wave wake catamaran designs.

In numerical terms the criteria for each river are:

10.1

NOOSA RIVER

11.

COMPARISON BETWEEN BANK

EROSION STUDIES

10.1(a) Energy Criterion

The energy per metre of crest length of the maximum
wave is to be less than 60 J/m, i.e.,
1962Hm2Tm2 60 J/m

(Equation 3)

2008 Royal Institution of Naval Architects

It is useful to compare the operating limits derived

during the Noosa River study (using the original
Gordon River erosion data of von Krusenstierna, 1990)
with those obtained from the ongoing Gordon River
studies.

The threshold (practically zero turbidity) values for

energy, power and period of the maximum wave at one
Gordon River site (Figures 7, 8 and 6) are:
Energy
30 J/m
Power
10 W/m
Period
1.1 to 1.2 seconds
When this threshold energy value of 30 J/m is
transposed from the 50 metre lateral distance
measurement used on the Gordon River to the 23 metre
distance used on the Noosa River using a - wave
decay exponent, the threshold energy becomes 50 J/m.
This is comparable to the 60 J/m applied to the Noosa
River. It must be remembered that the Noosa River
energy criterion was based on a low erosion data
grouping and not a zero threshold recorded in the
later Gordon River studies. It is also worth noting that
to completely avoid erosion the criteria must be
matched to the most erosion susceptible site under
conditions where erosion is most likely to occur.
Although the Gordon results presented indicate a
threshold of 30 J/m, the selected test site was
deliberately chosen to allow testing of the widest range
of vessel speeds because it was considered to be
relatively robust and capable of repeated wave impacts
without unnecessary bank degradation. Work at more
sensitive sites has concentrated simply upon defining
the threshold to erosion and a generalised limit as low
as 10 J/m at the measurement point may be more
appropriate. This would translate to approximately
16 J/m under the Noosa River test conditions.
When comparing wave periods, the Noosa criterion of a
maximum waterline length of 5.2 metres corresponds to
a period of the maximum wave at high speeds of about
1.4 to 1.5 seconds, taken by reviewing the vessel wave
wake data and comparing the recorded periods with the
recorded waterline lengths. This period is slightly
longer than the 1.1 to 1.2 seconds threshold recorded in
the Gordon data, but the variation may again be
explained by the difference between low erosion and
zero threshold.
In conclusion, it appears that two independent studies
separated by over a decade, using different equipment,
vessels, measurement techniques and analyses, have
derived very similar relationships between bank erosion
and high-speed, small craft wave wake.
12.

WATERWAY TYPES

As noted, unless the overall numbers of parameters are

reduced and those chosen are simplified, it would be
impossible to derive any useful erosion criteria. From
discussions with coastal engineers, it is believed that
there may only need to be as few as two, but probably
three, bank types studied. All are natural depositional
landforms. Artificial shorelines are more diverse and

should be engineered to withstand an appropriate wave

climate, although that has not always been the case.
The first is typical of very sheltered waterways that
may experience little or no tidal range and do not have
a beach structure since the energy climate is not wave
dominated. These low-lying banks tend to be
characterised by cohesive muds and substantial
sediment trapping riparian or saltmarsh vegetation
(usually not mangroves). The sediments are fine
enough to be transported in suspension by currents and
these deposits may represent the accumulation of
sediment over extensive (geomorphological) timescales. Once such natural features are disturbed by
erosion the damage is effectively permanent.
The second is characterised by some resemblance to a
beach, usually consists of fine sand and muddy
sediment (so-called muddy sands) but may not have
formed entirely (or at all) in response to wave driven
processes. Cohesive soil banks may lie at the head of
the beach, such that the beach represents an adjustment
of the bank which has been exposed to wind wave, tidal
and flood influences. These banks can withstand some
wave action, but they often do not have the support of
riparian vegetation. The upper bank structure can be
severely weakened if the riparian vegetation is removed
due to anthropogenic intervention such as land
development (such as the Brisbane and Parramatta
Rivers), cattle grazing (such as along the Patterson
River in NSW, and many others) and tidal influx
(Brisbane River).
The third possible bank type is what we would regard
as a sandy beach, with fine to coarse grained sand
(sandy muds or clean sand) that extend well above and
below the mean waterline. These beaches are normally
found in open areas in bays and at estuary mouths
where there is a substantial wind wave climate and/or
strong tidal flows. It is this third bank type which may
be surplus to the study requirements, as they are
somewhat dynamic by nature and are already
reasonably understood by current coastal engineering
science. Energetics are typically such that true beaches
are not susceptible to wave wake from small craft
although some may be affected by larger, high-speed
ferries.
For small craft operating in sheltered waterways, only
the first two bank types are considered to be of prime
importance.
13.

EVENTUAL APPLICATIONS

Any pragmatic observation would suggest that

widespread application of the double criteria of wave
wake energy and period limits will never happen. At
best it would be limited to specific instances, such as

2008 Royal Institution of Naval Architects

locations with sensitive banks and commercial vessel

operations.
If it were to be implemented, the simplest way to make
it work would be as follows:
Designate boating areas according to their threshold
wave wake limits. This could be as simple as
dividing waterways into zones, such as A,B,C,D,
each with particular wave wake threshold values
that suit the perceived erosion potential.
For each licenced vessel, generate a table of
relevant operating speeds for each zone using vessel
waterline length and displacement. The table could
be generated using any validated prediction tool,
such as the AMCs wave wake database, which can
generate maximum wave height and corresponding
period given simple input data (L, displacement, y
and speed).
Each vessel owner would simply have to operate
their vessel according to the zone they were in,
using the tabulated data.
As an example, Table 2 shows a typical speed limit
table that could be fixed to the dashboard of the
hypothetical vessel Minnow.
If builders of small craft were forced to register their
designs, where various design parameters are verified
and recorded, the information would be readily
available. The displacement in particular would have to
be at some specific loading, which might be the
registered capacity of the vessel (according to existing
registration requirements). Provided this information
was accurate, the means would then exist for regulatory
bodies to provide vessel owners with simplified
operating guidelines.
14.

CONCLUSIONS

Although limited, there has been sufficient work

undertaken to demonstrate the erosive components of
deep water wave wake from small craft. Small
recreational craft wave wake is easier to analyse as the
accompanying wave periods are low and the wave
wake behaves as deep water wave wake in all but the
shallowest of water. Threshold erosion values of a
number of wave wake parameters have been derived,
along with their relative ranking in terms of erosion
influence.
The salient conclusions reached are:
For the deep water condition, there is a clear
relationship between vessel waterline length and
the period of the maximum wave.
Wave height alone is a poor indicator of erosion
potential
Derived wave wake parameters such as energy,
power and energy per unit wave height (which is

2008 Royal Institution of Naval Architects

applicable to short period, small craft wave wake),

are better measures of erosion potential
A single operating criterion may either not
encompass all erosive wave wake components or
be overly restrictive for at least some vessel types.
Multiple operating criteria such as a combination
of wave energy and period limits appear to offer
the best solution.
Simplified operating criteria have been derived for
several rivers and, with existing knowledge, can be
applied to all small recreational vessels.

We have dispelled some of the myths regarding wave

wake and erosion and outlined a method by which
erosive effects might be reasonably controlled.
REFERENCES
1.

2.

3.

4.

5.
6.

7.

8.

9.

Bradbury, J., Lower Gordon River erosion

history
and
lessons
learned,
http://www.dpiw.tas.gov.au/inter.nsf/WebPages/LJ
EM-6J289T?open. Accessed 8 January 2009.
Bradbury, J., Lower Gordon River turbidity
monitoring, April 2003 - December 2004,
http://www.dpiw.tas.gov.au/inter.nsf/WebPages/LJ
EM-6J289T?open. Accessed 8 January 2009.
Bradbury, J., Cullen, P., Dixon, G. and Pemberton,
M., Monitoring and management of streambank
erosion and natural revegetation of the lower
Gordon River, Tasmanian Wilderness World
Heritage
Area,
Australia,
Environmental
Management, Vol. 19, No. 2, pp. 259-272, 2005
Cox, G., Sex, lies and wave wake, Proceedings of
the RINA International Conference on the
Hydrodynamics of High Speed Craft Wake Wash
and Motions Control, London, United Kingdom,
2000
Kogoy, P., Tidal Wave of Protest, Sun-Herald
(Sydney), 20th of September, 1998
Lesleighter, E. J., Hawksbury River - The effect of
speedboat activities on bank erosion, NSW
Department of Public Works, Report No. 106,
1964
Macfarlane, G.J., The measurement and
assessment of sub-critical vessel generated waves,
Master of Philosophy Thesis, Australian Maritime
College, 2002.
Macfarlane, G. J., Correlation of Prototype and
Model Wave Wake Characteristics at Low Froude
Numbers, RINA Transactions, Part A2,
International Journal of Maritime Engineering,
2006
Macfarlane, G.J. and Cox, G., The development of
vessel wave wake criteria for the Noosa and
Brisbane Rivers in Southeast Queensland, Fifth
International Conference on Coastal Environment
2004 Environmental Problems in Coastal
Regions, Alicante, Spain, 2004.

10. Macfarlane, G. J. and Cox, G., An introduction to

the development of rational criteria for assessing
vessel wash within sheltered waterways, IMarEST
Journal of Marine Design and Operations, Part BII,
2007.
11. Macfarlane, G.J. and Renilson, M.R., Wave wake
A rational method for assessment. Proceedings of
the International Conference on Coastal Ships and
Inland Waterways, Royal Institution of Naval
Architects, London, United Kingdom, 1999.
12. Macfarlane, G.J., and Renilson, M.R., When is low
wash low wash An investigation using a wave
wake database, Proceedings of the International
Conference on Hydrodynamics of High-Speed
Craft Wake Wave wake and Motions Control,
Royal Institution of Naval Architects, London,
United Kingdom, 2000.
13. Nanson, G.C., von Krusenstierna, A., Bryant, E.A.
and Renilson, M.R., Experimental measurements
of river bank erosion caused by boat-generated
waves on the Gordon River, Tasmania, Regulated
Rivers, Research and Management Vol. 9, 1-14,
1994.
14. Pattiaratchi, C. and Hegge, B., Impact of ferry and
large vessel traffic on Swan River foreshore

Fetch (m)
100
500
1000
10000

15.

16.

17.

18.

5m/s

erosion, Centre for Water Research Report WP 452

CP, University of Western Australia, 1990.
U.S. Army Corps of Engineers, Coastal
Engineering Manual, Volumes I, II, III, IV,
http://140.194.76.129/publications/eng-manuals/.
Accessed 9 January 2009.
Sorensen, R.M., Water waves produced by ships,
The Journal of the Waterways Harbours and
Coastal Engineering Division, Proceedings of the
American Society of Civil Engineers, Vo. 99. No.
WW2, pp.245-256, May 1973.
von Krusenstierna, A., River bank erosion by boatgenerated waves on the lower Gordon River,
Tasmania, Master of Science Thesis, Department
of Geography, University of Wollongong, 1990.
Warren, N., Wash from fast ferries, Ship and Boat
International, The Royal Institution of Naval
Architects, pp. 11-13, April 1991.

10m/s

20m/s

26mm / 0.5s
59 / 0.8

62 / 0.7
137 / 1.1

144 / 0.9
321 / 1.5

83

/ 1.0

194 / 1.4

452 / 1.9

586

250

2.0

2.8

1304

3.8

Table 1 Hindcast Wind Waves (wave height in mm / wave period in seconds)

Vessel Name
Registration No.
ZONE
A
B
C
D

: Minnow
: ABC123
Slow Speed
Up to 4 knots
Up to 5 knots
Up to 6 knots
Open Speed

High Speed
Do not operate
Over 18 knots
Over 14 knots
Open Speed

Speeds to Avoid
Over 4 knots
5 to 18 knots
6 to 14 knots
Nil

Table 2 Example table of speed restrictions

2008 Royal Institution of Naval Architects

800

AMC Wave Wake Database

Vessel Speed = 13 knots
Lateral Distance = 50 metres

700

Maximum Wave Height (mm)

600

Monohulls
Multihulls

500

400

300

200

100

0
4

Slenderness Ratio ( LWL / Volume

1/3

10

Figure 1 Data from AMC wave wake database

1800

AMC Wave Wake Database

Waterline Length = 45 metres
Vessel Speed = 32 knots
Lateral Distance = 200 metres

1600

Maximum Wave Height (mm)

1400

Monohulls
1200

Multihulls

1000

800

600

400

200

0
0

100

200

300

400

500

600

Vessel Displacement (tonnes)

Figure 2 Data from AMC wave wake database

2008 Royal Institution of Naval Architects

700

800

900

1000

1.2

Wave Period / L1/2

1.0

Full Scale Trials Data

Lateral Distance, y = 23 metres

0.8

0.6

Houseboat, LWL = 11.5m, Disp.= 15,500kg

0.4

Horizon V-Nose Punt, LWL = 3.9m, Disp.= 800kg

Beam Trawler, LWL = 7.3m, Disp.= 6,100kg
Water Bus, LWL = 8.2m, Disp.= 3,900kg
Sea Jay 5.2m Centre Console, LWL = 4.5m, Disp.= 1000kg
Pacific 7.7m Centre Console, LWL = 6.7m, Disp.= 2,500kg

0.2

Ski Boat (Inboard), LWL = 5.3m, Disp.= 1,450kg

Ski Boat (Outboard), LWL = 4.6m, Disp.= 1,030kg
12' Aluminium Dinghy, LWL = 3.3m, Disp.= 280kg
16' Aluminium Dinghy, LWL = 4.4m, Disp.= 510kg

0.0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

Length Froude Number

Figure 3 Data from full scale field tests in Queensland

Transition Speeds
Slow Speeds

High Speeds

Maximum Wave Height

Period of the Maximum Wave
Energy of the Maximum Wave
Power of the Maximum Wave

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Length Froude Number

Figure 4 Maximum wave energy, power and height against increasing vessel speed

2008 Royal Institution of Naval Architects

250
Full Scale Bank Erosion Data
200

Water Depth = 1.0 metres

Elevated Turbidity (NTU)

Water Depth = 0.25 metres

150

100

50

0
10

100

1000

Maximum Wave Height (mm)

Figure 5 Turbidity against maximum wave height

250
Full Scale Bank Erosion Data
200

Water Depth = 1.0 metres

Elevated Turbidity (NTU)

Water Depth = 0.25 metres

150

100

50

0
0.1

1
Period of the Maximum Wave (s)

Figure 6 Turbidity against period of maximum wave

2008 Royal Institution of Naval Architects

10

250
Full Scale Bank Erosion Data
Water Depth = 1.0 metres

Elevated Turbidity (NTU)

200

Water Depth = 0.25 metres

150

100

50

0
1

10

100

1000

Energy of the Maximum Wave (J/m)

Figure 7 Turbidity against Energy of maximum wave

250
Full Scale Bank Erosion Data
Water Depth = 1.0 metres

200

Elevated Turbidity (NTU)

Water Depth = 0.25 metres

150

100

50

0
1

10

100

1000

Power of the Maximum Wave (W/m)

Figure 8 Turbidity against power of maximum wave

2008 Royal Institution of Naval Architects

250
Full Scale Bank Erosion Data
200

Water Depth = 1.0 metres

Elevated Turbidity (NTU)

Water Depth = 0.25 metres

150

100

50

0
0.01

0.1

1
HmT2 (ms2)

Figure 9 Turbidity against energy per unit of wave height

2008 Royal Institution of Naval Architects

10

Gordon River Erosion Data

6
5

62.3

70.7

29.3

High erosion

35.1

39.7

34.1

14.9

Wilderness Seeker
Gordon Explorer

16.4

101.0

14.1

9
8
7
6

7.6

13.6
15.6
14.8

Moderate
erosion

Brisbane Limit

9.2

8.2

11.4

7.5

HmT

8.7

3.2
2

0.0

10

Values for each data point

correspond to erosion in mm

1.6

6.8
3.7

Noosa Limit

8
9
7
8
6
7
6
5

0.5

8.3

5.5

2.3

1.0

2.5

Low erosion

4
3

0.0

10-1.0

T1
Figure 10 Gordon River erosion data

2008 Royal Institution of Naval Architects