Content uploaded by Mark Patrick Taylor
Author content
All content in this area was uploaded by Mark Patrick Taylor on Aug 17, 2017
Content may be subject to copyright.
(2005) 22 EPLJ 193 193 ©
Up the creek: What is wrong with the
definition of a river in New South Wales?*
Mark Taylor** and Robert Stokes***
The legal definition of a river or stream has been largely determined from
a European perspective of what a river “should look like”. In New South
Wales, the definition of a river or watercourse is codified in a range of
legislative instruments. In most cases, the Rivers and Foreshores
Improvement Act 1948 (NSW) and the partially implemented Water
Management Act 2000 (NSW) provide the basis for legal decision-
making. The Rivers and Foreshores Improvement Act 1948 implies that a
bona fide river should have perennial and intermittent flow. However, it is
well known that the Australian landscape is characterised by river
systems that occupy the full range of discharge patterns from perennial to
intermittent through to ephemeral. Thus, the Rivers and Foreshores
Improvement Act 1948 defines a river in manner that is quite
uncharacteristic for many smaller watercourses in New South Wales, as
these are frequently ephemeral. Consequently, the difference between
the legal definition and the true physical form of a “watercourse” has
created numerous disputes that have had to be settled by the courts.
This article discusses the legal definition of river, and a range of
geomorphological descriptions and perceptions of what a river is. To
highlight the differences between legal and geomorphological definitions,
we discuss a recent legal dispute that arose with respect to the definition
of a watercourse at Lumley Park Farm, Bungonia, New South Wales
along with the Farm Dams Policy 1999 (NSW), which was relevant to the
case.
INTRODUCTION
The law is not so much carved in stone as it is written in water, flowing in and out with
the tide.1
Many environmental scientists and geographers will have an instinctive understanding of the
meaning of a “river” and what it represents in the landscape. However, for lawyers and land use
planners, the definition of a river is frequently a challenging issue. When considered in the light of
the full range of physical, social and economic issues that impinge on river environments, it is
apparent that the meaning of a river is as much a construct of individual perceptions as being based
upon a specific and definable physical reality. Accordingly, “river” can mean different things to
* This article is the result of several consultancies regarding the determination of a river and the authors would like to thank
the many colleagues from Macquarie University, the DIPNR, local government, engineering and planning organisations
with whom they have held discussions. Jennifer Manefield is thanked for her assistance the field and for suggestions
regarding the meaning of a river to the wider community. We are also grateful to Bill Dobbie for giving permission to use
the Lumley Creek case in this paper and for the many discussions we have had regarding his land. Peter Hoare is also
thanked for his fastidious proofreading of an earlier version of this article – his constructive comments have improved this
article. We are also grateful for the helpful comments provided by James Smith of PWCLegal.
** Senior Lecturer, Department of Physical Geography, Macquarie University, Sydney, mark.taylor@mq.edu.au
*** Lecturer, Department of Business Law, Macquarie University, Sydney, robert.stokes@law.mq.edu.au
1 Melvoin J, Northern Exposure, Crime and Punishment (1992) http://www.quotationspage.com/quotes/Jeff_Melvoin/
viewed 17 May 2005.
Taylor and Stokes
© 194 (2005) 22 EPLJ 193
different people. Such differences can be explained by an individual’s cultural beliefs and
childhood domicile, which may have had extremely diverse climatic characteristics and different
types of river. For example, a native of the Namib Desert may consider all the rivers which drain
into the Atlantic Ocean off the west coast of Africa to be bona fide rivers, even though they flow
only two or three times a year and rarely reach the sea. In contrast, a person from northern Europe
may think of rivers as continuously flowing bodies of water that may also be navigable for the
purposes of recreation or transport. In Australia, landscapes are characterised by contrasting
climate regimes: from the seasonally wet tropics of the north; to the Mediterranean climates of
South and Western Australia; to the deserts in the centre and the temperate climes in southeast
Australia and Tasmania. Consequently, people from different localities may have vastly different
perspectives of what a river is. For example, inhabitants of Alice Springs will think of the Todd
River as a watercourse that flows very infrequently while inhabitants of Launceston will never
have seen the North or South Esk rivers without a copious flow of water.
The gulf between perception and reality is further complicated when the definition of river is
subject to legal interpretation. In New South Wales, the definition of a river has caused numerous
conflicts between landowners and local councils, New South Wales Fisheries and the Department
of Land and Water Conservation, (now absorbed into the Department of Infrastructure, Planning
and Natural Resources (DIPNR)). Expert witnesses such as fluvial geomorphologists,
environmental scientists and engineers are often called upon to determine whether a particular
section of a drainage network is a river within the definition provided in the Rivers and Foreshores
Improvement Act 1948 (NSW) (RFIA). While the question of whether a river is a river may seem
to be of peripheral interest to most environmental scientists, the same cannot be said for
landowners, developers, government agencies, farmers and river managers who have to resolve
such issues using the services of geomorphologists and lawyers.
Thus, the precise definition of a watercourse can have significant impacts not only for
planning purposes but also for the ecological functioning associated with rivers and their riparian
vegetation. This is because even in a badly degraded catchment, riparian buffers offer a number of
geo-ecological functions including enhanced stream bank stability, refugia for flora and fauna
communities, maintenance of seed banks for downstream areas, shading and nutrient contributions
(leaf litter).2 Riparian systems also help remove agricultural and domestic pollutants, provide inter-
reach connectivity and linkages for terrestrial wildlife movement and provide environmental
aesthetics for the catchment and region.3
In urban environments, riparian corridors are often the only remaining sanctuary for flora and
fauna. Consequently, environmental specialists will often argue for the protection and
rehabilitation of riparian corridors in order to create optimal conditions for species survival.4 Thus,
not only from a legal perspective but also from a geo-ecological viewpoint, it is advantageous to
develop an unambiguous definition of what a river is. This article considers a range of definitions
used to describe a river along with the associated land management issues that arise from these
definitions. The RFIA, which provides a current legal definition of a river or stream, is examined
critically with respect to geomorphological interpretations.
LEGISLATIVE DEFINITIONS OF “RIVER”
The RFIA was initially enacted in 1948 to facilitate the removal of obstructions from rivers and
foreshores and to prevent erosion caused by tidal and non-tidal water. These developmental
2 Hughes FMR, and Rood SB, “Floodplains” in Warren A and French JR (eds), Habitat Conservation: Managing the
Physical Environment (John Wiley and Sons Ltd, 2001) pp 103-121.
3 Paul MJ and Meyer JL, “Streams in the Urban Landscape” (2001) Annual Review of Ecology and Systematics 32, at 333–
365.
4 See Lake PS, “Restoring Streams: Re-building and Reconnecting” in Rutherfurd I, Sheldon F, Brierley G, and Kenyon C
(eds), Third Australian Stream Management Conference (Brisbane, 2001) pp 369-371, and Paul and Meyer, n 3.
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 195 ©
objectives were supplemented by conservation objectives with the passage of the Rivers and
Foreshores (Amendment) Act 1991 (NSW). This amendment incorporated Pt 3A into the Act,
which requires a person to obtain a permit (commonly called a Pt 3A permit) for any activity that
obstructs or detrimentally affects the flow of a river, or is likely to do so.5 While the legislation
was intended to apply primarily in regional areas, the encroachment of cities into bushland, and the
insertion of Pt 3A, means that it is now frequently used in urban situations where developers are
required to apply for a permit to undertake any works within 40 m of a “river”.6 Applying for such
approvals can involve the imposition of new conditions on development, as well as slowing down
the development process, so that developers are keen to avoid the approvals process wherever
possible by arguing that drainage depressions and channels do not fit within the legislative
definition of a river.7
This inclusive definition is contained in s 2 of the RFIA, which provides that a river includes:
any stream of water, whether perennial or intermittent, flowing in a natural channel, or in a natural
channel artificially improved, or in an artificial channel which has changed the course of the stream
of water and any affluent, confluent, branch, or other stream into or from which the river flows and,
in the case of a river running to the sea or into any coastal bay or inlet or into a coastal lake,
includes the estuary of such river and any arm or branch of same and any part of the river
influenced by tidal waters.
This definition reflects a Eurocentric view of rivers by assuming that they will be characterised by
a perennial or intermittent flow. The definition was expanded by the 1991 amending Act to include
coastal lakes and lagoons on the basis that these bodies may significantly affect aquatic habitats.8
The reference to tidal waters is further defined as including:
the waters of the sea or of any lake, estuary, harbour, river, bay, lagoon or inlet of the sea in which
the tide ebbs and flows.
In order for any works to be carried out on rivers or tidal waters, permission has to be sought
from the Constructing Authority, which is either the Minister for Public Works or the Ministerial
Corporation.9 The types of work included in the RFIA are listed as follows:
(a) the removal of dead or growing timber, or other vegetation or aquatic plants, or of silt, shingle,
soil, sand, gravel, stone, rock or other matter or thing whatsoever, from the bed, banks or
foreshore of any tidal waters or coastal lake or lagoon, or from the water or the bed or banks of a
river or from any adjoining, adjacent or nearby lands;
(b) changing or preventing the changing of the course of a river;
(c) preventing the erosion of the bed or banks of a river or of adjoining, adjacent or nearby lands by
the waters of a river;
(d) preventing the siltation of the course of a river where such work is confined to the bed or banks of
a river and adjoining, adjacent or nearby lands;
(e) preventing the flooding of land by the waters of a river;
(f) deepening, widening, straightening, or improving the course of a river;
(g) preventing the inflow of sea water or saline water into the course of a river; or
(h) preventing the erosion of lands by tidal waters or by the waters of any coastal lake or lagoon.10
5 On 8 December 2004, the Minister issued a notice pursuant to s 22 of the RFIA that removed the requirement for a Pt 3A
permit in respect of certain types of development (ie dwelling houses and dual occupancy housing) within or adjacent to
Sydney Harbour and its tributaries.
6 Rivers and Foreshores Improvement Act 1948 (NSW), s 22B.
7 See, for example, Water Administration Ministerial Corporation v Auburn Council [2000] NSWLEC 76 (18 April 2000).
8 New South Wales Legislative Assembly, Debates (11 December 1991) p 6460 (Richard Amery).
9 Rivers and Foreshores Improvement Act 1948 (NSW), s 2. The Ministerial Corporation was constituted under the Water
Administration Act 1986 (NSW), which was repealed by the Water Management Act 2000 (NSW).
Taylor and Stokes
© 196 (2005) 22 EPLJ 193
The RFIA provides that any work that is carried out on protected land, ie
(a) land that is the bank, shore or bed of protected waters; or
(b) land that is not more than 40 m from the top of the bank or shore of protected waters (measured
horizontally from the top of the bank or shore); or
(c) material at any time deposited, naturally or otherwise and whether or not in layers, on or under
land referred to in paragraph (a) or (b).11
requires ministerial approval before works can take place. The meaning of “protected land” is
crucial because it affects the land adjacent to protected waters (eg a river) which are defined as:
a river, lake into or from which a river flows, coastal lake or lagoon (including any permanent or
temporary channel between a coastal lake or lagoon and the sea).12
The above definitions indicate the type of works and the type of land that might come under
the RFIA. The important elements here relate to the definition of protected waters and land which
includes rivers and land lying within 40 m of the bank top. Thus, if development is to be conducted
either on or within 40 m of a river, a permit is required under Pt 3A of the RFIA. Such a permit is
frequently refused because rivers are protected land and therefore so is the habitat immediately
adjacent. Herein lies the issue for developers and landowners.
If it can be demonstrated that the drainage line is not a river, then a Pt 3A permit is not
required, thus removing the obstacle for development, dam building or other engineering works. If,
on the other hand, the consent authority is satisfied that the watercourse is a river, then works
cannot take place within 40 m of the top of the bank without a Pt 3A permit. The issue of what
precisely is “top of the bank” is not discussed here, but is reviewed in Knighton.13 It is perhaps
sufficient to say that it is a complicated matter, particularly when channels have multiple inset
benches14 or poorly-defined banks.
In addition to the RFIA, Sch 3, Pt 4(38) of the Environmental Planning and Assessment
Regulation 2000 (NSW) provides the following legal description of a river or stream which is an
abridgement from the RFIA, within a wider definition of “waterbody”:
“waterbody” means:
(a) a natural waterbody, including:
…
(ii) a river or stream, whether perennial or intermittent, flowing in a natural channel with an
established bed or in a natural channel artificially modifying the course of the stream.
This definition is applied to the classification of certain types of development near rivers as
“designated development” for which additional processes must be undertaken before consent may
be issued. The Crown Lands Act 1989 (NSW), also incorporates the following definition of river,
that clearly echoes the definition contained in the RFIA:
“river” includes any stream of water, whether perennial or intermittent, flowing in a natural
channel, and any affluent, confluent, branch or other stream into or from which the river flows.
It is important to note, however, that this definition is solely applied to the title boundary of land
alienated by the Crown under the Crown Lands Act 1989.15
10 Rivers and Foreshores Improvement Act 1948 (NSW), s 2.
11 Rivers and Foreshores Improvement Act 1948 (NSW), s 22A.
12 Rivers and Foreshores Improvement Act 1948 (NSW), s 22A.
13 Knighton AD, Fluvial Forms and Processes (Arnold, 1998) p 383.
14 Erskine WD and Livingstone EA, “In-channel Benches: The Role of Floods in Their Formation and Destruction on
Bedrock Confined Rivers” in Miller AJ and Gupta A (eds), Varieties of Alluvial Form (John Wiley and Sons Ltd, 1999) pp
445-475.
15 Crown Lands Act 1989 (NSW), s 172(1).
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 197 ©
As part of the development of a national approach to conserving and managing Australia’s
water resources, the New South Wales Government has introduced the Water Management Act
2000 (NSW)(WMA).16 This legislation is being implemented in stages and will ultimately replace
the RFIA, which will be totally repealed upon the commencement of Sch 7 of the WMA. The
WMA provides a definition of a river that is very similar to that enshrined within the RFIA.
According to this definition, a “river” includes:
a) any watercourse, whether perennial or intermittent and whether comprising a natural channel or a
natural channel artificially improved; and
b) any tributary, branch or other watercourse into or from which a watercourse referred to in
paragraph (a) flows; and
c) anything declared by the regulations to be a river;
whether or not it also forms part of a lake or estuary, but does not include anything declared by the
regulations not to be a river.17
It is clear that the legislative definition of “river” in New South Wales is very consistent
between different instruments, being based on the definition contained in the RFIA. However, it is
worth noting that these similar definitions are applied to very different purposes. The long title of
the RFIA indicates that it is aimed to facilitate the “improvement” of rivers and foreshores, by the
“removal of obstructions.”18 The Environmental Planning and Assessment Act 1979 (NSW) and
the WMA, on the other hand, are directed to “ecologically sustainable development”,19 including,
in the case of the WMA, the protection and conservation of the water sources of the State.20 It may
be appropriate, therefore, to consider the development of a new definition of river based upon
environmental and climatic realities rather than adopting a definition that reflects the
anthropocentric and Eurocentric focus of older legislation.
JUDICIAL DEFINITIONS OF “RIVER”
In the same way as the legislative definition of “river” can be comprised in a definition of
“waterbody” or “watercourse”, the common law definition is similarly entwined. As Pearlman J
noted in Zouki v Water Administration Ministerial Corporation (2001) 118 LGERA 229 at 233,
the definition of “river” in s 2 of the RFIA is inclusive, so that it includes both “the definition there
set out, as well as any other definition derived by reference to the ordinary meaning of the word”.
Indeed, in many earlier cases the court was not assisted by any legislative definition of “river”.21
For these reasons, all of these terms (river, watercourse and waterbody) have been read
together in the case law, and only distinguished where required.22 In Azzopardi v Gosford City
Council (2002) 123 LGERA 118 at 122 it was argued that the trial judge had misdirected himself
by considering whether a watercourse had “defined banks and bed”, on the basis that this did not
16 Hamstead M and Gill J, “Implementing the Water Management Act 2000” (Water Management Division,
Department of Infrastructure, Planning and Natural Resources, September 2004) p 1,
http://www.dipnr.nsw.gov.au/water/legal/pdf/law_soc_conf.pdf viewed 20 December 2004.
17 Water Management Act 2000 (NSW).
18 Pearlman CJ in Zouki v Water Administration Ministerial Corporation (2001) 118 LGERA 229 at 233 found that the
purposes of the RFIA included the “protection of rivers an d lakes”. However, this purpose is not directly included in the
legislation. “Ecologically sustainable development” is not an objective of the Act.
19 Environmental Planning and Assessment Act 1979 (NSW), s 5(a)(vii); Water Management Act 2000 (NSW), s 3(a).
20 Water Management Act 2000 (NSW).
21 See Windeyer J in Gartner v Kidman (1962) 108 CLR 12 at 23.
22 “River” and “watercourse” were distinguished in Warringah Council v Ardel Ltd (2000) 106 LGERA 340, where
Talbot J determined that a channel may be define d as a “watercourse”, yet not meet the legislative definition of a “river”
provided in the RFIA. Relevantly, he said (at 346): “The Rivers Act does not seek to control the sites of all watercourses,
only those which answer the ordinary meaning of a river or where there is a stream of water flowing in a natural channel
that may not immediately be recognised as a river in the accepted sense”.
Taylor and Stokes
© 198 (2005) 22 EPLJ 193
form part of the definition in the RFIA. Handley JA dismissed this argument, finding that the
statutory definition included these elements by implication.
As Windeyer J noted in Gartner v Kidman (1962) 108 CLR 12, the law relating to
watercourses as developed in England through “innumerable cases” has been received into
Australian law. Under this common law, a watercourse was defined to consist of:
bed, banks and water: yet the water need not flow continuously and there are many watercourses
which are sometimes dry. There is, however, a distinction to be taken between a regular flowing
stream of water which at certain seasons is dried up and those occasional bursts of water which at
times of freshet or melting of ice and snow descend from the hills and inundate the country.23
This Eurocentric definition shaped judicial reasoning throughout the nineteenth century. For
example, in Lyons v Winter (1899) 25 VLR 464, Hood J (at 465) held that a watercourse must
comprise “a course, marked on earth by visible signs, along which water usually flows.” Similarly,
in Cooper v The Corporation of Sydney (1853) 1 Legge 765, the New South Wales Supreme Court
found that the occasional overflow of a swamp did not constitute a watercourse in the legal sense,
which consisted of a regular stream between banks.
This approach was generally adopted by the court in Knezovic v Shire of Swan-Guildford
(1968) 118 CLR 468. According to the judgment of Barwick CJ (at 475-476):
a watercourse consists of a stream with a bed, with banks, and water. That the flow of the water in
the stream is intermittent or seasonal will not prevent what would otherwise be a watercourse from
being accounted such: but though it is quite true that a watercourse may exist though its bed be dry
for some periods, the watercourse, in my opinion, must exhibit features of continuity, permanence
and unity, best seen, of course, in the existence of a defined bed and banks with flowing water. It
must, in my opinion, essentially be a stream and be sharply distinguished from a mere drain, or a
drainage depression in the contours of the land which serves to relieve upper land of excess water in
times of major precipitation. It is not enough that the water, when it does flow, does so in what may
be seen as a defined course or channel. In the case of a drainage depression, the water being drained
off can be expected to flow in the lowest portion of the contours confined by the rising levels of the
adjacent land: thus water can be seen when flowing to do so in what could be called a defined
channel.
In reference to the “bed and banks” required under this definition, Lee J in Latta v Klinberg
(unreported, Sup Ct NSW, Lee J, 1 July 1977) (at 21) found that a watercourse must also have
“identifiable margins”. These margins must be more than “observable margins which are really no
more than the natural surface of the ground rising gently on either side”, which do not “constitute
banks and bed”.
The Knezovic definition of “bed, banks and water” and “continuity, permanence and unity”
has been applied in more recent decisions of the New South Wales Land and Environment Court.
In Jugiong Quarries Pty Limited v Water Administration Ministerial Corporation [1995]
NSWLEC 74 (11 May 1995) Talbot J noted (at 22) the inclusive nature of the definition in s 2 of
the RFIA. However, he asserted that the “normal” understanding of the word river:
is the flow or stream of water evident for the most part together with the area of land contained
within a defined bank or shore associated directly with that body of water. The area of land would
encompass the bed of the river. Land beyond the bed is not generally referred to as part of the river
even though, in some cases it might be recognised as a related flood plain inundated in times of
extreme high flow.
23 Angell JK, A Treatise on the Common Law, In Relation to Watercourses: Intended More Particularly as an Illustration
of the Rights and Duties of the Owners and Occupants of Water Privileges (5th ed, Hilliard, Gray and Co, 1854) p 3.
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 199 ©
He went on to find that a floodrunner described as a depression on the inside of a meander
near the Murrumbidgee River could not be considered to be a “river” within the definition
provided by the RFIA.
In Narrambulla Action Group Inc v Mulwaree Council No. 40168 of 1995 [1996] NSWLEC
199 (30 July 1996), Bannon J also applied the definition provided by Barwick CJ in Knezovic. In
that case he determined that the watercourse under consideration could, at best, “be classified as a
drainage line with gullies to the East and West, together with intermittent ponds and flood plane
[sic], where water flows are rare intervals, under the influence of rain”. This was not enough to
meet the test of “continuity, permanence and unity” developed by Barwick CJ in Knezovic.
In Mitchell v Vella (1998) 101 LGERA 333, Sheahan J determined that a series of channels
which flowed only in times of heavy rain were not a “waterbody” within the definition of Pt 3 of
Sch 3 of the Environmental Planning and Assessment Regulation 1994 (NSW). These channels
were described (at 347-348) in the following terms:
[they] do not have a sandy or gravelly bed, or even a muddy one, and they do not have steep banks
with different vegetation on them. Yet they have a bed, which is the bottom of the depression, and
they have banks, which are the sides of the depression.
Adopting the same approach as the court in Narrambulla, Sheahan J determined that these features
were not enough to meet the threshold requirements identified by Barwick CJ in Knezovic.
In Don Burke v Hawkesbury City Council [2001] NSWLEC 222, Sheahan J followed the
reasoning developed in previous cases to determine that the waterbody under consideration did not
exhibit the features of continuity, permanence and unity, and did not display identifiable margins.
He also found that the waterbody held water only “empherally, if at all”, so that it did not contain
the permanent or intermittent flow required by the definitions in the RFIA and the Environmental
Planning and Assessment Regulation 2000 (NSW).24
These cases indicate that the Eurocentric definitions received into Australian law continue to
inform and influence judicial decision-making. However, there has been some judicial recognition
of the changed circumstances to which these definitions are applied. Windeyer J in Gartner noted
that
the conditions of settlement, of climate and of geography in which this body of customary law
developed are very different from those prevailing in many parts of Australia. And this is to be
borne in mind when particular decisions of English courts are brought forward as analogies.
However, despite recognising these climatic differences, he found that the English definitions
of watercourses formed part of Australian law and that, “it is beyond doubt that these rules are a
part, and an important part, of the common law that Australia has inherited”.
The judgment of Bannon J in Narrambulla Action Group Inc v Mulwaree Council [1996]
NSWLEC 1999 also indicates an awareness of the dichotomy between common law definitions
and the environmental realities of Australian watercourses. He noted the definition of “river”
provided by one expert witness in the case which referred to the “arid conditions of Australia”,
where rivers are best understood as being in a “process of evolution” whereby “a continuous
channel or bank may not be apparent”, and flow may be extremely intermittent, “but this does not
negate the fact that the creeks and tributaries are flowing in a natural channel with an established
bed”.
Bannon J considered that such views “deserve respect”, but that such a “scientific” definition
did “not fit the definitions given by the Courts, taken as they are, from European conditions”.
Thus, in that case, he determined that the watercourse under consideration could, at best, “be
24 Don Burke v Hawkesbury City Council [2001] NSWLEC 222 at [53]-[79].
Taylor and Stokes
© 200 (2005) 22 EPLJ 193
classified as a drainage line with gullies to the east and west, together with intermittent ponds and
flood plane [sic], where water flows are rare intervals, under the influence of rain”. Ultimately, he
concluded that the watercourse did not “measure up” to the definition provided by Barwick CJ in
Knezovic.
Definitions of rivers and watercourses have traditionally been framed in largely economic
terms. For instance, Farnham, writing at the turn of the 20th century, wrote that a “water course is
a stream of water of such well-defined existence as to make its flow valuable to the owners of land
along its course”.25 Judicial definitions of “rivers” have been developed in the context of
determining the existence of riparian rights. Thus, as Madden CJ observed in Vinnicombe v
MacGregor (1902) 28 VLR 144, that case was unusual since the question was not “to whom do the
blessings of natural waters belong, but who must bear its curses”.
However, as rivers have become the subject of environmental protection as much as economic
exploitation, this contrary perspective has become more common. Consequently, as identified in
Warringah Council v Ardel Limited (2000) 106 LGERA 340 at 347, the application of the same
definition of “river” to very different situations, such as environmental conservation on one hand,
and the protection of riparian rights on the other, can have profound consequences. Yet, despite
this acknowledgment, there has been little change to the definition of “river” in response to this
shift in perception.
There is also a level of judicial disagreement over the defining characteristics of a river. In
Azzopardi v Gosford City Council (2002) 123 LGERA 118, Davies AJA (at 24-26) disagreed with
the leading judgment of Handley AJA to determine that the watercourse in question was a river
within the definition of the RFIA. Handley AJA (with whom Foster AJA agreed) found that, at
best, the watercourse could be described “as a drainage depression or a swampy creek”, that did
not meet the legislative definition of “river”.26 Davies AJA, on the other hand, determined the
existence of a “clear watercourse and a stream”, relying on the following uncontested evidence:
• it was marked on the relevant contour map;
• its course was identified by a tree line, consistent with the existence of a watercourse;
• the existence of a spring, which was likely to be the commencement of the stream;
• a description of the watercourse as a “swampy creek” with running water; and
• photographs showing water running in the watercourse.
While he conceded that the watercourse was perennial and did not carry any running water in
very dry conditions, he argued that the definition of “river” in the RFIA was expansive and should
not be read narrowly. In particular, he noted that the legislation was aimed to ensure that
excavations (such as the ones subject of the dispute at hand) did not detrimentally affect the flow
of protected waters. He argued that excavations taking place close to a stream forming part of a
river (such as that under consideration) would be likely to have such a detrimental effect.
Redefining the meaning of a river or stream
The definition of a river or stream can be viewed from a variety of perspectives: legal, personal,
technical, cultural, political and economic. As Lloyd J observed in Maule v Liporoni (2002) 122
LGERA 140 at 171, the fact that a “river” is shown on a topographical map does not prove the
existence of a river at law, as the criteria for determining the presence of a river used by one
government department may not be the same as that contained in the RFIA. On the other hand,
moving away from a purely technical and physical construct of rivers, Howitt argues that we often
25 Farnham HD, The Law of Waters and Water Rights (Lawyers Cooperative Publishing Company, 1904)Vol III, p 2556.
26 Azzopardi v Gosford City Council (2002) 123 LGERA 118 at 121.
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 201 ©
detach the environment from societal concerns. He suggests that the environment, (including, inter
alia, a river system) should not be defined in terms of a particular discipline or to suit political or
legal purposes, but should be defined more holistically to embrace the character of the affected
area.
The definition of a river should not only be technical, but should also incorporate the web of
relationships that arise from multiple vantage points of meaning, perception and utility such that it
is recognised by all affected groups.27 A review of several significant geomorphological texts
revealed they did not contain a specific meaning of what a river or a stream actually is.28 However,
there are several definitions of a river contained within English language dictionaries:
• copious natural stream of water flowing in a channel to the sea or a lake etc;29
• a considerable natural stream of water flowing in a definite course or channel or series of
diverging and converging channels; and30
• a stream of water which flows in a channel from high ground to low ground and ultimately to
a lake or the sea, except in a desert area where it may dwindle away to nothing.31
The above definitions all indicate that a river or stream is defined as having a channel and
water that flows permanently or intermittently. However, despite the fact that such definitions are
regularly cited in judgments,32 geomorphologists are aware that the flow characteristics of stream
do not fit neatly within these two flow regimes.33 Thomas and Goudie identify three types of river
and stream flow regime: perennial, intermittent and ephemeral.
• Perennial stream: A stream which flows all year. A dynamic drainage network also includes
intermittent streams and ephemeral streams but there should always be flow in a perennial
stream channel. For much of the time this flow may be in the form of base flow or delayed
flow except when quick flow occurs after rainstorms.34
• Intermittent stream: A stream is classified as intermittent if flow occurs only seasonally when
the watertable is at maximum level. The drainage network is composed of ephemeral,
intermittent and perennial streams and the network expands during rainstorms and extends to
limits affected by antecedent conditions especially antecedent moisture. Flow may occur
along intermittent streams for several months each year but will seldom occur when the water
table is lowered during the dry season.35
• Ephemeral stream: A stream which is often one of the outer links of the drainage network and
which contains flowing water only during and immediately after a rainstorm which may be
fairly intense. As the water flows along the ephemeral channel it may infiltrate into the
27 Howitt R, “Local and Non-specialist Participation in Impact Assessment” in Liu CQ, Zhao Z, Xiao T, and Guha J (eds),
Strategic Management of Environmental and Socio-economic Issues – A Handbook (Guizhou Science and Technology
Publishing House, 2003) pp 27-36.
28 See Leopold LB, Wolman MG, and Miller JP, Fluvial Process in Geomorphology (W. H Freeman and Company, 1964) p
522; Thomas DSG and Goudie AS (eds), The Dictionary of Physical Geography (Blackwell Publishers, 2000) p 610; and
Bridge JS, Rivers and Floodplains (Blackwell Publishing, 2003) p 491.
29 Allen RE (ed), The Concise Oxford Dictionary of Current English (Oxford University Press, 1990) p 1040.
30 Delbridge A and Bernard JRL (eds) The Macquarie Concise Dictionary (Macquarie Library Pty Ltd, 1993) pp 857-858.
31 Whittow JB, The Penguin Dictionary of Physical Geography (Penguin Books, 1986) p 457.
32 For example, Talbot J in Warringah Council v Ardel Limited (2000) 106 LGERA 340 at 346 cited the Oxford Dictionary
definition of a river as having “a copious natural stream” and “a copious flow”, but noted that “neither of which expressions
could be applied to the subject site”. In Zouki v Water Administration Ministerial Corporation [2001] NSWLEC 258 at [13]
it was noted that the definition of “river” in s 2 of the RFIA is inclusive, including both “the definition there set out, as well
as any other definition derived by reference to the ordinary meaning of the word”.
33 See Don Burke v Hawkesbury City Council [2001] NSWLEC 222 at [60] – [61].
34 Thomas and Goudie, n 28, p 363.
35 Thomas and Goudie, n 28, p 271.
Taylor and Stokes
© 202 (2005) 22 EPLJ 193
channel bed as a transmission loss by influent seepage and therefore the peak discharge may
decrease downstream along the ephemeral channel by as much as 5% per km of channel. In
arid and semi-arid areas of the world ephemeral streams are very extensive and represent the
major channel type.36
These definitions support Erskine, who argued that there were three types of river as opposed
to the two that are described within the RFIA.37 The three aforementioned river flow regimes are
clearly distinct and should form the basis for describing the characteristics of a drainage line or
depression that transfers water from hill slopes to lower ground and ultimately to the ocean or an
inland basin.
Is there a difference between a river, a stream and a creek?
Several terms are commonly used to describe watercourses. The most common are brook, beck,
creek, stream and river. The words “brook” and “beck” are often used in England to describe a
small stream. Such small watercourses typically form the lowest order streams that ultimately flow
into larger systems commonly referred to as streams or rivers.38 The terms “creek”, “stream” and
“river” are frequently used interchangeably.39 However most geomorphologists would probably
use these terms to describe systems of different orders of magnitude, ie a stream is smaller than a
river and would represent a lower order watercourse. Where one draws the line between the end of
a stream and the start of a river is arbitrary; eg Boulton and Brock suggest that “you can ford a
stream in waders”.40 Fieldwork by one of the authors on the perennially flowing spring fed rivers
of the Barkly Tablelands, eg the O’Shanassy and Gregory rivers and Louie, Carl and Lawn Hill
creeks revealed that it was possible to ford all the rivers, especially where rock outcrops or
freshwater calcium carbonate deposits have accumulated.41 The ability to wade across a
watercourse is largely dependent on the width to depth ratio of a channel, whether the system is
experiencing base or flood flow and the nature of the riverbed (gravels versus deep, soft
sediments). For example, the upper Shoalhaven River in New South Wales has shallow a gravel
bed and can be waded across in many places. In contrast, rivers with a fine grained bed and banks
such as the lower Hunter River (140 km north of Sydney) often have a smaller width to depth
channel ratio42 and cannot be so easily forded.
Dictionary definitions are unhelpful and somewhat circular. For example, a “stream” is
defined as “a flowing body of water, especially a small river”.43 Further, the application of
watercourse related terms varies between regions and countries making it difficult to determine
their precise meaning. For example, the term creek was originally used to refer to a small coastal
inlet, but in Australia and New Zealand it is used to describe a stream or a brook.44 In Australia,
the term “creek” is more commonly used than stream and, in general though not exclusively, it is
frequently used to describe smaller watercourses. Therefore, it may be argued that a stream is
broadly equivalent to a creek. However, the point at which either becomes a river is debateable,
simply because geomorphic process and form are not exclusive to systems of a particular size and
36 Thomas and Goudie, n 28, p 177.
37 Don Burke v Hawkesbury City Council [2001] NSWLEC 222 at [60].
38 See Strahler AN, “Hypsometric (Area-altitude) Analysis of Erosional Topography” (1952) Bulletin of the Geological
Society of America 63, 1117-1142.
39 Boulton AJ and Brock MA, Australian Freshwater Ecology: Processes and Management (Glen Eagles Publish ing, 1999)
p 300.
40 Boulton and Brock, n 39, p 91.
41 Carthew KD, Taylor MP, and Drysdale RN, “Are Current Models of Tufa Sedimentary Environments Applicable to
Tropical Systems? A Case Study From the Gregory River, Northern Australia” (2003) Sedimentary Geology 162, at 199 -
218.
42 Schumm SA, “The Shape of Alluvial Channels in Relation to Sediment Type” United States Geological Survey
Professional Paper (1960) 352B, pp 17-30.
43 Allen, n 29, p 1205.
44 Allen, n 29, p 1205.
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 203 ©
therefore cannot be used as recognition criteria. A quantitative definition of these terms, taking
into account factors such as drainage size, discharge or channel dimensions could be used to help
define creeks and streams over rivers, but nevertheless, specific classifications would ultimately
still be a qualitative decision.45 To circumnavigate this problem, river systems have been
categorised according to their order, rendering any inferred relationship between terminology and
magnitude redundant.46 This method overcomes one of the main failings of the Strahler system
(Figure 1A) which ignores downstream changes in catchment area and discharge when calculating
stream order (Figure 1B).47
Figures 1A and 1B*
* (A) The stream order method devised by Strahler (1952). First order streams are those with no tributaries, second order
streams form following the confluence of two first order streams, third order streams form where two second order
streams meet, etc. “Excess streams” are tributaries which are of a lower order with their adjacent trunk stream and
therefore do not alter the final stream order.
(B) A modification of Strahler’s (1952) where channel links are ordered by magnitude (Knighton, 1998). Stream order
is calculated by adding confluent stream order values.
The terms used to describe a watercourse are often colloquial; for example in the north of
England a small watercourse is often referred to as a beck, and therefore cannot be readily
standardised. This is further complicated by the distinctly different perceptions of what these
features are in reality, ie one landowner’s ditch is another’s river.48 The terms used to describe
creeks, streams, rivers and rivulets are often artefacts of history which typically describe either
European experiences or names associated with these rivers. Names such as Dinnertime Creek,
Happy Jacks Creek and Licking Hole Creek reveal much about European historical events, while
creek names such as Gordon Creek, Smiths Creek and Hewitts Creek reveal the names of early
settlers who “discovered” the creeks. With respect to the naming of creeks and rivers analogous
situations also occur across much of the new world, such as the United States for example.49
A review of a topographic atlas of Australia confirms that watercourse terminology is quite
variable. Around the south east coast of Tasmania not only are the terms “creek” and “river” in
45 Taylor MP and Stokes R, “When is a River Not a River? Considerations of the Legal Definition of a River for
Geomorphologists Practicing in New South Wales, Australia.” Australian Geographer, forthcoming publication.
46 Strahler, n 38.
47 Knighton, n 13.
48 Howitt, n 27.
49 Riley AC, Restoring Streams in Cities: A Guide for Planners, Policymakers, and Citizens (Island Press, 1998) p 423.
Taylor and Stokes
© 204 (2005) 22 EPLJ 193
common use but so is the term “rivulet” as in Guy Fawkes Rivulet, Peak Rivulet and White
Kangaroo Rivulet. In Western Australia, the term “river” is more common although the word
“brook” is often used for small streams in the southwest of the State as in Bickley Brook, a
tributary of the Canning River 12 km south-east of Perth. The term “creek” tends to be favoured in
the dry interior and the term “river” at the coast. A similar situation prevails in northwest
Queensland, whereas along the wetter eastern areas of northern Queensland the principal drainage
lines are generally referred to as rivers: Mitchell River, Tully River and Burdekin River. Around
Melbourne, Victoria, the terms “creek” and “river” are widespread, with the term “river” being
used more commonly at coastal outlets for major systems. Along the New South Wales coast
major watercourses are generally referred to as rivers as in the Colo River, the Parramatta River
and the Hunter River although the term “rivulet” is occasionally used such as the Macquarie
Rivulet just south of Wollongong. In the drier areas of New South Wales, the term “creek” is more
commonly used, as in Stephens Creek (Broken Hill), Bland Creek, west of Temora, and Lumley
Creek near Bungonia (Figure 2).
Figure 2*
* Map of Lumley Creek. Numbers indicate stream order using the Strahler (1952) method.
However, larger watercourses in western New South Wales, such as those that have perennial
water flow are commonly called rivers, as in Murray River, Darling River and the Murrumbidgee
River. However, to emphasise the contradictory application of the terms “creek” and “river” would
be the names of several major ephemeral watercourses in central Australia that are called rivers:
Finke River, Todd River, Palmer River. A good example of the mixing of watercourse terms
occurs where the Thomson and Barcoo rivers in southwest Queensland combine to become Cooper
Creek! Therefore, while it is possible to make some broad generalisations regarding a correlation
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 205 ©
between flow regime and the term used to describe a watercourse, there are numerous exceptions
to this rule.
A case example of the difficulty in defining a watercourse – Lumley Creek,
Bungonia, New South Wales
Lumley Creek, approximately 220 km southwest of Sydney is a tributary of the Bungonia River
which forms part of the Shoalhaven River catchment (Figure 2). Bungonia has a mean rainfall of
672 mm pa (Australian Bureau of Meteorology) with slightly drier months in June and July and
wetter months in November and January. The creek drains for the most part through Lumley Park
farm. The creek drains a catchment which is predominantly underlain by granite with some
limestone also occurring at the periphery of the catchment. The upper parts of Lumley Creek are
usually dry except during rainfall events as there are no springs upstream or ground water seepage
above the dam (Figure 2).
The DIPNR alleged that the owner of Lumley Park, Bill Dobbie, had infringed the RFIA by
building a dam across the upper reaches of Lumley Creek in January 2001 (Figure 2). A further
instrument that was important to the case was the Farm Dams Policy (1999) which allows farmers
a “Harvestable Right”. This policy, announced by the New South Wales Minister for Land and
Water Conservation in late 1998 and subsequently enshrined in ss 53 and 54 of the WMA, permits
landholders to capture 10% of the runoff from their property for any purpose, without a licence.
The Farm Dams Assessment Guide was released in August 1999 in order to implement the Farm
Dams Policy. It provides that:
Harvestable Right dams can be placed on 1
st and 2
nd order watercourses (except where these
watercourses have permanent flow) as shown on topographic maps and where no watercourse is
shown. Dams on 3rd, 4th and higher order watercourses or any permanently flowing watercourse
require a licence.50
The Farm Dams Assessment Guide cites Strahler’s method for determining stream order and
differentiates a river, creek and minor watercourse in the following way:
In order to have an objective and simple method of deciding between rivers, creeks and minor
watercourses, a method involving ordering of water courses using topographic maps has been
adopted. (Note: Watercourses are shown on topographic maps as broken or continuous blue lines
and are deemed to be continuous even if they lose definition and then reappear).51
The Farm Dams Assessment Guide explains that the stream order system should be applied to
watercourses shown on the most detailed topographic maps available, which in the case of Lumley
Creek is the 1:25,000 Kooringaroo 8828-11-S map sheet. The document also states that Depatment
of Land and Water Conservation (now DIPNR) will rule on ambiguities. Therefore, while the
document purports to explain the differences between rivers, creeks and minor watercourses all it
actually does is to say that it uses the Strahler (1952) method (Figure 1A) to determine their order;
it does not state which of these orders are minor watercourses vis-á-vis creeks or rivers. With
respect to the Dobbie dispute, the Department served notice to remove the dam arguing that it was
built on a river and thus was an infringement of the RFIA which prohibits any unauthorised works
being carried out on protected land without a Pt 3A permit.
As a consequence of the DIPNR serving a removal notice with respect to the dam, Bill Dobbie
decided to challenge the demand to remove his dam in the New South Wales Land and
Environment Court. As a consequence expert witness reports were called from DIPNR and a
representative for the landowner.52 The basis of the argument revolved around two main issues: (1)
50 NSW Government Farm Dams Assessment Guide (1999) http://www.dlwc.nsw.gov.au/care/
water/wr/pdfs/Farm_dams_assessment_guide.pdf p 2, viewed 9 May 2005.
51 See n 50, p 10.
52 Outhet D and Taylor MP, “Lumley Creek Dam Joint Expert Witness Report” (Macquarie Research Limited, 2004) p 15.
Taylor and Stokes
© 206 (2005) 22 EPLJ 193
is Lumley Creek a river in the sense of the RFIA; and (2) if Lumley Creek is a river, what is the
correct stream order?
Assessment of the watercourses marked on the 1:25,000 Kooringaroo 8828-11-S map sheet
revealed the dam is located on a 5th order stream (Figure 2). Even though the legend for the
Kooringaroo map sheet describes the blue lines of Lumley Creek as “intermittent watercourses”,
the Farm Dams Policy states that discontinuous watercourses are to be included in the stream
ordering procedure. Field observations of Lumley Creek upstream of the dam revealed that the
drainage line was poorly defined along much of its network and in many places lacked features
that one would expect to find in a “normal” river system such as fluvial sediments, well-defined
beds and banks or habitat for aquatic flora and fauna (Table 1). Observations of the creek revealed
that the majority of the tributaries marked as 1st and 2nd order (Figure 2) had no physical
expression and that they could be more accurately described as drainage depressions. Subsequent
to field observations of Lumley Creek, it was necessary to establish the validity of applying the
Strahler (1952) system to all or only part of the drainage system upstream if the contested dam.
To determine this, the Strahler (1952) system was investigated for its application to
dry/ephemeral river systems. According to Strahler:
The smallest, or “finger tip” channels constitute the first-order segments. For the most part these
carry wet-weather streams and are normally dry. A second order segment is formed by the junction
of any two first-order streams; a third-order stream is formed by the joining of any second order
streams, etc.53
Further clarification was provided by Strahler (1964):
Assuming that one has available a channel-network map including all intermittent and permanent
flow lines located in clearly defined valleys, the smallest fingertip tributaries are designated
order 1.54
The latter quote clearly states that a bona fide river has to have intermittent or perennial flow.
In addition to field observations revealing Lumley Creek has all the characteristics of ephemeral
flow, water balance modelling using the previous 20 years’ precipitation record indicates that
channel flow probably only occurs in the creeks for a maximum of 102 days or part thereof per
year (28% of the year) and lasts only during rainfall or for a few hours or a day or so after.55 Thus,
it could be argued that it may be more precise to apply the stream order method to the channels of
Lumley Creek which have intermittent and perennial flow, ie those which occur to fall outside the
strict classification proposed by Strahler (1952).
Determination of stream order on Lumley Creek
The DIPNR argued that the blue lines on the map indicate the dam to be located on a 5th order
watercourse, but if field evidence revealed the 1st order streams shown on the topographic map did
not exist, and the 2nd and higher order streams are characterised as intermittent, Lumley Creek at
the dam site would be a 4th order stream.56 DIPNR also suggested that if the 1st order streams
shown on the topographic map do not exist and the 2nd order streams are characterised as
ephemeral, and the 3rd and higher order streams are intermittent, Lumley Creek at the dam site
would be a 3rd order stream. Evidence presented by Taylor in the Lumley Creek expert report
demonstrated that the 1st and 2nd order streams as identified using the 1:25,000 topographic map
were not bona fide rivers according to the RFIA.57 Further, because the upper part of the catchment
is characterised by ephemeral flow, it was not possible for immediately downstream areas which
53 Strahler, n 38, p 1120.
54 Strahler AN, “Quantitative Geomorphology of Drainage Basins and Channel Networks” in Ven Te chow (ed) Handbook
of Applied Hydrology: A Compendium of Water Resources Technology (McGraw-Hill, 1964) pp 4-43.
55 Mark Tooker, Patterson Britton and Partners Pty Ltd, person al communication.
56 Outhet and Taylor, n 52.
57 Outhet and Taylor, n 52.
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 207 ©
have no springs or ground water sources to sustain them in the absence of rainfall, to be defined as
anything other than ephemeral. This would therefore invalidate DIPNR’s claim that the creek is a
3rd order system at the dam. Two principal points arise from this information: first, because field
observations show that Lumley Creek catchment upstream of the dam are ephemeral, then these
drainage lines should not be considered for the purposes of stream ordering because they cannot be
defined as a river or stream according to the RFIA. Second, Strahler (1952, 1964) indicates that
only channel networks with intermittent and perennial stream flow should be used in stream
ordering, and observations of Lumley Creek upstream of the dam reveal the drainage network is
ephemeral.
Several criteria were used to assess the prevailing flow regime in Lumley Creek and its
tributaries above the dam (Table 1). The presence or absence of these physical characteristics was
used to establish if the drainage lines might be described as a river according to the RFIA.
Table 1*
Are there definable channel banks and a channel bed?
Are there fluvial bedforms eg pools, riffles, sediment point bars etc and if so what are they?
Is there any evidence for substantial erosion from water flow within the drainage feature?
Are there any spring lines that may indicate seasonally intermittent or perennial flow?
Is the catchment large enough to sustain perennial or intermittent groundwater flow?
Are there any indicators of prolonged wetness within the drainage feature?
If surface flow is present, is it continuous and how extensive across the base of the drainage
feature is it?
Are there any visible habitats that might sustain aquatic fauna?
Are there any aquatic flora present that would require periods of uninterrupted mois ture?
* A list of the observational questions used in the field to assess presence/absence of fluvial features that are typically
present in a bona fide river or stream. The observations can be used during site inspections to help differentiate
between a river or stream and a drainage depression (modified from Burke v Hawkesbury City Council).
An assessment of all of the evidence indicated that the majority of Lumley Creek’s drainage
network upstream of the dam lacked the features listed in Table 1. At only a small number of
localities did the tributaries have defined banks while in the majority of places there was no
semblance of a bed (ie the channel was vegetated and no fluvial sediments or other bedforms were
present). The Lumley Creek catchment does possess areas of gullied creeks, eg immediately
downstream of a metalled road to the south of the dam at site A (Figure 3). Along with site B
(Figure 3) these watercourses have the greatest numbers of features present that indicate they are
bona fide channels (Table 1). It is possible that the gullying and channel erosion may have been
exacerbated by enhanced runoff from the adjacent metalled road. Here, there are unequivocal
channel banks and bed and fluvial bedforms such as bars and sandy-silty drapes. In addition, there
are hydrophilic plants (Cumbungi) growing in the bed of the channel at site B (Figure 3).
However, beyond the confines of these principal drainage lines the smaller tributary channels lack
the full range of features that are commonly associated with rivers and therefore do not qualify as
“rivers” in the broad sense of the RFIA.58 The vast majority of locations in Lumley Creek above
the dam have no evidence for bed scour or other channel erosion features. Leaving discharge
patterns aside, observations of the physical characteristics of Lumley Creek and its tributaries
indicate that only the drainage lines marked in bold (on Figure 3) have the majority of features
listed in Table 1. Discounting the fact that these channels are discontinuous and ephemeral flow
58 Outhet and Taylor, n 52.
Taylor and Stokes
© 208 (2005) 22 EPLJ 193
and considering only the physical form of the watercourse, this would render the stream order of
Lumley Creek at the dam to be a maximum of two, thus falling within the guidelines as set out in
the Farm Dams Assessment Guide.
Further, the disputed dam with volume of 20 megalitres is well within Lumley Park’s
maximum permissible “Harvestable Right” allowance of 264 megalitres as determined via the
equation as set out in the Farm Dams Assessment Guide.
Figure 3*
* Re-assessment of the stream order using only the physical characteristics (excluding flow conditions) of Lumley Creek
and its tributaries. Field observations revealed that only the drainage lines marked in bold have the majority of the
features associated with in a bona fide river (Table 1). Notwithstanding the fact that these channels are discontinuous
and ephemeral, this would render the stream order of the channel at Lumley Creek dam to be a maximum of two and
thus be permitted under the Farm Dams Policy (1999).
Discussion
The identification and determination of watercourses as per the RFIA has significant practical and
legal implications for land management practice as well as for developers, landowners, natural
resource and government organisations. The DIPNR currently uses two informal methods to
determine if a watercourse is considered a bona fide river or stream under the RFIA. The presence
of a blue line (indicating a watercourse) on a 1:25,000 topographic map sheet and/or if the
catchment has a minimum area of 20 hectares then the DIPNR anticipate that a river or stream as
per the RFIA channel would be present. If the catchment is less than 20 hectares then it is
anticipated that the area is too small to support a natural channel as per the act. This latter
definition would generally exclude the lower order streams, which when viewed in the field may
be more accurately described as drainage depressions rather than as river or stream per se. In most
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 209 ©
situations, unless they are fed by springs, such small, lower order (Strahler, 1952) watercourses are
ephemeral.
Watercourses fed by catchments less than 20 hectares could be considered as “rivers” where
the channel is fed by a spring or contains significant permanent flow along with typical riverine
habitat and fluvial sediments (Table 1). The Urban Development Institute for Australia (UDIA)
also considers the “blue lines” on a 1:25 000 topographic map sheet combined with a minimum
catchment area of 20 hectares as an appropriate starting point for the definition of a river in an
urban environment.59 The UDIA opinion is based on numerous Land Development Inquiry
Workshops for the north-west and south-west Sydney Sectors, and on experience gained in many
environmental conflicts where there has been a significant difference of professional opinion
regarding the true nature of a watercourse.60 The 20 hectare definition would exclude not only the
lowest order streams, but many of the urban streams which have been heavily modified or diverted
and which now have limited or no ecological value.61 While it is common for small depressions
and channels to be piped and culverted in urban situations, this does not negate their value as part
of a drainage network. The application by DIPNR of the “20 hectare rule” is clearly arbitrary.
Geomorphic studies have established an empirical estimate of the relationship between channel
length and catchment area:62 L= 1.4Ad0.6
where L is channel length in kilometres, and Ad is drainage area in square kilometres. When this
relationship is applied to a 20 hectare catchment the expected channel length is ~533 m. This
figure is only a broad estimate because the actual relationship will vary according to climate
(precipitation amount and intensity), geology, relief and any associated land use changes
(deforestation, increases in impervious areas, etc). However, the relationship is useful because it
provides a guideline for the potential length of channel that can be used to compare with field
observations.
In addition to the criteria in Table 1, it is useful to deconstruct the meaning of river and
channel so that it is unequivocal what these represent in the landscape. While it has been argued
that a “river” can mean different things to different people,63 it is suggested here that the specific
term “river” or “stream” relates only to the body of water that flows across the lowest point of the
land. The flow of a river over the landscape generates physical evidence of its activity and
presence, for example, by forming a definitive channel with a bed and bank containing sediments
and aquatic flora and fauna (Table 1). Although the terms the terms “channel” and “river” are
often used interchangeably they are frequently not specifically defined as to what they actually
represent in the landscape, and thus, establishing the difference will help clarify when a river is
truly a river. Without a flow of water in a channel, ie a river, the features (erosion, transport,
sedimentation) and activities (fishing, navigation, irrigation) that are usually associated with a river
will be absent and thus there can be no river per se. However, a channel may still be present, but
its size and morphology will be dependent on how effective earlier flows have been at eroding the
landscape.
In areas characterised by low rainfall it is likely that lower order ephemeral streams will have
discontinuous channels, or more simply, may form depressions which transfer storm flow into the
larger channel network during and shortly after precipitation events. Where the flow is more
consistent over time, channel forms will be extant. Therefore, a river can only be a river when
there is water because it is the body of water that forms the river or stream. Where the flow in
59 Don Fox Planning, “Discussion paper – Urban Riparian Corridors – Categories, Criteria and Requirements” (Urban
Development Institute of Australia, 2003) p 37.
60 Don Fox Planning, n 59.
61 Don Fox Planning, n 59.
62 Hack JT, “Studies of Longitudinal Stream Profiles in Virginia and Maryland” (1957) United States Geological Survey
Professional Paper 294B, p 97.
63 See, eg Howitt, n 27.
Taylor and Stokes
© 210 (2005) 22 EPLJ 193
watercourses occurs through all seasons of the year the river is perennial, whereas if the flow is
“intermittent” the flow occurs at intervals and will cease during the dry season,64 implying that
flow continues for longer than it is absent. Watercourses are also ephemeral, such as Lumley
Creek, which has flows only during or immediately after rainfall for a period that lasts for a few
days at most. Even if flow is ephemeral, a watercourse could still be described as a “river” if it has
the majority of the criteria listed in Table 1. The features listed in Table 1 are common to most
rivers and are formed by the action of water on the landscape and are not scale limited.
It is evident that the current law is inadequate in representing the true nature of New South
Wales (and Australian) rivers. The current interpretation and application of the RFIA and the
WMA allows landowners and developers to disregard the lower order reaches of river networks
even though they may be important for drainage, riparian connectivity and other biophysical
processes including organic matter recycling, seed and sediment dispersal. However, with respect
to land designated for agricultural purposes, such as Lumley Park farm, a sufficient water supply is
clearly required to be able sustain a viable agricultural business, which in this example is sheep
farming. In this case, in order to maintain productivity and economic viability the owner needed a
sufficient water supply which precipitated the construction of the disputed dam. With sporadic
rainfall and runoff at Lumley Creek and sheep grazing requiring between two and six litres of
water per head per day,65 then sensible land management practice, from the owner’s perspective,
involves securing a regular water supply by building a dam. Thus, dam building continues to be a
common solution to water supply issues in our dry landscape, even though they disconnect the
energy flux (water and sediment) and biological connectivity (flora, fauna and organic matter)
within a drainage network.
The current structure of the law relating to the definition of a river in New South Wales
coupled to the fact that water harvesting via dam construction is often the only viable option for
dependable and reliable agriculture, will ensure that drainage networks and their ecology continue
to be irrecoverably damaged.66
CONCLUSION
It is unfortunate that the definition of a river in the RFIA and the WMA fails to address the true
spectrum of rivers in New South Wales. These statutes are inadequate because they fail to
encompass all river system types and flow behaviour that characterise the New South Wales (and
Australian) landscape. The insufficient definition of a river will perpetuate disagreements between
authorities and landowners regarding the true meaning of a river or stream and will bring about
further litigation in the Land and Environment Court. Retention of the definitions contained within
the RFIA and the WMA will also encourage poor land use and land management practices which
will lead to further degradation and, in some cases, inappropriate use of the environment. The use
of blue lines on a map to indicate the presence of rivers or streams is also inadequate for
determining if a watercourse is a “river” because they are often little more than a cartographer’s
interpretation of reality. The same also applies to the 20 hectare rule of thumb that DIPNR has
employed to ascertain if a watercourse should be described as a “river” under the RFIA. The
20 hectare catchment rule has limited application since climate, geology and relief produce a
natural variability in stream formation and morphology and, as a result, each case needs be
reviewed separately. The RFIA combined with the Farm Dams Policy and the Farm Dams
Assessment Guide is a flawed combination because discontinuous rivers as described in the latter
document are not “rivers” under the RFIA since they are ephemeral and furthermore, they lack the
physical features of a bona fide river (Table 1). By retaining a similar definition of “river” as that
64 Allen, n 29; Thomas and Goudie, n 28.
65 Markwick G, “Water Requirements for Sheep and Cattle” Agfact A0.5.4 (4th ed, NSW Agriculture, 2002).
66 McCully P, Silenced rivers: The Ecology and Politics of Large Dams (Zed Books, 1996) p 350.
Up the creek: What is wrong with the definition of a river in New South Wales?
(2005) 22 EPLJ 193 211 ©
contained in the RFIA, the WMA will fail to provide a legislative description that accurately
accommodates the geographic and climatic realities of watercourses in New South Wales.
