Sie sind auf Seite 1von 16

REGULATED RIVERS: RESEARCH & MANAGEMENT

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

HYDRAULIC GEOMETRY OF NEW ZEALAND RIVERS AND ITS USE


AS A PRELIMINARY METHOD OF HABITAT ASSESSMENT
I.G. JOWETT*
National Institute of Water and Atmospheric Research, PO Box 11 -115, Hamilton, New Zealand

ABSTRACT
Downstream hydraulic geometry relationships describe the variation of water depth, velocity, and water surface
width between rivers of different size at a characteristic discharge, whereas at-a-station geometry describes the
variation of hydraulic geometry with discharge within a reach. The instream flow incremental methodology (IFIM)
also predicts the variation in water depth and velocity with discharge at a reach scale, so that hydraulic geometry
relationships can potentially be used as a preliminary method of habitat assessment. Hydraulic geometry relationships
were calculated from instream habitat surveys of 73 New Zealand river reaches with mean flows varying from 0.6 to
204 m3 s 1 and an average gradient of 0.0047. The exponents of both at-a-station and downstream hydraulic
geometry relationships were within the range of values reported in other international studies, although the exponents
indicated that New Zealand rivers tended to experience greater changes in velocity and less in depth than the
international average, probably because of high average gradient. The frequency distributions of water depth and
velocity were positively skewed in most rivers, and on average the modal velocity was 90% of the mean velocity and
the modal depth was 80% of mean depth. The use of at-a-station hydraulic geometry relationships for instream
habitat assessment was compared to depth and velocity predictions using habitat simulation techniques (IFIM) in two
streams. Measurements of stream width and depth at five cross-sections at two calibration discharges were used to
establish at-a-station hydraulic geometry relationships. These predicted mean depth and velocity within 8% of the
reach average values of the IFIM surveys within the range of calibration discharges and within 10 15% of the IFIM
reach average when extrapolated beyond the calibration discharges. Hydraulic geometry can be used to indicate
whether hydraulic conditions approach a threshold such as a minimum acceptable depth or velocity, thus predicating
the need for more extensive habitat survey and analysis. 1998 John Wiley & Sons, Ltd.
KEY WORDS:

stream morphology; hydraulic geometry; habitat; instream flow assessment

INTRODUCTION
The hydraulic geometry of stream channels is described by three basic equations between discharge and
water surface width, depth, and velocity (Leopold and Maddock, 1953; Park, 1977; Kellerhals and
Church, 1989; Mosley, 1992). There are two forms of hydraulic geometry relationships. Downstream
relationships describe the variation in hydraulic geometry between reaches at one discharge, whereas
at-a-station relationships describe the variation of hydraulic geometry with discharge within a reach. The
reaches used to develop downstream hydraulic geometry relationships can be on different rivers, as well
as at different locations along a river, providing the characteristic discharge has the same frequency at
each reach. The characteristic discharge used for downstream geometry relationships is controversial and
depends on the availability of data and the conceptual framework (Bray, 1982), but most researchers have
used some form of flood flow, either a flood with a return period of 23 years or the bankfull flood.
However, there is agreement between hydraulic geometry exponents derived using discharges between
mean and bankfull (Griffiths, 1980; Kellerhals and Church, 1989). The equations used to describe the
change in water surface width, average velocity, and depth with flow at a single site or in a short reach,
the at-a-station geometry, have the same form as those used to describe the downstream geometry.
* Correspondence to: National Institute of Water and Atmospheric Research, PO Box 11-115, Hamilton, New Zealand.
Contract grant sponsor: Foundation for Research, Science and Technology (New Zealand); Contract grant number: CO 1519

CCC 08869375/98/050451 16$17.50


1998 John Wiley & Sons, Ltd.

Recei6ed 3 June 1998


Re6ised 7 July 1998
Accepted 7 July 1998

452

I.G. JOWETT

Methods of assessing instream flow requirements have been divided into three broad categories: (1)
historic flow regime; (2) hydraulic geometry; and (3) habitat (Karim et al., 1995; Jowett, 1997).
Historic flow regime methods use the historic flow record to determine flows that will meet the
ecological requirements of a river, hydraulic geometry methods use the wetted perimeter to determine
flow requirements, and habitat methods determine flows that provide suitable habitat for target
species. A habitat based method, the instream flow incremental methodology (IFIM, Bovee, 1982) and
its physical habitat component (PHABSIM, Milhous et al., 1989), is the most widely used method in
the US (Reiser et al., 1989), and these or similar methods are becoming widely used internationally
(Dunbar et al., 1997). A disadvantage of habitat based methods is their relatively high data requirement and cost compared to historic flow methods. The Tennant or Montana method (Tennant, 1976)
is an example of how the best features of habitat and historic flow methods can be combined.
Tennant (1976) analysed cross-section data from rivers in Montana, Nebraska, and Wyoming and
found that 10% of the mean annual flow would maintain an average depth of 0.3 m and a velocity of
0.25 m s 1; criteria that were considered to be lower limits for aquatic organisms, particularly fish.
Tennant also introduced the concept of different levels of environmental protection by suggesting that
a higher flow, 30% of the mean annual flow, would provide water depths and velocities that were in
the good to optimum range for aquatic organisms. The Tennant method is the second most used
method in the US (Reiser et al., 1989) and has been proposed for New Zealand (Fraser, 1978).
However, Tennants (1976) habitat criteria will be met only where rivers are similar to those used in
the development of the method, because water depth and velocity at a given proportion of mean
annual flow vary with stream size. For example, water depths at 10% of the mean annual flow in a
small stream will be less than water depths at 10% of the mean flow of a large river.
The aim of both instream habitat surveys and at-a-station hydraulic geometry relationships is to
predict how stream depth and velocity change with flow. The difference between the two methods is
that instream habitat surveys predict how the depth and velocity at each measurement point changes
with flow, whereas hydraulic geometry relationships predict how the cross-section average depth and
velocity change with flow. Kellerhals and Church (1989) recognised that at-a-station hydraulic geometry could be used as a habitat description, but thought that habitat characterisation required a more
refined assessment of the distribution of depths and velocities. However, many assessments of instream
flow requirements cannot afford the extensive sampling involved in instream habitat surveys, and an
adequate preliminary assessment of habitat suitability could be made from the average stream depth
and velocity predicted using at-a-station hydraulic geometry relationships. Milhous et al. (1989) describe such a method (HABVD) as part of their PHABSIM computer programme. Such a method,
when applied to individual streams and generalised to a region where rivers are hydrologically and
morphologically similar, could be used to develop regional methods similar to that proposed by
Tennant (1976). Application of a regional method would involve selecting an appropriate target
community and level of protection, according to management goals and the perceived value of the
resource (Jowett, 1997), and applying a formula to determine the flow required to meet the habitat
requirements.
A set of studies of New Zealand rivers, popularly termed the 100 rivers study (Biggs et al., 1990),
described the flow variability, chemistry, water colour and clarity, and periphyton, benthic invertebrate, and trout communities of a group of New Zealand rivers. The present study derives at-a-station
and downstream hydraulic geometry relationships using instream habitat survey data from the 100
rivers and several more recent surveys to provide a description of average characteristics of New
Zealand rivers and their channel shape. A simple method of developing at-a-station hydraulic geometry relationships for instream habitat assessment was tested in two streams by comparing depth and
velocity predicted by at-a-station hydraulic relationships with depth and velocity calculated using IFIM
survey techniques. Repeated applications of this method would allow relationships between the mean
flow of rivers and environmental flow requirements to be developed into regional formulae, similar to
the Tennant method (1976).
1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

HYDRAULIC GEOMETRY OF NEW ZEALAND RIVERS

453

METHODS
The methods used to develop hydraulic geometry relationships in this study were similar to those
described by Bray (1982), in that rivers with good hydrological records (\ 5 years) were selected and the
characteristics of representative reaches rather than single cross-sections were measured. Instream habitat
surveys (Jowett, 1989, 1990) were made at 73 river reaches around New Zealand (Figure 1; Table I). Each
reach included at least one pool/run/riffle sequence and comprised 847 cross-sections, the number of
sections increasing with the variability of stream geometry. The rivers were usually surveyed at between
median and mean annual low flow the average ratio of survey flow to median flow was 0.7590.30 S.D.
Forty-two rivers were located in the North Island and 31 in the South Island, at altitudes ranging from
4 to 846 m above mean sea level. The mean flow of these rivers ranged from 0.6 to 204 m3 s 1, with a
median mean flow of 9.7 m3 s 1 and average gradient of 0.0047.

ANALYSIS
Reach average depths and velocities were calculated by hydraulic simulation for a range of flows,
following methods described by Milhous et al. (1989) and Jowett (1989, 1990). Briefly, a hydraulic model
of each reach was developed by fitting roughness coefficients (Mannings N) between each pair of
cross-sections so that the water surface profile predicted by the standard step method (Henderson, 1966)
matched the measured water surface profile. A relationship between water level and discharge at the
downstream cross-section was developed from calibration measurements at several flows, as is customary
in hydrology (Chow, 1964). The water surface profile and water level at each cross-section was predicted
for a range of flows using the stage-discharge relationship at the downstream cross-section and the
standard step method. Morphological characteristics (water surface width, mean depth, and mean
velocity) were calculated for each cross-section and then averaged for the reach, with each cross-section
weighted by the distance between adjacent cross-sections.

Figure 1. Location of 73 study reaches (circles)


1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

454

1998 John Wiley & Sons, Ltd.

Table I. Hydraulic characteristics of 73 New Zealand river reaches at mean annual and mean annual minimum discharge
River

Mean annual minimum discharge

Flow
(m3 s1)

Width
(m)

Depth
(m)

Velocity
(m s1)

Flow
(m3 s1)

Width
(m)

Depth
(m)

Velocity

23.7
5.6
71.1
58.0
7.7
12.9
203.6
5.5
26.6
53.8
5.9
80.0
25.0
22.4
16.3
5.2
1.1
1.8
6.5
3.1
2.6
3.0
15.3
6.7
2.1
17.2
9.9
34.2
18.8
4.0
61.0
17.6
6.4
5.7
19.0
9.9
30.3
11.2
28.1

46.9
21.0
70.4
46.5
21.7
24.4
88.4
16.4
31.0
77.4
17.9
37.6
40.4
30.2
26.4
23.1
8.4
12.4
33.6
13.4
8.4
14.9
37.2
18.4
13.7
35.6
18.7
36.0
29.5
16.4
56.4
26.2
27.8
19.1
33.1
18.2
54.5
20.8
34.6

0.58
0.48
1.31
1.17
0.53
0.53
2.13
0.48
0.87
0.85
0.46
1.23
0.72
0.85
0.72
0.60
0.33
0.34
0.49
0.42
0.59
0.33
0.63
0.76
0.44
0.84
0.67
0.95
0.81
0.72
1.02
0.91
0.39
0.58
0.61
0.53
0.51
0.50
1.06

0.86
0.54
0.71
0.99
0.63
0.88
1.03
0.65
0.82
0.86
0.62
1.56
0.76
0.82
0.78
0.40
0.40
0.48
0.40
0.54
0.52
0.55
0.60
0.44
0.35
0.55
0.75
1.00
0.80
0.35
0.94
0.66
0.55
0.48
0.84
0.94
0.98
0.96
0.73

8.5
1.0
11.3
22.5
1.6
3.9
73.6
2.1
9.5
12.8
1.0
6.0
8.2
2.1
2.1
0.3
0.3
0.4
0.7
0.7
0.3
0.6
1.6
1.2
0.3
1.4
2.0
8.0
5.3
0.9
9.3
4.1
1.0
1.1
3.4
2.9
5.0
1.1
4.7

37.5
13.2
52.4
41.6
18.4
18.3
79.4
14.3
24.3
66.9
12.9
30.7
33.1
20.8
19.0
16.3
7.2
9.2
22.9
10.6
6.8
10.9
28.3
15.0
11.1
25.7
14.5
26.1
25.2
13.7
42.9
19.5
17.2
14.5
25.1
14.2
31.7
13.4
24.3

0.39
0.34
0.78
0.77
0.26
0.33
1.49
0.33
0.56
0.45
0.24
0.59
0.46
0.41
0.41
0.37
0.18
0.24
0.25
0.25
0.3
0.17
0.36
0.36
0.21
0.46
0.40
0.65
0.52
0.56
0.43
0.72
0.23
0.36
0.31
0.31
0.27
0.19
0.73

0.58
0.27
0.35
0.67
0.32
0.58
0.62
0.44
0.59
0.47
0.32
0.34
0.45
0.29
0.31
0.10
0.23
0.29
0.16
0.27
0.18
0.28
0.19
0.20
0.13
0.17
0.37
0.54
0.47
0.15
0.46
0.34
0.29
0.25
0.41
0.61
0.55
0.41
0.33

Gradient

Channel
shape
exponent

0.0050
0.0027
0.0026
0.0013
0.0046
0.0084
0.0012
0.0023
0.0029
0.0033
0.0034
0.0042
0.0017
0.0051
0.0045
0.0012
0.0180
0.0081
0.0031
0.0099
0.0021
0.0054
0.0012
0.0072
0.0043
0.0019
0.0040
0.0048
0.0019
0.0016
0.0026
0.0042
0.0051
0.0034
0.0034
0.0076
0.0038
0.0044
0.0045

0.90
1.03
0.84
0.95
0.81
0.82
0.78
0.77
0.77
0.97
0.77
0.69
0.79
0.89
0.79
0.82
1.14
0.98
0.98
0.99
0.67
0.99
1.01
1.14
1.18
0.65
0.75
0.87
0.79
0.53
0.95
1.05
1.03
0.89
0.79
0.84
1.07
0.92
0.93

I.G. JOWETT

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

Ahuriri
Akatarawa
Aorere
Arnold
Baton
Buller
Clutha
Esk
Gowan
Grey
Hakataramea
Hawea
Hurunui
Hutt
Inangahua
Kakanui
Kapoaiaia
Kapuni
Kauaeranga
Kaupokonui
Kopuaranga
Maerewhenua
Mangahao
Manganui
Mangaoraka
Mangatainoka
Mangles
Mararoa
Mataura
Maowhango
Motueka
Ngaruroro
Ohau
Opihi
Opihi
Orari
Oreti
Oroua
Otaki

Mean annual discharge

1998 John Wiley & Sons, Ltd.

Table I. (continued)
River

Mean annual discharge


Flow
(m3 s1)
7.6
5.2
0.7
20.6
17.3
17.5
20.1
2.5
10.1
22.5
3.4
1.7
6.4
43.3
14.9
9.0
9.9
0.7
11.6
32.2
16.1
4.2
5.5
5.1
7.6
10.4
2.7
31.2
2.8
2.8
16.5
7.8
27.5
1.7

Width
(m)

Depth
(m)

Velocity
(m s1)

Flow
(m3 s1)

Width
(m)

Depth
(m)

Velocity

27.2
13.1
5.2
45.3
27.9
29.5
43.2
15.4
19.9
39.9
18.0
11.3
15.8
42.2
33.7
23.2
20.6
4.8
34.3
42.6
35.8
9.8
11.8
11.9
28.4
22.2
11.9
47.6
12.5
9.6
30.7
20.8
55.0
13.4

0.43
0.64
0.40
1.05
0.71
0.93
0.52
0.36
1.09
0.66
0.42
0.43
0.55
0.79
0.62
0.58
0.80
0.50
0.53
0.87
0.54
1.01
0.76
0.56
0.45
0.63
0.53
0.77
0.71
0.58
0.79
0.64
0.60
0.34

0.56
0.61
0.41
0.45
0.88
0.64
0.71
0.42
0.58
0.80
0.48
0.38
0.61
1.16
0.59
0.60
0.60
0.27
0.57
0.86
0.68
0.41
0.60
0.66
0.59
0.70
0.45
0.78
0.36
0.65
0.61
0.57
0.77
0.34

1.7
0.8
0.1
2.1
2.2
1.5
4.9
0.6
1.1
2.5
0.7
0.1
2.2
12.8
1.6
1.2
2.8
0.2
3.2
27.2
3.4
3.8
4.8
3.9
1.4
1.2
0.5
8.2
0.4
2.0
0.4
2.1
8.7
0.3

20.7
9.7
3.8
31.5
21.1
22.3
26.4
12.4
14.7
30.9
14.3
7.1
11.2
32.5
19.4
15.9
18.3
4.3
20.6
42.1
22.4
9.7
10.8
11.3
18.5
13.7
9.5
39.1
9.0
21.6
8.1
17.0
45.1
9.3

0.25
0.31
0.23
0.60
0.36
0.46
0.31
0.23
0.87
0.29
0.24
0.32
0.47
0.51
0.26
0.32
0.53
0.3
0.39
0.81
0.32
0.98
0.76
0.49
0.27
0.34
0.32
0.52
0.53
0.47
0.41
0.38
0.36
0.21

0.32
0.27
0.16
0.16
0.40
0.19
0.52
0.21
0.18
0.30
0.25
0.09
0.43
0.71
0.30
0.24
0.31
0.16
0.41
0.78
0.45
0.38
0.61
0.61
0.32
0.31
0.19
0.42
0.14
0.24
0.15
0.33
0.49
0.17

Gradient

Channel
shape
exponent

0.0043
0.0093
0.0028
0.0020
0.0045
0.0020
0.0039
0.0035
0.0021
0.0092
0.0050
0.0015
0.0118
0.0093
0.0081
0.0042
0.0008
0.0015
0.0062
0.0071
0.0016
0.0009
0.0043
0.0028
0.0075
0.0076
0.0078
0.0032
0.0117
0.0028
0.0017
0.0158
0.0050
0.0047

0.81
1.07
0.78
0.83
0.75
0.76
0.76
0.95
0.92
1.02
1.03
0.90
0.93
0.78
0.96
0.94
0.71
0.63
1.06
0.78
0.85
0.50
0.67
0.81
1.07
0.95
0.93
0.92
1.01
0.82
1.04
1.08
1.06
0.99

HYDRAULIC GEOMETRY OF NEW ZEALAND RIVERS

455

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

Otematata
Patea
Pauatahanui
Pelorus
Pohangina
Rai
Rangitikei
Riwaka
Ruahamanga
Ruahamanga
Selwyn
Shag
Stony
Taipo
Takaka
Tauherenikau
Tauranga-Taupo
Tawhiti
Tongariro
Tongariro
Tutaekuri
Waiari
Waihou
Waimakariri
Waimana
Waingawa
Waingongoro
Waiohine
Waiongona
Wairoa (Auckland)
Wairoa (Nelson)
Waiwakaiho
Wanganui
Whakatiki

Mean annual minimum discharge

456

I.G. JOWETT

The mean, median, mean annual minimum, and 80% of mean annual minimum flow were simulated.
The mean annual minimum flow was calculated as the arithmetic average of the annual minimum daily
flows over the full period of flow record. Annual minimum flows were usually positively skewed and were,
on average, 7% higher than the median annual minimum flow. Bankfull or dominant discharges were not
used because of the difficulties in defining bank full in New Zealand rivers (Mosley, 1992) and potential
errors in extrapolating instream habitat survey measurements made at less than median flow to a higher
flow, such as the dominant discharge. In addition, instream habitat assessments are usually concerned
about flows that are less than the median flow.
Hydraulic geometry
The hydraulic geometry of a river is described by three basic equations between flow (Q) and water
surface width (W), mean depth (D), and mean velocity (V) (Leopold and Maddock, 1953):
W=aQ b

(1)

D = cQ

(2)

V = kQ

(3)

The mean depth is defined as the cross-section area divided by the water surface width. As flow is the
product of the mean velocity and cross-section area, the relationship between flow, width, depth, and
velocity is:
Q=WDV

(4)

To maintain continuity (Equation (4)), the sum of the exponents (b + f + m) and the product of the
constants (a c k) equal 1.0 so that the velocityflow Equation (3) can be expressed:
V=

Q 1bf
ac

(5)

Water surface width, mean depth, and mean velocity were calculated at mean annual discharge for the 73
study reaches. Downstream exponents (b, f, and m in Equations (1)(3)) were derived by linear regression
of the logarithm of flow versus logarithms of width, depth, and velocity. The use of geometric mean
regression (Ricker, 1973) was considered, because both flow and measures of hydraulic geometry have
some degree of error. However, simple regression that minimised the deviations of the hydraulic geometry
variable from the regression line, was used so that results were comparable with other studies. As river
depth varies inversely with slope when flow is uniform (Henderson, 1966), a steep river will be shallower
and swifter than a low gradient river for the same channel shape and discharge. This suggests that
gradient will influence downstream hydraulic geometry relationships for depth and velocity. The influence
of slope on hydraulic geometry relationships was tested by multiple regression.
At-a-station exponents were calculated by linear regression of the logarithm of flow versus logarithms
of water surface width, depth, and velocity at four levels of flow; mean annual, median, mean annual
minimum, and 80% of mean annual minimum flow. If at-a-station hydraulic geometry relationships are
to be useful as a predictor of instream habitat, the mean depth and velocity predicted by the relationships
should be a good indicator of the most frequently occurring or modal values of depth and velocity. The
surveyed distributions of water depth and velocity for all study reaches were examined to determine
whether mean values of depth and velocity were consistently related to median and modal values.
Channel shape
Channel shape, along with the nature of the hydraulic control, such as friction or downstream channel
constrictions, determines how the cross-sectional area, and thus velocity and depth, change with
discharge. The shape of a channel is usually assessed subjectively by comparison with a geometrical form,
such as a parabola or rectangle. However, an objective measure of the channel shape can be derived from
the power relationship between channel width (W) and height or elevation (Y):
1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

HYDRAULIC GEOMETRY OF NEW ZEALAND RIVERS

457

Figure 2. Variation of average channel width and height above the lowest point in the cross-section for the Waitaki River, an
extensively braided river with a shape exponent (bc in Equation (6)) of 1.67, the spring-fed Waiari River with a shape exponent of
0.54, and the Maerewhenua River, a typical alluvial river with a shape exponent of 0.95

W = ac (YYmin)bc,

(6)

where bc, is the shape exponent and Ymin the height of the lowest point in the cross-section (W= 0). The
exponent bc was calculated for each cross-section and averaged for the reach. The shape exponent (bc ) is
less than 1 where channels are U-shaped and there is little variation of width with height and is greater
than 1 where channels are alluvial and there are large increases in channel width with height. If the
relationship is plotted with height on the y-axis and width on the x-axis, the curve resembles the average
geometrical form of the channel (Figure 2). For example, a triangular cross-section has a shape exponent
of 1 and a semi-circular cross-section a shape exponent of about 0.5. The exponent of the power
relationship between channel width and depth (b/f in Equations (1) and (2)) also describes channel shape.
However, the shape exponent (bc ) provides a better description of channel shape because it is derived from
measurements of width over the full channel geometry, whereas the exponents of hydraulic geometry
relationships are usually derived from measurements over the range of naturally occurring flows and can
be influenced by local bank effects.
Use of at-a-station hydraulic geometry in instream flow assessments-a simple method of field
determination
If the channel width coefficients (a, b) and the depth coefficients (c, f ) can be determined, it is possible
to calculate the flow required to meet average depth and velocity criteria using Equations (2) and (5). The
channel width and depth coefficients can be estimated from stream measurements of width (W1, W2) and
mean depth (D1, D2) at two flows (Q1, Q2):
b=

log(W1/W2)
log(Q1/Q2)

(7)

a=

W1
Q1b

(8)

1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

458

I.G. JOWETT

f=

log(D1/D2)
log(Q1/Q2)

c=

(9)

D1
Q1 f

(10)

The survey method used in this study was to estimate mean water surface width and depth from
measurements of five randomly selected cross-sections in run habitats in two streams that contained a
mixture of pool/run/riffle habitats. Measurements were made in run habitats because water depths and
velocities in runs are intermediate between the depths and velocities in pools and riffles and are
indicative of average depths and velocities in the reach. The water surface width (W1) was measured at
each cross-section and the average depth (D1) estimated from depth measurements at five points across
the cross-section (d1, . . , d5) using the trapezoidal rule, D1 = (d1 + 2d2 + 2d3 + 2d4 + d5)/10. The discharge was measured where the hydraulic characteristics were best for accurate flow measurement, i.e.
uniform depth and velocity across the section. A temporary staff gauge was installed at each cross-section and the water level noted. The cross-sections were revisited when the flow had changed and the
discharge, water surface width (W2), and change in water level (DL) were measured at each cross-section. To allow for the effect of the change in width on average depth, the average depth at the second
flow (D2) was calculated as:
W (D1 +DL) +
D2 =

DL (W2 W1)
2

W2

(11)

With steep banks, the difference between W2 and W1 is small and D2 = D1 + DL.
The method of field determination of hydraulic geometry was tested in the Kuratau River and
Cosseys Creek (see below). The exponents of hydraulic geometry equations were determined from
measurements of water surface width, depth and water level at two calibration flows using Equations
(7)(11). Mean water depths and velocities were calculated for a range of flows and compared to
mean depths and velocities predicted from instream habitat (IFIM) surveys (Jowett, 1989; Milhous et
al., 1989).
The Kuratau River is a small river in the central North Island with a mean flow of ca 5 m3 s 1
and a stable flow regime. The IFIM instream habitat survey reach was in a well-confined section of
river and comprised 15 cross-sections randomly selected in five runs, five pools, and five riffles. The
survey was carried out at a flow of 5.32 m3 s 1 and calibration at flows of 1.3 and 10 m3 s 1.
Stage-discharge relationships at each cross-section were derived from water levels measured at the
three discharges. Mean depth and velocity at each cross-section were calculated from the stage-discharge relationship and cross-section geometry for flows of 115 m3 s 1. Reach mean depth and
velocity were calculated from the mean cross-section depths and velocities weighted by the proportion
of pool, run, or riffle habitat represented by each cross-section. In contrast, the simple method
determined hydraulic geometry exponents from five cross-sections in runs, with measurements of width
and depth at five points taken at a flow of 5.32 m3 s 1, and measurements of width and water level
change measured at 10 m3 s 1 (Table II). The hydraulic geometry relationships were used to predict
mean depth and velocity at flows of 1 15 m3 s 1 and these were then compared to predictions made
using IFIM techniques.
Cosseys Creek is a small stream in the north of the North Island. Its natural mean flow is ca 0.4
m3 s 1 but most of this is diverted for water supply. The stream is well-confined in a steep sided
valley. Fifteen randomly chosen cross-sections were surveyed in Cosseys Creek at a flow of 0.122 m3
s 1 in pool, run, and riffle habitats. Water level measurements were repeated at flows of 0.0236 m3
s 1 and 0.249 m3 s 1 and mean depth and velocity calculated for flows of 0.010.49 m3 s 1 as for
the Kuratau River. The simple method of determining hydraulic geometry relationships was applied
using data from five cross-sections in runs (Table II) and mean depth and velocity predicted for flows
of 0.010.49 m3 s 1 for comparison with predictions made using IFIM survey techniques.
1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

459

HYDRAULIC GEOMETRY OF NEW ZEALAND RIVERS

RESULTS
Downstream hydraulic geometry
The average river width, depth, and velocity of New Zealand rivers at mean discharge increased with
the mean discharge to the power of 0.488, 0.239, and 0.241 respectively, with mean discharge explaining
86% of the variance in river width and 56% and 62% of the variance in depth and velocity respectively
(Figure 3, Table III). The sum of the width, depth, and velocity exponents was less than one, and
apportioning the difference equally (Park, 1977) gave exponents of 0.50, 0.25, and 0.25, respectively.
When reach gradient was included in multiple regressions, its coefficient was significantly different from
zero in relationships with velocity (P = 0.001) and with depth (PB 0.001), but not significantly different
from zero in the relationship with width (P = 0.55). Reach gradient increased the variance explained in the
velocitydischarge relationship from 62 to 73% and the variance explained in the depthdischarge
relationship from 56 to 62%. The signs of the regression coefficients indicated that velocity increased and
depth decreased as gradient increased. There was no significant correlation between the logarithms of
mean annual discharge and reach gradient (R 2 = 0.041, P= 0.089).
At-a-station hydraulic geometry
The average exponents for at-a-station relationships for width, depth, and velocity were 0.176, 0.305,
and 0.427, with width exponents ranging from 0.052 to 0.438, depth exponents from 0.092 to 0.448, and
velocity exponents from 0.154 to 0.693 (Figure 4, Table III). As with downstream relationships, the sum
of the exponents was less than 1.0 and apportioning the difference equally between the exponents gave
width, velocity, and depth exponents of 0.207, 0.458, and 0.335, respectively.
Mean values of depth and velocity did not necessarily predict the most commonly occurring values of
depth and velocity. The distribution of velocity was positively skewed (mean\ median) in all reaches
except one, and the distribution of water depths was positively skewed in 93% of reaches. There were
good linear relationships (R 2 =0.92 for both, P 0.001) between mean and median values of depth and
velocity:
Vm =0.93V( 0.065
Dm =0.86D( 0.007
where Vm is median velocity and Dm is median depth.
In the reaches surveyed in this study, the distribution of velocity was often bi-modal as a result of
channel configuration, with one modal peak at low velocity (B 0.05 m s 1) and the other at a higher
velocity. In the 53 cases where the modal velocity was higher than 0.05 m s 1, the modal velocity was
Table II. Mean water depth and width at a flow of 5.32 m3 s1 in the Kuratau River and a flow of 0.122 m3 s1
in Cosseys Creek, and the width and water level change at each cross-section when the flow increased to 10 m3 s1
in the Kuratau River and to 0.249 m3 s1 in Cosseys Creek
Kuratau River

Cosseys Creek

Mean depth
(m)

Width
(m)

1
2
3
4
5

0.64
0.65
0.60
0.86
0.37

15.8
16.3
16.55
18.0
20.11

Mean

0.62

17.35

Run
Run
Run
Run
Run

1998 John Wiley & Sons, Ltd.

Water level change


(m)

Mean

Width
(m)

Water level change


(m)

16.15
16.85
17.27
18.4
21.43

0.170
0.223
0.217
0.204
0.172

0.16
0.22
0.28
0.17
0.16

3.82
4.90
3.00
3.40
4.30

4.57
5.02
4.52
3.82
4.30

0.036
0.049
0.071
0.074
0.055

18.02

0.197

0.20

3.88

4.45

0.057

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

460

I.G. JOWETT

Figure 3. Reach average downstream hydraulic geometry at mean annual discharge for 73 New Zealand river reaches shown as
circles with solid regression line and 25 gauging sites from Griffiths (1980) shown as squares with dashed regression line (exponents
and coefficients in Table III)

closely related to both mean and median (R 2 \0.79, PB0.001), and on average, the modal velocity was
0.87 times the mean velocity. The modal depth was also closely related to both mean and median depth
(R 2 \0.73, PB0.001) and, on average, was 0.79 times the mean depth.
Average at-a-station hydraulic geometry relationships (Table III) and relationships between modal
depth and velocity and mean depth and velocity can be used to estimate the flow required to provide
maximum adult brown trout drift feeding habitat (depth\ 0.5 m, velocity= 0.5 m s 1; Hayes and Jowett,
1994) in an average New Zealand river. Modal habitat requirements of 0.5 m depth and 0.5 m s 1
velocity are equivalent to a mean depth of 0.63 m and a mean velocity of 0.57 m s 1. This mean depth
Table III. Exponents and coefficients of downstream hydraulic geometry at mean annual discharge and average
at-a-station hydraulic geometry of 73 New Zealand river reaches
Variable

Downstream
coefficients9S.E.*

Downstream
exponents9 S.E.*

At-a-station
coefficients 9S.D.*

At-a-station
exponents9 S.D.*

Width
Depth
Velocity

7.76 90.49
0.379 0.02
0.36 9 0.02

0.4889 0.025
0.2399 0.026
0.24190.021

15.8 9 7.5
0.319 0.12
0.249 0.09

0.17690.066
0.3059 0.088
0.4279 0.090

* S.E. is used in this table to represent the uncertainty in downstream regression coefficients and S.D. is used to indicate variability
in at-a-station geometry among river reaches.
1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

HYDRAULIC GEOMETRY OF NEW ZEALAND RIVERS

461

Figure 4. Frequency of occurrence of at-a-station hydraulic geometry exponents and shape exponents for 73 New Zealand rivers

and velocity, in the average New Zealand river, is provided by a flows of 10.2 m3 s 1 and 7.6 m3 s 1,
respectively using the average at-a-station relationships in Table III. Estimates of the flow required to
provide maximum adult trout drift feeding habitat in New Zealand rivers made using IFIM techniques
are typically between 5 m3 s 1 and 14 m3 s 1 (unpublished NIWA reports).
Relationship between channel shape and hydraulic geometry exponents
The average shape exponent (bc in Eq. (6)) was 0.89 90.14 S.D., with 86% of values between 0.7 and
1.1 (Figure 4). The modal peak was between 0.9 and 1.0, indicating that the channel shape of many rivers
was close to triangular, with width increasing linearly with level. Rivers with low exponents (U-shaped)
were often stable spring- or lake-fed, whereas rivers with high exponents had either braided or unconfined
alluvial channels (Figure 2).
When the channel shape exponents were compared with the at-a-station hydraulic geometry exponents
(Figure 5), there was a trend for the width exponent to increase as the shape exponent increased, as would
be expected because there is less variation of width with discharge in U-shaped channels (shape
exponent B 1, geometry exponent: 0.1) than in V shaped or braided channels (shape exponent ] 1,
geometry exponent: 0.2). The depth exponent did not vary with channel shape, but the velocity exponent
increased as the shape exponent decreased and channels became more U-shaped (Figure 5). The linear
relationships between the shape and width and shape and velocity exponents were poor and significant at
the 0.1 level (P= 0.06 and P = 0.1, respectively).
1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

462

I.G. JOWETT

Figure 5. Variation of at-a-station hydraulic geometry exponents with channel shape, showing linear regression lines fitted to
hydraulic geometry exponents for velocity (upper dashed line), depth (centre solid line), and width (lower dashed line)

Use of at-a-station hydraulic geometry in instream flow assessments


The at-a-station hydraulic geometry depth and velocity exponents calculated from measurements in
Table II according to the procedure listed in Table IV were 0.44 and 0.50, respectively, for Cosseys Creek
and 0.38 and 0.56 for the Kuratau River. These were within the range of measured values (Figure 4), but
were slightly higher than average (Table III) as would be expected for rivers with confined channels. Over
the range of the two calibration flows used to calculate the hydraulic geometry relationships, predicted
depth and velocity was within 2.8% on average (maximum difference less than 8%) of the mean depth and
velocity predicted by IFIM techniques, i.e. habitat-weighted means in 15 cross-sections located in pool,
run, and riffle habitats (Figure 6). The difference between predictions by the two methods increased with
extrapolation, with less than 15% difference at half of the lower of the two calibration flows and less than
10% difference at twice the higher calibration flow.
Table IV. Steps in the application of at-a-station hydraulic geometry to instream flow assessments
Step
1
2
3
4
5
6
7
8
9
10

Process
Measure width and average depth at five cross-sections in intermediate (run) habitats
Establish temporary staff gauge on each cross-section and record water levels
Measure flow Q1 at the cross-section with the most uniform flow characteristics
Calculate mean depth D1 and width W1 for reach
When flow has changed, measure flow Q2 again at the most uniform cross-section, measure widths and
water levels at each cross-section from staff gauges
Calculate mean width W2 and mean change in water level DL
Calculate mean depth D2 using Equation (11)
Calculate a, b, f, c using Equations (7)(10) and use them in Equations (2) and (5) to give depth vs.
discharge and velocity vs. discharge relationships
Determine mean depth and velocity habitat requirements for the reach by increasing the target depth
habitat requirement by 25% and target velocity habitat requirement by 10% to compensate for positively
skewed distributions about the mean
Calculate the discharge required to provide mean depth and velocity for habitat requirements

1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

463

HYDRAULIC GEOMETRY OF NEW ZEALAND RIVERS

Figure 6. Comparison of mean depth and velocity in the Kuratau River (left) and Cosseys Creek (right) predicted by an instream
habitat survey (IFIM) with mean depth and velocity predicted by hydraulic geometry relationships estimated by the quick survey
method at flows of 5.32 and 10 m3 s 1 in the Kuratau River and flows of 0.122 and 0.249 m3 s 1 in Cosseys Creek

DISCUSSION
Downstream hydraulic geometry
The hydraulic geometry exponents derived in this study were similar to exponents in two other New
Zealand studies (Griffiths, 1980; Henderson and Ibbitt, 1996) and exponents summarised by Park (1977)
for 72 sites world-wide (Table V). Griffiths (1980) derived downstream hydraulic geometry exponents
from 25 flow gauging sites at mean annual discharge in 6 large river systems and Henderson and Ibbitt
(1996) carried out extensive surveys of downstream geometry in five upland catchments at several flows
less than mean annual discharge. The depth exponent (0.25) in this study was lower than Parks (1977)
modal range of 0.3 0.4, but in the range of values that occurred with the second highest frequency. It was
also lower than the value reported by Griffiths (1980), but similar to that found by Henderson and Ibbitt
(1996). Conversely, the velocity exponent of 0.25 in this study was higher than values reported by Park
(1977) and Griffiths (1980), but similar to that of Henderson and Ibbitt (1996). Differences in stream
gradient may explain the variations in depth and velocity exponents. Hydraulic theory (Henderson, 1966)
indicates that a steep river will be shallower and swifter than an equivalent low gradient river and, in this
study, gradient explained 6 11% of the unaccounted variance in relationships between discharge and
depth and velocity. There was a high correlation between gradient and discharge in the data of both
Table V. Comparison of downstream hydraulic geometry exponents with other New Zealand (Griffiths, 1980;
Henderson and Ibbitt, 1996) and world-wide (Park, 1977) exponents
Variable

This study
(N =73)

Griffiths
(N= 25)

This study plus Griffiths


data (N= 98)

Henderson and
Ibbitt

Park

Width
Depth
Velocity

0.50
0.25
0.25

0.48
0.43
0.11

0.48
0.30
0.22

0.450.55
0.200.35
0.200.35

0.40.5
0.30.4
0.10.2

1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

464

I.G. JOWETT

Griffiths (1980) and Henderson and Ibbitt (1996), but no significant correlation in this study. Although
there was no significant difference between the average slope of the sites in Griffiths study and the
reaches in this study (Mann Whitney U test, P= 0.31), the gradients of the six highest discharge sites in
Griffiths (1980) study were less than half the average gradient of reaches in this study, and these sites
were deeper and more slowly flowing than indicated by regression lines fitted through the reach data
(Figure 3). This suggests that values of depth and velocity exponents may be influenced by river gradient
in cases where there are large differences between the gradients of the largest and smallest rivers. When
Griffiths 25 sites were added to the 73 reaches in this study, the downstream hydraulic geometry
exponents for width, depth, and velocity were 0.48, 0.30, and 0.22, respectively.
The relationship between mean annual discharge and water depth and velocity shows that that
hydraulic habitat (depth and velocity) varies with river size. If fish community composition and
abundance is controlled by hydraulic habitat, one might expect fish populations to vary with river size.
This change in habitat with river size may be partially responsible for the longitudinal gradients that are
commonly observed in fish communities (e.g. Sheldon, 1968; Rahel and Hubert, 1991; Oberdorff and
Porcher, 1992; Aadland, 1993). New Zealand fish communities also show strong longitudinal gradients, as
a result of diadromy (McDowall, 1993; Jowett and Richardson, 1996) and stream size (Jowett et al., 1996,
1998).
At-a-station hydraulic geometry
Park (1977) summarised at-a-station exponents for 139 sites from around the world. Average at-a-station hydraulic depth and velocity exponents of 0.335 and 0.458 for reaches in this study were within
Parks modal class ranges of 0.3 0.4 and 0.4 0.5 for depth and velocity, respectively, but the average
width exponent of 0.203 was slightly than higher Parks modal range of 0.00.1. The cross-sectional shape
of 48% of the reaches in this study was triangular, with shape exponents of 0.91.1, 49% were U-shaped,
with shape exponents B0.9, and few had shape exponents \ 1.1. Rivers with triangular cross-sections
were usually unconfined alluvial rivers with exposed gravel banks. Some rivers with U-shaped channels
were associated with stable flow regimes that allowed the development of steep stable banks in the
alluvium, whereas other U-shaped rivers were confined by either valley sides or dense riparian vegetation
on one or both banks. Rivers with shape exponents \ 1.1 were probably poorly represented in this study,
because it is difficult to carry out instream habitat surveys in the braided rivers that exponents \ 1.1
characterise.
There were poor relationships between channel shape and hydraulic geometry, especially depth and
velocity, and this would make it difficult to estimate habitat response to flow changes from the
appearance of the river alone. Channel shape had no influence on the way depth changed with discharge,
but the velocity exponent tended to be higher in confined channels than in unconfined channels. Thus, a
change in flow in a confined channel tends to cause a greater variation in velocity than in an unconfined
channel, so that during floods, environmental conditions in confined channels may be harsher than in
unconfined channels. This is similar to the hydraulic differences between pools and riffles. Andrews (1984)
found that at-a-station hydraulic geometry depth exponents in pools and riffles of gravel-bed rivers were
similar, but that the velocity exponent was higher in pools than in riffles. During floods, velocities in pools
exceeded those in riffles and pools became scouring zones and riffles depositional zones (Andrews, 1984).
Use of at-a-station hydraulic geometry in instream flow assessments
Instream flow assessments can be made using at-a-station hydraulic geometry relationships derived
from relatively few measurements compared to IFIM survey techniques. In the two trials of this method,
mean depths and velocities were within 15% of the values predicted by IFIM surveys over a range of flows
from half to twice the calibration flows. Milhous et al. (1989) compared predictions of habitat suitability
(weighted usable area) calculated using hydraulic geometry relationships with habitat suitability calculated
from IFIM surveys and found they predicted similar patterns of variation of habitat with flow, with
habitat values that were usually within 20% of each other. Hydraulic geometry provides a promising
1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

HYDRAULIC GEOMETRY OF NEW ZEALAND RIVERS

465

method for making an initial assessment of environmental impact of proposed flow changes, provided
habitat requirements can be specified in terms of mean velocity and depth. Although Kellerhals and
Church (1989) thought that habitat assessment required a more detailed description of the distribution of
depth and velocity, mean depth and velocity may be an adequate indicator of habitat suitability in many
situations. This is an issue of scale; micro-habitat, meso-habitat, or macro-habitat. While many fish and
stream insects undoubtedly make use of habitat on a micro scale, many of the features that create
micro-habitat, such as substrate, bed, and bank forms, vary little with flow and a flow requirement that
provides suitable macro-habitat should also provide suitable micro-habitat. For example, Jowett et al.
(1996) found that the density of some New Zealand native fish species was highest in streams where the
mean depth and mean velocity were similar to the preferred mesohabitat of those species. Mean values of
depth and velocity will usually overestimate modal velocity and depth, but this could be taken into
account in the habitat criteria. Velocity and depth distributions in the reaches of this study suggested that
modal depth is about 80% of mean depth and modal velocity about 90% of mean velocity.
Calculation of velocity and depth from at-a-station hydraulic relationships assumes that the power law
relationships between stream width and average depth and between discharge and average depth hold
over the required range of flows. The relationships will be inaccurate where there are changes in
cross-section geometry that were not in the range of calibration flows. For example, if a trapezoidal
channel were calibrated with water level and flow contained within sloping banks, the equations would
not be accurate when the water level was either above or below those banks. Habitat suitability criteria
usually specify a range of suitable water depths and velocities (Jowett and Richardson, 1995) and this,
combined with the distribution of depth and velocity, means that there is a range of flows over which the
area of suitable habitat is relatively constant. Habitat suitability criteria can be used to define threshold
levels of depth and velocity, below which habitat quality begins to decline. A simple method of flow
assessment based on hydraulic geometry can be used as a preliminary means of indicating whether mean
or modal hydraulic conditions that result from a change in flow are safe or approaching a threshold
such as a minimum acceptable depth or velocity, thus predicating the need for more extensive habitat
survey and analysis.

ACKNOWLEDGEMENTS

I thank the hydrological staff of NIWA, then DSIR Water Resources Survey, and staff of the Regional
Councils for the habitat survey work, Kathy Walter for obtaining the hydrological data, and G. McBride
for reviewing an early draft of this paper. This research was carried out under contract (CO 1519) from
the Foundation for Research, Science and Technology (New Zealand).

REFERENCES
Aadland, L.P. 1993. Stream habitat types: their fish assemblages and relationship to flow, N. Am. J. Fish. Mgmt, 13, 790 806.
Andrews, E.D. 1984. Bed-material entrainment and hydraulic geometry of gravel-bed rivers in Colorado, Geol. Soc. Am. Bull., 95,
371 378.
Biggs, B.J.F., Duncan, M.J., Jowett, I.G., Quinn, J.M., Hickey, C.W., Davies-Colley, R.J., and Close, M.E. 1990. Ecological
characterisation, classification, and modelling of New Zealand rivers: an introduction and synthesis, N. Z. J. Mar. Freshw. Res.,
24, 277 304.
Bovee, K.D. 1982. A guide to stream habitat analysis using the instream flow incremental methodology, US Fish and Wildlife
Ser6ice Biological Ser6ices Program, FWS/OBS-82/26.
Bray, D.I. 1982. Regime equations for gravel-bed rivers, in Hey, R.D., Bathurst, J.C., and Thorne, C.R. (Eds), Gra6el-bed Ri6ers,
Wiley, Chichester, UK. pp. 517552.
Dunbar, M.J., Gustard, A., Acreman, M., and Elliot, C.R.N. 1997. Overseas approaches to setting river flow objectives, Interim
Technical Report, Environmental Agency, Bristol, Environmental Agency Project W6B(96)4. pp. 67.
Fraser, J.C. 1978. Suggestions for developing flow recommendations for instream uses of New Zealand streams, Water and Soil
Miscellaneous Publication 6, Ministry of Works and Development, Wellington.
Griffiths, G.A. 1980. Hydraulic geometry relationships of some New Zealand gravel-bed rivers, N. Z. J. Hydrol. 19, 106 118.
1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

466

I.G. JOWETT

Hayes, J.W., and Jowett, I.G. 1994. Microhabitat models of large drift-feeding brown trout in three New Zealand rivers, North Am.
J. Fish. Mgmt, 14, 710725.
Henderson, F.M. 1966. Open Channel Flow, MacMillan, London. pp. 522.
Henderson, R., and Ibbitt, R. 1996. When the going gets tough, the tough get going: field work in a sub-alpine basin on the West
Coast, South Island, Water Atmos., 4(4), 810.
Jowett, I.G. 1989. River Hydraulic and Habitat Simulation, RHYHABSIM Computer Manual, N. Z. Fisheries Miscellaneous
Report No. 49. New Zealand Ministry of Agriculture and Fisheries, Christchurch. 39 pp.
Jowett, I.G. 1990. Factors related to the distribution and abundance of brown and rainbow trout in New Zealand clear-water
rivers, N.Z.J. Mar. Freshwater Res., 24, 429440.
Jowett, I.G. 1997. Instream flow methods: a comparison of approaches, Regul. Ri6ers: Res. Mgmt, 13, 115 127.
Jowett, I.G., Hayes, J.W., Deans, N., and Eldon, G.A. 1998. Comparison of fish communities and abundance in unmodified
streams of Kahurangi National Park with other areas of New Zealand, N. Z. J. Mar. Freshw. Res., 307 322.
Jowett, I.G., and Richardson, J. 1995. Habitat preferences of common, riverine New Zealand native fishes and implications for flow
management, N. Z. J. Mar. Freshw. Res., 29, 1323.
Jowett, I.G., and Richardson, J. 1996. Distribution and abundance of freshwater fish in New Zealand rivers, N. Z. J. Mar. Freshw.
Res., 30, 239 255.
Jowett, I.G., Richardson, J., and McDowall, R.M. 1996. Relative effects of instream habitat and land use on fish distribution and
abundance in tributaries of the Grey River, New Zealand, N. Z. J. Mar. Freshw. Res., 30, 463 475.
Karim, K., Gubbels, M.E., and Goulter, I.C. 1995. Review of determination of instream flow requirements with special application
to Australia, Water Resour. Bull., 31, 10631077.
Kellerhals, R. and Church, M. 1989. The morphology of large rivers: characterization and management, in Dodge, D.P. (Ed),
Proceedings of the International Large Ri6er Symposium, Can. Spec. Publ. Fish. Aquat. Sci., 106, 31 48.
Leopold, L.B. and Maddock, T. 1953. The hydraulic geometry of stream channels and some physiographic implications, U. S. G.
S. Professional Paper, 252. pp. 57.
McDowall, R.M. 1993. Implications of diadromy for the structuring and modelling of riverine fish communities in New Zealand,
N. Z. J. Mar. Freshw. Res., 27, 453462.
Milhous, R.T., Updike, M.A., and Schneider D.M. 1989. Physical habitat simulation system reference manual-Version II, US Fish
and Wildlife Ser6ice Instream Flow Information Paper No. 26, Biological Report, 89(16).
Mosley, M.P. 1992. River morphology, in Mosley, M.P. (Ed.), Waters of New Zealand, New Zealand Hydrological Society,
Wellington. pp. 285304.
Oberdorff, T. and Porcher, J-P. 1992. Fish assemblage structure in Brittany streams (France), Aquat. Li6ing Resour., 5, 215 223.
Park, C.C. 1977. World-wide variations in hydraulic geometry exponents of stream channels: an analysis and some observations,
J. Hydrol., 33, 133146.
Rahel, F.J. and Hubert, W.A. 1991. Fish assemblages and habitat gradients in Rocky Mountain Great Plains streams: biotic
zonation and additive patterns of community change, Trans. Am. Fish. Soc., 120, 319 332.
Reiser, D.W., Wesche, T.A., and Estes, C. 1989. Status of instream flow legislation and practices in North America, Fisheries,
14(2), 22 29.
Ricker, W.E. 1973. Linear regressions in fishery research, J. Fish. Res. B. Can., 30, 409 434.
Sheldon, A.L. 1968. Species diversity and longitudinal succession in stream fishes, Ecology, 49, 193 198.
Tennant, D.L. 1976. Instream flow regimens for fish, wildlife, recreation, and related environmental resources, in Orsborn, J.F. and
Allman, C.H. (Eds), Proceedings of the Symposium and Speciality Conference on Instream Flow Needs II, American Fisheries
Society, Bethesda, Maryland. pp. 359373.
Chow, V.T. 1964. Handbook of Applied Hydrology, McGraw Hill, New York. 1418 pp.

1998 John Wiley & Sons, Ltd.

Regul. Ri6ers: Res. Mgmt. 14: 451 466 (1998)

Das könnte Ihnen auch gefallen