close

Вход

Забыли?

вход по аккаунту

?

40737%282004%29409

код для вставкиСкачать
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
APPLICABILITY OF GEOMORPHOLOGICAL
PROCEDURES FOR RIVER RESTORATION
R.D.Hey
School of Environmental Sciences, University of East Anglia,
Norwich, NR4 7TJ, UK.
Tel: +44(0)1603 593118. Fax: +44 (0)1603 591327
e-mail: r.hey@uea.ac.uk
Abstract
A geomorphological procedure, based on scaling reference reach data, is now
being widely adopted for developing natural stable channel designs for river restoration
and riverine habitat creation. Considerable controversy exists about the appropriateness
or otherwise, of the method. This paper critically examines the various key steps in the
procedure. Success depends on choosing an appropriate reference reach. Not only has it
to be stable and of the same stream type as that required for the restoration site, it also
has to have the same bankfull width/depth ratio and sinuosity as these values are
transferred to the restoration reach. For these dimensionless shape parameters to be
equivalent at the reference and restoration sites, the ratios of their bankfull discharges,
bed material sizes and loads, valley slopes and equivalent bank vegetation densities are
prescribed. UK regime equations for the stream type in question define these critical
ratios. The use of reference reaches that do not match these critical ratios will result in
failure of the restoration design. These ratios are confirmed by US field data.
NATURAL STABLE CHANNEL DESIGN
The aim of river restoration is either to restabilise reaches of river that have been
destabilised or to naturalise rivers that have been canalised but which have essentially
remained stable. The former can be regarded as active rivers as they regularly transport
bed material load, while the latter are passive ones as they rarely, if ever, do so.
Restoring a river to its original, pre disturbed state is not normally feasible, as the
controlling conditions may have irreversibly changed or because of continuing
1
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
engineering constraints, rehabilitation to a natural state is probably a better term.
Nevertheless, because of its long term usage, restoration is used herein to describe this
process.
There are three basic approaches to natural stable river design i) analytical ?
using rational equations ii) empirical ? using regime equations and iii) analogue ? using
geomorphological procedures. Currently neither rational or regime equations are
adequate to prescribe a river?s three dimensional morphology. The lack of sufficient
equations restricts their application to straight channels and to a limited range of stream
types. Equally they fail to provide information to design natural variations in widths,
depths and slopes that are typical of natural reaches of river: namely detail that is
required to achieve a sustainable natural design and provide appropriate habitat variety.
The geomorphological procedure offers an alternative solution. Unlike previous
methods, it is not restricted by stream type, it prescribes the plan form of the river and it
enables variations in the morphology of the river to be determined. This paper critically
reviews the geomorphological procedure developed by Rosgen (1998), evaluates its
predictive capability and provides guidelines for its successful application.
FLUVIAL GEOMORPHOLOGICAL METHODOLOGY
The fluvial geomorphological procedure uses the three dimensional shape of an
appropriate natural stable reference reach as a template for the design. This is
subsequently scaled to prescribe the morphology for the restored river (Rosgen 1998).
The following steps are required:i) River classification
First it is necessary to classify the stream types in the reach to be restored and
adjacent upstream and downstream reaches. Rosgen?s (1994) classification is used for
this purpose. This is a morphological classification based on dimensionless measures or
a river?s cross sectional dimension (riffle bankfull width/depth ratio; W/d) pattern
(sinuosity: p), profile (bankfull slope: S) and entrenchment (ratio of valley width, at
twice maximum flow depth in river at bankfull stage, to the bankfull width: W blt /W ).
These ratios reflect the boundary conditions that control channel morphology and the
associated governing flow processes. Identical channels have the same boundary
conditions and flow processes. Similar channels, those with the same dimensionless
shape ratios but a different scale, have specific combinations of boundary condition
ratios (Hey 2004).
Although there is a longstream continuum of river morphologies, distinct spatial
and temporal breaks occur which reflect local variations in boundary conditions. Over
time unstable rivers will self stabilise and there will be associated systematic change in
stream types. By observing the streamwise variation in stream types and identifying the
direction instability is migrating, space-time substitution enables the evolving stable
stream type to be forecast.
2
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
ii) Determine stream type required for the restoration
The stream type required for the restoration design depends on: a) the cause of the
instability and the departure of the river either from its original stable state or its evolving
one, b) on any new constraints imposed on the river by highway construction, river
crossings, flood alleviation schemes etc. In some cases there may be more than one
possible solution. Choices could include re- instating the original river/flood plain system
or creating new river/flood plains system within the confines of an incised channel.
Where there are significant constraints, for example due to a highway or urban
encroachme nt on the flood plain, then the loss of flood conveyance would necessitate a
different type of solution.
iii) Choice of reference reach.
As the reference reach provides the template for the restoration design, it is
essential that it corresponds to the stream type required for the restoration reach. To
qualify it has to be both vertically and laterally stable, be in the same valley type (width,
slope, depositional materials, landforms, landtypes/associations) and in the same
hydrophysiographic region (same geology, topography, land use, vegetation, climate and,
hence, flow and sediment transport regime). Each hydrophysiographic region is
characterised by a specific relation between the river?s bankfull cross-sectional area and
its basin area (Rosgen 1998) : referred to as a regional curve.
iv) Design shape ratios.
Dimensionless morphological ratios from the stable reference reach are used to
develop the restorative design. The cross-sectional shape is defined by the bankfull
width to mean depth ration at the riffle section (W/d) and its associated ratio of maximum
depth to mean depth (dm /d). These variables are sufficient to uniquely define its shape
(Hey 1978). Planform measures for the reference reach include ratios of meander
wavelength, radius of curvature and meander arc length to channel width (?/W, r/W, and
z/W repectively). Because of local variability in those ratios, maximum and minimum as
well as mean values are obtained. To incorporate local morphological variability into the
design, ratios are also obtained on bankfull pool width, depth and maximum depth
relative to associated riffle values and for low flow slopes at riffle, run, pool and glide
sections relative to the average bankfull slope (Rosgen 1998).
v) Scaling the design
The required bankfull cross-sectional area (A) at the restoration site is used to
scale the restoration design. This is obtained from an adjacent stable section provided it
is of the required stream type and provided that no tributaries join the river between it
and the restoration reach. Alternatively it is obtained from a regional curve for its
hydrophysiographic region.
3
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
W/d
The design riffle width (W) is obtained from:
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
W = A(
)
and the associated mean depth (d) from A/W.
Similarly maximum riffle depth (dm) is given by:
dm = d (dm /d)
(1)
(2)
The planform ratios for the reference reach ? /W, r/W and z/W when multiplied
by the predicted width (W) define the wavelength ( ? ), radius of curvature (r) and
meander arc length (z).
The channel slope is prescribed by the ratio of the local valley slope divided by
the sinuosity. The latter is derived from the predicted plan form.
vi) Design evaluation
Checks are undertaken to determine whether the proposed design will transmit
the largest clast supplied from upstream at bankfull flow (Rosgen 1998). Experience
indicates that this is prescribed by the largest exposed stone on the lower third of a
depositional point bar in the affected reach. Should the design be incompetent to
transmit this stone, or it is computed it can transmit larger material, then it would be
necessary to adjust the bankfull depth to achieve the required competence. Care has to
be exercised to ensure that the change in width/depth ratio does not, as a consequence,
fall outside the observed range for that stream type. In the event that this occurs, the
width would also need to be adjusted and, thereby, the plan form and slope would be
altered.
EVALUATION OF FLUVIAL GEOMORPHOLOGICAL PROCEDURE
Provided the shape parameters at the reference reach correspond to those required
at the restoration reach and the scaling parameter is correctly determined, the procedure
must accurately prescribe the design dimensions, pattern and profile of any stable river.
UK regime equations (Hey and Thorne 1986) can be used to evaluate the
procedures for defining the scaling and shape parameters for natural meandering gravel
and cobble-bed rivers flowing through alluvial flood plains; respectively C3 and C4
stream types (Hey 2004). As these rivers exhibit small scale roughness and
predominantly transport their bed material as bedload, the flow processes controlling
their morphology are identical and, therefore, one set of equations will apply to both C3
and C4 stream types.
i) Scaling parameter.
The use of regional curves to define the bankfull cross-sectional area of the restoration
reach assumes that there is no change in its value following a period of instability
4
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
provided that there is no systematic change in the flow and sediment transport regime of
the river. This assumption can be evaluated using the UK regime equations. The
bankfull cross-sectional area (A) is given by:-
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
A = h Q 0.87 D50 -0.11
(3)
where Q is the bankfull discharge (m3 s-1 ), D50 is the median grain size (m) of the surface
bed material and h a coefficient which depends on bank vegetation density (0.95 for ? 1%
trees and shrubs, category I; 0.73 for 1-5%, category II; 0.60 for 5-50%, category III and
0.52 for ? 50%, category IV).
Provided there is no change in the streamtype following a period of instability,
then there must be either no change in any of these boundary condition variables or any
changes must compensate each other if the cross sectional area is variable or unchanged.
If the river establishes a new stable state without the flow and sediment transport
regimes being affected, then the effective discharge will be unchanged, and, thereby, the
bankfull flow. Similarly the bed material size will be invariant if sediment delivery from
the catchments is maintained. Bankside vegetation is likely to alter following
destabilisation, however provided the river is restored with a comparable bank vegetation
density, the original bankfull cross-sectional area can be re- instated.
If a different stream type is required for the restoration or the same stream type
but with a different effective bank vegetation density, the cross-sectional area would
differ. This indicates that regional curves need to be stratified by stream type and by
bank vegetation densities, since both influence W/d ratios. Given the bankfull discharge
for a particular location in the basin, the cross sectional area will be dependent on the
velocity of flow ie. directly vary with W/d ratio.
ii) Transference of W/d ratio
As dimensionless shape ratios from the reference reach are imposed on the
restoration reach, it is important to determine the particular circumstances that enable this
transfer to occur. To assess whether this constrains the choice of reference reach, the UK
regime equations were analysed to determine under what specific conditions the W/d at
the reference reach would match that at a restoration site. These equations indicate that:W/d = kQ 0.13 D50 0.11
(4)
where Q and D50 are as defined previously. The coefficient k depends on bank
vegetation density: 19.68 for category I, 15.14 for category II, 12.41 for category III, and
10.64 for category IV.
Equation 4 indicates that there are various combinations of bankfull discharge
(Q r), bed material size (Dr), and bank vegetation category ratios (kr) between the riffle
sections at the reference and restoration reaches which would result in equivalence of
W/d at the two sites viz:
5
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
Qr = kr7.69 Dr-0.846
(5)
Similarly for maximum depth to mean depth ratios, the UK regime equations
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
give:
dm/d = 0.91 Q-0.01 D50-0.45 D840.35
(6)
where 84% of the surface bed material is less than or equal to D84 (m). As it is virtually
independent of discharge, equivalence of d m/d at the two sites is prescribed by:
D 50r = D84 r 0.555
(7)
iii) Transference of plan form ratios.
As the plan form is obtained by transferring ratios of ( ?/W), ( r/W) and ( z /W)
from the reference back to the restoration site, the sinuosity of the reference reach is
actually imposed on the restoration site. This follows from the definition of sinuosity (p
= channel length/valley length).
Hence
z
2( ) W
2z
2z
W
p=
=
=
z
?
?
( )W
W
(8)
As sinuosity is also equivalent to the ratio of valley to channel slope, then its value will
depend on the gradient of the valley and the boundary condition variables controlling
channel slope. From the UK regime equation for slope, it follows that for C3/C4 stream
types
Sv
S v Q 0.43
p=
=
0.10
0.75
S
0.081Q s
D 50
(9)
where Sv is the valley slope and Qs the bankfull bed material load (kg s-1) estimated by
the Parker et al. (1982) equation.
The sinuosity of the reference reach will match that at the restoration site when
the combinations of bankfull discharge, bed material load, bed material size and valley
slope ratios are prescribed by:
Q r = D r 1.744 S v r ?2.326 Qs r 0.233
(10)
6
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
iv) Predictive capability
The foregoing suggests that the UK regime equations could provide guidance on
the choice of an appropriate reference reach for restoration design (eqs 5 and 10). To
evaluate these equations, US data for stable C3 and C4 reference reaches were obtained
from Rosgen. Five sites were on C3 rivers and four on C4. The sites were drawn from a
range of hydrophysiographic regions, valley types and basin areas. Each site was chosen
in turn as the ?restoration? reach and the others used as reference reaches to predict its
dimension, pattern and profile. This gave 20 different combinations for the C3 stream
types and 12 for the C4 (Hey, 2004).
First the geomorphological procedure was applied using the actual cross-sectional
area at the restoration site in order to isolate the error in transferring the shape parameters
(W/d and p) from the reference to restoration reach. The following conclusions can be
drawn:(a) bankfull width and mean depth can be predicted within � provided the W/d ratios
at the two sites were within �%.
(b) where the W/d ratio at the two sites were within �%, their boundary condition
ratios (Q r, Dr and bank vegetation density) conform to those prescribed by the UK regime
equations.
(c) it is preferable to design the planform of the river by transferring sinuosity, rather
than scaling it from ?/W, r/W and z/W, in order to avoid additional errors associated with
any inaccuracies in predicting channel width.
(d) sinuosity can be successfully transferred from the reference reach to the restoration
site provided that the boundary condition ratios (Q r, Q sr , Sv r , Dr) are in general.
conformance with UK regime equations
(e) for sites where the transferred p was within �7% of the value at the restoration site,
back calculated values of Qr, given Q s , S vr and Dr, were within +248 and -71% of the
r
actual Qr ratio (average +37%). Better predictions would be expected with an improved
regime slope equation.
(f) slope is predicted from the transferred sinuosity and the local valley slope. Its error
corresponds to that for sinuosity. Compared to more conventional engineering
approaches for predicting channel slope, this is commendably accurate.
Second, to evaluate scaling errors, the exercise was repeated using the bankfull
cross-sectional area at the restoration site obtained from regional curves. This indicated
that:a) width and depth are, inevitably, overestimated if the cross-sectional area is overpredicted and vice versa. This can exacerbate or mitigate overall error depending on
shape discrepancies.
b) scaling error does not affect predicted sinuosity.
7
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
CHOICE OF REFERENCE REACH
The choice of reference reach is obviously crucial to the success of the restoration
design using the geomorphological procedure. Published practice stipulates that the
reference reach should be of the same stream type as that required for the restoration site
be vertically and laterally stable, be in the same hydrophysiographic region and have the
same valley type. The forego ing analysis indicates that these criteria do not necessarily
fulfil the requirements for transference of the shape parameters. For transference, the
boundary condition ratios have to be unity or have a specific set of values. Provided the
basin areas of the two sites are comparable, then these ratios would approximate unity
and the W/d and p could be transferred with minimum error. Should the bankfull crosssectional area at the two sites be identical, then the reference reach could be copied
without scaling. If the reference reach has a different basin area, significant errors could
arise. This will occur should their boundary condition ratios preclude equivalence of
W/d or p at the two sites.
It is apparent that existing criteria for choosing a reference reach are overly
restrictive. Provided that the boundary condition ratios conform to those prescribed by
an appropriate set of regime equations for equivalence of W/d and p there are no
restrictions with regard to hydrophysiographic region and valley type or even for basin
area. These equations are likely to vary by stream type as they reflect specific channel
forming processes.
Clearly the reference reach has to be chosen such that the boundary condition
ratios fulfil the requirement for the transfer of both W/d and p. For C3/C4 stream types
the following steps should be taken:
i) Sinuosity equivalence
Using eq (10), calculate the Qr ratio given the observed Sv r and Dr and the
predicted Q sr (Parker et al. 1982 eq). Compare this with the actual Qr. Ideally it should
closely correspond for the reference reach to be suitable. However, because the regime
slope equation has a relatively low r2 (68%), equivalence is shown to occur when the
error between the calculated and actual Qr is within �0%. Should the error exceed
this, reject the site as a reference reach.
Although the design sinuosity obtained from the reference reach represents the
best practical estimate, it is evident that it may not perfectly match that required at the
restoration site. Provided the banks of the restored river are protected from erosion, the
river will adjust its depth to accommodate the imposed pattern and profile. To minimise
such adjustments, the design sinuosity needs to match that required as closely as
possible.
8
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
ii) W/d equivalence
Using eq (5) establish whether Qr, Dr and bank vegetation density ratios lie within �%
limits for equivalence of W/d ratio. If not, reject as a reference reach. This indicates
that, given the bank vegetation density at the reference reach, equivalence of W/d can be
achieved with a wide range of combinations of Qr and Dr. This occurs because bank
protection at the restoration site can ensure that the effective bank vegetation density can
be ascribed any value. If it is necessary to minimise the W/d ratio at the restoration site
(vegetation density IV), then the Dr and Qr ratios, plus the bank vegetation density at the
reference site are more closely prescribed.
iii) Specifying bankfull cross-sectional area at restoration reach.
Ideally this should be obtained from the section immediately upstream of the restoration
reach provided it is of the required stream type. Failing that it should be derived from a
regional curve for its particular hydrophysiographic region. To minimise error, the
regional curve should be stratified by stream type and, for C3-C4 stream types, also by
bank vegetation density. This recognises that channels having a lower W/d ratio can
transmit their bankfull discharge with a higher velocity and, hence, smaller bankfull
cross-sectional area. Should its cross-sectional area correspond to that at the reference
reach, the latter is a blueprint for the restoration site and can be copied. If it is larger, the
cross-sectional dimensions are scaled up and, vice versa, should it be smaller.
CONCLUSIONS
Although this analysis was undertaken for C3-C4 stream types, the geomorphological
procedure will enable the dimension, pattern and profile of all types of stable river to be
designed provided that the chosen reference reach is of the requisite stream type, stable
and similar to that required for the restoration. For this to be achieved, the rivers must
have specific boundary condition ratios. These are prescribed by appropriate regime
equations.
Until regime equations are developed for other stream types, alternative
guidelines are required to identify suitable reference reaches. In the light of this study,
Rosgen (personal communication) no longer requires the reference reach to be in the
same hydrophysiographic region as the restoration site. In addition to being stable, of the
same stream type as the restoration reach and be in the same valley type (width, slope,
depositional materials, landform/landtype associations), it now has to have a dense
riparian vegetation (i.e. bank vegetation density IV). Empirical evidence indicates that
these constraints ensure that the range of possible W/d ratios and sinuosities is very
limited. Any small errors that may occur in their transferred values will be
accommodated by slight bed scour or fill provided that the design pla nform is maintained
by bank protection measures.
9
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
REFERENCES
Downloaded from ascelibrary.org by University Of Florida on 10/28/17. Copyright ASCE. For personal use only; all rights reserved.
Hey, R.D. (1978) Determinate hydraulic geometry of river channels. Jour. Hydraulics
Div., ASCE 104(HY6), 869-885.
Hey, R.D. (2004) Fluvial geomorphological methodology for natural stable channel
design. Jour. American Water Resource Association
Hey, R.D. and Thorne, C.R. (1986) Stable channels with mobile gravel-beds. Jour.
Hydraulic Eng. ASCE, 112(8), 671-689.
Parker, G., Klingeman, P.C. and McClean, D.G. (1982) Bedload and size distribution in
paved gravel-bed streams. J. Hydraulic Div., ASCE, 103(HY4), 544-571.
Rosgen, D.L. (1994) A classification of natural rivers. Catena, 22, 169-199.
Rosgen, D.L. (1998) The reference reach - a blueprint for natural channel design. In:
Proc. Wetland Engineering and River Restoration Conference, ASCE, CDROM.
10
Copyright ASCE 2004
Critical Transitions in Water and Environmental Resources Management
World Water Congress 2004
Документ
Категория
Без категории
Просмотров
2
Размер файла
92 Кб
Теги
29409, 40737, 282004
1/--страниц
Пожаловаться на содержимое документа