Вы находитесь на странице: 1из 33

National Center for Earth-surface Dynamics

Stream Restoration Program

Enabling Landscape Sustainability

A tool from the NCED


Stream Restoration
Toolbox:

The Gravel River


Bankfull Discharge
Estimator
Gary Parker, 10/2004

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


CAVEAT

This tool is provided free of charge.


Use this tool at your own risk.
In offering this tool, none of the following accept responsibility or liability for
its use by third parties:
the National Center for Earth-surface Dynamics;
any of the universities and institutions associated with the National Center
for Earth-surface Dynamics; or
any of the authors of this tool.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability

The Gravel River Bankfull


Discharge Estimator
This tool consists of an equation to estimate bankfull discharge in an
undisturbed (reference) reach of a single-thread, mobile-bed gravel-bed
stream from measured channel characteristics.
River bankfull discharge is a key parameter for estimating channel geometry.
A knowledge of bankfull discharge is necessary for the evaluation and
implementation of many river restoration projects.
The best way to measure bankfull discharge is from a stage-discharge
relation. Bankfull discharge is often estimated in terms of a flood of a given
recurrence frequency (e.g. 2-year flood, or a flood with a peak flow that has a
50% probability of occurring in a given year; Williams, 1978).
In some cases, however, the information necessary to estimate bankfull
discharge from a stage-discharge relation or from flood hydrology may not be
3
available. The tool presented here provides an alternative estimator.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


RIVERS ARE THE AUTHORS OF THEIR OWN GEOMETRY
Given enough time, rivers construct their own channels.
A river channel is characterized in terms of its bankfull geometry.
Bankfull geometry is defined in terms of river width and
average depth at bankfull discharge.
Bankfull discharge is the flow discharge when the river is just
about to spill onto its floodplain.
A river restoration scheme is likely to become more successful in
a shorter amount of time if it takes into account the natural
bankfull geometry of a channel.
This tool allows estimation of bankfull discharge of a singlethread gravel-bed river with a definable floodplain that
actively moves the gravel on its bed from time to time.
Bankfull discharge is estimated from measured bankfull
channel characteristics.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


THIS TOOL IS FOR GRAVEL-BED STREAMS

Raging River, Washington,


USA: a gravel-bed river

Little Wekiva River,


Florida, USA: a sandbed river.

This tool addresses gravel-bed streams. Typical gravel-bed streams


have bed surface median sizes Ds50 in the range from 8 to 256 mm.
Boulder-bed streams have median sizes in excess of 256 mm. Sand5
bed streams have median sizes between 0.062 and 2 mm.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


THIS TOOL ADDRESSES SINGLE-THREAD RATHER
THAN MULTIPLE-THREAD RIVERS

Raging River, Washington,


USA: a single-thread
gravel-bed river

Sunwapta River, Canada: a


multiple-thread (braided)
gravel-bed river

This tool addresses single-thread streams. A single-thread stream has


a single definable channel, although mid-channel bars may be present.
A multiple-thread, or braided stream has several channels that intertwine
6
back and forth.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


THIS TOOL ADDRESSES MOBILE-BED RATHER THAN
THRESHOLD CHANNELS

Raging River, Washington,


USA: a mobile-bed river

Trinity Dam on the Trinity River,


California, USA. A threshold channel
forms immediately downstream.

This tool addresses mobile-bed gravel streams. Such streams are


competent to modify their beds because they mobilize all or nearly all
gravel sizes on the bed from time to time during floods. Threshold
channels are defined in the next slide.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


THRESHOLD CHANNELS
Threshold gravel-bed channels
are channels which are barely not
able to move the gravel on their
beds, even during high flows.
These channels form e.g.
immediately downstream of dams,
where their sediment supply is cut
off. They also often form in urban
settings, where paving and
revetment have cut off the supply
of sediment. Threshold
channels are not the authors of
their own geometry. The
relations presented in this tool do
not apply to them.

Trinity Dam on the Trinity River,


California, USA. A threshold
channel forms immediately
downstream.
8

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


PARAMETERS USED IN THIS TOOL
This tool uses the following parameters:
1. Bankfull discharge Qbf in cubic meters per second (m3/s) or cubic
feet per second (ft3/s);
2. Bankfull channel width Bbf is meters (m) or feet (ft);
3. Bankfull cross-sectionally averaged channel depth Hbf in meters (m)
or feet (ft);
4. Down-channel slope S (meters drop per meter distance or feet drop
per feet distance).
5. Bed surface median grain size Ds50. This parameter is usually
measured in millimeters (mm); the value must be converted to
meters or feet in using the tool presented here.
These parameters are defined before the tool is introduced. If you are
familiar with the parameters, click the hyperlink to jump to the tool.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


BANKFULL PARAMETERS: THE RIVER AND ITS
FLOODPLAIN
floodplain

A river constructs its own


channel and floodplain.

channel

At bankfull flow the river is on the verge of spilling out onto its
floodplain.

10

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


THE DEFINITION OF BANKFULL DISCHARGE Qbf

Let denote river stage (water surface elevation in


meters or feet relative to an arbitrary datum) and Q
denote volume water discharge (cubic meters or feet
per second). In the case of rivers with floodplains,
tends to increase rapidly with increasing Q when all the
flow is confined to the channel, but much less rapidly
when the flow spills significantly onto the floodplain.
The rollover in the curve defines bankfull discharge Qbf.
The floodplain is
often somewhat
poorly-developed in
mountain gravel-bed
streams. Bankfull
stage, however, can
often still be
determined by direct
field inspection.

Qbf

Minnesota River and flooded


floodplain, USA, during the
record flood of 1965
11

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


CHARACTERIZING BANKFULL DISCHARGE Qbf
Bankfull discharge Qbf is used as a shorthand for the characteristic flow
discharge that forms the channel.
One way to determine it is by means of direct measurement of the flow in
a river. Since bankfull flow is not frequent, this method may be
impractical in a river restoration scheme.
Another way to estimate it is from a stage-discharge curve, as described
in the previous slide. In order to implement this, the river must be
gaged near the reach of interest.
Another way is to estimate it using stream hydrology. It has been found
that in gravel-bed streams bankfull flow is often reasonably
estimated in terms of a peak flood discharge with a recurrence
of 2 years (e.g. Williams, 1978 ). This corresponds to a flow
discharge that has a 50% probability of occurring in any given year.
When none of the above methods can be implemented, Qbf can be estimated
12
from bankfull channel characteristics, as described in this tool.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


CHARACTERIZING BANKFULL CHANNEL GEOMETRY:
BANKFULL WIDTH Bbf AND BANKFULL DEPTH Hbf
Bankfull geometry is here defined in terms of the average characteristics
of a channel cross-section at bankfull stage, i.e. when the flow is at
bankfull discharge. Here the key parameters are:
bankfull width Bbf and
cross-sectionally averaged bankfull depth Hbf.
These parameters should be determined from averages of values
determined at several cross-sections along the river reach of interest.

Bbf
Hbf
13

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


CAVEAT: NOT ALL RIVERS HAVE A DEFINABLE
BANKFULL GEOMETRY!
Rivers in bedrock often have no
active floodplain, and thus no
definable bankfull geometry.
Highly disturbed alluvial rivers are
often undergoing rapid
downcutting. What used to be the
floodplain becomes a terrace that
is almost never flooded. Time is
required for the river to construct a
new equilibrium channel and
floodplain.

Wilson Creek, Kentucky: a


bedrock stream. Image
courtesy A. Parola.

Reach of the East


Prairie Creek,
Alberta, Canada
undergoing rapid
downcutting due to stream straightening. Image
courtesy D. Andres.

The tool presented here cannot be used to estimate bankfull discharge from bankfull
channel characteristics if a) there is no floodplain or b) the channel is so disturbed
14
that it is no longer interacting morphologically with its floodplain.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


FIELD MEASUREMENT OF BANKFULL CHANNEL
GEOMETRY
Not all field channels have definable bankfull geometries. Even when a
channel does have a definable bankfull geometry, some experience and
judgement is required to measure it. In the future a worked example
complete with photographs and data files will be added to the toolbox.
Until this is done, the user is urged to spend some time to determine how
bankfull geometry should be determined.

Bbf
Hbf
15

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


CHARACTERIZING BED SEDIMENT IN GRAVEL-BED
STREAMS: MEDIAN SURFACE SIZE Ds50
Armored surface

Gravel-bed streams usually show a


surface armor. That is, the surface
layer is coarser than the substrate
below.

substrate
Bed sediment of the River Wharfe,
U.K., showing a pronounced surface
armor. Photo courtesy D. Powell.

16

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


SURFACE AND SUBSTRATE MEDIAN SIZES
Here the surface median size is denoted as Ds50 and the substrate median
size is denoted as Dsub50. The surface is said to be armored when Ds50/Dsub50
> 1.
This ratio also provides a rough estimate of ability of the stream to move its
own gravel. Low values of Ds50/Dsub50 (e.g. < 1.3, i.e. relatively weak
armoring) are generally indicative of relatively high mean annual sediment
transport rates, whereas high values of Ds50/Dsub50 (e.g. > 4, relatively strong
armor) are generally indicative of relatively low mean annual sediment
transport rates (Dietrich et al., 1989).
Notes on bed sampling, grain size distributions and the determination of
median sediment size are given in and Appendix (slides 21-27) toward the
end of this presentation. To jump to them click the hyperlink bed sampling.
17

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


CHARACTERIZING DOWN-CHANNEL SLOPE S
Down-channel bed slope should be determined from a survey of the long
profile of the channel centerline. The reach chosen to determine bed slope
should be long enough to average over any bars and bends in the channel,
which are associated with local elevation highs and lows.
plan view
A
B
A
bed
elevation

S
long profile of centerline bed elevation
down-channel distance

B
18

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


DATA BASE FOR THE TOOL
The bankfull discharge predictor presented here was developed by Parker
(2004) from the following data base for gravel-bed streams.
16 stream reaches flowing from the Rocky Mountains in Alberta, Canada
(Kellerhals et al., 1972);
23 mountain stream reaches in Idaho (Parker et al., 2003);
23 upland stream reaches in Britain (mostly Wales) (Charlton et
al.1978);
10 reaches along the upper Colorado River, Colorado (Pitlick and Cress, 2002)
(Each reach represents an average of several subreaches.)
The original data for Qbf, Bbf, Hbf, S and Ds50 for each reach can be found in
the companion Excel file, ToolboxGravelBankfullData.xls.
The predictor was further tested with a set of 11 reaches in Maryland/
Pennsylvania, USA (McCandless, 2003) and 62 reaches of British streams (Hey
and Thorne, 1986).
19

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


RANGE OF PARAMETERS USED TO DEVELOP THE TOOL

Among all four sets of data, the range of parameters is as given below:
Bankfull discharge

Qbf (in meters3/sec)

2.7 ~ 5440

Bankfull width

Bbf (in meters)

5.24 ~ 280

Bankfull depth

Hbf (in meters)

0.25 ~ 6.95

Channel slope

0.00034 ~ 0.031

Surface median size

Ds50 (in mm)

27 ~ 167

These ranges define the range of applicability of the tool.


20

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


THE TOOL
The tool consists of the following relation. Where g = the acceleration of
gravity (9.81 meters/second2 or 32.2 feet/second2),

Qbf = 3.732 BbfHbf

Hbf

gHbf S
Ds50

0.2645

That is, if Bbf, Hbf, S and Ds50 can be determined from field measurements,
Qbf can be estimated from the above relation.
The tool is implemented as an Excel spreadsheet in the next slide.
21

Caution: use the relation subject to the caveats of Slides 5, 6, 7, 8 and 14!

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


IMPLEMENTATION OF THE TOOL
Qbf = 3.732 BbfHbf

Hbf

gHbf S
Ds50

0.2645

Stop the slide show and double-click the Excel spreadsheet to activate it. If you type
in the indicated input parameters in the indicated units, Qbf is computed as output.
SI Units
Input
Hbf
2.50
Bbf
30.00
S
1.50E-03
Ds50
60.0
Calculated
Ds50
0.0600
Output
Qbf
144.0

m
m
mm
m
m3/s

English Units
Input
Hbf
8.20
98.4252 Bbf
98.40
S
1.50E-03
Ds50
60.0
Calculated
Ds50
0.1969
Output
Qbf
5082.0

ft
ft
mm
ft
ft3/s
22

Caution: use the relation subject to the caveats of Slides 5, 6, 7, 8 and 14!

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


ACCURACY OF THE TOOL
The plot shows the values of Qbf predicted by the tool versus the reported
(observed) values for the data set used to develop the tool. In 93% of all
reaches the predicted value is between half and twice the reported value.

Predicted Qbf (m3/s)

10000

1000

predicted
equality
1/2
2

100

10

1
1

10

100

1000

Observed Qbf (m3/s)

10000

23

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


APPENDIX: SEDIMENT SIZE DISTRIBUTIONS IN GRAVELBED STREAMS
Armored surface
Implementation of the regression relations requires a
knowledge of the median size of the surface armor
Ds50. This value must be determined by sampling the
bed. In order to characterize the bed sediment of the
stream the surface and substrate should be sampled
separately. The results of sampling are plotted in
terms of percent finer versus grain size (mm) as
illustrated below.
100

Bed sediment of the River Wharfe,


U.K., showing a pronounced surface
armor. Photo courtesy D. Powell.

Percent Finer

substrate

90
80
70
60

Substrate
Surface

50
40
30
20
10
0
0.1

10
D (mm)

100

1000

24

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


WOLMAN COUNT OF SURFACE SEDIMENT
The simplest way to sample a gravel bed surface is by means of a Wolman
count (Wolman, 1954). The gravel surface is paced, and at set intervals a
particle next to the toe of ones foot is sampled. The sampling should be
chosen so as to capture the spatial variation in bed texture. Grain size is
characterized in terms of the b-axis of a grain (middle axis as measured
with a caliper) or the size of the smallest square through which the grain
will fit. A series grain size ranges is set for estimating the grain size
distribution. In analyzing a Wolman sample, it is necessary to determine
the number of grains in each range. These numbers are used to
determine the grain size distribution. A sample calculation is given in the
live spreadsheet of the next slide.
Wolman sampling is not practical for sand-sized or smaller grains. More
specifically, grains finer than about 4 mm should not be included in a
sample. It should be understood that this method misses the finer grains
25
in the surface.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


GRAIN SIZE DISTRIBUTION FROM WOLMAN COUNT

Percent Finer

The live
spreadsheet to the
right shows a
worked example
for a Wolman
count. Stop the
slide show and
double-click to
activate it. It is
customary to plot
grain size on a
logarithmic scale
when presenting
grain size
distributions.

Measured
Computed
Range mm No. of grains Percent
Size mm Percent finer
128-256
46
24.21%
256
100.00%
64-128
48
25.26%
128
75.79%
32-64
37
19.47%
64
50.53%
16-32
26
13.68%
32
31.05%
8-16
27
14.21%
16
17.37%
4-8
6
3.16%
8
3.16%
4
0.00%
Total no.
190 100.00%
100%
80%
60%
40%
20%
0%
1

10

100

Grain Size mm

1000

26

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


KLINGEMAN SAMPLE OF SURFACE SEDIMENT
The methodology for a Klingeman sample of the surface sediment is
outlined in Klingeman et al. (1979). A circular patch of sediment is specified
on the bed. The largest grain that shows any exposure on the bed surface
is located and removed. All of the bed material (including sand) is then
sampled down to the level of the bottom of the hole created by removing
the largest grain. The resulting sample is analyzed by mass (weight) rather
than number.
A Klingeman sample captures the sand as well as the gravel in the surface
layer. Sampling is, however, more laborious than that required for a
Wolman sample. In addition, several Klingeman samples at different
locations may be needed to characterize the spatial variability of the
surface sediment. A sample calculation is given in the live spreadsheet of
the next page.
27

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


KLINGEMAN SAMPLE OF SURFACE SEDIMENT contd.
18
16
12
13
5
7
2
3
6
2
84

Computed
Percent
Size mm Percent finer
21.43%
256
100.00%
19.05%
128
78.57%
14.29%
64
59.52%
15.48%
32
45.24%
5.95%
16
29.76%
8.33%
8
23.81%
2.38%
4
15.48%
3.57%
2
13.10%
7.14%
1
9.52%
2.38%
0.5
2.38%
kg
0.25
0.00%

100%
Percent Finer

The live spreadsheet to


the right shows a
worked example for a
Klingeman sample.
Stop the slide show and
double-click to activate
it.

Measured
Range mm Mass kg
128-256
64-128
32-64
16-32
8-16
4-8
2-4
1-2
0.5-1
0.25-5
Total mass

50%
0%
0.1

10
Grain Size mm

1000

28

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


BULK SAMPLE OF SUBSTRATE

The substrate may be sampled in bulk. The surface layer is first carefully
stripped off down to the depth of the bottom of the largest particle exposed
on the surface. A bulk sample (e.g. cubical) volume of substrate is then
excavated. According to the guidelines of Church et al. (1987), the mass
(weight) of the sample should be at least 100 times the mass (weight) of the
largest grain contained in the sample. Several such samples may be
needed to characterize the spatial variability of the substrate.
The sample is analyzed in terms of mass (weight) rather than number.

29

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


MEDIAN SIZE
It is useful to characterize a sample in terms of its median size D50, i.e. the
size for which 50% of the material is finer. To do this, find the grain sizes D1
and D2 such that the percentage content F1 is the highest value below 50%
and the percentage content F2 is the lowest percentage above 50%. The
median size D50 is then estimated by log-linear interpolation as:

[n(D 2 ) n(D1 )] (50 F )


D 50 = exp n(D1 ) +
1
F
F

2
1

For example, in the Klingeman sample of slide 13:


D1 = 32 mm, F1 = 45.24%, D2 = 64 mm and F2 = 59.52%. The calculation
of D50 is illustrated in terms of the live spreadsheet below. Stop the slide
show and double-click to activate it.
Input
D1
D2
F1
F2

32 mm
64 mm
45.24
59.52

Output
40.3 mm
D50
30

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


REFERENCES
Charlton, F. G., Brown, P. M. and R. W. Benson 1978 The
hydraulic geometry of some gravel rivers in Britain. Report INT 180,
Hydraulics Research Station, Wallingford, England, 48 p.
Church, M. A., D. G. McLean and J. F. Wolcott 1987 River bed gravels: sampling
and analysis. In Sediment Transport in Gravel-bed Rivers, Thorne, C. R.,
J. C. Bathurst, and R. D. Hey, eds., John Wiley & Sons, 43-79.
Dietrich, W. E., J. W. Kirchner, H. Ikeda and F. Iseya 1989 Sediment supply and
the development of the coarse surface layer in gravel-bedded rivers.
Nature, 340, 215-217.
Ferguson, R. I. 1987 Hydraulic and sedimentary controls of channel pattern. In
Rivers: Environment and Process, K. Richards. ed., Blackwell, Oxford,
129-158.
Hey, R. D. and Thorne, C. R. 1986 Stable channel with mobile gravel bed. Journal
of Hydraulic Engineering, 112(8), 671-689.
Kellerhals, R., Neill, C. R. and D. I. Bray 1972 Hydraulic and
geomorphic characteristics of rivers in Alberta. River Engineering
and Surface Hydrology Report, Research Council of Alberta, Canada,
31
No. 72-1.

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


REFERENCES contd.
Klingeman, P. C., C. J. Chaquette, and S. B. Hammond 1979 Bed Material
Characteristics near Oak Creek Sediment Transport Research Facilities,
1978-1979. Oak Creek Sediment Transport Report No. BM3, Water
Resources Research Institute, Oregon State University, Corvallis, Oregon,
June.
McCandless, T. L., 2003, Maryland Stream Survey: Bankfull Discharge and
Channel Characteristics of Streams in the Allegheny Plateau and the
Valley and Ridge Hydrologic Regions. Report CBFO-S03-01, U.S. Fish
and Wildlife Service, Chesapeake Bay Field Office, May, 33 p.
Parker, G., Toro-Escobar, C. M., Ramey, M. and S. Beck 2003 The effect of
floodwater extraction on the morphology of mountain streams. Journal of
Hydraulic Engineering, 129(11).
Parker, G. 2004 Quasi-universal relations for bankfull hydraulic geometry of
single-thread gravel-bed rivers . In preparation.
Pitlick, J. and R. Cress 2002 Downstream changes in the channel of a large
gravel bed river. Water Resources Research 38(10), 1216,
doi:10.1029/2001WR000898, 2002.
32

National Center for Earth-surface Dynamics


Stream Restoration Program

Enabling Landscape Sustainability


REFERENCES contd.
Williams, G. P. 1978 Bankfull discharge of rivers. Water Resources Research,
14, 1141-1154.
Wolman, M.G. 1954. A method of sampling coarse river bed material.
Trans. Am. Geophys. Union, 35, 951956.

33