PROGRAMS & PLANS--Estimating flow characteristics from channel size

In Reply Refer To:                                         March 26, 1975
EGS-Mail Stop 415


Subject:  PROGRAMS & PLANS--Estimating flow characteristics from
                            channel size

Channel size has been used by WRD in estimating mean flow of
streams in arid and semiarid regions of western U.S. for 6 or 8
years.  More recently the value of channel size for estimating
flood-peak characteristics has been demonstrated.  More than half
a dozen Districts have used the method and several applications
have been published.  Most of this work was done by interested
hydrologists working more or less independently of each other.
Two years ago the Surface Water Branch began coordinating this
effort.  This coordination consisted of field studies, reanalysis
of data, convening of a meeting on standardization of methods and
terminology in Denver, April 22-24, 1974, and organization of a
test in northern Wyoming to determine how consistently channel
measurements could be made by different individuals.

This memo (1) suggests appropriate uses for the method, (2)
presents guidelines for data collection, analysis, and
application, (3) emphasizes that field guidance must be obtained
from a qualified consultant, (4) encourages improvements in these
methods, and (5) reaffirms the coordinating function of the
Surface Water Branch.

The channel-width method is appropriate for use where flow
characteristics are poorly related to drainage area and other
basin characteristics.  These include arid and semiarid areas and
those areas where geology plays a dominant role in determining the
flow characteristics.

In areas having these characteristics, specific uses include:

1.  Estimation of mean flows of perennial streams when time does
not permit making monthly discharge measurements for a year.

2.  Estimation of mean flows of ephemeral streams.

3.  Estimation of flood-peak characteristics.

4.  Presentation of a relation of a flood-peak characteristic to
channel width in a regional flood-frequency report.  The relation
might be the only one presented for a particular part of the
regional study, or it might be given as an alternate to regression
on basin characteristics.  Users of such a report should be
advised to obtain field training before applying the relation, or
to call on experienced hydrologists in WRD to estimate the
characteristics at specific sites.

Before making channel measurements the attached "Guidelines for
Using Channel Size as an Indicator of Flow Characteristics, 1975"
should be studied, and a few days of field instruction should be
obtained from a man experienced in the technique.

Districts should inform the Surface Water Branch when they begin a
project of this type.  And all reports using channel size for
estimating flow characteristics are to be reviewed in the Surface
Water Branch.

The variability of channel measurements made by 7 experienced
individuals was defined on 22 test reaches in Wyoming in Autumn
1974.  A tentative analysis was based on the assumption that the
average of the logarithms of the 7 widths in each reach is the
best estimate of the width in that reach.  Using this assumption
the analysis indicates that variability in measuring channel width
would result in a standard error in discharge of 30-35 percent if
the model error were zero.  Since model error is not zero, the
standard error of a discharge estimate would be appreciably

Further research is needed; promising aspects include use of
channel slope, an index of channel shape, and some index of bed
and bank material in the relations.  Districts are encouraged to
review published work on channel morphology and to seek improved

                                       Harry H. Barnes, Jr.
                                       Chief, Surface Water Branch


WRD Distribution:  A, B, FO

Guidelines for Using Channel Size as an Indicator of Flow
Characteristics, 1975

Selection of Reach

The selection of a suitable reach in which to measure channel
dimensions is important.  The following guidelines represent the
ideal reach and can seldom be satisfied; there can be no
substitute for good judgment in selecting the best reach

1.  A straight, narrow reach in which flows are approximately
uniform is preferred.

2.  Bed and banks in the reach should be stable but should be of
material that has permitted the channel to develop into a normal
size and shape for the flow regimen (but the flow regimen should
not be known at time of reach selection and channel measurement!).
Bedrock channels may not be representative; they should not be

For large rivers, low-level aerial photography may be helpful in
locating a suitable reach.

Selection of Cross Section

Channel dimensions at 3 levels in a channel may be measured.  The
section defined by the lowest channel bars is most commonly
related to mean flows, and the sections defined by the active
channel width and the main channel width are indices to flood
characteristics.  The terminology and descriptions of these 3
levels are still tentative.

Section defined by lowest channel bars.--This section has been
described by Moore (1968), Hedman (1970), and others.  The
reference level is defined by the tops of point or channel bars
where these are (a) bars of which the particles are moved annually
and (b) the lowest prominent channel bars.  In some channels these
bars do not exist.  In some other channels these bars will be
below the water surface for most of the year.

The above description refers to perennial streams.  In ephemeral
streams, particles may not be moved annually but will be moved by
significant flows.  Care must be taken to avoid selection of
erosional features in ephemeral streams.  Consistency in
identifying this section in ephemeral channels is difficult to

Active channel section.--This section has been described by Hedman
and Kastner (1974) and by Hedman, Kastner, and Hejl (1974).

The active channel is the lower portion of the channel
entrenchment in the flood plain that is actively involved in the
transportation of water and sediment during the usual regime of a
stream.  Depositional features within the active channel are
altered and shifted regularly during the normal.fluctuation of
streamflow.  Beyond the boundaries of the active channel, the
geomorphic features are relatively permanent and generally are
vegetated.  The sides of the active channel occur as steeply
sloping banks in straight reaches and as stabilized point bars on
the inside or convex side of a channel bend.  The reference point
used in measuring the geometry of the active channel is selected
at the upper edge of the banks or point bars where they abruptly
change to a flatter slope.  This reference point is above and
shoreward from the reference point defined by the lowest channel
bars.  Annual vegetation generally is present above the reference
point and can often be used as a clue in identifying this point.
However, judgment is required in using the vegetation line as a
guide.  In regions where high flows are infrequent, some grasses
and sedges may grow part of the way down the banks or even into
the water.  On the other hand, in extremely arid regions the banks
may be completely devoid of vegetation.

If the selected reach meets the listed criteria, the active
channel width will be the width of the low-water channel because
there will be no. bars in a narrow channel.  As the stream widens
bars will form on one or both sides of the channel and the width
between these bars or between one bar and the opposite bank (if
only one bar is present) will be about the same as the width of
the low-water channel in a restricted section.  These bars will
become vegetated in contrast with those describing the lower

Main-channel section.--The reference level for this section is
variously defined by breaks in bank slope, by the streamward edges
of the flood plain, or by the lower limits of permanent
vegetation.  On perennial streams it is virtually the same as the
flood-plain level described by Emmett (1972, 1973) and by Leopold,
Wolman, and Miller (1964).  In selecting the main channel width,
one should avoid a high reference level (such as a terrace) that
does not reflect the present flow regimen.  This is most often a
possibility on ephemeral streams, especially if the channel has
been downcutting.


1.  Measurements should be made by someone with field experience
in selecting the features; and measurements should be made without
knowledge of flow characteristics at the sites.

2.  At selected reference level for each type of section measure
width to two significant digits.

3.  Measure depth at sufficient verticals to define shape of

4.  Measure at 2 or 3 cross sections separated by at least one
channel width if suitable ones are found.

5.  If practical, measure local channel bed-slope or water-surface
slope and record how determined (hand level, transit-stadia,

6.  Photograph the reach and cross section, including the feature
defining the edges of the section.

7.  Record type of bed and bank material; sand, gravel, broken
rock, clay, etc.  For gravel or rock, determine approximate size

8.  Rate the suitability of the section.

9.  Record location of the section.

Analysis of Data

The purpose is to define an estimating relation.

1.  For preliminary work, graphically relate the flow
characteristic to width.  If the flow characteristic is not well
defined, plot the probable range rather than the point value.
Identify on the graph any width measurements that are of
questionable reliability.  Draw the relation line so that most
weight is given to the reliable data; a least squares fit to the
data points is not necessarily the best interpretation.  Channel
depth is ordinarily not a significant variable but the ratio of
depth to width may be if channels of unusual shapes are included
in the data.

Flood characteristics should be related to active-channel width or
to whole-channel width, not to the width of the lowest channel

2.  Compare relation obtained in (1) with relations applicable to
nearby regions.  If the relations and the streams are similar,
consider using one relation applicable to a large region.  A
single relation for flood characteristics has wide areal
applicability whereas a relation to mean flow is much more

3.  Define separate relations of flood characteristics to widths
of active-channel and whole-channel if data are available.

4.  Do not combine basin and channel parameters in one relation.

5.  Include basic data in the report.

6.  Identify channels that don't fit the relations and look for
reasons.  Keep a record of these.

7.  Original ideas and analyses are encouraged, but a user report
should only document defined relations; it should avoid describing
how to measure a channel.

8.  Reports for field use will be reviewed in the Surface Water
Branch to maintain some uniformity.

Application of Relations

1.  This should be done by someone with field experience in
selecting the features; if this experience is not available
consulting services should be sought.

2.  Look for suitable reach.  If there is not a suitable one near
the point of interest, it may be possible to find ones some
distance upstream and downstream; then the discharge at the point
of interest can be interpolated.

3.  Select one to three cross sections in each reach and measure
as previously described at one or more reference levels in each
cross section if these levels can be identified.  Commonly it will
not be possible to identify all three levels at one cross section;
in some sections only one reference level can be identified with

4.  Determine discharge from the appropriate relation.


(Some describe recent work but these should not be used as models
because they do not conform strictly to the guidelines.)

Emmett, W. W., 1972, The hydraulic geometry of some Alaskan
streams south of the Yukon River:  USGS open-file report, 102 p.

Emmett, W. W., 1973, The channels and waters of the upper Salmon
River area, Idaho:  USGS open-file report, 467 p.

Hedman, E. R., 1970, Mean annual runoff as related to channel
geometry of selected streams in California:  USGS Water-Supply
Paper 1999 E.

Hedman, E. R. and Kastner, W. M., 1974, Progress report on
streamflow characteristics as related to channel geometry of
streams in the Missouri River basin:  USGS open-file rept., 24 p.

Hedman, E. R., Kastner, W. M., Hejl, H. R., 1974, Selected
streamflow characteristics as related to active-channel geometry
of streams in Kansas:  Kansas Water Resources Board Tech. Rept.
no. 10, 21 p.

Hedman, E. R., Moore, D. 0., Livingston, R. K., 1972, Selected
streamflow characteristics as related to channel geometry of
perennial streams in Colorado:  USGS open-file rept.

Langbein, W. B., 1971, Estimation of annual runoff from channel
surveys:  USGS Water Resources Bull., Apr. - Sept. 1971, p. 3-5.

Leopold, L. B. and Maddock, Thomas, Jr., 1953, The hydraulic
geometry of stream channels and some physiographic implications:
USGS Prof. Paper 252.

Leopold, L. B., Wolman, M. G., Miller, J. P., 1964, Fluvial
processes in geomorphology:  San Francisco, W. H. Freeman, 522 p.

Moore, D. 0., 1968, Estimating mean runoff in ungaged semiarid
areas:  IAHS Bull., V. XIII, no. 1, p. 28-39.

Moore, D. 0., 1974, Estimating flood discharges in Nevada using
channel-geometry measurements:  Nevada Highway Dept., Hydrologic
Rept. no. 1, 43 p.

Riggs, H. C., 1974, Flash flood potential from channel
measurements:  IAHS Publ., no. 112, Proc. of Paris Symposium, p.

Wahl, K. L., 1974, Relating streamflow characteristics to channel
cross-section properties - a literature review:  Unpublished but
available on request from Surface Water Branch, 36 p.