Hydrology – From Measurement To Hydrological Information (5)

CHAPTER 5

SURFACE WATER QUANTITY AND SEDIMENT MEASUREMENT

5.1 WATER LEVELS OF RIVERS, LAKES
AND RESERVOIRS
5.1.1 General
Water level, or stage, is the elevation of the water
surface of a stream, lake or other water body relative to a datum (ISO, 1988b), and may be used
directly in forecasting flows, to delineate flood
hazard areas and to design structures in or near
water bodies. When correlated with stream
discharge or with the storage volumes of reservoirs
and lakes, water levels become the basis for computation of discharge or storage records. An expanded
discussion of this topic is given in the Manual on
Stream Gauging (WMO-No. 519).


The site selected for observation of stage should
be governed by the purpose for which the records
are collected and by the accessibility of the site.
Hydraulic conditions are an important factor in site
selection on streams, particularly where water levels
are used to compute discharge records. Gauges on
lakes and reservoirs are normally located near their
outlets, but sufficiently upstream to avoid the influence of drawdown.


5.1.2 Gauges for measurement of stage
[HOMS C71]
5.1.2.1 Non-recording gauges
Several types of non-recording gauges for measuring stage are used in hydrometric practice. The
common gauges are of the following types:
(a) Graduated vertical staff gauge;
(b) Ramp or inclined gauge;
(c) Wire-weight gauge installed on a structure
above the stream;
(d) Graduated rod, tape, wire or point gauge for
measuring the distance to the water surface;
(e) Maximum-stage gauge for obtaining the
elevation of the flood crest by the adherence of regranulated cork to a graduated staff
held in a fixed position with relation to the
datum.


5.1.2.2 Recording gauges
Many different types of continuously recording
stage gauges are in use. They may be classified
according to both mode of actuation and mode of
recording.
A commonly used installation consists of a stilling
well connected to the stream by pipes and a float in
the stilling well connected to a wheel on a recorder
by a beaded wire or perforated tape. In high-velocity streams, it may be necessary to install static
tubes on the end of the intake pipes to avoid drawdown of the water level in the well.


The recorder can either be mechanical or electronic.
Recorders with the wheel linked to a pencil or pen
and the pencil or pen placed on a strip chart moved
by a mechanical clock are still widely used and have
proved to be reliable. The timescale and stage scale
chosen for a particular station will depend on the
range in stage, sensitivity of the stage-discharge
relation, and runoff characteristics of the basin.
Back in the main office the strip chart can be digitized so that the data can be entered into a computer.


The wheel can also be connected directly to an
encoder. The encoder will give out analogue or
digital values that can be read and stored by a data
logger.


Various pressure-actuated recording gauges in
common use operate on the principle that static
pressure at a fixed point in the stream is directly
proportional to the head of liquid above the point.
This relation is described by the following
equation:

Water level = (Pstatic – Patm) C

where Pstatic is the pressure in bar on a fi xed spot in
the water column (one has to make sure that any
dynamic pressure from water movement is not
measured), Patm is the atmospheric pressure in bar
on the surface of the water column, and C is a factor
of the water’s net weight (C = 10.2 for freshwater at
20°C), which changes with the temperature and
salinity of the water. Some gauges use a gas-purge
system to transmit the pressure to the gauge. A
small quantity of air or inert gas (for example, nitrogen) is allowed to bubble through a pipe or tubing
to an orifice in the stream. The pressure of the air or
gas that displaces the liquid in the pipe is then
measured and recorded. Other gauges use pressure
transmitters placed directly into the riverbed.

Compensating for the atmospheric pressure is done
by taking air down a small ventilation tube in the
cable or by measuring it with another pressure
transducer on the surface. The main advantage of
pressure-actuated recorders is that they do not
require a stilling well, although any unfortunate
alignment of the pressure-transducer with respect
to flow can cause significant error, and the gas purge
systems in particular are not sensitive to sediment if
its concentration is in the range normally encountered in a natural setting. Care has to be taken when
placing the pressure transducer or bubble gauge on
the riverbed. It is important to make sure it does
not move and that it is exposed only to static pressure. Compensating for changes in temperature
and atmospheric pressure on the surface is also
critical.


Two kinds of recording gauge that have come into
recent use are those that use ultrasonic or radar
sensors. The ultrasonic sensor is based on the speed
of transit of a pulse of ultrasonic frequency
(>20 kHz), which is emitted by a transmitter located
in a structure over the lake or the river. When the
pulse hits the surface of the water body, it echoes
back to the sensor. The time T which passes from
the moment of emission of the pulse and the
moment of reception of the echo by the sensor is
directly proportional to the distance d between the
sensor and the water surface, and inversely proportional to the speed of the pulse in the air. It can be
calculated as:
T=2d/v (5.2)


As sound speed depends on air temperature, it is
necessary to compensate with a correction factor
to obtain a precise value. The radar sensor is similar to the ultrasonic sensor, but uses high
frequencies (around 20 GHz). It has the advantage
that at the higher frequency the transit speed of
the pulse is not affected by air temperature.
River stage may be recorded on graphical
(analogue) recorders. Alternatively the stage can
be recorded digitally at fi xed or action-triggered
intervals.


5.1.3 Procedures for measurement of
stage
5.1.3.1 Establishment of gauge datum
To avoid negative readings, the gauge should be set
so that a reading of zero is below the lowest anticipated stage. The gauge datum should be checked
annually by levelling from local benchmarks. It is
important to maintain the same gauge datum
throughout the period of record. If feasible, the
local gauge datum should be tied to a national or
regional datum. The precise locations of the benchmarks should be carefully documented.


5.1.3.2 Recording gauges
The graphical, digital, electronic, or telemetering
device recorder is set by reference to an auxiliary
tape-fl oat gauge or to a staff gauge located inside
the stilling well. In addition, a staff, ramp or wireweight gauge set to the same datum is necessary to
compare the water surface elevation in the stilling
well with that of the river. For gauges with gas-purge
systems and no stilling well, the staff, ramp or wireweight gauge in the river should serve as the
reference gauge. Small differences usually will occur
because of velocity past the ends of the intake pipes.
Large differences indicate that the intake pipes may
be obstructed.
5.1.3.3 Winter operation of recording
gauges
(a) Float-actuated – This type of installation
requires a stilling well that must be kept icefree in winter. This can be done by heating the
well with, for example, electricity or gas. Other
devices to prevent freezing within a stilling
well are a temporary fl oor within the well at an
elevation just below the frost line, and a vertical, open-ended tube, large enough in diameter
to receive the fl oat, and containing a layer of
fuel oil on the water surface;
(b) Pressure-actuated air bellows and transducers
– These types of installations require neither
a stilling well nor an operating medium
subject to freezing. However, the tube or
cable going into water has to be protected
from ice.
5.1.4 Frequency of stage measurement
The frequency of recording of water level is
determined by the hydrological regime of the water
body and by the purposes for collecting the data.
At continuous-record gauging stations hourly
recordings are normally sufficient for most rivers.
For measurement in small or flashy streams and
urban catchments, stage has to be recorded more
frequently in order to obtain a sufficiently accurate
hydrograph. In general, it is recommended to record
stage as frequently as possible within the limitations
given by the available battery capacity and data
memory. Installation of water level recorders is essential for streams where the level is subject to
abrupt fluctuations. The non-recording gauge is
frequently used as a part of flood forecasting
systems, where a local observer is available to report
on river stage. For purposes such as fl ood forecasting
or fl ood management, telemetering systems may be
employed to transmit data whenever the stage
changes by a predetermined amount.
For some purposes, the recording of only the maximum stages during fl oods is suffi cient and
maximum-stage gauges are used.

A daily measurement of stage is usually sufficient in lakes and
reservoirs for the purpose of computing changes in
storage. The recording time interval for a particular
station is selected on the basis of the rapidity with
which the stage can change and its signifi cance to
change in discharge. Flashy streams require shorter
time intervals, and large streams allow longer time
intervals (ISO, 1981).


Output from pressure transducers, shaft encoders,
or other devices that provide electronic outputs
representing the stage can also be recorded on electronic data loggers (2.5), or with appropriate
interfaces the data can be telemetered from remote
locations.


5.2 ICE ON RIVERS, LAKES AND
RESERVOIRS
5.2.1 General
Observations of ice conditions on rivers, lakes and
reservoirs are of great interest in regions where ice
formation affects navigation or results in damage to
structures, and where ice jams may form (even to
the extent of damming a major river). The obstruction of streamfl ow by ice can also cause serious local
fl ooding. Long-term data on ice conditions in rivers
are extremely valuable in designing various structures, in studying processes of ice formation and
dissipation, and in developing methods of ice
forecasting.


5.2.2 Elements of ice regime
The most important elements of ice regime to be
recorded are the following:
(a) Dates on which fl ows of fl oating ice are fi rst
observed each winter;
(b) Ratio of the surface area of drifting ice to the
open–water surface (ice cover ratio);
(c) Ratio of the surface area of drifting ice to the
stationary ice surface;
(d) Dates on which ice becomes immovable;
(e) Thickness of ice;
(f) Features of ice destruction;
(g) Dates of ice break-up;
(h) Dates on which the ice on rivers and reservoirs
vanish completely.
5.2.3 Methods of observation
Many of the elements given in 5.2.2 cannot be
measured instrumentally and must be evaluated
subjectively and recorded in descriptive language.
For this reason, it is very important that observers
be well trained and that instructions be clearly
prepared.
The thickness of ice is measured by means of an
auger and a ruler at representative sites. To minimize errors caused by spatial variability in ice
thickness, measurements should be made at a minimum of three points spaced over a distance of at
least 5 m, and the measurements should be averaged. The depth of any snow on top of the ice
should also be measured.
The kilometre signs of navigable rivers or dykes
may be used to identify the locations at which ice
surveys are routinely conducted. Particularly
dangerous conditions (for example, ice jams) must
be identified in relation to other landmarks (for
example, bridges, river regulation structures and
harbours).


Determining some of the characteristics of ice
phenomena can be made by means of regular
photogrammetric surveys from a location on the
shore or by aerial photography. In the case of large
rivers, reservoirs or lakes, aircraft observations of ice
formation or break-up are of great value. They are
also useful in the case of ice gorges when fl ood
warnings are required.


For surveying ice conditions over a reach, a strip
width, s, and a fl ying height, hƒ, can be determined
as a function of focal length, Lƒ, of the camera being
used and the effective width, l, of the fi lm frame,
hƒ = s (Lƒ/l). Because Lƒ is a camera constant that is
approximately equal to 1.0, the strip width is
approximately equal to the fl ying height. By repeat
aerial photography at intervals of a few minutes,
the velocity of the ice drift can be determined along
with the density of cover. If the average ice thickness is known, the ice discharge (throughput) can
also be calculated.


Television and IR remote-sensing data from meteorological and Earth-resource satellites are also

useful for estimating ice conditions on lakes and
reservoirs (Prokacheva, 1975).
5.2.4 Times and frequency of
observations
Observations of the state of the ice are made at
times when the water level is observed, while ice
thickness and snow depth on major rivers, lakes
and reservoirs should be measured at intervals of
5 to 10 days during the critical periods of ice formation and break up. Aircraft observations should be
made, as required, to meet special purposes.
5.2.5 Accuracy of measurement
The measurement of ice cannot be very accurate
because of diffi cult conditions. However, uncertainty of ice thickness measurement should not
exceed 10 to 20 mm or 5 per cent, whichever is
greater.


5.3 DISCHARGE MEASUREMENTS AND
COMPUTATION
5.3.1 General [HOMS E70]
River discharge, which is expressed as volume per
unit time, is the rate at which water fl ows through
a cross-section. Discharge at a given time can be
measured by several different methods, and the
choice of methods depends on the conditions
encountered at a particular site. Normally, the
discharge shall be related to a corresponding water
stage at a gauging station.

The accuracy of the discharge measurement depends
on the length of time required to make the measurement, and the extent to which the stage and the
discharge change during the measurement. Changes
in the downstream conditions during the measurement can influence the result and should be avoided.
5.3.2 Measurement of discharge by
current meters [HOMS C79, C85,
C86, C88, E79]


Measurement of discharge by the velocity-area
method is explained by reference to Figure I.5.1.
The depth of fl ow in the cross-section is measured
at verticals with a rod or sounding line. As the depth
is measured, observations of velocity are obtained
with a current meter at one or more points in the
vertical. The measured widths, depths and velocities permit computation of discharge for each
segment of the cross-section. The summation of
these segment discharges is the total discharge (ISO,
1979b).


5.3.2.1 Selection of site
Discharge measurements need not be made at the
exact location of the stage gauge because the
discharge is normally the same throughout a reach
of channel in the general vicinity of the gauge. Sites
selected for measurements should ideally have the
following characteristics (ISO, 1979b):
(a) The velocities at all points are parallel to one
another and at right angles to the cross-section
of the stream;
(b) The curves of distribution of velocity in the
section are regular in the vertical and horizontal planes;

Figure I.5.1. View of a stream cross-section showing the location of points of observation

(c) The velocities are greater than 0.150 m s–1;
(d) The bed of the channel is regular and stable;
(e) The depth of flow is greater than 0.300 m;
(f) There is no aquatic growth;
(g) There is minimal formation of slush or frazil ice
(5.3.2.5.1).
5.3.2.2 Measurement of cross-section
The accuracy of a discharge measurement depends
on the number of verticals at which observations of
depth and velocity are obtained. Observation verticals should be located to best defi ne the variation in
elevation of the stream bed and the horizontal variation in velocity. In general, the interval between
any two verticals should not be greater than 1/20 of
the total width and the discharge of any segment
should not be more than 10 per cent of the total
discharge.

Figure I.5.2. Relationship between correct depth d
and observed depth


Channel width and the distance between verticals
should be obtained by measuring from a fixed reference point (usually an initial point on the bank),
which should be in the same plane as the crosssection. Normally, the distance between verticals is
determined from graduated tape or beaded wire
temporarily stretched across the stream or from
semi-permanent marks, for example, painted on a
bridge handrail or a suspension cable (ISO, 1979b).
For large rivers, telemetry systems or triangulation
practices can be used for measuring widths.
Depth may be read directly on a graduated rod set
on the stream bed if measurement is by wading.
If the drum-wire-weight system is used for measurement, the current meter and weight are lowered until the bottom of the weight just touches the water surface, and the depth dial reading is set
at zero. The weight is then lowered until it rests on
the stream bed, and the depth is read on the dial.
If the weight on the sounding line is not sufficient
to keep the line perpendicular to the water surface,
the angle between the line and the vertical should
be measured to the nearest degree with a protractor.
The relationship between the correct depth, d, and
the observed depth, dob, based on the observed
angle, ϕ, and the distance from the water surface to
the point of suspension of the sounding line, x, is
shown in Figure I.5.2 and is given below:


d = [dob – x (secϕ – l)]1 – k

Values of k as given in Table I.5.1 are based on the
assumptions that the drag pressure on the weight
in the comparatively still water near the bottom
can be neglected and that the sounding wire and
weight are designed to offer little resistance to the
water current. The uncertainties in this estimation
are such that signifi cant errors may be introduced if
the vertical angle is more than 30°.


5.3.2.3 Measurement of velocity
[HOMS C79, E79]
5.3.2.3.1 Meters for measurement of velocity
Velocity of flow at a point is usually measured by
counting revolutions of a current meter rotor
during a short-time period measured with a stopwatch (ISO, 1979b). Two types of current meter
rotors are in general use: the cup type with a vertical shaft and the propeller type with a horizontal
shaft. Both types use a make-and-break contact to
generate an electric pulse for indicating the revolutions of the rotor (ISO, 1988a). Optical, non-contact type counters are also in use with
cup-type meters.

Table I.5.1. Correction factor k for given values

Current meters are calibrated to cover the range in
velocity of fl ow to be measured. Detailed calibration procedures are described in ISO 3455 (ISO, 1976). Current meters may be calibrated individually or a group rating may be used. Individually
calibrated meters should be recalibrated after three
years or 300 hours of use or if their performance is
suspect (Technical Regulations (WMO-No. 49),
Volume III, Annex).


5.3.2.3.2 Measurement of velocity using the
current meter
Velocity is observed at one or more points in each
vertical by counting revolutions of the rotor during
a period of not less than 30 seconds. Where the
velocity is subject to large periodic pulsations the
exposure time should be increased accordingly
(Technical Regulations (WMO-No. 49), Volume III,
Annex).


For shallow channels, the current meter should be
held in the desired position by means of a wading
rod. For channels too deep or swift to wade, it
should be positioned by suspending it from a wire
or rod from a bridge, cableway or boat. When a boat
is used
, the meter should be held so that it is not
affected by disturbances to the natural fl ow caused
by the boat. After the meter has been placed at the
selected point in the vertical, it should be allowed
to become aligned with the direction of fl ow before
readings are started. If oblique fl ow is unavoidable,
the angle of the direction of the fl ow normal to the
cross-section should be measured and the measured
velocity should be corrected. If the measured angle
to the normal is γ, then:

Vnormal = Vmeasured cos γ (5.4)

The meter on cable suspension will automatically
point in the direction of the current owing to the
tail vanes built into the meter. In some cases, such
as using an oblique bridge as the measuring section,
the horizontal distances should be corrected as:


dnormal = dmeasured cos γ (5.5)

The current meter should be removed from the
water at intervals for examination. For measuring
very low velocities, special current meters may be
used if they have been tested in this range of velocities for repeatability and accuracy.


The horizontal axis of the current meter should not
be situated at a distance less than one and one-half
times the rotor height from the water surface, nor
should it be at a distance less than three times the
rotor height from the bottom of the channel.
Furthermore, no part of the meter should
break the surface of the water (Technical Regulations
(WMO-No. 49), Volume III, Annex).


5.3.2.3.3 Determination of mean velocity in a
vertical
The mean velocity of the water in each vertical can
be determined by one of the following methods:
(a) Velocity distribution method;
(b) Reduced point methods;
(c) Integration method.
Selection of the appropriate method depends on
the time available, the width and depth of the
water, the bed conditions, the rate of change of
stage, the velocity of the water, the existence of ice
cover and the required accuracy.


Velocity distribution method
The measurement of the mean velocity by this
method is obtained from velocity observations
made at a number of points along each vertical
between the surface of the water and the bed of the
channel. The velocity observations at each position
should be plotted in graphical form and the mean
velocity should be determined by dividing the area
of this plot by the depth. In developing the graph it
may be necessary to estimate the velocities near the
stream bed by assuming that the velocity for some
distance up from the bed of the channel is proportional to the logarithm of the distance x from that
boundary. If the observed velocity at points
approaching the bed are plotted against log x, then
the best-fi tting straight line through these points
can be extended to the bed and the velocities close
to the bed read from this graph.
The velocity distribution method may not be suitable for discharge measurements made during
significant variations of stage because the apparent
gain in precision may be more than offset by errors
resulting from the longer period required to make
the measurement.


The velocity distribution method is valuable in determining coefficients for application to the results
obtained by other methods, but it is not generally
adapted to routine discharge measurements because
of the extra time to compute the mean velocity.
Reduced point methods
(a) One-point method – Velocity observations
should be made at each vertical by placing the

current meter at 0.6 of the depth below the
surface. The value observed should be taken
as the mean velocity in the vertical. Where
measurements are made under ice cover, this
method is applicable with a correction factor of
0.92 for depths shallower than 1 m. Under ice
conditions, the current meter may be placed at
0.5 of the depth. A correction factor of 0.88 is
then applied to this result;
(b) Two-point method – Velocity observations
should be made at each vertical by placing the
current meter at 0.2 and 0.8 of the depth below
the surface. The average of the two values should
be taken as the mean velocity in the vertical;
(c) Three-point method – Velocity observations
are made by placing the current meter at each
vertical at 0.2, 0.6 and 0.8 of the depth below
the surface. The average of the three values
may be taken as the mean velocity in the vertical. Alternatively, the 0.6 measurement may
be weighted and the mean velocity may be
obtained from the equation:
v
_
= 0.25 (v0.2 + 2v0.6 + v0.8) (5.6)
(d) Five-point method – It consists of velocity measurements on each vertical at 0.2, 0.6 and 0.8 of
the depth below the surface and as near as possible to the surface and the bottom. The mean
velocity may be determined from a graphical
plot of the velocity profi le as with the velocity
distribution method or from the equation:
v
_
= 0.1 (vsurface + 3v0.2 + 3v0.6 + 2v0.8 + vbed) (5.7)

(e) Six-point method – Velocity observations are
made by placing the current meter at 0.2, 0.4,
0.6 and 0.8 of the depth below the surface
and as near as possible to the surface and the
bottom. The velocity observations are plotted in graphical form and the mean velocity
is determined as with the velocity distribution
method or from the equation:
v= 0.1 (vsurface + 2v0.2 + 2v0.4 + 2v0.62v0.8 + vbed ) (5.8)
(f) Two-tenths method – In this method, the velocity is observed at 0.2 of the depth below the
surface. A coefficient of about 0.88 is applied to
the observed velocity to obtain the mean in the
vertical;
(g) Surface velocity method – In this method,
velocity observations are made as near as possible to the surface. A surface coefficient of 0.85
or 0.86 is used to compute the mean velocity in
the vertical.
The two-point method is used where the velocity
distribution is normal and depth is greater than
about 60 cm. The one-point method is used for
shallower depths. The three-point method should
be used for measurements under ice or in stream
channels overgrown by aquatic vegetation. The
fi ve-point method is used where the vertical
distribution of velocity is very irregular. The sixpoint method may be used in difficult conditions,
where, for instance, there is aquatic growth, or
there is covering ice. Also it can be used where
the vertical distribution of velocity is very irregular. The two-tenths method is principally used
when it is not possible to position the meter at
the 0.8 or 0.6 of the depth. The surface velocity
method may be used for measuring flows of such
high velocity that is not possible to obtain depth
soundings. In this case a general knowledge of
the cross-section at the site or a cross-section
measured as soon as possible can be used to obtain
the depths.
The accuracy of a particular method should be
determined, if possible, by observing the velocity at
6 to 10 points in each vertical for the fi rst few
discharge measurements made at a new site.
Integration method
In this method, the current meter is lowered and
raised through the entire depth at each vertical at a
uniform rate. The speed at which the meter is
lowered or raised should not be more than 5 per
cent of the mean velocity of fl ow in the crosssection, and it should be between 0.04 and
0.10 m s–1. The average number of revolutions per
second is determined. Two complete cycles are
made in each vertical and, if the results differ by
more than 10 per cent, the measurement is
repeated. This method is seldom used in water
having a depth of less than 3 m and velocities of
less than 1 m s–1. The integration method should
not be used with a vertical axis current meter
because the vertical movement of the meter affects
the motion of the rotor.


5.3.2.4 Computations of discharge
Arithmetical methods
(a) Mean-section method – The cross-section
is regarded as being made up of a number of
segments bounded by two adjacent verticals. If
v
_
1 is the mean velocity at the fi rst vertical and v
_
2
the mean velocity at the second vertical, and if
d1 and d2 are the total depths measured at verticals 1 and 2, and b is the horizontal distance between verticals, then the discharge q of the
segment is:

The total discharge is obtained by adding the
discharge from each segment;
(b) Mid-section method – The discharge in each
segment is computed by multiplying vd in each
vertical by a width, which is the sum of half
the distances to adjacent verticals. The value of
d in the two half-widths next to the banks can
be estimated. Referring to Figure I.5.1, the total
discharge Q is computed as:

Graphical methods
(a) Depth-velocity integration method – The fi rst
step consists in drawing, for each vertical, the
depth velocity curve, the area of which represents
the product of the mean velocity and the total
depth. The value of this product at each vertical
is then plotted versus lateral distance and a curve
is drawn through the points. The area defined by
this curve is the discharge in the cross-section;
(b) Velocity-contour method – Based on the velocity distribution curves of the verticals, a velocity
distribution diagram for the cross-section is
prepared showing curves of equal velocity. Starting
with the maximum, areas enclosed by the equal
velocity curves and the water surface should be
measured and then plotted in another diagram,
with the ordinate indicating the velocity and the
abscissa indicating the area. The area enclosed by
the velocity area curve represents the discharge of
the cross-section (ISO, 1979b).
5.3.2.5 Measurement of discharge under
ice cover
Measurement of discharge under ice cover requires
general knowledge of instruments and procedures
described in 5.3.2.1 to 5.3.2.4. These sections deal
only with equipment and procedures peculiar to
the measurement of discharge under ice cover.
5.3.2.5.1 Selection of site
It is advisable to select alternate cross-sections
during the open water season when channel
conditions can be evaluated. At some stations, the
same measuring section may be used during winter
and summer, but it is more important that winter
measurements be made under suitable conditions
than it is to use the same measuring section. After
initial selection, exploratory holes may be cut at
quarter points along the section to detect the presence of slush ice or poor distribution of fl ow. Frazil
ice should be avoided whenever possible because
ice particles impede the operation of the meter and
because of diffi culty in determining ice thickness.
Also, a small fl ow may occur through the frazil ice
which cannot be measured by usual methods.
Winter freshets often lead to water breaking through
the ice and forming two independent currents, one
above and the other below the ice. Such locations
should be avoided.


5.3.2.5.2 Equipment
(a) Cutting holes – When ice is thick, a mechanical ice auger, drill or chainsaw is desirable for
cutting holes. For thin ice, an ice chisel may be
used;
(b) Determination of effective depth – Effective
depth of water below ice cover is the total
depth of water minus the distance from the
water surface to the underside of the ice. The
distance between the water surface in the ice
hole and the underside of the ice may be measured using an ice-measuring stick or “ice stick”,
which is an L-shaped graduated bar of appropriate length. The short projection of the L-shaped
stick is held against the underside of the ice,
and the depth to that point is read at the ice
surface on the graduated portion of the stick.
If there is slush under solid ice at a hole, the
depth at which it ends may be determined by
suspending the current meter below the slush
ice with the meter rotor turning freely and then
raising it slowly until the rotor stops. This point
is assumed to be the interface between water
and slush;
(c) Current meter and weight assembly – If an ice
auger or drill is used to cut holes through ice,
a special current meter and sounding weight
assembly is passed through the ice hole, which
is generally about 150 mm in diameter. The
assembly may consist of two teardrop-shaped
lead weights, one above and one below the
meter, or one teardrop-shaped weight below
the meter. When the hole can be made large
enough, the standard current meter and weight
assembly can be used as described in 5.3.2.3.1;
(d) Meter suspension – The meter suspension may
be by a rod, handline or sounding reel. If the
total depth of water under ice cover is greater

than 3 or 4 m, a reel or handline is usually used.
The reel is mounted on a collapsible support
set on runners. In extremely cold weather, the
support may be equipped with a heated water
tank or hot air chamber to keep the meter from
freezing while moving the equipment from
one position to the next. For shallower depths,
where a meter without tail vanes is suspended
by a-rod through a drilled hole, the direction of
current must be determined so that the meter
can be properly aligned.


5.3.2.5.3 Discharge measurement
(a) Spacing of verticals – The information in 5.3.2.2
is also applicable to the spacing of verticals
under ice. However, in addition to the variation in elevation of the stream bed, variation in ice cover and slush ice thickness must also be
taken into account in selecting the number and
location of verticals. If the current is divided
into different channels by slush ice, not less
than three verticals should be used in each
channel;
(b) Measurement of velocity – Ideally, velocity
curves should be determined from velocity
observations at every tenth of the effective
depth in at least two verticals to determine what
coefficients, if any, are necessary to convert
the average velocity obtained by any standard
open-water method of observation to an average velocity in a vertical under the ice cover.
In shallow water, velocity may be observed at
one point at either 0.5 or 0.6 of the effective
depth, but a coefficient is normally required to
convert the observed velocity to mean velocity.
In deeper water (1 m or more), velocity observations could include two observations at 0.2 and
0.8 of the effective depth, three observations
at 0.15, 0.5 and 0.85 of the effective depth,
six observations at 0.2, 0.4, 0.6 and 0.8 of the
effective depth, and at points close to the top
and bottom. The average velocity observed in
the two- and three-point methods may be used
as the mean in the vertical. For the six-point
method, see 5.3.2.3.3;
(c) General notes – When measuring discharge
from an ice cover, appropriate safety precautions should be observed. For example, the
safety of ice should always be tested by probing
ahead with an ice chisel while moving across
the ice. If the velocity measured under ice
conditions is less than the accepted lower limit
of the current meter, the cross-section should
be moved to another reach of the river where
the velocity is higher. Care must be taken to
ensure that the meter is rotating freely and is
not impeded by ice that can accumulate on
the meter and freeze while moving from one
vertical to another. At the time the measurements are taken, a record should be kept of a
complete description of weather and ice conditions on the river, particularly at the control
sections. This will aid in the later computation
of discharge between measurements.
5.3.2.5.4 Computation of discharge
The computation of discharge under ice cover is the
same as for open-water conditions described in
5.3.2.4 except that effective depth is used instead of
total depth of water.
5.3.2.6 Accuracy of measurement
The accuracy of discharge measurements depends
on the reliability of the meter rating, on the conditions of flow, on the skill of the hydrometrist, and
on the number of observations of depth and velocity obtained (ISO, 1981; 1985). Measurements are
normally made by observing the depth and the
velocity at two points, in 20 to 25 verticals in the
cross-section. For this type of measurement, under
the flow conditions that are usually encountered,
the standard error at the 95 percent confidence
level is about 5 per cent (ISO, 1979b).
5.3.3 Measurement of discharge by the
fl oat method [HOMS C86]
This method should be used in the following
instances: it is impossible to use a current meter
because of unsuitable velocities or depths, or where
there is the presence of a large amount of material
in suspension, or when a discharge measurement
must be made in a very short time.
5.3.3.1 Selection of sections
Three cross-sections should be selected along a
reach of straight channel. The cross-sections should
be spaced far enough apart for the time that the
fl oat takes to pass from one cross-section to the
next to be measured accurately. A travel time of
20 seconds is recommended, but a shorter time may
have to be used on small rivers with high velocities
where it is often impossible to select an adequate
length of straight channel.

5.3.3.2 Floats
Surface floats or rod floats may be used. A surface
float has a depth of immersion of less than one-quarter the depth of the water. Surface floats should not be used when they are likely to be affected by wind. A rod fl oat has a depth of immersion exceeding one quarter the depth of the water. Rod floats
must not touch the channel bed. Floating trees or
ice cakes may serve as natural fl oats during periods
when it is unsafe to be on the river.
5.3.3.3 Measuring procedure
Float observations must be uniformly distributed
over the width of the stream. The float should be
released far enough above the upper cross-section
to attain a constant velocity before reaching the
first cross-section. The time at which the fl oat
crosses each of the three cross-sections should be
noted with a stopwatch. This procedure should be
repeated with the floats at several locations across
the stream. The width of the channel should be
divided into segments of equal width or of approximately equal discharge. The number of segments
should be not less than three, but where possible a
minimum of five should be used. Distances of the
float from the bank as it passes each cross-section
maybe determined by suitable optical means, for
example, a theodolite.
The depth of flow at points in the cross-section may
be determined by surveying methods.
5.3.3.4 Computation of velocity
The velocity of the float is equal to the distance
between cross-sections divided by the time of travel.
At least five values of the float velocity should be
taken at each segment and the mean of these values
should be multiplied by a coefficient to obtain the
mean water velocity for each segment. This coeffi –
cient is based on the shape of the vertical velocity
profile and the relative depth of immersion of the
fl oat. The coefficient to be applied to the measured
velocity should be determined, if possible, for each
site by an analysis of discharge measurements that
have been made by current meter. When such
measurements are not available, an adjustment
factor, F, from Table I.5.2 may be used for rough
estimation.

Table I.5.2. Float velocity adjustment factor F as a
function of R, the ratio of the immersed depth of
fl oat to depth of water


Alternatively the float velocity may be plotted
as a function of the corresponding distance
from the bank, and the mean surface velocity
across the river should be determined from this
plot. The mean velocity of flow in the crosssection is equal to the mean surface velocity
multiplied by a coefficient, K, the value of
which is deduced, if possible, from preceding
measurements made with a current meter for
smaller discharges.

5.3.3.5 Computation of discharge
Discharge in each segment is computed by multiplying the average area of the cross-section of the
segment by the mean velocity of fl ow in the
segment. The total discharge is the sum of these
discharges (ISO, 1979b).
5.3.4 Measurement of discharge by
dilution methods [HOMS E73]
The measurement of discharge by this method
depends on determining the degree of dilution by
the flowing water of an added tracer solution. The
method is recommended for sites with excessive
turbulence flows. The two principal tracer methods used for discharge measurements are the
constant-rate-injection method and the sudden injection method. The general requirements
(5.3.4.1) for both methods are the same (ISO,
1973a; 1987).
The dilution method is a fully acceptable method
for discharge measurement at sites where the conditions for this method are good.
5.3.4.1 General requirements
A solution of a stable tracer is injected into the
stream at either a constant rate or all at once.
Computation of the stream discharge requires
knowledge of the following factors:
(a) The rate of injection for the constant-rate-injection method or the total amount injected for
the sudden-injection method;
(b) The concentration of the tracer in the injected
solution;
(c) The calibrated relationship between tracer
concentration and the recorded property (for
example, conductivity, colour and radioactivity) at the measuring site after it has been well mixed laterally.

The accuracy of these methods critically depends
upon:
(a) Adequate mixing of the injected solution
throughout the stream cross-section at the
sampling section. If the tracer solution is
continuously injected, the concentration of the
tracer should be essentially constant throughout the sampled section. If the tracer is injected
all at once, cdt
0
r
∫ should essentially be the same
at all points in the section, where c is the
concentration and T is the time for all of the
tracer to pass a particular point in the section;
(b) No absorption or adsorption of the added
tracer by stream bottom materials, sediments, plants or organisms, and no decomposition of the added tracer in the stream
water. The concentration should be determined at the sampling section and at least
one other cross-section downstream to verify
that there is not a systematic difference in
the mean concentration from one sampling
section to another.


5.3.4.2 Selection of site
The primary criterion for the selection of sites for
measurement of discharge by dilution is adequate
mixing of the injected solution with the stream
water in a short length of channel. Mixing is
enhanced by high boundary roughness and features
that cause the channel flow to be highly turbulent,
such as at waterfalls, bends or abrupt constrictions.
A small injection of rhodamine dye or fl fluorescein
can help to assess the mixing condition at the measuring site. Large dead-water zones between the
injection site and the sampling site will often affect
the mixing so that the tracer will not be adequately
mixed in the cross-section at the sampling site.
5.3.4.3 Tracers and detection equipment
Any substance may be used as a tracer if:
(a) It dissolves readily in the stream’s water at ordinary temperatures;
(b) It is absent in the water of the stream or is
present only in negligible quantities;
(c) It is not decomposed in the stream’s water and
is not retained or absorbed by sediment, plants
or organisms;
(d) Its concentration can be measured accurately
by simple methods;
(e) It is harmless to humans, animals and vegetation in the concentration it assumes in the
stream.


The cheapest tracer is common salt. Where the
tracer is instantaneously injected into the stream,
the required quantity is not particularly large and
detection by conductivity methods is relatively
simple.
Sodium dichromate is used extensively in the
dilution method. Its solubility in water is relatively high (600 kg m–3), and the salt satisfies
most requirements of 5.3.4.1. Colourimetric
analysis (ISO, 1987) permits the measurement
of very low concentrations of sodium
dichromate.
Lithium chloride has solubility in water of 600 kg m–3
and its concentrations down to 10–4 kg m–3 can be
detected using flame photometric analysis.
Other chemicals used for dilution gauging are
sodium iodide, sodium nitrite and manganese
sulphate.


Rhodamine WT dye is widely used in the United
States in the dilution method. Its absorptive characteristics are much better than those of other
rhodamine dyes. The concentration of the dye can
be measured using commercially available fluorometers that can measure concentrations of 5 to
10 parts per billion.
Radioactive elements such as bromine-82, gold-198,
iodine-131 and sodium-24 have been used as tracers. Concentrations of these elements as low as 10–9
may be determined accurately with a counter or
count rate meter with the sensing probe suspended
in the stream or in a standard counting tank.
Although radioactive elements are ideal tracers for
the dilution method, the health hazards may limit
their use in measurement of stream discharge in
some localities.
5.3.4.4 Computation of discharge
Equations used to compute the stream discharge, Q,
are based on the principle of continuity of the tracer:
(continuous injection) (5.11)
and
(sudden injection) (5.12)
where Qtr is the rate of injection, ci
is the concentration of injection solution, cs
is the concentration in
the stream at the sampling section, V is the volume
of injected solution and t is time.

5.3.5 Computations of discharge by
indirect methods [HOMS E70]
5.2.5.1 General
During fl ood periods, it may be impossible to measure discharge directly because of the excessive rate
of change of discharge, excessive velocities, debris,
depths or widths, or because flooded conditions
make roads impassable or measuring structures
inaccessible. When such conditions occur, the peak
discharge may be determined after the flood has
subsided by computations that combine well-established hydraulic principles with field observations
of channel conditions and flood profiles. All the
methods involve the simultaneous solution of
continuity of mass and energy equations. Such
computations may be made for reaches of river
channel, through roadway culverts and bridge
openings, and over dams and highway embankments. Although the hydraulic formulae differ for
each type of waterway, all the methods involve the
following factors:


(a) Geometry and physical characteristics of the
channel and boundary conditions of the reach
used;
(b) Water-surface elevations at time of peak stage
to define the cross-sectional areas and the head
difference between two significant points;
(c) Hydraulic factors, such as roughness coeffi –
cients based on physical characteristics.

5.3.5.2 Field survey
A reconnaissance study, from maps, by air or by
travel in the region, is made to select the most
favourable site for determining discharge by one of
the indirect methods. The site should be as close as
possible to the desired measuring point, and large
intervening tributaries or diversions should be
avoided. The site must contain good high-water
marks defining the water-surface profile during the
peak.
A detailed survey is made to define channel geometry adjacent to and within the selected reach, the
channel cross-sections, the dimensions and details
of culverts, bridges, dams, roadways or other artifi –
cial structures, and the positions and locations of
high-water marks left by the flood. All factors that
affect channel roughness are noted and roughness
coefficients are selected. Photographs should be
taken of the cross-sections and reach to facilitate
office evaluations of site conditions.
From the field survey notes, drawings are made
showing the plan, the profiles of the channel
bottom and high-water surface on both banks, the
cross-sectional areas and details of any artificial
structures. Computations are made of hydraulic
factors and the discharge is computed.
5.3.5.3 Slope-area measurements
Slope-area measurements require a reach of river
channel that is selected for uniformity or uniform
variation in hydraulic properties (ISO, 1973b).
Discharge is computed on the basis of a uniform
fl ow equation, such as the Manning equation,
involving channel characteristics, water-surface
profiles and roughness coefficients.
5.3.5.4 Measurement of flow through
culverts
Peak discharge through culverts can be determined
from high-water marks that define the headwater
and tailwater elevations, culvert geometry and
slopes, and cross-sections that define approach
conditions. The head-discharge relationships of
culverts have been defined by laboratory investigations and field verification. Peak discharge is
determined by the application of continuity and
energy equations between the approach section
and a section within the culvert barrel. For convenience in computation, culvert fl ow has been
classified into six types on the basis of the location
of the control section and the relative heights of
the headwater and tailwater elevations.
5.3.5.5 Measurement of flow through width
contractions
The contraction of a stream channel by a roadway
crossing creates an abrupt drop in water surface
elevation between an approach section and the
contracted section under the bridge. The contracted
section formed by bridge abutments and the channel bed may be used as a discharge control to
compute flood flows. The head on the contracted
section is defined by high-water marks (upstream
and downstream), and the geometry of the channel
and bridge is defined by field surveys. The discharge
equation results from a combination of the energy
and continuity equations for the reach between
these two sections.
5.3.5.6 Measurement of flow over weirs,
dams and highway embankments
A weir, dam or embankment generally forms a
control section at which the discharge may be
related to the upstream water-surface elevation.
The peak discharge at the control section can be determined on the basis of a field survey of highwater marks and the geometry of the structure. The
methods are derived from laboratory and field studies of the discharge characteristics of weirs, dams
and embankments.
The fieldwork consists of a survey of headwater and
tailwater elevations from high-water marks, an
approach cross-section to define velocity of
approach, and an exact determination of the profile
of the control structure to assign the proper
discharge coefficient. Coefficients are available for:
(a) Thin-plated weirs, either discharging freely or
submerged;
(b) Broad-crested weirs, not submerged;
(c) Ogee or design-head dams, submerged or not
submerged;
(d) Many irregular shapes.
5.3.6 Measurement of discharge under
diffi cult conditions
General discussion on the measurement of discharge
under difficult conditions is provided in the Level
and Discharge Measurements under Difficult Conditions
(WMO-No. 650).


5.3.6.1 Unstable channels
Channel instability is characterized by systematic
shifts of the bed, high silt content and the presence
of various kinds of debris in the fl ow. Channel
instability is a hindrance to the operation of a
permanent gauging structure and/or measurement
section. This problem can be minimized by selecting
a site midway along a straight reach of the river
with a uniform section remote from various
obstructions (bridges, etc.). The greatest stability in
the banks is usually found at places where the
channel narrows. On small rivers, the site should be
convenient for the construction of a permanent
measurement section.


On small streams, where there is no transport of
large stones and debris, portable or permanently
installed fl umes may be used to measure fl ow. On
small rivers, it is desirable, in some cases, to have an
artificial section for measurements to improve the
stage-discharge relationship. Improvements may
take the form of a low weir or fl ume depending on
the specific conditions at the site. The structure
should be high enough to remove variable backwater from downstream but not so high as to cause
excessive disturbances downstream. At low water,
the structure should provide a sensitive relationship between discharges and water levels. To clean
the crests of large structures and to provide a means
for making current-meter measurements, a footbridge may be provided. Because of the large silt
content of unstable channels, it is desirable to use
current meters with a sealed contact chamber.
Sounding rods should be provided with a foot to
prevent them from sinking into the silt.
When measuring discharge by the velocity-area
method, the depth is usually determined before
and after measurement of the velocity. When the
velocity is high, the presence of various kinds of
debris in the stream may lead to external damage to
the current meter. In such cases, it is advisable to
compare the current-meter readings, before and
after measuring the discharge, with the readings
from a separate current meter not used in the
measurement.


In rivers with intensive channel shifts, the distribution of velocity in a cross-section varies periodically.
The choice of velocity verticals must be made by
taking into account the velocity distribution at the
time of measurement. The use of permanent verticals may lead to systematic errors. If there is an intensive
shifting of the channel, it is also desirable to use a
reduced point method of velocity measurement and
a reduced number of verticals (ISO, 1979b).
If soundings have been made twice (before and
after velocity measurements), the area of water
cross-section is computed on the basis of the mean
depths from the two soundings. On wide rivers,
where the location of sounding verticals usually is
determined by distances from an initial point on
the shore, the verticals obtained on the two runs
may not coincide. In this case, an average crosssection profile of the measurement site is used to
select depth values for the discharge computation.
5.3.6.2 Mountain streams
Mountain streams are characterized by high flow
velocities, shallow and uneven beds blocked by
boulders and debris, transverse and uneven water surface slopes, and transport of large but varying
quantities of stones and pebbles. Measurement or
gauging locations with these characteristics should
be avoided if possible.
Due to very turbulent flows, it is desirable to use
one of the dilution methods of fl ow measurement
on small mountain streams (5.3.4).
Improvements in the channel to make better
measurements may be advisable. It may also be
desirable to equip the site with a gauging bridge
(5.3.2). If it is possible to build a reach with acceptable conditions for current-meter measurements these should be comprised of at least
20 verticals. Measurement of depth by wading rod
in mountain streams does not lead to systematic
errors. However, the use of a sounding weight with
tailfin may lead to underestimates of the depth if
the depth is small. For depths of about 1 m, these
differences from measurements made by wading
rod may amount to about 2.5 to 3 per cent, while
for depths of 0.4 to 0.8 m, the difference may be as
much as 10 to 15 per cent.


It is best to use the two-point method to measure
velocities by current meter. The discharge is calculated as explained in 5.3.2.4.


5.3.6.3 Measurement of unsteady flow
5.3.6.3.1 Measurement of discharge during
floods and on large rivers
Flood measurements are best made from bridges,
cableways or boats. Portable electromechanical
winches are available, which can be set up on special
trucks, motorcars and tractors. On large rivers, where
there are no bridges or cableways, boats, large vessels
or ferries are used. Optical or telemetric equipment
may be set up on board the vessel and on the bank
to determine the position in the channel. Ferries
using a cable for the crossing are equipped with electric or mechanical engines for traction by the cable
and for lifting and lowering the equipment.


Generally, sounding weights of up to 200 kg are
necessary because maximum velocities on large
rivers may be as great as 3 to 5 m s–1. Soundings of
depth also may be made by echo sounder.
For flood measurements on small rivers, remote
control or bank-operated traversing systems are
particularly suitable. These systems may be portable
and can be used at several sites, which need merely
to be equipped with a main carrying cable across
the river. If such systems are not available, easily
transportable duraluminium boats or inflatable
rubber rafts with outboard motors and equipment
platforms can be used. Locations that are difficult
to access may have to be reached by helicopter.
For very high velocities, surface fl oats or stroboscopic instruments for measuring velocities may be
used. The stroboscope has a telescope that is directed
towards the surface of the water and a number of
rotating mirrors. The speed of rotation of the
mirrors is chosen so that a stationary image of the
surface of the water is obtained. The velocity of the
flow is determined from the speed of rotation of the
mirrors. The maximum speed measurable by this
method is 15 m s–1, but this maximum is dependent on the height of the observation point above
the water surface. Measurements by stroboscope
can be made in very turbid flow with floating ice
and other solid matter preventing the use of a
current meter. The coefficient for converting the
surface velocity to the mean velocity at a vertical,
determined by similar measurements under less
difficult conditions, is usually equal to 0.85–0.90.
Measurement of depth is commonly made by echo
sounder or a standard cross-section is used.
For wide rivers (3 to 20 km) with several subchannels, measurements by current meter become extremely difficult. In this case, the moving boat method (5.3.7.2) or discharge measurement by
acoustic Doppler instruments (5.3.7.5) may be used.
Moreover, these are convenient methods when there
are short breaks in the ice run or if there is debris. If
there is ice or debris in some particular part of the
flow, measurements may be made by the float
method and by current meter during breaks in occurrence of such debris. Aerial photography using floats
may also be employed for wide river measurements.

5.3.6.3.2 Measurement of discharge in tidal
reaches
Where a measurement section is affected by ocean
tides, the following effects must be taken into
account:
(a) Continuous change of water level, with and
without change of direction of the current;
(b) Continuous change of velocity with time, even
at a single point in a vertical with considerable
velocity gradients;
(c) Change in the time-distribution of velocity;
(d) Change of direction of the current for the tidal
cycle with zero velocity;
(e) Presence of stratifi ed fl ow with varying density
and direction of fl ows;
(f) Considerable change in the width and crosssection of the fl ow;
(g) Presence of large-scale turbulence (for example,
fl uctuations with a period of more than
30 seconds and the amplitude of velocity
variations up to 50 per cent) and of seiches.
The discharge of tidal river is generally determined
by one of the following methods (ISO, 1974):
velocity-area method, volumetric method, or by
solving the equation for unsteady fl ow. The moving
boat method (5.3.7.2) or the acoustic Doppler
method (5.3.7.5) may also be used, particularly at
times when the distribution curve of velocities is
close to its usual shape. Other methods, such as the
ultrasonic method (5.3.7.3), may also be suitable.

In the method of computation of discharge by the
velocity-area method, the velocity is measured
during the entire fl ood-ebb cycle. Measurements
are usually made at several points to be able to
account for the different directions of fl ow. At the
same time, the water level and the depths at verticals are measured continuously. Then, all
measurements are reduced to a single time for
which the discharge is calculated.

The accuracy of the velocity-area method is greater if:
(a) The tidal cycle during which the measurement
is made is periodic or nearly periodic;
(b) Currents, particularly during the period of
maximum fl ow, are parallel to each other and
at right angles to the gauging site at all points;
(c) Curves of horizontal and vertical velocity distributions are of the regular shape encountered at
the gauging site;
(d) The transverse profi le of the gauging site is
uniform and lacks shallow areas.
The site selected should meet as closely as possible
the following requirements:
(a) The river bed section should be straight and of
regular shape;
(b) The depth of the water at the site should be such
that current meters can be used effectively;
(c) The channel section should be stable during
the tidal cycle;
(d) The discharge should be concentrated within
channels the cross-sections of which can be
determined with a fair degree of accuracy;
(e) The site should not be near artificial or natural
obstacles causing non-parallel flows;
(f) The gauging site should be clear of vegetation;
(g) Oblique fl ow, backflow and dead zones should
be avoided.
The site should be conspicuously marked on both
banks.
To determine discharge during the rise and recession of floods, measurements are made at each
vertical during the entire tidal cycle. To determine
accurately the moment of zero velocity, measurements begin and end half an hour before and after
the tidal cycle. Depending on the equipment available and on the physical characteristics of the
selected site, different procedures can be adopted
for velocity measurements:
(a) If a sufficient number of boats are available,
measurements are made simultaneously at all
verticals during the entire tidal cycle;
(b) If only a limited number of boats are available,
the chosen verticals are marked by anchored
buoys. One or two boats are necessary to
carry out the measurements, proceeding
successively from one vertical to the next, at
intervals of not more than one hour between
each vertical. At least one additional boat
remains permanently at one reference vertical,
carrying out measurements continuously
during the entire tidal cycle. In this case, the
curves of velocity changes occurring over
time at each vertical are plotted by using the
concurrent velocities at the reference vertical
as a basis of comparison;
(c) If the shape of the tidal curve does not change
considerably from day to day and if at least
two boats are available, then one of the boats
is stationed at the reference vertical to carry out
measurements during the whole tidal cycle for
each day. The other boat carries out measurements during the whole cycle at each vertical,
moving to a new vertical each day. In this case,
the number of days required for the whole cycle
of observations is equal to the number of velocity verticals;
(d) If there are different tidal amplitudes and if it
is not possible to make measurements in many
verticals, measurements are carried out at each
vertical for the entire cycle at different tidal
amplitudes during a lunar month and at spring
and neap tides;
(e) If there is considerable pulsation, measurements
should be carried out at each vertical with
the aid of several current meters set at different heights for periods of 10 to 15 minutes.
The mean velocity is determined for the mean
period of time;
(f) In the case of oblique currents, use must be
made of direct reading current meters or of
instruments capable of measuring the angle of
deviation.
Where rapid velocity changes occur, the velocity
values at the various points in the vertical must be
adjusted to a specific time. For this purpose, velocity measurements are either repeated at all points in
the vertical by moving from the bottom to the
surface, or are measured only at one point at the
surface.
For the computation of the discharge at each
vertical, a curve of velocity changes with time is
plotted, from which the value for a specified time
is taken.


For the computation of discharge by the volumetric
method, synchronous measurements of the water
level are made at the boundaries of the measuring
section or sections after their geometrical characteristics (cross-sections, lengths and flooded areas) are determined. An additional gauging station is located
on the river above the area of tidal effects so that
the discharge attributable to the river can be determined. Where there are transverse slopes in wide
estuaries, levels are measured at both banks. The
the difference in volumes of the tidal prisms during the
accounting interval is computed from the change
in mean depths and areas of water surface between
the boundaries. To determine the mean discharge,
the difference in the volume of the total prism is
divided by the accounting period minus the inflow
into the river.


In the method of computation of discharge from
equations of unsteady motion, the solution of the
equations of unsteady motion for the section under
consideration is simplified by certain assumptions,
such as parallel fl ow and uniform density, and that
the channel is prismatic. Measurements are usually
made for two typical (high and low) tidal cycles.
The measurements are used to calibrate the parameters of the equations.
5.3.6.4 Weed growth in stream channels
Weed growth in rivers can cause relatively large
errors. For small rivers, it is advisable, if possible, to
construct artificial controls. If this is not possible,
discharges should be measured by the velocity area
method. For this purpose, a reach of the river 6 to
10 m long should be kept clear of weed growth
during the entire season. In addition, the banks
should be kept clear of shrubs and high grass over a
somewhat larger reach.
The use of toxic substances to impede the growth of
vegetation is effective for a short time only. Frequent
clearing of the bed may be the most practical
method. The weeds growing in the bed may be cut
by a special machine attached to a mechanized
chainsaw or by the aid of an ordinary scythe.
Flow velocity in each vertical should be measured
at three points (at depths of 0.15, 0.5 and 0.85).
Where the depth of the vertical is less than 0.40 m,
velocity is measured by the single-point method.
In the discharge measurement notes, a short description of the actual state of weed growth should be
given.


Because algae and weeds could become entwined in
the propeller of the current meter, the instrument
should be inspected and cleaned frequently during
the measurement process. Where measurements
are made at one point only, the regularity with
which signals are received must be carefully
checked. Experience has been acquired with the use
of the electromagnetic method for gauging under
such conditions (5.3.7.4).


5.3.7 Non-traditional methods of
stream gauging
5.3.7.1 General
Determination of discharge by the velocity-area
method, the dilution method and by means of a
hydraulic structure (5.4) have certain limitations
and are not applicable in some instances. Four relatively new methods of fl ow measurement in open
channels are the moving boat method, the ultrasonic method, the electromagnetic method and the
Acoustic Doppler method.
5.3.7.2 Moving-boat method [HOMS E79]
In this method, a boat is fitted with a specially
designed component current-meter assembly that
indicates an instantaneous value of velocity. A
measurement is made by traversing the stream
along a preselected path that is normal to the fl ow.
During the traverse, which is made without stopping, an echo sounder records the geometry of the
cross-section, and the continuously operating
current meter measures the combined stream and
boat velocities. These data, collected at some 30 to
40 observation points (verticals) across the path,
are converted to discharge. The velocity recorded at
each of the observation points in the cross-section
is a vector quantity that represents the relative
velocity of flow past the meter assembly. This
assembly consists of a vane attached to a stainless
steel shaft, which, at its upper end, incorporates a
dial and pointer for reading the angle between the
direction of the vane and the true course of the
boat. This is performed by sighting on carefully
located markers on the banks. About six runs, in
alternate directions, are usually taken and the measurements are averaged to give the discharge (ISO,
1979a; Smoot and Novak, 1969).
The discharge is calculated in a similar manner to
the conventional velocity-area method by summing
the products of the segment areas and average
velocities. Because the current meter is located
about 1 m below the surface, a coefficient is required
to adjust the measured velocity. In large rivers, the
coeffi cient is usually uniform across the section.
Investigations on several rivers have shown that
the coefficient generally lies between 0.85 and 0.95.
The moving boat method provides a single measurement of discharge, and an accuracy of ±5 per
cent is claimed at the 95 per cent confidence level.

5.3.7.3 Ultrasonic (acoustic) method
[HOMS C73]
The principle of the ultrasonic method is to measure
the velocity of fl ow at a certain depth by simultaneously
transmitting sound pulses through the water from
transducers located on either side of the river. The
transducers, which are designed both to transmit and
receive sound pulses, are located on opposite banks,
so that the angle between the pulse path and the
direction of fl ow is between 30° and 60°. The difference
between the time of travel of the pulses crossing the
river in an upstream direction and those travelling
downstream is directly related to the average velocity
of the water at the depth of the transducers. This
velocity can be related to the average velocity of fl ow
of the whole cross-section. The incorporation of an
area computation into the electronic processor allows
the system to output discharge.
Ideally, the transducers are set at a depth such that
they measure the average velocity of fl ow. In practice, they are ultimately fi xed in position so that for
any change in stage, they probably will not be at
the point of average velocity, and a coeffi cient is
necessary to adjust the measured velocity.
There are two types of ultrasonic systems commonly
in use, the fi rst where the transducers are fixed in
position and the station is calibrated by current
meter, and the second where the transducers are
designed to slide on either a vertical or inclined
assembly. In the latter method, the system is selfcalibrating and therefore no current-meter
measurements are necessary. By moving the transducers through a number of paths in the vertical
(generally 7 to 10), velocity readings are obtained
along these paths. From each set of the readings,
vertical velocity curves are established over as large
a range in stage as possible. It is then possible first,
to estimate a suitable position for the fi xing of the
transducers in the vertical and, second, to establish
a curve of stage against the coefficient of discharge
as in the first method.
In rivers with small range in stage, a single-path
transducer system may be acceptable. For rivers
with large variations in stage, a multipath system
with several pairs of transducers may be necessary.
The accuracy of the ultrasonic method depends on
the precision with which the travel times can be measured. The several techniques available at the present
time are capable of measuring time to very high accuracy (Smoot and Novak, 1969; Herschy and Loosemore,
1974; Smith, 1969; 1971; 1974; Botma and Klein,
1974; Kinosita, 1970; Holmes and others, 1970;
Halliday and others, 1975; Lenormand, 1974).
5.3.7.4 Electromagnetic method
The motion of water flowing in a river cuts the
vertical component of the Earth’s magnetic field,
and an electromotive force (emf) is induced in the
water that can be measured by two electrodes. This
emf, which is directly proportional to the average
velocity in the river, is induced along each traverse
fi lament of water as the water cuts the line of the
Earth’s vertical magnetic field.


Figure I.5.3 shows diagrammatically an electromagnetic gauging station where the coil is placed in the bed and the magnetic field is in the x direction, the
emf is in the y direction and the streamfl ow is in
the z direction. Faraday’s law of electromagnetic
induction relates the length of a conductor moving
in a magnetic field to the emf generated by the
equation (Herschy and Newman, 1974).
In practice, most river beds have significant electrical conductivity that will allow electric currents to
fl ow in the bed. From practical considerations, the
induced fi eld will be spatially limited and electric
currents fl owing in the area outside the fi eld will
have the effect of reducing the output potential.
Both of the above factors have the effect of reducing the signal and hence the voltage recorded. At
an electromagnetic gauging station, it is necessary
to measure both the bed and water conductivity.
The most suitable current for the coil is a direct
current, the direction of which is reversed a few times
per second and an alternating square wave with a
frequency of about 1 Hz should be used. A typical
installation may have a coil of 12 turns, each of
16 mm² double PVC insulated cable, and be supplied
with 25 A with a voltage across the coil of about 20 V
(Herschy and Newman, 1974).


The electromagnetic method will be suitable for use
in rivers with weed growth, high sediment concentration or unstable bed conditions. It gives a
continuous record of the average velocity in the
cross-section that can be combined with stage to
give an on-site output of discharge.


The accuracy depends on the signal processing
equipment detecting and measuring small potentials sensed at the voltage probes. It is possible to
detect a signal of 100 nV, which represents a velocity of approximately 1 mm s–1. The electromagnetic
gauging station requires on-site calibration by
current meter or other means and a relation established between discharge and output.
5.3.7.5 Measurement of discharge by
acoustic Doppler instruments


5.3.7.5.1 General
Developments in acoustic Doppler technology have
made these instruments a viable alternative for
making measurements of discharge in rivers and
large streams. During recent years the instruments
and techniques have changed appreciably and it
has become possible to use Dopper instruments in
small and shallow rivers. All instruments use the
Doppler principle to measure velocity from particles (scatters) suspended in the water in order to
compute discharge. An acoustic Doppler instrument contains transducers and temperature sensors
that are made for operating in water. None of the
instruments requires periodic calibrations, unless
there is physical damage to the instrument.

Figure I.5.3. Basic system of the electromagnetic method


5.3.7.5.2 Doppler principle
An acoustic Doppler instrument (see Figure I.5.4)
measures the velocity of the water using a physical
principle called the Doppler shift. This states that if
a source of sound is moving relative to the receiver,
the frequency of the sound at the receiver is shifted
from the transmit frequency. The instrument transmits an acoustic pulse of energy into the water much
like a submarine’s sonar but at much higher frequencies. This energy is refl ected off particles suspended
in, and moving with, the water and some of it
returns to the instrument. The instrument measures
the Doppler shift (change in frequency) of the
refl ected energy and uses this to compute the velocity of the water relative to the instrument. The
refl ected pulses have a frequency (Doppler) shift
proportional to the velocities of the scatterers they
are travelling in along the acoustic beam:
V = Fd
2 F0




C (5.13)
where Fd is the Doppler shifted frequency received
at the transducer, F0 is the transducer transmit
frequency, C is the sound speed, and V is the scatterer (water) velocity.

Figure I.5.4. Transducer of an acoustic Doppler
instrument installed on a boat


All Doppler instruments operate within a pre-set
frequency. The frequency determines under which
conditions they are best equipped to measure. An
instrument that operates on a lower frequency has a greater range of distance than an instrument with
a higher frequency. The amount and type of particles in the water will also determine the range of
the instrument and the quality of the measurements. If there are too few particles in the water,
the range will be noticeably shorter and the quality
of the data might be compromised.
These principles are true for all of the acoustic
Doppler instruments, but different instruments
compute discharge in different ways.
5.3.7.5.3 Acoustic Doppler Current Profi lers
The use of Acoustic Doppler Current Profi lers (ADP/
ADCPTM) has become a common method of measuring river discharge. There are a handful of instruments
on the market today designed for use in larger or
smaller rivers. They have several traits in common.
ADCP instruments can be mounted on a moving
vessel, such as an inflatable boat (see Figure I.5.5).
The instrument measures water velocity, depth and
vessel path simultaneously to compute discharge.
This method computes the discharge as the vessel is
crossing the river. The total discharge measurement
(ΣQ1) is completed in a few minutes. The result
from one measurement is not enough to give an
accurate value of the water fl ow/discharge; it only
gives a freeze-frame picture of the flow. To get an
accurate value of the discharge of the river, it is
important to take the average of several transects.
At least four transects are recommended to calculate the discharge at a site. The actual river discharge
estimate will then be the average of the N individual transects discharge values


(5.14)
There is need for the instrument to communicate
with a computer that computes the discharge. As
an ADCP instrument processes the signal refl ected
off the particles in the water, it divides the water
column into a number of discrete segments stacked
in the vertical. These segments are called depth
cells. An ADCP instrument determines the velocity
and direction of each depth cell. At the same time
the signal from the bottom, called bottom-track,
measures the speed and direction of the boat. This
means that the boat does not have to cross perpendicularly to the flow.

Figure I.5.5. The layout of a typical acoustic Doppler measurement
(Source: United States Geological Survey, http://www.usgs.gov)


The procedures for collecting good data are becoming more standardized worldwide. The number of
transects depends on the difference between the
discharge measurements. If the discharge for any
of four transect differs more than 5 per cent, a
minimum of four additional transects should be
obtained and the average of all eight transects will
be the measured discharge. Sometimes even more
transects are made to reduce potential directional biases. The user must confi gure the instruments
before starting the measurements. The choice
among different modes of confi guration is based
on the conditions at the site (water depth, water
speed, etc.) at the time of measurement. Use of the
correct mode is important for greater accuracy in
discharge measurements. The user has to set proper
ADCP depth, distance to the banks and make sure
that the pitch and roll and the speed of the boat/
instrument is within accept able limits during the
measurements. A bias in any of these can result in
a signifi cant bias in the resulting measured
discharges.
Another kind of acoustic Doppler profi ler instrument
makes discharge measurements without using a
bottom track. Instead, it measures by use of sections
or “verticals”. Depending on the characteristics of
the river, the instrument takes 10–20 verticals, each
measured for 30–60 seconds, to make a discharge
measurement. Such instruments measure the full
vertical velocity profi le and can easily be suspended
from a bridge or suspended with a tag line across
the river.


The beams are all oriented in the direction of a twodimensional (2D) system that makes it possible to
measure close to the banks of the river (channel).
The user has to set the distance from the bank and
the software calculates the cross-sectional area.
Since there is no bottom tracking, the instrument
must be oriented in the direction of the fl ow and
move across the river in pre-defi ned segments/verticals. Failure to do this results in inaccurate discharge
measurements.


5.3.7.5.4 Acoustic Doppler Velocimeter
An Acoustic Doppler Velocimeter (ADV) is a singlepoint current meter designed specifi cally for
low-power measurements in slow-moving water.
These meters require much smaller water sampling
volumes than traditional current meters.
One type of ADV is Flowtracker, which is currently
the only hand-held ADV on the market. The instrument is an alternative to mechanical current meters
for making wading discharge measurements. The
Flowtracker consists of a probe head attached to a
top-setting wading rod with an interface. The interface allows entering the basic parameters required
to make a discharge measurement: station, distance,
depth and vertical location of the measurements
(0.6, or 0.2 and 0.8 of the depth). By using the
velocity-area method, it computes discharge by
multiplying the channel area and the mean channel velocity.
True 2D or 3D velocity data are output in Cartesian
coordinates (XYZ) relative to probe orientation.
Only the X component of velocity (Vx) is used for
river discharge measurements. The probe direction
has to be perpendicular to the tag line to ensure
proper discharge calculations. The operator does
not have to estimate the fl ow angle as is required
for 1D current meters.


5.3.7.5.5 Discharge measurements from fi xed
platform
In addition to use for vessel mounted discharge
measurements, an acoustic Doppler instrument can
be used on fi xed platforms to compute the discharge
in rivers. The instrument is normally mounted from
an underwater structure facing perpendicular to the
river fl ow, and measures water velocity in a twodimensional plane at multiple points. These
instruments are often called Acoustic Doppler
Velocity Meters (ADVM) (Gotvald, 2005).
The water velocity meaured by the ADVM is used to
compute the mean velocity of the river channel.
This is called the index velocity of the river. By
using the index velocity, the discharge can be
computed in different ways. This is called the
index-velocity method. An ADVM gives the opportunity to measure discharge in a river with no or
poor stage/discharge relationship. The index-velocity method is basically computing the discharge
from the equation Q = VA, where Q is the total
discharge, V is the mean velocity and A is the channel area. Use of ADVMs on fi xed platforms to
provide index velocity measurements for river
discharge has increased recently.


5.4 STREAM-GAUGING STATIONS
5.4.1 Purpose of stream-gauging
stations
The purpose of stream-gauging stations is to provide
systematic records of stage and discharge. Continuous
streamfl ow records are necessary in the design of
water supply and waste systems, in designing hydraulic structures, in the operations of water management
systems, and in estimating the sediment or chemical
loads of streams, including pollutants.
Since continuous measurement of discharge is not
usually feasible, unless one of the methods in
5.3.7.3 and 5.3.7.4 is used, records of discharge are
computed from the relationship between stage and
discharge, as defi ned by periodic discharge measurements and a systematic record of stage, or
from a measuring structure that has been calibrated
in either a laboratory or the field.

5.4.2 Selection of site
The selection of streams to be gauged should be
governed by the principles of network design (2.4)
and the proposed use of the data. The selection of a
particular site for the gauging station on a given
stream should be guided by the following criteria
for an ideal gauge site:
(a) The general course of the stream is straight for
about 100 m upstream and downstream from
the gauge site;
(b) The total fl ow is confi ned to one channel
at all stages and no fl ow bypasses the site as
subsurface fl ow;
(c) The stream bed is not subject to scour and fi ll
and is free of weeds;
(d) Banks are permanent, high enough to contain
fl oods, and free of brush;
(e) Unchanging natural controls are present in the
form of a bedrock outcrop or other stable riffl e
during low fl ow, and a channel constriction for
high fl ow, or a fall or cascade that is unsubmerged
at all stages to provide a stable relationship
between stage and discharge. If no satisfactory
natural low-water control exists, then installation
of an artifi cial control should be considered;
(f) A site is available, just upstream from the
control, for housing the stage recorder where
the potential for damage by drifting ice or
water-borne debris is minimal during fl ood
stages. The elevation of the stage recorder itself
should be above any fl ood likely to occur during
the life of the station;
(g) The gauge site is far enough upstream from the
confl uence with another stream or from tidal
effect to avoid any variable infl uences which
the other stream or the tide may have on the
stage at the gauge site;
(h) A satisfactory reach for measuring discharge at
all stages is available within reasonable proximity of the gauge site. It is not necessary that low
and high fl ows be measured at the same stream
cross-section;
(i) The site is readily accessible for ease in the installation and operation of the gauging station;
(j) Facilities for telemetry or satellite relay can be
made available, if required;
(k) If ice conditions occur, it would still be possible
to record stage and measure discharge;
(l) The fl ow in the channel section containing the
gauging site is subcritical at all stages;
(m) There are no waves and ripples on the water
surface in the vicinity of the gauging site.

In many instances, it may be impossible to meet all
of these criteria. Judgement is then required to
select the most suitable site for the gauge.
5.4.3 Stage-discharge controls
The physical element or combination of elements
that control the stage-discharge relationship is
known as a control. The major classification of
controls differentiates between section control
and channel control. Another classification
differentiates between natural and artificial
controls.


Section control exists when the geometry of a single
cross-section is such as to constrict the channel, or
when a major downward break in bed slope occurs
at a cross-section. The constriction may result from a
local rise in the stream bed, as at a natural riffle e or
rock ledge outcrop or at a constructed weir or dam. It
may also result from a local constriction in width,
which may occur naturally or may be caused by
some man-made channel encroachment, such as a
bridge with a waterway opening that is considerably
narrower than the width of the natural channel.
Channel control exists when the geometry and
roughness of a long reach of channel downstream
from the gauging station are the elements that
control the relationship between stage and
discharge. The length of channel that is effective as
a control increases with discharge. Generally, fl atter
stream gradients will result in longer reaches of
channel control.
A low dam, weir or fl ume is often built in the channel to provide an artifi cial control. Such controls
are usually submerged by high discharges, but they
provide a stable stage-discharge relationship in the
low to medium fl ow range.
The two attributes of a good control are resistance
to change – ensuring stability of the stage-discharge
relationship – and sensitivity, whereby a small
change in discharge produces a signifi cant change
in stage.


5.4.4 Measuring structures
At some gauging sites it is feasible to utilize an artifi cial control of such shape that head-discharge
relationships can be determined without calibration, that is, by the application of a discharge
formula. There is a set of weirs and flumes that have
well-established relationships between head and
discharge. However, only under favourable fi eld
conditions can the established formulae for some types of weirs and flumes be applied accurately. If
these structures are used to measure flow directly
from water level readings, it is important that care
be taken in their construction and operation and
that the most suitable formulae be used (WMO,
1986b; ISO, 1977b, 1980, 1983, 1984, 1989).
Under less favourable conditions, in situ calibration
is necessary to establish the extent of the departures
from the standard formulae or to develop the headdischarge relationship. It is particularly important
at low fl ow to measure periodically the discharge by
other means in order to detect changes in the
discharge coefficient caused by sediment deposits
in the pool or growth of algae on the weir or
fl ume.
The material in this Guide is limited to the general
considerations involved in the selection and use of
weirs and fl umes at gauging stations. Specific information on their geometries and head-discharge
formulae are presented in the Use of Weirs and
Flumes in Stream Gauging (WMO-No. 280).
5.4.4.1 Scope
Weirs and fl umes for use at gauging stations may be
catalogued into three groups:
(a) Thin-plate weirs generally used on small,
clear-fl owing streams or small research
watersheds;
(b) Flumes used on small streams and canals
conveying sediment and debris or in other
situations where the head loss associated with
thin-plate weirs is unacceptable;
(c) Broad-crested, triangular-profi le and roundshaped weirs used on larger streams.
Weirs and fl umes may be free-fl owing or submerged.
In the first case, the discharge is a function of the
headwater elevation, and accurate calibrations are
possible. For submerged conditions, the discharge
is a function of both the headwater and tailwater
elevations, and less accuracy is obtained by use of
laboratory calibrations. At many sites, weirs or
fl umes are used to measure only the lower range of
discharge, and the stage-discharge relationship for
the upper range of discharges is determined by
direct methods.
5.4.4.2 Selection of structure
The choice of a measuring structure depends on
costs, the characteristics of the stream and channel
at the site, the range of discharges, the accuracy
desired and the potential head loss. Criteria to be
considered in choosing a structure include:
(a) Cost is usually the major factor in deciding
whether or not a measuring structure is to be
built. The cost of the structure is affected most
by the width of the stream and the type or
condition of the bed and bank material. Stream
width governs the size of the structure, and bed
and bank material govern the type of construction that must be used to minimize leakage
under and around the structure;
(b) Channel characteristics and fl ow conditions
influence the design of the measuring structure.
Factors controlling velocity or Froude number,
sediment loads and the stability of the bed need
to be considered in the structure design;
(c) The range of discharge, range of stage, desired
sensitivity and allowable head loss must also be
considered in structure design and positioning.
Submergence by high flows or from backwater
infl uence both the design and elevations of the
structure. The sensitivity, that is, the change in
stage corresponding to change in discharge at
very low fl ows, may dictate whether a V-crest or
fl at crest is appropriate;
(d) Cheap, portable weirs made of canvas and light
metal plates, for example, may be used on small
rivers for limited periods of time.
5.4.4.3 Measurement of head
The head over the structure is usually measured at a
distance upstream from the structure equal to about
three times the depth of water, hmax, on the control
at the maximum stage for which the section control
is effective. Some special weir shapes and all fl umes
require that stage be measured at specifi c distances
from the control section that differ from the general
rule of 3 x hmax. The locations for the gauge or gauge
intake for these special cases are described in the Use
of Weirs and Flumes in Stream Gauging (WMONo. 280). The zero of the gauge should be set at crest
elevation and should be checked regularly.
5.4.4.4 Operation of measuring structures
Both the channel and structure are subject to
changes with time that may affect the head discharge relationship. Sand, rocks or debris may be
deposited in the approach section or on the structure
itself. Algae may grow directly on the crest of the
structure during summer and ice may form on the
structure during winter.
For optimum accuracy the approach channel to
weirs should be kept clean and free of any accumulation of silt or vegetation. The structure must be
kept clean and free of debris, algae and ice. Damage
to critical parts of the structure should be repaired.

The datum of the gauge should be checked periodically. Periodic discharge measurements should also
be made to define possible changes in the original
calibration.
5.4.5 Stage-discharge relationships
5.4.5.1 General
The stage-discharge relationship for most gauging
stations is defined by plotting the measured
discharges as the abscissa and the corresponding
stage as the ordinate (ISO, 1981). The shape of the
stage-discharge relationship is a function of the
geometry of the downstream elements of the channel that act as the control. When plotted on
rectangular coordinate paper, the relationship is
generally concave downwards (depends on the
exponent value) since discharge often can be
described by a power function of the fl ow depth.
Hence, when plotted on logarithmic coordinate
paper, the medium- and high-stage sections of the
relationship are often approximately linear if the
stage represents the effective head on the control
for medium and high stages. If this is not linear, the
stage-discharge relationship is typically comprised
of two or more segments because of shifts in geometry and/or channel resistance. The stage-discharge
relationship can readily be expressed by a mathematical equation derived from the available
measurements. This equation can be determined by
graphical methods or regression methods.
Independent of what method is used for deriving
the stage-discharge relationship, its accuracy is
determined by:
(a) The number of available measurements;
(b) The spread of the measurements;
(c) The average discharge measurement uncertainty.
An estimated stage-discharge relationship should
not be extrapolated. Where it is desirable to extrapolate, the application of indirect methods based on
the physical conditions of the actual channel and
hydraulic control is recommended.
At many sites, the discharge is not a unique function of stage, and additional variables must be
measured continuously to obtain a discharge
record. For example, in situations where variable
backwater at the gauge is caused by a downstream
tributary, by tidal effect or by downstream reservoir operation, an auxiliary stage gauge must be
installed to measure continuously the fall of the
water surface in the gauged reach of the channel.
Where fl ow is unsteady and the channel slopes
are fl at, the rate of change of stage can be an
important variable, and a given discharge that
occurs on a rising stage will have a lower gauge
height than the same discharge occurring on a
falling stage.
5.4.5.2 Stability of stage-discharge
relationships
The stability of a stage-discharge relationship is
directly related to the stability of the control. For
natural section controls, a rock-ledge outcrop will
be unaffected by high velocities. Boulder, gravel
and sandbar riffles are likely to shift. Boulder riffles
are the most resistant to movement, and sandbars
are the least. Of the natural channel controls, those
found in sand-channel streams are the most likely
to change as a result of velocity-induced scour and
deposition.


The growth of aquatic vegetation on section
controls increases the stage for a given discharge,
particularly in the low-fl ow range. Vegetal growth
on the bed and banks of channel controls also
affects the stage-discharge relationship by reducing velocity and the effective waterway area. In
temperate climates, accumulation of water-logged
leaves on section controls during autumn may
clog the interstices of alluvial riffl es and raise the
effective elevation of natural section controls. The
first ensuing stream rise of any significance usually
clears the control of leaves.
Ice cover also affects the stage-discharge relationship of a stream by causing backwater that varies
in effect with the quantity and nature of the ice.
If the section control remains open and if the
gauge is not too far from the control, there probably will be little or no backwater effect even
though the entire pool is ice covered. The only
effect of the ice cover will be to slow the velocity
of approach, and that effect probably will be
minor. However, if the gauge is a considerable
distance upstream from the riffl e, surface ice on
the pool may cause backwater when the covered
reach of the pool becomes a partial channel
control.
Surface ice forming below a section control may
jam and dam water suffi ciently to cause backwater
effects at the control. Anchor ice may build up the
bed or control to the extent that a higher than
normal stage results from a given discharge. The
magnitudes of ice effects can be determined accurately only by measuring the discharges, observing
the corresponding stages and analysing the differences between the observed stage and the discharge
corresponding to the open-water stage-discharge
relationship.

The various additional conditions that have to be
taken into account in making discharge measurements under ice conditions and the procedures for
making such measurements are described in
5.3.2.5.
Artificial controls to eliminate or alleviate many of
the undesirable characteristics of natural section
controls. Not only are they physically stable, but
also they are less subject to the cyclic or progressive growth of aquatic vegetation. Algal slimes that
sometimes form on artificial controls can be
removed with a wire brush, and the controls can
be self-cleaning with regard to fallen leaves. In
moderately cold climates, artificial controls are
less likely to be affected by the formation of winter
ice than are natural controls. However, even when
the artificial control structure is unchanged, the
stage-discharge relationship may be affected by
changes in the velocity of approach caused by
scour and/or fill, or by vegetal growth in the
approach channel.


5.4.5.3 Frequency of discharge
measurements
Factors to be considered in scheduling the number
and distribution of discharge measurements
throughout the year include:
(a) Stability of stage-discharge relationship;
(b) Seasonal discharge characteristics and variability;
(c) Accessibility of the gauge in various seasons.
Many discharge measurements are necessary at a
new station to define the stage-discharge relationship throughout the entire range of the stage.
Periodic measurements are then necessary to define
changes in the stage-discharge relationship. A minimum of 10 discharge measurements per year is
recommended.


Adequate definition of discharge during flood and
under ice conditions is of prime importance. It is
essential that the measurement programme provides
for non-routine measurement of discharge at these
times.
Where it is important to record streamflow
continuously throughout the year, discharge
measurements should generally be made more
frequently when the stream is under ice cover.
During freeze-up and break-up periods, measurements should be obtained as often as possible
because of the extreme variability of flow. In midwinter, the frequency of the measurements will
depend on climate, accessibility, size of stream,
winter runoff characteristics and the required accuracy. In very cold climates, where discharge follows
a smooth recession curve, fewer measurements are
required than for a stream in a climate of alternate
freezing and melting.
5.4.6 Computation of mean
gauge height of a discharge
measurement [HOMS E71]
Stage and corresponding time should be noted at
intervals to identify segments of total
discharge with time and stage. Usually the stage at
the mid-time of the measurement or the average of
the stage at the beginning and end of the
measurement can be used as the mean stage corresponding to the measured discharge. If the
stage does not change linearly with time the
following weighting procedure should be used,
where h_ is the weighted stage and Q1, Q2,… QN are
segments of discharge corresponding to stages:

5.5 SEDIMENT DISCHARGE AND YIELD
5.5.1 General [HOMS E09]
Sediment is transported by flowing water in different ways. The sediment grains may be moved by
saltation, rolling or sliding on or near the bed or
may be swept away from it and kept in suspension. The type of movement experienced by the
grains depends upon their physical characteristics (size and form of particles, specific weight,
etc.) and upon the grain-size composition of the
sediment, as well as upon fl ow velocities and
depths. The different phases of sediment transportation generally occur simultaneously in
natural streams, and there is no sharp line of
demarcation between them. For convenience,
sediment discharge is divided into two categories:
suspended-sediment and bed-material discharge.
The latter consists of grains sliding, rolling or
saltating on or near the bed.


This chapter provides guidance on the collection
of sediment-discharge data. For each phase of
transport, a more in-depth discussion of this
topic can be found in the Manual on Operational
Methods for Measurement of Sediment Transport
(WMO-No. 686).

5.5.2 Selection of site
The same criteria used for the selection of a site for
a water-discharge measurement should be used in
selecting a site for measuring sediment transport
(5.3.2.1 and 5.4.2).
5.5.3 Measurement of suspended sediment discharge
5.5.3.1 Sampling instruments and in situ
gauges [HOMS C10]
Several types of suspended-sediment samplers are
in use, for example, instantaneous, bottle, pumping or integrating. However, only some of these are
designed so that the velocity within the cutting
circle of the sampler intake is equal to the ambient
stream velocity. This feature is essential so that the
samples obtained are truly representative of the
suspended-sediment discharge at the point of measurement. The well-designed sampler faces the
approaching fl ow, and its intake protrudes upstream
from the zone of disturbance caused by the presence of the sampler.
Instantaneous samples are usually taken by trap
samplers consisting of a horizontal cylinder
equipped with end valves that can be closed
suddenly to trap a sample at any desired time and
depth. The very simple bottle sampler is corked or
provided with an orifice of variable diameter, or
wide open. As soon as the bottle is opened and air
within the bottle is being displaced by the sample,
bubbling takes place at the mouth, which slows the
filling process. Consequently, bottle-sampling is
not actually instantaneous.


The pumping sampler sucks the water-sediment
mixture through a pipe or hose, the intake of
which is placed at the sampling point. By regulating the intake velocity, the operator can obtain a
sample that is representative of the sediment
concentration at the point of measurement. The
integrating sampler consists of a metallic streamlined body equipped with tail fi ns to orient it into
the flow. The sample container is located in the
body of the sampler. An intake nozzle of variable
diameter projects into the current from the sampler
head. An exhaust tube, pointing downstream,
permits the escape of air from the container. Valve
mechanisms enclosed in the head are electrically
operated by the observer to start and stop the
sampling process.


A relatively new method of in situ determination of
suspended-sediment concentration is the use of
optical or nuclear gauges. The working principle of
these instruments is that a visible light of X-ray
emitted by a source with constant intensity is scattered and/or absorbed by the suspended-sediment
particles. The decrease of the intensity measured by a
photoelectric or nuclear detector situated at
constant distance from the source is proportional
to the sediment concentration, if other relevant
characteristics of water and sediment (chemical,
mineral composition, etc.) remain unchanged.
The overall design of suspended-sediment samplers
should be checked by towing them in still water at
a known velocity or by holding them in flowing
water of known velocity. The optical and nuclear
gauges must be calibrated by simultaneous and
repeated sampling in sediment-laden fl umes and
natural streams.


5.5.3.2 Measurement procedure
Samples of suspended sediment in streams are taken
in the discharge-measuring cross-sections, but not
necessarily in the velocity-measuring verticals. In
lakes, the locations of sampling verticals are scattered over an area, because here the measurements
are usually aimed at the determination of distribution of sediment concentration in time and space.
The samplers are suspended in the water on a rod or
on a wire.


In streams, there are two methods that give comparative results:
(a) Equal discharge increment (EDI) method: The
cross-section is divided into 3 to 10 subsections
of about equal discharge. A depth-integrated
sample is taken at each vertical in the centroid
of each subsection by lowering the sampler
from the stream surface to the bed and back at
a uniform transit rate. This gives a discharge weighted sample for each centroid;
(b) Equal transit rate (ETR) method: The stream
width is divided into 6 to 10 equal distances
separated by the verticals and one depth integrated sample is taken at each vertical at a constant transit rate. In the latter case, all
samples can be composited into a single representative discharge-weighted sample (ISO,
1977b).


By using a point sampler, samples may also be taken
at evenly spaced points at each vertical mentioned
above, and the sediment concentrations obtained
are weighted by the ratio of the velocity at the given
point to the mean velocity in the vertical. In practice, this procedure can be combined with the
mid-section method of discharge measurement

(5.3.2.4) because the velocity measuring and
sampling verticals coincide.
The optical and nuclear sediment gauges may be
used both for point- and depth-integrating measurements, provided the electrical signals from the
detector are summarized by a scalar. Depending
upon the statistical characteristics of counting by a
particular instrument, the usual counting period is
three to five minutes.


5.5.3.3 Determination of sediment
concentration
Suspended-sediment samples are usually processed
and analysed in special laboratories for the determination of the sediment concentration. Evaporation,
filtration or displacement methods are generally
used for this purpose. In general, the evaporation
method is suitable for use with low concentrations.
Filtering may be used for samples with medium to
high concentrations. The displacement method,
however, is suitable only when the concentration is
high (WMO, 1989). The sample is usually allowed a
settling time of one to two days, the water is then
carefully drained off and the remaining sediment is
oven dried at a temperature of about 110°C, and
weighted. If the sediment is separated by evaporation, a correction must be made for dissolved solids.
The concentration of suspended sediment is the
weight of dried sediment contained in a unit
volume of the sediment-water mixture and is
expressed in mgl–1, gl–1 m–3 or in kg m–3.
Sediment samplers have been standardized in
some countries to have a container capacity of
one litre or less. In such cases, sampling should
be repeated until the required volume of sediment sample is obtained (ISO, 1977b).
The intensities of light or X-ray indicated by the
submerged photoelectric or nuclear probes of in
situ gauges should be divided by the intensity
measured in clear water and the sediment
concentration corresponding to this ratio is
read from the calibration curves of these
instruments.


5.5.3.4 Computation of suspended-sediment
discharge
For the EDI method, the weighted mean sediment
concentration, c_s, in kg m–3 for the entire crosssection is computed as:

where qp is the partial discharge in the subsection in
m3 s–1, and cq is the discharge weighted concentration in the vertical at the centroid of the subsection in kg m–3 (ISO, 1977b).

For the ETR method the concentration of the
composite sample is the weighted mean concentration in the entire cross-section. The suspended sediment discharge, Qs
, is computed as:

where Qs is in kg s–1 and Q is the stream discharge in m3 s–1.

5.5.3.5 Continuous record of suspended sediment discharge
A continuous record of suspended-sediment
discharge may be computed from a record of
stream discharges and systematic samples of
suspended-sediment concentration. The samples
should be taken daily during periods of low and
mean flow and more frequently during floods.
The most valuable information concerning the
time-variation of concentration and its peak
values can be obtained by the continuous recording of signals supplied by the photoelectric or
nuclear suspended-sediment gauges during flood
periods. The peak in concentration usually
precedes peak flow, and loops can be observed on
plots of the water discharge versus sediment
discharge, similar to those in stage-discharge
rating curves during floods.


The samples or observation records are collected at
a single vertical in the cross-section, preferably
using the depth-integrating procedure. The relation
between the concentration at this vertical and the
mean concentration in the section must be established by detailed measurements of the distribution
of sediment in the cross-section, as outlined in
5.5.3.2. This relation is not necessarily linear and
constant throughout the year, nor in all ranges of
sediment concentration.


5.5.3.6 Use of remote-sensing techniques
The determination of the amount of sediment in
water is based on the reflectance of radiation in
the visible and IR parts of EMS (WMO, 1972). In
general, reflection is a non-linear function of the
concentration of suspended sediments with maximum reflectance dependent on wavelength and
suspended sediment concentration. Because turbidity and suspended sediments are closely linked in most water bodies, estimates of turbidity can also
be made. A limitation on the use of this technique
is the need to collect field data to calibrate the
relationship between suspended sediments and
reflectance. Furthermore, scanner data can be used
without calibration data to map relative suspended
sediment concentrations in river plumes and draw
conclusions about sediment deposition patterns in
lakes and estuaries. A good review of applications of
remote-sensing to estimation of suspended sediments can be found in Dekker and others (1995).


5.5.4 Measurement of bed-material
discharge
5.5.4.1 Instrumentation [HOMS C12]
The field measurement of bed-material discharge is
difficult because of the stochastic nature of the sediment movement and because the phenomenon
takes place in the form of ripples, dunes and bars.
No single apparatus has proved to be completely
adequate for trapping the largest and smallest sediment particles with the same efficiency, while
remaining in a stable, fl ow-oriented position on the
stream bed, and still not altering the natural flow
pattern and sediment movement. Available samplers
can be classifi ed into three types: basket, pan and
pressure-difference (ISO, 1977c). Another type of
sampler is the slot or pit-type sampler which is
adaptable for use mainly in relatively small rivers
and particularly for experimental study or calibration of samplers (Emmett, 1981).
Basket samplers are generally made of mesh material with an opening on the upstream end,
through which the water-sediment mixture
passes. The mesh should pass the suspended
material but retain the sediment moving along
the bed.


Pan samplers are usually wedge-shaped in longitudinal section and are located so that the point
of the wedge cuts the current. The pan contains
baffles and slots to catch the moving material.
Pressure-difference samplers are designed to produce
a pressure drop at the exit of the sampler which is
sufficient to overcome energy losses and to ensure an
entrance velocity equal to that of the undisturbed
stream. A perforated diaphragm within the sampler
forces the flow to drop its sediment into the retaining chamber and to leave through the upper exit.


It is necessary, because of several uncertainties
involved in sampling, to determine an efficiency
coefficient for each type of sampler. The calibration
generally takes place in a laboratory flume, where
the bed-material discharge can be directly measured
in a sump at the end of the fl ume, although uniformtransport conditions over the width and length of
the fl ume are difficult to maintain. Even under
favourable conditions, efficiency factors are not
easily determined because they vary according to,
among others, the grain-size composition of the
bed material and the degree of fullness of the
sampler. An efficiency of 60 to 70 per cent can be
regarded as satisfactory.


5.5.4.2 Measurement procedure
Bed-material discharge is determined from the
amount of sediment trapped per unit time in a
sampler located at one or more points on the stream
bed. There should generally be 3 to 10 measurement points in a cross-section, depending on the
width of the cross-section and the sediment concentration distribution. In determining the distribution
of sampling points, it should be noted that, except
during flood periods, bed-material transport takes
place only in a part of the stream width.
The inclusion of a zero measurement in the computation of bed-material discharge can lead to
uncertainties in the result even though the sampling
point may be situated between two moving strips
of the stream bed. Uncertainties can also occur if a
measured rate of transport is extended over a
segment of the cross-section with low or zero sediment movement.
On gravel-bed streams, of which partial bed-material movement is most characteristic, different types
of acoustic detectors can help to solve this problem.
Submerged to a depth near the bed, these detectors
pick up the sound of moving gravel, indicating the
movement of bed material at this particular point.
Moreover, the intensity of the sound and that of
the sediment transport may be correlated.
The samplers (see, for example, Figure I.5.6) are
lowered to the bottom and held in position by a rod
or a wire. The duration of the sampling period is
usually a few minutes, depending on the dimensions of the sampler and on the intensity of the
sediment transport. When low-flow velocities exist
near the bed, the downstream forces are reduced
and the sampler tends to dive into the stream bed
and scoop up bed material that is not in transport.
A similar tendency can develop during an abrupt or
incautious lifting of the sampler.


Measurements should be made at various stream
discharges so that a rating may be prepared showing

the relationship between stream discharge and bedmaterial discharge. Owing to the highly complex
mechanism and random nature of sediment
transport and to the errors of sampling, one single
catch at a measuring point can provide a very
uncertain estimate of the true bed-material
transport. Therefore, repeated sampling should be
carried out at each point. The number of repetitions
depends on the local circumstances. However,
statistical analyses of field data resulting from up to
100 repetitions have shown that only the bedmaterial discharge can be measured with restricted
accuracy, unless an impracticably large number of
samples are taken at each point.


5.5.4.3 Computation of bed-material
discharge
The sediment collected in the sampler is dried and
weighed. The dry weight, when divided by the time
taken for the measurement and the width of the
sampler, gives the bed-material discharge per unit
width of stream at the point of measurement, qb. A
curve showing the distribution of qb in the stream
width can be constructed based on data obtained at
the sampled points. The area enclosed between this
curve and the water-surface line represents the total
daily bed-material discharge over the entire crosssection Qb. The value of Qb can also be computed by
using the measured qb data as:

where Qb is in kg s–1, qb is in kg s–1 m–1 and x is in
metres. The variable x represents the distance
between sampling points, between a marginal point
and the edge of the water surface, or that of the
moving strip of stream bed.

Figure I.5.6. Delft Nile sampler consisting of a
bed-load and suspended-load sampler as well as
an underwater video camera

The existence of dams trapping most of the sediment transported by upstream river reaches offers
the possibility of estimating the annual or
seasonal sediment discharge by successively
surveying suitable selected profiles of the reservoir and by computing the volumes occupied by
the trapped sediment. This method, combined
with regular suspended-sediment sampling
upstream and downstream of the dam, can
provide acceptable estimates of bed-material
discharge.
5.5.4.4 Continuous record of bed-material
discharge
A continuous record of bed-material discharge can
be obtained by relating bed-material discharge to
stream discharge or other hydraulic variables with
available records. This relationship can be assumed
approximately linear for water discharges above the
limiting value corresponding to the beginning of
sediment movement because the tractive force of
the flow increases in direct proportion to the
increase in stream discharge. Bed-material transport
is of primary interest in all investigations concerning stream bed-changes.


References and further reading
Botma, H.C. and R.E. Klein, 1974: Some notes on the
research and application of gauging by electromagnetic and ultrasonic methods in The Netherlands.
Proceedings of Symposium on River Gauging by
Ultrasonic and Electromagnetic Methods,
16–18 December 1974, Water Research Centre,
Department of the Environment, Water Data Unit,
University of Reading.
Dekker, A.G., T.J. Malthus and H.J. Hoogenboom, 1995:
The remote-sensing of inland water-quality. In:
Advances in Environmental Remote Sensing,
Chapter 11, John Wiley, New York, pp. 123–142.
Emmett, W.W., 1981: Measurement of bed-load in rivers.
Proceedings of the Florence Symposium,
22–26 June 1981, Florence, International Association
of Hydrological Sciences Publication No. 133,
Wallingford, pp. 3–15.
Gotvald, A.J., 2005: The use of hydroacoustic current
meters to measure the fl ow of Georgia streams.
Proceedings of the 2005 Georgia Water Resources
Conference, 25–27 April 2005, Athens, Georgia
(http://ga.water.usgs.gov/pubs/other/gwrc2005/pdf/
GWRC05_Gotvald.pdf).
Halliday, R.A., W.M. Archer and P.I. Campbell, 1975:
The Niagara River acoustic streamflow measurement
system. Technical Bulletin No. 86, Inland Waters
Directorate, Environment Canada, Ottawa

Herschy, R.W. and W.R. Loosemore, 1974: The ultrasonic
method of river fl ow measurement. Proceedings
of Symposium on River Gauging by Ultrasonic and
Electromagnetic Methods, 16–18 December 1974,
Water Research Centre, Department of the
Environment, Water Data Unit, University of
Reading.
Herschy, R.W. and J.D. Newman, 1974: The electromagnetic method of river fl ow measurement. Proceedings
of Symposium on River Gauging by Ultrasonic and
Electromagnetic Methods, 16–18 December 1974,
Water Research Centre, Department of the
Environment, Water Data Unit, University of
Reading.
Holmes, H., D.K. Whirlow and L.G. Wright, 1970: The
LE (Leading Edge) flowmeter: a unique device for
open channel discharge measurement.
Proceedings of the International Symposium on
Hydrometry, 13–19 September 1990, Koblenz,
UNESCO/ WMO/ International Association of
Hydrological Sciences Publication No. 99,
pp. 432–443.
International Organization for Standardization,
1973a: Liquid Flow Measurement in Open Channels:
Dilution Methods for Measurement of Steady Flow.
Part 1: Constant rate injection method. ISO 555-1,
Geneva.
International Organization for Standardization, 1973b:
Liquid Flow Measurement in Open Channels: Slope-area
Method. ISO 1070, Geneva.
International Organization for Standardization, 1974:
Measurement of Flow in Tidal Channels. ISO 2425,
Geneva.
International Organization for Standardization,
1976: Liquid Flow Measurement in Open Channels:
Calibration of Rotating-element Current-meters in
Straight Open Tanks. ISO 3455, Geneva.
International Organization for Standardization, 1977a:
Liquid Flow Measurement in Open Channels by Weirs
and Flumes: End-depth Method for Estimation of Flow
in Rectangular Channels with a Free Overfall. ISO 3847,
Geneva.
International Organization for Standardization, 1977b:
Liquid Flow Measurement in Open Channels: Methods
for Measurement of Suspended Sediment. ISO 4363,
Geneva.
International Organization for Standardization, 1977c:
Liquid Flow Measurement in Open Channels: Bed
Material Sampling. ISO 4364, Geneva.
International Organization for Standardization, 1979a:
Measurement of Liquid Flow in Open Channels: Movingboat Method. ISO 4369, Geneva.
International Organization for Standardization, 1979b:
Liquid Flow Measurement in Open Channels: Velocityarea Methods. Second edition, ISO 748, Geneva.
International Organization for Standardization, 1980:
Water Flow Measurement in Open Channels Using Weirs
and Venturi Flumes. Part 1: Thin-plate weirs. ISO
1438-1, Geneva.
International Organization for Standardization, 1981:
Measurement of Liquid Flow in Open Channels. Part 1:
Establishment and operation of a gauging station
and Part 2: Determination of the stage-discharge
relation. ISO 1100, Geneva.
International Organization for Standardization,
1983: Liquid Flow Measurement in Open Channels:
Rectangular Trapezoidal and U-shaped Flumes. ISO
4359, Geneva.
International Organization for Standardization, 1984:
Liquid Flow Measurement in Open Channels by Weirs
and Flumes: Triangular Profi le Weirs. Second edition,
ISO 4360, Geneva.
International Organization for Standardization, 1985:
Liquid Flow Measurement in Open Channels: Velocityarea Methods – Collection and Processing of Data
for Determination of Errors in Measurement. Second
edition, ISO 1088, Geneva.
International Organization for Standardization, 1987:
Liquid Flow Measurement in Open Channels: Dilution
Methods for Measurement of Steady Flow. Part 2:
Integration method. Second edition, ISO 555-2,
Geneva.
International Organization for Standardization, 1988a:
Liquid Flow Measurement in Open Channels: Rotating
Element Current-meters. Third edition, ISO 2537,
Geneva.
International Organization for Standardization,
1988b: Liquid Flow Measurement in Open Channels:
Vocabulary and Symbols. Third edition, ISO 772,
Geneva.
International Organization for Standardization, 1989:
Liquid Flow Measurement in Open Channels by Weirs
and Flumes: Rectangular Broad-crested Weirs. Second
edition, ISO 3846, Geneva.
Kinosita, T., 1970: Ultrasonic measurement of discharge
in rivers. Proceedings of the International Symposium
on Hydrometry, 13–19 September 1990, Koblenz,
United Nations Educational, Scientifi c and Cultural
Organization, World Meteorological
Organization and International Association of
Hydrological Sciences Publication No. 99,
pp. 388–399.
Lenormand, J., 1974: Débimètre à ultrasons mdl 2 compte
rendu d’essais: ponts et chaussées. Service des voies
navigables du Nord et du Pas-de-Calais, Service
hydrologique centralisateur, Lambersant.
Prokacheva, V.G., 1975: Otsenka prigodnosti televizionnoj
informatsii meteorologiceskih ISZ ‘Meteor’ dlya opredeleniya ledovoj obstanvki na ozerah i vodokhraniliscakh
(Estimate of the suitability of television data from
the ‘Meteor’ meteorological satellite for
determining ice conditions on lakes and
reservoirs). Proceedings of the State Hydrological
Institute, St Petersburg, No. 205, pp. 115–123.

Smith, W., 1969: Feasibility study of the use of the acoustic
velocity meter for measurement of net outfl ow from the
Sacramento-San Joaquin Delta in California. United
States Geological Survey Water-Supply Paper 1877,
Reston, Virginia.
Smith, W., 1971: Application of an acoutic streamfl ow
measuring system on the Columbia river at The
Dalles, Oregon. Water Resources Bulletin, Volume 7,
No. 1.
Smith, W., 1974: Experience in the United States of
America with acoustic fl owmeters. Proceedings
of Symposium on River Gauging by Ultrasonic and
Electromagnetic Methods, 16–18 December 1974,
Water Research Centre, Department of the
Environment, Water Data Unit, University of
Reading.
Smoot, G.F. and C.E., Novak, 1969: Measurement of
Discharge by the Moving-boat Method. Book 3, Chapter
A11, United States Geological Survey, Techniques of
Water-Resources Investigations.
World Meteorological Organization, 1971: Use of Weirs
and Flumes in Streamgauging: Report of a Working
Group of the Commission for Hydrology. Technical
Note No. 117, WMO-No. 280, Geneva.
World Meteorological Organization, 1980:
Manual on Stream Gauging. Volumes I and II,
Operational Hydrology Report No. 13,
WMO-No. 519, Geneva.
World Meteorological Organization, 1986a: Level and
Discharge Measurements under Diffi cult Conditions
(Ø.A. Tilrem). Operational Hydrology Report No. 24,
WMO-No. 650, Geneva.
World Meteorological Organization, 1986b: Methods of
Measurement and Estimation of Discharges at
Hydraulic Structures (Ø.A. Tilrem). Operational
Hydrological Report No. 26, WMO-No. 658,
Geneva.
World Meteorological Organization, 1989: Manual on
Operational Methods for the Measurement of Sediment
Transport. Operational Hydrology Report No. 29,
WMO-No. 686, Geneva.
World Meteorological Organization, 1992: Remote Sensing
for Hydrology: Progress and Prospects (R. Kuittinen).
Operational Hydrological Report No. 36,
WMO-No. 773, Geneva.
World Meteorological Organization, 2006: Technical
Regulations, Volume III, WMO-No. 49,
Geneva.