7. Head Measurements

The head is usually sensed either in the channel itself or in a stilling well located to one side of the channel. The stilling well is connected by a small pipe to the channel. Many methods can be used to detect the water surface in a stilling well or in the flume channel. Some methods exploit the electrical conductance of water and capacitance of immersed insulated wires. Sonic sensors depend on timing sound pulses reflected from the water surface. Measurement heads can also be determined with a variety of pressure sensing devices. The most frequently used methods are wall-mounted staff gages in the entrance section of the flume or in a stilling well or float-operated recorders placed in a stilling well.

(a) Location for Head Measurement

The measuring station for short-form flumes must be installed as specified to match closely the location used when the flume was empirically calibrated. For example, the measuring station of a Parshall flume is in the convergence water surface drawdown. For long-throated measurement structures, the gaging or head-measuring station should be located sufficiently far upstream to avoid detectable water surface drawdown, but close enough for the energy losses between the gaging station and approach section to be negligible. This placement is particularly critical if the ratings are based on coefficient values in a discharge equation as discussed in Bos (1989). In the computer derived ratings, drawdown and friction losses caused by the gage location are an integral part of the calculation. Therefore, the gage should be located as indicated in the precomputed long-throated structure design and selected tables.

(b) Selection of Head-Measurement Device

The success or failure of the structure and the value of the collected data depend closely on the proper selection of a suitable head measurement device. The three most important factors that influence this selection are: (1) frequency of discharge measurement, (2) allowable error in the head detection, and (3) type of measurement structure under consideration.

The usual expected reading errors in the sill-referenced head are listed in table 8-1 for some common head measurement devices. The listed errors are higher than the expected random errors, partly to compensate for the effects of several systematic errors, such as zero-setting, instrument lag, reading error, temperature, and stilling well leakage. If no device with sufficient accuracy is found from this procedure, two choices are available: (1) allow greater error in the measured discharge for the minimum required head loss or (2) redesign the structure with a narrower bottom width, resulting in a higher value of minimum measurement head.

(c) Staff Gages

Periodic readings on a calibrated staff gage may serve adequately when continuous information on the flow rate is not required. Examples are canals where the flow changes are gradual. The gage should be placed so that the water level can be read from the canal bank. The gage should be easy to clean.

Table 8-1. Common reading errors in flat crest reference head as detected by various devices


Reading error deltah1, ft

If head detection is in


Open channel

Stilling well

Point gage Not applicable 0.0015 Commonly used for research
Dipstick Not applicable 0.003 Good for research/field use

Staff gage
> 0.050
beta1 <0.1
beta1 = 0.2
beta1 = 0.5
Pressure bulb + recorder 0.066 Not required Suitable for temporary installations (Error = +2% h1max).
Bubble gage + recorder 0.033 Not required Stilling well is not required but can be used
Float-operated recorder Not applicable 0.016 Stilling well is required
Float totalizer attached to recorder - - Some additional random and systematic error is possible

For concrete-lined canals, the gage can be mounted directly on the canal wall. The value for measuring head on the sloping walls of trapezoidal-shaped canals must be appropriately converted to vertical head values before entering the discharge tables. These tables are usually made for stilling well use or vertical gage applications. The sloping gage can be marked to read direct values or equivalent values of vertical head. Sometimes, sloping staffs are marked to display discharge directly, but the discharge gradations are not equally spaced. The gage may be mounted onto a vertical support for unlined canals.

Most permanent gages are enameled steel, cast aluminum, or some type of plastic resin. Enameled linear scales marked in metric or English units are available from commercial sources. An example staff gage is shown on figure 8-4. Important flow rates can be noted on these scales by separate markings, allowing convenient adjustment of control gates to desired discharges without requiring tables. For convenience, the gages can be marked directly in discharge units rather than in measuring head units.

Figure 8-4 -- Typical staff gage for measuring head or water stage.

(d) Stilling Wells

For accurate discharge measurements, the effective head in flumes, accurately referenced to a known elevation in the flume, must be measured. Head readings on staff gages attached directly to the inside channel walls may be only estimates because of waves and turbulent fluctuations on the scale face. Thus, stilling wells are connected by holes and pipes to the body of water at the measuring station to translocate head and dampen water surface fluctuations by throttling, which increases head measurement accuracy.

The pipe connecting the stilling well to the flume/canal should be large enough to allow the stilling well to respond quickly to water level changes. Usually, this pipe diameter is about one-tenth the diameter of the stilling well. However, special cases may require more dampening using smaller connecting pipe diameters.

The pipe connection to the stilling well should be perpendicular and carefully cut flush with both the canal and the stilling well walls. Otherwise, the translocated water surface elevation in the well can deviate considerably from the actual elevation in the flume because of flow velocity impact or aspiration. Connections that are not flush and/or have rough edges have different head losses depending on direction of flow in the connecting piping. This causes buildup or reduction of head in stilling wells compared to the actual head in the measurement device.

The size of the stilling well depends on the method used to measure the head. The diameter, if circular shaped, ranges from a recommended minimum size of 4 in for hand-inserted dipsticks to 18 in to accommodate larger diameter floats. Wells may be much larger to provide access for cleaning or to make the reading of wall attached staff gages at sight angles at least as flat as 30 degrees. It is recommended that well walls have a 2-in clearance from floats and weights used with recorders.

A stilling well may need to house the float and recorder system or other surface detecting equipment. The wells may need to be tall enough to provide convenient access to recorders for reference setting and maintenance. The wells may also need to be tall enough to keep counterbalance weights from interfering with float movement.

Before making a measurement, the wells should be flushed with fresh water to be sure they are free of sediment, foreign material, or blockages, which could cause erroneous head readings. Recording equipment should be checked and serviced regularly. Cross-checks should be made between the staff gages, hook gages, plumb bobs, recorder values, and any other discharge indicators to expose system errors. Thus, even when using stilling wells, staff gages should still be used on the insidewalls of flumes for cross-checking. Further details on stilling wells can be found in chapter 6 and Bos (1989), Bos et al. (1991), and Brakensiek et al. (1979).

(e) Gage Installation and Zero Setting

The most important factor in obtaining accurate discharge measurements is the accurate determination of the sill-referenced head, h1. The upstream sill-referenced head can be measured by a gage or recorder only if the observed water level is known with respect to the weir sill (or flume crest) level at the control section. The method used to set (zero register) the gage, recorder, etc., depends on the structure size, the flow rate in the channel during the setting procedure, and available equipment. Standard surveying techniques are practical for accurate setting of most wall or staff gages.

The canal side slopes usually only approximate the intended slope. Mounting sloped gages so that a selected scale reading in the most frequently used range of the gage coincides with the corresponding elevation for that reading will partially compensate for deviation from design slope. Thus, the greatest reading errors will occur in the flow ranges that are seldom used. If this procedure causes the zero end of the scale to be displaced by more than about 0.015 ft, the actual side slope should be determined for adjustments to the calibration. This determination also should be made if accuracy over the full flow range is required.

Several methods can be used to zero a water level recorder; three are particularly suitable. The recorder can be set: (1) when the canal is dry, (2) when water is ponded over the flume, or (3) when water is flowing through the flume. These zero-setting methods assume that the sill-referenced elevation can be determined during the procedure. This determination is not always possible, especially on wide structures. A stable and permanent surveying bench mark, such as a bronze cap placed in concrete, should be added in an acceptable location near the measuring structure. Its elevation should have been previously established relative to the sill elevation. More detailed information on zero-setting procedures is presented in Bos et al. (1991).