Geoscience Reference
In-Depth Information
3.2 Discharge Data
The availability of discharge data is important for the model calibration process. Discharge data are,
however, generally available at only a small number of sites in any region. It is also an integrated measure in
that the measured hydrograph will reflect all the complexity of flow processes occurring in the catchment.
It is usually difficult to infer the nature of those processes directly from the measured hydrograph, with the
exception of some general characteristics such as mean times of response in particular events. Rainfall-
runoff modelling for sites where there are no discharge data is a very much more difficult problem.
This ungauged catchment problem is one of the real challenges for hydrological modellers and has been
the subject of the ten-year Prediction of Ungauged Basins (PUB) research initiative of the International
Association of Scientific Hydrologists (see Chapter 10).
There are many different ways of measuring discharges (see, for example, Herschy, 1995). Except
for very small flows, it is difficult to make a direct measurement. The level of water in a channel is,
however, relatively easy to measure and most methods for estimating discharges require a conversion
of a water level measurement to flow. If it is done as the water flows through a well-maintained weir or
flume structure, this conversion can be accurate to better than 5%. If there is no such structure, or if a
structure is overtopped in a high flow, then the accuracy may be very much worse than this, particularly
where the cross-section and effective roughness at the gauging site might change over time as a result of
either sediment transport or seasonal vegetation growth (e.g. Westerberg
et al.
, 2010a). In the worst case
of extreme floods, the water-level measuring device may itself be washed away and then the only resort
is to try to estimate the maximum flow using the
slope-area
method, in which the cross-sectional area
of the flow and the slope of the water surface are estimated from the trash lines indicating the maximum
extent of the flow and a uniform flow roughness equation is used to determine an average velocity. Since
at the crest of a flood, the flow may be non-uniform, highly turbulent, and with a high sediment load in
a dynamically changing cross-section, it may be difficult to estimate an effective roughness coefficient
and cross-sectional area, and hence the average velocity and discharge. Errors in discharge estimates will
then be much higher.
These potential errors tend to get forgotten when the discharge data are made available as a computer
file for use in rainfall-runoff modelling. There is always a tendency for the modeller to take the values
as perfect estimates of the discharge. To some extent this is justified: the data are the only indication of
the true discharges and the best data available for calibrating the model parameters. However, if a model,
any model, is calibrated using data that are in error, then the effective parameter values will be affected
and the predictions for other periods, which depend on the calibrated parameter values, will be affected.
This is an additional source of uncertainty in the modelling that we return to in Chapter 7.
For now, it is worth stressing that, prior to applying any model, the rainfall-runoff data should be
checked for consistency. Some errors, of course, may not be obvious, but the following types of simple
check can be made (see also Section 7.17).
•
If possible, check the discharge rating curve for consistency, and for how far the “observed” discharges
are the result of extrapolation of the rating curve far beyond the range of the directly measured discharge
values. It is very difficult to obtain direct measurements of discharges during flood conditions (even
at sites with a gantry or cableway extending across the flood plain) so high discharges determined by
extrapolation might be very uncertain.
•
Calculate the total volumes of rainfall and runoff for different periods in the record, choosing periods
separated by similar low flows where possible so that the calculated volumes are not greatly affected
by recession discharges. Is the runoff coefficient (the ratio of runoff to rainfall volume) consistent
with expected seasonal changes? Lower values would be expected in the summer, higher values in
the winter.
