7910

• UCREFRP Catalog Number: 7910
• Author: Bovee, K. D. and R. T. Milhous.
• Title: Hydraulic Simulation In Instream Flow Studies
• USFW Year: 1978.
• USFW - Doc Type: Theory And Techniques, Instream Flow Information Paper No. 5.
• Copyright Material: NO

system with various numbers of data points, was determined from
data obtained from 11 streams. Rating curves for each stream were
obtained from U.S. Geological Survey gaging stations. These curves
served not only as potential data sources, but were also used to define
the measured stage for a given discharge.
In the case of Manning's equation, one stage-discharge point was
taken from a rating curve and used to calibrate the equation. Predic-
tions of the stage at different discharges were then compared to the
measured value at each of the other discharges on the rating curve.
Then another point was taken from the rating curve, the equation re-
calibrated, and the process repeated. A similar technique was used in
testing the direct determination approach, first using two points, then
three, four, and so on. By selecting different points and combinations
of points, it was possible to make several thousand comparisons between
predicted and measured stages at different discharges over a wide range
of hydrologic conditions.
. For each stage-discharge prediction, the absolute value of the
error was determined by:
Error M = IQm - D? X 100% C18)
Qm
where,
Q = the measured discharge of the section
Qp = the predicted discharge of the section
For each river, and for each approach used, the mean error was
calculated as the arithmetic average of the absolute errors for each
trial. Table 1 shows the mean error of the stage-discharge predictions,
using all combinations and permutations of data points and no limits to
the range of extrapolations made from measured calibration flows. Table
2 shows the mean error of the stage-discharge predictions, with the
following limitations placed on selection of data pairs and range of
extrapolations:
1. When Manning's equation was used, the range of extrapolations
was limited to a minimum flow of 40% of the calibration flow;
to a maximum flow of 250% of the calibration flow.
2. If a two-point rating curve were used:
a. The second point in the rating curve must be outside the
range of 0.8 times the first flow to 1.25 times the first
flow.
b. The slope of the regression line, Q = a(S - ZF)b, must be
between 1.4 and 4.0. If the slope of the regression line
was outside this range, a third point was added by
estimating the point of zero flow.
32