FLOOD05191

• County: Statewide
• Community: State of Colorado
• Basin: Statewide
• Title: Flood Characteristics of Urban Watersheds in the United States
• Date: 12/9/1983
• Prepared By: Sauer, Thomas, Stricker, and Wilson
• Floodplain - Doc Type: Flood Mitigation/Flood Warning/Watershed Restoration

shows the coefficients of determination, R', and the
standard errors of regression for equations 3-9.

Floodch<lfacteristic

Statistic

UQ2 UQ5 UQ10 UQ25 UQ50 UQlOO UQ500

Coefficient of
determination, R I

.93 .93 .93 .93 .92

_92

.90

Standard error of
regression:
Log units

.1630 .1584 .1618 .1705 .1774 .1860 .2071

Average percent

:1::38 :1:37 :i:38 :1::40 :t42 :1:44 :1:49

Because of their suitability and accuracy, these
equations provide a good method of estimating the
effects of urbanization on magnitude and frequency of
peak discharge. From the 269 sites available for analysis,
55 were omitted because of known detention storage, 10
were omitted because detention storage effects were
uncertain, and 5 were omitted because of missing data.
Therefore, the equations are derived from 199 sites. Fig-
ures 3, 4, and 5 compare the 2-year, IO-year, and 100-

10.000

Cl
Z
o
u
W
<Jl
0:
w
0-
f-
w
~ 1000
u
'"
::J
U
Z

EXPLANATION
0-1 OBSERVATION
0-2 OBSERVATIONS
~-3 OBSERVATIONS

w
r.:J
0:
<l:
I
U
<Jl
Cl
Cl
w
f-
<l:
::;;
i=
<Jl
W

100

10

10

year observed peak discharges to the respective peaks
estimated from equations 3, 5. and 8.
All independent variables in equations 3-9 are sta-
tistically significant at the I-percent level with tbe fol-
lowing exceptions. The percent of impervious area, lA,
was statistically significant at the I-percent level in
equation 3 and at the 2-percent level in equation 4, but
was not significant at the 5-percent level for equations
5-9. The change in significance of the variable IA sug-
gests that impervious area in a basin will effectively
increase runoff (primarily volumes) for low-order fioods,
but will rapidly become less effective during large floods
when soils become saturated and approach a runoff
condition similar to that produced by impervious sur-
faces. Even though IA is not higbly significant for equa-
tions 5-9, it was retained to provide continuity with
equations 3 and 4. Storage, ST, and slope, SL, for equa-
tions 8 and 9 were significant at the 2-percent level.
The most significant variable in each of the equa-
tions is the equivalent rural discharge, RQ, because it is
closely related to the urban peak discharge. Rural dis-
charge is the key for explaining geographical variations
in runoff in different parts of the country. Consequently,
the equations are suitable for use in urban areas through-
out the United States, with no expecte': geographical
bias. The tests made to substantiate this conclusion are

o
o

o

o

100 1000
OBSERVED DISCHARGE, IN CUBIC FEET PER SECOND

10,000

Figure 3. Comparison of observed 2-year urban peak discharge to peak discharge estimated from equation 3.

Estimating Procedures for Ungaged Urban Sites 11