FLOOD08489

• County: Statewide
• Community: State of Colorado
• Basin: Statewide
• Title: Determination of Urban Watershed Response Time
• Date: 12/1/1974
• Prepared By: E.F. Shulz and O.G. Lopez
• Floodplain - Doc Type: Flood Mitigation/Flood Warning/Watershed Restoration

the direction of Dr. John Geyer. During this period a
number of cities in the United States began programs
to gage the runoff from urban catchments ~n order to
develop the data base in order to apply the hydro graph
design methods of Hicks (1944) or Tholin and Kiefer.
Many of these watersheds are described by Tucker
(1969a and 1969b).

Houston, Texas, was one of the metropolitan re-
gions assembling urban runoff data. The Houston metro-
politan region is a relatively flat plain which is
drained by six streams which discharge toward the east
into the San Jacinto River. One of these streams is
Brays Bayou which has an 88.4 square mile catchment.
This gaging station has been in operation since 1937.

One of the first studies to actually document the
effect of urbanization on the unit hydro graph was pre-
sented by Van Sickle (1962) in a paper to the Texas
Section of the American Society of Civil Engineers.
Van Sickle presen~ed six unit hydrographs derived from
Brays Bayou during the period from July 1939 to June
1960. In this period of time, the watershed evolved
from a rural watershed to an urban watershed and peak
of the unit hydrograph increased from about 1500 cfs
to about 4800 cfs-- more than a 300 percent increase.
The watershed response time measured as the time to
peak decreased from 12 hours to 3 hours.

The U. S. Geological Survey in Cooperation with
the City of Houston has undertaken a comprehensiveef-
fort to obtain data on the flood response of urban
watersheds in this region. Johnson and Sayre (1973)
have presented an analysis of the data for the Houston
region. Dempster (1974) has presented a similar anal-
ysis of data obtained in the Dallas, Texas area.

Rattapan (1968) applied the Chicago
Method of Tholin and Kiefer in developing a
computing an urban storm runoff hydro graph
for Bangkok, Thailand.

Hydrograph
method of
procedure

InC4eahe in Flood Peak V~~haxge. Johnson and
Sayre found that the peak discharge of a T-year flood
could be estimated for the Houston area using the
equation:

OT = aAble

where

is the peak discharge of a flood having
return period of T years in cfs,

A is the watershed area in square miles,
I is the percent of impervious watershed,
a, b, and c are constants for the region.

QT

The values of the constants are given in the following
table.

Table I

Johnson and Sayre also discussed the storm drain-
age patterns in use in the Houston area. They ana-
lyzed the network data from 28 watersheds having
between 8 floods and 19 floods for each station. A log
transformed multiple regression equation was found for
these flood events. This is similar to a presentation
by Dooge (1973). The regression equation is:

Op = apbOe Md
85

",'here

~
p

is the peak discharge for the flood in cfs.

is the causative rainfall areally avera~ed
over the watershed in inches,
is the storm duration in hours during which
85 percent of the rainfall, P, occurred,
is the soil-moisture index found from the
relation:

085
M

M = (Mo + Po)kt ,

where M is the soil moisture index in inches,
M is the soil moisture index computed or mea-
0 sured t days preceding M,
p is the precipitation on the day when M
0 was determined in inches, 0
k is soil moisture depletion factor dependent
upon the season,
t is the number of days between M and M,
0

The values of the constant a and exponents b,
e and d are given by Johnson and Sayre. The values
have a complex relationship between the watershed area,
surface storage, channel network and local topography.
Because of the flat topography around Houston, there
is watershed piracy between adjacent watersheds during
some of the events.

Unlike the Houston metropolitan region, the Dallas
region is bisected by a major river-the Trinity River.
Dempster had data from 19 storms over six smaller
basins which discharge into the Trinity River. Dempster!s
data base contained 205 flood events on 19 sub-basins.
These data were used to calibrate a USGS digital model
developed by Dawdy, Lichty and Bergmann (1972). The
57-year rainfall record of climatological data was
used to simulate a 57-year runoff record for the urban
watersheds. The log-pearson Type III distribution was
fitted to the derived record.

a

A regional flood-frequency equation relating the
flood peak for an assumed return period with storm and
physical watershed characteristics was developed using
a stepwise multiple regression equation. The pro-
cedure operates on the input data by successively
discarding the independent variables which are the

Regional Relations for T-Year Flood
Houston, Texas Region

Recurrence Interval Constant Exponents Standard Error
percent
T, in years (a) (b) (e) lOR: units (averaee)
2 38,8 0.86 0.62 0.111 26
5 62.7 .87 .57 .119 28
10 82.0 .87 .54 .129 30
25 109 .88 ,50 .141 33
50 132 .88 .48 .150 35
100 156 .89 .45 .159 37
from Johnson and Sayre (1973)
5