FLOOD06431

• County: Statewide
• Basin: Statewide
• Title: Estimating Design-Flood Discharches for Streams in Iowa Using Drainage-Basin and Channel-Geometry Characteristics
• Date: 1/1/1993
• Prepared For: Iowa City
• Prepared By: USGS
• Floodplain - Doc Type: Educational/Technical/Reference Information

region (fig. 2). Consequently, both data sets
were split into regional data sets, and additional
multiple-regression analyses were performed
for two regions in Iowa.

The State was divided into two hydrologic
regions using information on areal trends of the
residuals for the statewide regression
equations, the Des Moines Lobe landform
region, and topography a3 guides. The
delineation of channel-geometry Regions I and
II is shown in figure 2. The topography of the
Des Moines Lobe landform region (Region II) is
characteristic of a young, postglacial landscape
that is unique with respect to the topography of
the rest of the State (Region Ii (Prior, 1991,
p. 30-47). The region genprally comprises
low-relief terrain, accentuated by natural lakes,
potholes, and marshes, where surface-water
drainage typically is poorly defined and
sluggish. The shaded area between hydrologic
Regions I and II (fig. 2) represents a transitional
zone where the channel morphology of one
region gradually merges into the other. This
regionalization process served to compensate for
the geographic bias observed in the statewide
residual plots, which was not accounted for
otherwise in the 111- and 157-station channel-
geometry regression equations listed in table 3.

Using the OLS and WLS multiple-
regression techniques previouE,ly described, two
sets of flood-estimation equations were
developed for each channel-geometry region. Of
the ll1-station data set, 78 stations were in
Region I and 33 stations were in Region II. Of
the 157-station data set, 120 stations were in
Region I and 37 stations were in Region II.
Gaging stations located within the regional
transition zone (fig. 21 were compiled into either
Region I or Region II data sets on the basis of
residuals from the statewide regression
equations and on the regional locations of their
stream channels. The best equations developed
in terms of PRESS statistics, coefficients of
detennination, and standard errors of estimate
for the Region I data sets are listed in table 4
and the best equations developed for the Region
II data sets are listed in table I;.

The channel-geometry characteristic that
was identified as most significant in the Region
I 78-station bankfull equations was bankfull
width IBFW1. The characteri,;tic identified as

most significant in the Region I 120-station
active-channel equations was active-channel
width (ACW). The channel-geometry character-
istics that were identified as most significant in
the Region II 33-station bankfull equations were
bankfull width (BFW) and bankfull depth
(BFD), and the most significant characteristic in
the Region II 37 -station active-channel
equations was active-channel width (ACW).
Appendix C (at end of this report) outlines the
procedure for conducting channel-geometry
measurements of these characteristics.

Comparison of Regional and Statewide
Channel-Geometry Equations

Comparison of the Region I and II equations
with the statewide equations shows an
improvement in the average standard errors of
prediction for all of the regional equations
except the 25-, 50- and 100-year recurrence
intervals of the Region II active-channel
equations. The regional equations listed in
tables 4 and 5 may provide improved accuracies
for estimating design-flood discharges based on
channel-geometry measurements. The
statewide equations listed in table 3 also can be
used to estimate design-flood discharges,
although their accuracies may be less than for
the regional equations. Comparison of the
bankfull equations with the active-channel
equations listed in tables 3-5 shows an
improvement in the average standard errors of
prediction for all of the bankfull equations. The
bankfull equations may provide improved
estimation accuracies in comparison to active-
channel equations for estimating design-flood
discharges for channels unaffected by
channelization.

Bankfull depth (BFm was identified as a
significant channel-geometry characteristic in
the statewide bankfull equations (table 3). It is
also a significant channel-geometry character-
istic in the estimation of design-flood discharges
for stream sites located within the Des Moines
Lobe landform region (fig. 2, Region Ill. While
bankfull depth was not identified as significant
in estimating flood discharges in Region I, it
appears to be a significant morphologic feature
distinguishing stream channels in Regions I and
II.

24 ESTIMATING DESIGN.FLOOD DISCHARGES FOR STREAMS IN IOWA

~