FLOOD04705

• County: Statewide
• Community: Foothills Areas of Colorado
• Basin: Statewide
• Title: A Plan for Study of the Flood Hydrology of Foothill Streams in Colorado
• Date: 9/1/1979
• Prepared By: USGS
• Floodplain - Doc Type: Floodplain Report/Masterplan

I
~

In 1976, the Colorado Water Conservation Board, in cooperation with the U.S.
Geological Survey, publ ished a statewide flood report (McCain and Jarrett, 1976)
for estimating flood characteristics of gaged and ungaged sites on natural-flow
streams in Colorado. Although this report addressed the problem of transition
between rainfall- and snowmelt-type flooding, the suggested interim procedure for
determining flood frequencies in foothill areas is highly subjective because of
the lack of flood records in these areas.

Flood characteristics derived through synthetic hydroZogic methods.--The
other general method used for estimating flood characteristics of ungaged and,
sometimes, gaged sites involves the use of various synthetic unit-hydrograph
techniques and a selected rainfall pattern that is assumed to occur over the
basin. Flood characteristics are then computed from the generated rainfall
excess. The limitations of this method can generally be summarized as follows:

1. The inherent assumption in all synthetic hydrograph methods that the T-
year rainfall produces the corresponding T-year flood.

2. Lack of continuous
intensity-duration-frequency
mountainous areas.

rainfall records to adequately
information and areal distribution

define rainfall
in foothill and

3. The assumption that rainfall occurs uniformly over all parts of a basin.

4. The lac~ of adequate understanding of rainfall abstractions because of
the variability of the soil-forest cover complexes in the foothills.

5. Inadequate verification of this technique on foothi 11 streams.

During the preparation of the report by McCain and Jarrett (1976), it was
noted that when using the annual flood array, the flood-frequency characteristics
of foothill streams were significantly different from either mountain streams or
plains and plateau streams. One major difference was that skew coefficients
computed for the Log-Pearson Type-I I I statistical distribution were large positive
numbers--sometimes greater than 1.0. In contrast, plains streams had skew
coefficients near zero and mountain streams had slightly negative skews, about
-0.3. Another shortcoming was that the computed flood-frequency relations failed
to fit the larger flood peaks for most stations and were significantly lower in
some instances. Additionally, the analysis indicated that total basin area was
inadequate for use in estimating the flood potential of foothill streams. Thus,
the problem was twofold: (1) How to compute reliable flood characteristics from
long-term flood records of foothiil streams; and (2) what method should be used
for estimating the flood characteristics for ungaged sites on streams in foothill
areas.

Several long-term gaging-station records for foothill streams were used to
develop annual arrays for both rainfall and snowmelt floods. After fitting the
separate annual flood arrays to the Log-Pearson Type-I I I distribution, it was
noted that the snowmelt flood-frequency relations conformed to the typical pattern
of mountain streams and the rainfall flood-frequency relations approximated that

3