Stormwater Calculations - Mathematics and Statistics for the Environment - page 767

25.3.1.2 Intensity-Duration-Frequency (IDF) Curves
To establish the importance of the relationship between average intensity, duration, and frequency,
the U.S. Weather Bureau compiled intensity-duration-frequency (IDF) curves based on historic
rainfall data for most localities across the country. The rational method uses the IDF curves directly,
while SCS methods generalize the rainfall data taken from the IDF curves and create rainfall dis-
tributions for various regions of the country. There is debate concerning which combinations of
storm durations and intensities are appropriate to use in a hydrologic analysis for a typical urban
development. Working within the limitations of the methodology as described later in this section,
small drainage areas (1 to 20 acres) in an urban setting can be accurately modeled using ether SCS
or rational methods. The belief that the short, very intense storm generates the greatest need for
stormwater management often leads engineers to use the rational method for stormwater manage-
ment design, because this method is based on short duration storms. However, the SCS 24-hour
storm is also appropriate for short duration storms because it includes short storm intensities within
the 24-hour distribution.
25.3.1.3 SCS 24-Hour Storm Distribution
The SCS 24-hour storm distribution curve was derived from the National Weather Bureau's
Rainfall Frequency Atlases of compiled data for areas less than 400 square miles, for durations
up to 24 hours, and for frequencies from 1 to 100 years. Data analysis resulted in four regional
distributions: Type I and IA for use in Hawaii, Alaska, and the Coastal side of the Sierra Nevada
and Cascade Mountains in California, Washington, and Oregon; Type II distribution for most of
the remainder of the United States; and Type III for the Gulf of Mexico and Atlantic coastal areas.
The Type III distribution represents the potential impact of tropical storms which can produce
large 24-hour rainfall amounts.
Note: For a more detailed description of the development of dimensionless rainfall distributions,
refer to Section 4 of the Soil Conservation Service's National Engineering Handbook .
The SCS 24-hour storm distributions are based on the generalized rainfall depth-duration-fre-
quency relationships collected for rainfall events lasting from 30 minutes up to 24 hours. Working in
30-minute increments, the rainfall depths are arranged with the maximum rainfall depth assumed to
occur in the middle of the 24-hour period. The next largest 30-minute incremental depth occurs just
after the maximum depth; the third largest rainfall depth occurs just prior to the maximum depth,
etc. This continues with each decreasing 30-minute incremental depth until the smaller increments
fall at the beginning and end of the 24-hour rainfall (see Figure 25.3). Note that this process includes
all of the critical storm intensities within the 24-hour distributions. The SCS 24-hour storm distri-
butions are, therefore, appropriate for rainfall and runoff modeling for small and large watersheds
for the entire range of rainfall depths.
One of the stated disadvantages of using the TR-55 method for hydrologic modeling is its restric-
tion to the use of the 24-hour storm. The following discussion, taken directly from Appendix B of
the TR-55 manual (USDA, 1986), addresses this limitation:
To avoid the use of a different set of rainfall intensities for each drainage area's size, a set of synthetic
rainfall distributions having “nested” rainfall intensities was developed. The set “maximizes” the rain-
fall intensities by incorporating selected short-duration intensities within those needed for larger dura-
tions at the same probability level.
For the size of the drainage areas for which SCS usually provides assistance, a storm period of 24
hours was chosen for the synthetic rainfall distributions. The 24-hour storm, while longer than that
needed to determine peaks for these drainage areas, is appropriate for determining runoff volumes.
Therefore, a single storm duration and associated synthetic rainfall distribution can be used to represent
not only the peak discharges but also the runoff volumes for a range of drainage are sizes.