Flood Frequency Analysis: International Edition: Produced by The COMET Program Print Version

Flood Frequency Analysis: International Edition file:///H:/hydro/basic_int/flood_frequency/print.
htm
Flood Frequency Analysis: International Edition

Produced by The COMET® Program
Print Version
Introduction to Flood Frequency Analysis
Role in Flood Prediction Process
Basic Terminology
Time Distribution of Floods
Flood Frequency Misconceptions
Estimate of Flood Magnitude
Estimate of Return Period
Review Questions
Statistical Representation of Floods
Period of Record Issues
Confidence in Return Period Estimates
Exceedance Probability
Example of Exceedance Probability
Independent and Homogeneous Data
Impact of Basin Alterations
Review Questions
Analyzing Flood Data
Annual Series
Annual Maximum Daily Flow
Annual Series Rank
Annual Series Probability
Annual Series Display
Partial Duration Series
Partial Duration Series Exercise
Preparing Partial Duration Series
Partial Duration Display
Review Questions
Design Events
Purpose of Design Events
What is a Design Event?
Types of Design Events
Flood Frequency Exercise
Review Questions
Summary
1 of 54 05-03-2020, 19:51
Flood Frequency Analysis: International Edition file:///H:/hydro/basic_int/flood_frequency/print.htm
Introduction to Flood Frequency Analysis

Flood frequency analysis uses historical records of peak flows to produce guidance about the expected
behavior of future flooding. Two primary applications of flood frequency analyses are:
1. To predict the possible flood magnitude over a certain time period

2. To estimate the frequency with which floods of a certain magnitude may occur
In this section you will learn to:
Define the meaning of return period (i.e., the 100–year flood).

Explain the exceedance probability and its relationship to return period.
Understand the two primary applications of flood frequency analyses.
2 of 54 05-03-2020, 19:51
Role in Flood Prediction Process
Flood frequency analysis provides information about potential floods based on either current or forecast
conditions. It also uses information from past floods to update statistical representations of possible future
flood events.
3 of 54 05-03-2020, 19:51
Basic Terminology
Many people have heard terms like the "100–year flood" or the "500–year flood" used in the news media
and within professional organizations. Although many are familiar with these terms, not everyone really
knows what they mean.
Let's look at the term "return period," also referred to as the "recurrence interval." The return period is the
time period over which it is likely that a particular magnitude flood will occur. Thus, a 100–year flood is
defined as a flood that can occur on average once every 100 years. In this example, 100 years is
considered the return period. However, floods do not occur in exact cyclic events. That is, they don't
occur at nicely spaced 100–year intervals.
Another way to think of the 100–year flood is of a flood that has a one one–hundredth, or 1 percent,
chance of occurring in any given year. This is called the exceedance probability or the probability of
occurrence, and is simply 1 divided by the return period.
A 100–year flood is less frequent but larger in magnitude than a 25–year flood.
The 25–year flood occurs on average once every 25 years and has an exceedance probability of 1 over 25,
or 4 percent, in any given year. That's a 4 times greater chance than the 100–year flood, which only has a
1 percent exceedance probability in any given year.
The 100–year flood is smaller in magnitude and 5 times more frequent than the rare 500–year flood,
which has a 1 over 500, or 0.2 percent chance of occurring in any given year.
Additional Resources:
The "100–Year Flood"
U.S. GEOLOGICAL SURVEY Fact Sheet 229–96
by Karen Dinicola
http://pubs.usgs.gov/fs/FS-229-96/
4 of 54 05-03-2020, 19:51
Time Distribution of Floods
The 2–year flood occurs on average 25 times over a 50–year period. To meet this definition, the time
distribution of the 2–year flood could be evenly spaced, that is, every two years. The 2–year flood
definition could also be met if the flood occurs every year for the first 25 years in a row, and then not at
all for another 25 years. However, either of these scenarios is very unlikely. The most likely time
distribution of the 2–year flood would be unevenly distributed over the 50–year period. The 2–year event
may have several consecutive years of occurrences, or there may be multi–year periods with no
occurrences.
5 of 54 05-03-2020, 19:51
Flood Frequency Misconceptions
Just as the 25–year flood would occur on average once every 25 years, the 25–year rainstorm would also
occur on average once every 25 years. A common misconception is that a 25–year rainstorm will always
produce a 25–year flood.
While this may be a good first guess, the conditions on the ground greatly affect the runoff generated
from any particular storm.
Click to play animation.
For example, if the soil is saturated, it is possible that a 25-year rainfall would become runoff so
efficiently that it could result in a runoff event larger than the 25–year flood. Likewise, dry ground
conditions could mitigate the impact of a large rainstorm and produce a smaller flood.
Another common misconception is that a 100–year flood occurs only once every 100 years or will occur
only once during a 100–year time span. In truth, there is an equal (and equally unlikely) chance for a
flood of this magnitude to occur in any year or even multiple times in a single year.
6 of 54 05-03-2020, 19:51
Estimate of Flood Magnitude
There are two basic applications for flood frequency analysis. The first is to estimate the potential flood
magnitude that can occur in a given time interval.
For example, we may want to know the peak flow expected over a 100–year period at a certain location.
Using a fictitious example, an engineer may ask, "What is the magnitude of a 100–year flood at the Main
Street Bridge over Spring Creek?" The information from flood frequency analysis provides an answer
such as: the 100–year flood has an estimated magnitude of 840 cubic meters per second.
7 of 54 05-03-2020, 19:51
This information can then used by a hydrologist or by a civil engineer to determine the size of a bridge
clearance. It's possible that the opening under the bridge may need to be large enough to accommodate
large, infrequent flows.
Here, we see photographic examples of floodwaters overrunning a bridge through Kruger National Park,
South Africa during the floods of 2008. This was an event that was obviously greater than the flood for
which the bridge was designed.
8 of 54 05-03-2020, 19:51
Estimate of Return Period
The second primary goal of flood frequency analysis is to estimate the return period for a certain
magnitude flood.
To continue with our example, let's assume that we know flooding will occur at the newly proposed Plaza
Mall if Spring Creek reaches a peak discharge of 560 cubic meters per second.
We may ask, "What is the return period, or recurrence interval, of 560 cubic meters per second?" Flood
frequency analysis can then tell us that such a flow has a 50–year return period. In other words, 560 cubic
meters per second is a 50–year flood. This information can be used by people such as urban planners and
civil engineers to determine the best locations and construction practices for new development.
9 of 54 05-03-2020, 19:51
Review Questions
Flood frequency analysis is an important element in _____. (Choose the best answer.)
a) designing transportation infrastructure such as bridges and culverts

b) estimating how often a certain magnitude flood may occur
c) estimating the flood magnitude of a particular probability
d) All of the above
The correct answer is d.
The 100–year flood is _____. (Choose all that apply.)
a) less frequent than the 500–year flood

b) lower magnitude than the 500–year flood
c) one that occurs once every 100 years
d) one with a 1 percent chance of occurring in any given year
The correct answers are b and d.
A 100–year rainstorm will always produce a 100–year flood. (True or False.)
a) True
b) False
The correct answer is b.
10 of 54 05-03-2020, 19:51

The statistical representation of flood events is affected by both length of the period of record and
consistency of the hydrologic conditions within the drainage basin. Unrepresentative data for either of
these reasons can lead to inaccurate guidance from flood frequency analysis.
This section will introduce basic concepts used in flood frequency analysis and demonstrate calculation of
flood statistics.
Explain how the period of record impacts flood frequency guidance.

Calculate the probability of exceedance or non–exceedance for a given flood magnitude over a
specified duration.
Understand how basin changes may impact the behavior and frequency of floods, thus reducing the
length of the period of record.
11 of 54 05-03-2020, 19:51
Period of Record Issues
Sometimes the record length of river flows may be insufficient for representing the river's entire flow
history. Floods or droughts that occur infrequently may be under–represented in a limited streamflow
record.
For example, the Mississippi river flood over the central United States in 1993 was a rare event that may
only occur once every 100 years. Thus, if the period of record is only 40 years, there is a good chance that
a flood like 1993 is not represented in the statistics. On the other hand, if the period of record is 300 years,
it is likely that more than one flood like 1993 is in the flood frequency statistics. Thus, a longer period of
record results in more representative flood statistics.
For accurate and reliable statistical hydrologic guidance about possible events, it is necessary to use a
dataset that includes a representative sample of as many different events as possible.
The longer the period of record, the better the likelihood of capturing the range of possible events.
There are several different return periods traditionally used by hydrologists.
Common return periods include the 2–, 10–, 25–, 50–, 100–, and even 500–year flood. Values for each of
these return periods can be calculated based on the statistics of the flow record. However, there is a
question of just how representative extreme values, such as the 500–year flood value, might be.
If possible, it is best to avoid estimating return period flood values that are greater than twice the record
length. So you might not want to put much faith in your 500–year flood estimate unless you have at least
250 years of data.
12 of 54 05-03-2020, 19:51
Confidence in Return Period Estimates
Estimates of flood return periods can be made with relatively short periods of record. But the associated
confidence level in the flood frequency statistics is much higher with a longer period of data. For
example, to estimate a 10–year flood with no more than a ±10 percent error, one would need 90 years of
record. If you are willing to accept a ±25 percent error, then only 18 years of record is needed. Here we
can see the length of data record needed to be within either ±10 percent or ±25 percent errors for the 10–,
25–, 50–, and 100–year floods.
13 of 54 05-03-2020, 19:51
Exceedance Probability
Sometimes a hydrologist may need to know what the chances are over a given time period that a flood
will reach or exceed a specific magnitude. This is called the probability of occurrence or the exceedance
probability.
Let's say the value "p" is the exceedance probability, in any given year. The exceedance probability may
be formulated simply as the inverse of the return period. For example, for a two-year return period the
exceedance probability in any given year is one divided by two = 0.5, or 50 percent.
But we want to know how to calculate the exceedance probability for a period of years, not just one given
year. To do this, we use the formula
Exceedance probability = 1 – (1 – p)n
In this formula we consider all possible flows over the period of interest "n" and we can represent the
whole set of flows with "1." Then (1–p) is the chance of the flow not occurring, or the non–exceedance
probability, for any given year.
(1–p)n is all the flows that are less than our flood of interest for the whole time period.
Finally, "1," all possible flows, minus (1–p)n, all flows during the time period than are lower than our
flood of interest, leaves us with 1 – (1–p)n, the probability of those flows of interest occurring within the
stated time period.
14 of 54 05-03-2020, 19:51
This table shows the relationship between the return period, the annual exceedance probability and the
annual non–exceedance probability for any single given year.
So, if we want to calculate the chances for a 100–year flood (a table value of p = 0.01) over a 30–year
time period (in other words, n = 30), we can then use these values in the formula for the exceedance
probability.
We can also use these same values of p and n to calculate the probability of the event not occurring in a
30–year period, or the non–exceedance probability.
15 of 54 05-03-2020, 19:51
Example of Exceedance Probability
Let's say you want to know what the probability is for a 50–year flood over a 50–year period. It's not 100
percent!
Calculation for Probability of 50–Year Flood Over 50–Year Period

1 – (1 – p)n
n = 50
p = 0.02
We know that n = 50 since we are looking at a 50–year period of time and using the probability of
occurrence table we see that p=0.02 for a 50–year return period.
1 – (1 – 0.02)50
= 1 – (0.98)50
So, applying these values in the equation, the (1–p) value is (1–0.02), or 0.98.
= 1 – 0.36
= 0.64 or 64%
(1–p) to the n is 0.98 raised to the 50th power. That comes out to 0.36.
Now we have (1–0.36), which is 0.64.
There is a 64 percent chance of a 50–year flood in a 50–year period. That means there is a 36 percent
chance we won't see a 50–year flood in the 50–year period.
Now let's determine the probability of a 100–year flood occurring over a 30–year period of a home
mortgage where the home is within the 100–year floodplain of a river.
Calculation for Probability of 100–Year Flood Over 30–Year Period

1 – (1 – p)n
n = 30
p = 0.01
n=30 and we see from the table, p=0.01 .
16 of 54 05-03-2020, 19:51
1 – (1 – 0.01) 30
= 1 – (0.99) 30
= 1 – 0.74
(probability of non–occurrence = 0.74)
= 0.26 or 26% probability of occurrence
The 1–p is 0.99, and .9930 is 0.74.
There is a 0.74 or 74 percent chance of the 100–year flood not occurring in the next 30 years.
But 1–0.74 is 0.26, which shows there is a 26 percent chance of the 100–year flood in that time.
17 of 54 05-03-2020, 19:51
Independent and Homogeneous Data
Flood frequency analysis requires that data be independent and homogeneous. The independence
requirement means that floods occur individually and do not influence each other. For example, two peak
flows that were above flood stage are independent floods if the flow completely returned to baseflow
between those two events.
On the other hand, if there is not a return to baseflow level between the peaks, the floods are not
independent because the first flood has influenced the second flood.
The homogeneous requirement means that each flood needs to occur under the same type of conditions.
Two flood events are homogeneous if both are caused by rainfall only.
18 of 54 05-03-2020, 19:51
An example of non–homogeneity is when one flood is caused by rainfall while another flood is caused by
a dam failure.
19 of 54 05-03-2020, 19:51
Impact of Basin Alterations

Basin changes such as land development, water diversions and reservoir construction may alter the
hydrologic properties of a drainage basin over time, thus changing the way a river responds to storms.
Such trends may result in future flow behavior that is different from that observed in the past, and violate
the requirement of homogeneity for flood frequency analysis. Consequently, flood frequency statistics
generated prior to the basin changes no longer apply.
This can have the effect of dramatically reducing the length of the period of record where homogeneous
conditions exist. In other words, if you have a 100–year record, but major urbanization took place 20
years ago, then you really only have a 20–year homogeneous record for your now urbanized basin.
20 of 54 05-03-2020, 19:51
Review Questions
It is best to avoid estimating return period flood values that are greater than twice the length of the data
record. (True or False.)
a) True
b) False
The correct answer is a.
To be within a 25 percent error for a 50–year flood estimate, the length of your data record should be at
least _____. (Please refer to the table below, and choose the best answer)
a) 18 years
b) 39 years
c) 50 years
d) 110 years
Land use changes within a watershed can violate the basic assumption of homogeneity and can change
flow behavior within the stream. (True or False.)
a) True
b) False
What is the probability of occurrence of a 25–year flood in the current year if that same magnitude flood
just happened last year? (Choose the best answer.)
a) Less than 0.01

b) 0.01
c) 0.04
d) 0.25
21 of 54 05-03-2020, 19:51
The correct answer is c.
The 25–year flood has a 1/25 = 0.04, or 4% chance of occurring in any given year.
The chance of a flood of a specific return period not occurring is called the _____. (Choose the best
answer.)
a) non–exceedance probability
b) exceedance probability
c) null frequency probability
d) exceedance non–probability
Two large, but separate, peak flows occur at the same gauging station one day apart during the passing
of a tropical storm remnant. The flow remained above baseflow during the minimum flow level between
the peaks. Using both values in flood frequency analysis would violate the principle of _____. (Choose
the best answer.)
a) return period
b) homogeneity
c) confidence limits
d) independence
22 of 54 05-03-2020, 19:51

Flood data are analyzed to gain an understanding of a river's past behavior and to provide guidance on
expected future floods.
Explain the basic concepts underlying both annual and partial duration time series.
Conduct a frequency analysis given peak flow data for a river.
23 of 54 05-03-2020, 19:51
Annual Series
Often a hydrologist is interested in a specific flow event – typically the largest of the year. When they
consult only the maximum flow within a single year, the dataset is called an annual series. The letter A on
the figure shows the events that comprise an annual series. Within an annual series, only the largest value
per year is allowed, even if an additional significant peak flow occurred such as we see in 1998.
24 of 54 05-03-2020, 19:51
Annual Maximum Daily Flow
The meaning of the term "peak flow" can vary in practice.
Ideally, an instantaneous maximum peak flow would be used. The instantaneous peak flow would be
the peak flow at any one moment during a day. Unfortunately, such values are not always available.
Instead, the annual maximum daily flow is often used. This is based on the average daily flow values.
The average daily flow is the average flow over a 24–hour period. In a year we would have 365 (or 366)
daily flow values. Each of these represents the average flow over the calendar day.
The annual maximum daily flow is the greatest of those values. So in the case of an annual series using
annual maximum daily flow, we are using the greatest 24–hour average flow within each year.
25 of 54 05-03-2020, 19:51
Annual Series Rank
The first step in a frequency analysis is to examine the streamflow data at the gauging station of interest
and to create a ranked annual series. Here is an example of the annual maximum daily flows for a gauging
location to be used in an annual series. There are 33 years in this dataset.
Now, we rank the data from the year with the highest value of annual maximum daily flow to the year
with the lowest value.
26 of 54 05-03-2020, 19:51
The annual maximum daily flow in 1977 was 8120 and represents not only the largest daily flow in 1977,
but also the largest daily flow within the entire observed record. The annual maximum daily flow in 1983
was 925. This represents the smallest annual maximum daily value of the record.
27 of 54 05-03-2020, 19:51
Annual Series Probability
Note: Data above is abbreviated for ease of viewing.
Once we have the annual maximum daily flow values of the 33–year dataset ranked from largest to
smallest, we can compute the probability associated with that rank. That is, what is the probability of that
magnitude flood occurring in any year based on this 33–year dataset?
There are several formulas for calculating this probability value, or plotting position as it is frequently
known when plotted on special probability paper. Here, we have used the Weibull formula, because of its
ease of use for illustration purposes. The US Geological Survey, among others, also use this formula in
their plotting–position diagrams. Other commonly used formulas for plotting positions include Cunnane
and median.
Other, more complex methods of calculation include the Pearson type III, Log–Pearson type III, 3
parameter lognormal and Wakeby. Each of these statistical fits to flow distributions is best suited to
certain situations, and they may be combined together to better describe flows on certain streambeds.
In the above equation "n" is the number of years in the dataset and "m" is the rank.
Since the return period, Tr , is estimated by taking the inverse of the probability, we can estimate the
return period of each annual peak flow using an inverse of the Weibull formula.
The probability of 50 percent is the same as a 2–year return period and is the median value.
However, the average of all the annual maximum daily flows in the table is 3205.
If we locate where the value of 3205 would be in the table, we can see that the probability is about 33
percent and the return period about 3 years. This shows that the average is not always the midpoint.
If our dataset covered a longer duration, the average values would tend more toward the midpoint.
28 of 54 05-03-2020, 19:51
Annual Series Display
There are a couple of ways that annual peak flood frequency data are typically displayed, but first we will
look at how data are often plotted with spreadsheet programs. These plots typically have linear scales on
both the X and Y axes, which is unsuitable for the probabilities that need to be displayed.
Here we have a plot of the annual series with the annual maximum daily flow per year on the Y axis
versus the annual exceedance probability on the X axis. The X and Y axes both use linear scales, which
presents a problem for the annual exceedance probability. It shows a 0 percent value, implying with
certainty that a flood above at a particular large value would never happen. For example, if you
extrapolate from the curve to the Y axis, this suggests that a flood of 9000 flow units per second or greater
will never happen. Extremely large flows are rare, but in reality there is some small chance of such an
event.
Sometimes a second X–axis variable is displayed to show the peak flow data as they relate to return
period.
29 of 54 05-03-2020, 19:51
To avoid a graph that will suggest either a zero or a one–hundred percent exceedance probability,
hydrologists use a variable probability scale along the X–axis. This type of scale allows for more detail at
the high and low probability events and does not show a 0 or 100 percent value.
30 of 54 05-03-2020, 19:51
The annual peak flow information may also be presented with a logarithmic rather than a linear scale.
This is often done to make the curve appear as a straight line. Straight line curves more easily allow
extrapolation beyond the data extremes.
31 of 54 05-03-2020, 19:51
Partial Duration Series
There are situations when it is advantageous for the flood statistics to represent more than one peak flow
per year. For example, for floods that occur frequently enough, there is a reasonable chance that this will
happen more than once per year.
It is possible that the second or third largest peak in a particular year may be larger than the maximum
flood of some other year. In an annual series, such additional events are ignored since only the largest
annual event is allowed.
The partial duration series includes all events above some arbitrary threshold value. Typically this is the
smallest annual maximum series value. For example, the letter A shows only the peak annual flows in this
4–year graph of flows including the relatively small peak flow of 10 units in 1997. The letter P shows all
events above a flow threshold of 10 units. All of these events would be included in a flood frequency
analysis based on a partial duration series.
Note that the second largest peak in 1998 is greater than the annual peaks in 1996, 1997, and 1999 but is
not part of the annual series. This second largest value in 1998 can only be included if a partial duration
series used.
32 of 54 05-03-2020, 19:51
Partial Duration Series Exercise

As discussed earlier, it may be desirable to include numerous peak flows in a year that are lower than the
annual maximum, but higher than a given threshold. Although the threshold can be arbitrary, often the
lowest annual peak flow in the dataset is used as the threshold.
To answer the following questions, refer to the table below.
Notice the data table includes the peak flows for each year from 1970 through 2002.
What is the greatest flow of the whole table? What year did it occur? (Choose the best answer.)
a) 6600 in 1972
b) 8120 in 1976
c) 8120 in 1977
d) 9250 in 1983
33 of 54 05-03-2020, 19:51
What is the lowest peak flow and in which year was this the only peak flow value? (Choose the best
answer.)
a) 925 in 1972
b) 925 in 1973
c) 875 in 1981
d) 925 in 1983
Which year had the most peak flows and how many were there? (Choose the best answer.)
a) 1973 – 10 peaks
b) 1977 – 5 peaks
c) 1989 – 11 peaks
d) 1999 – 6 peaks
In 1973, what was the greatest of the numerous observed peak flows. (Choose the best answer.)
a) 8120
b) 3110
c) 3020
d) 925
In 1973, only the value of 3110 would be used in the annual series, but 9 additional peak values from
that year are also part of the partial duration series dataset. Which dataset do you think would be able to
represent frequent small flood peaks? (Choose the best answer.)
a) Annual maximum series

b) Partial duration series
34 of 54 05-03-2020, 19:51
Preparing Partial Duration Series

Because the partial duration series can include more than one data point per year, there are more data
points than the number of years in the dataset. While the annual series had 33 points for this 33–year
dataset, the partial duration series has 113 points for the same period.
Therefore, these points cannot be plotted as flow versus annual probability. They must be plotted as flow
versus return period, as seen below.
To plot these points on a flow versus return period graph, the rank and flow are used to compute the return
period.
35 of 54 05-03-2020, 19:51
To compute the return period, first the data are ranked from the highest to lowest flow value.
The return period is then computed using the inverse of the Weibull formula as we did with the annual
maximum series.
Inverse Weibull Formula:

Return Period = (1+n) / m
m = rank
n = number of years in the dataset
36 of 54 05-03-2020, 19:51
Once the table is ranked by flow return period, we no longer need to keep track of the specific years that
these flows occurred, as shown in the table below.
37 of 54 05-03-2020, 19:51
38 of 54 05-03-2020, 19:51
Partial Duration Display
Here we see the partial duration series displayed as flow versus the return period. Both the Y and X axes
are shown with logarithmic scales. The annual maximum series from the previous section is plotted for
comparison. Note that the partial duration series has many more data points and quite a few with less than
a one–year return period. This shows that these peak flows are occurring with a greater frequency than
once per year.
Note that for return periods below 10 years, the annual series and the partial duration series tend to
diverge.
39 of 54 05-03-2020, 19:51
Review Questions
When using the annual maximum series in flood frequency analysis, _____. (Choose all that apply.)
a) probabilities can be computed for each data point

b) multiple peak flows can be used for each annual time period
c) the number of data points equals the number of years in the dataset
d) only instantaneous peak flows can be used
The correct answers are a and c.
The annual maximum daily flow is the instantaneous peak flow within a particular day within the year.
(True or False.)
a) True
b) False
Which statement about using a partial duration series in a flood frequency analysis is NOT true?
(Choose the best answer.)
a) More data points than an annual maximum series are used in the analysis.
b) Better estimates on rare, large floods are possible.
c) Data must be plotted using the computed return period.
d) Annual maximum series data are included within the partial duration data.
40 of 54 05-03-2020, 19:51
Design Events
It is difficult to estimate flood frequency values for ungauged basins or basins with limited data. To help
address this problem, the concept of a design storm was developed. The design storm is a "what–if"
approach that may be helpful when planning for potential floods.
Identify the reason for using design events.

Understand limitations and constraints when using design events.
Explain the concept of probable maximum flood, standard project floods, and storm transposition.
41 of 54 05-03-2020, 19:51
Purpose of Design Events
Design events are used to evaluate flood risks to structures and facilities. To meet set safety levels and/or
reduce damage costs, structures must be able to withstand certain natural events.
A design storm is a precipitation event of a specific magnitude and duration; for example, 150 millimeters
in 24 hours. A design storm results in a design flood of a specific magnitude that is used in a "what– if"
scenario to estimate the impact.
Engineering projects vary in the extent of cost–benefit analysis undertaken. That is, some projects will
accept occasional flooding in order to reduce the costs of making structures more flood resistant. Other
projects aim for little or no flood losses and may allocate extra planning and construction costs for flood
resistance. Knowing the potential magnitudes of floods is important, and that's where design events may
be helpful.
42 of 54 05-03-2020, 19:51
What is a Design Event?
The basic principle behind design events is that a precipitation event of a given return frequency will
produce an estimated flood having the same return frequency. For example, a 25–year rainfall would
result in a 25–year flood.
However, there are some limitations with this approach. Design events have been developed by statistical
analysis of long–term precipitation records. But these records do not allow the design storm to define the
rainfall areal coverage or its temporal and spatial distribution, both of which can be very important for
runoff. In addition, basin conditions can greatly influence runoff.
As a result, a 25–year rainfall won't necessarily produce a 25–year flood event. For example, if the ground
is already saturated from recent rainfall, a 25–year storm will likely produce a flood with greater than a
25–year return period magnitude.
Despite these limitations, the concept of a design event as a "what–if" approach is beneficial to
engineering projects.
43 of 54 05-03-2020, 19:51
Types of Design Events
Many countries construct design storms based on depth–duration–frequency curves, which show the
relationship between the depth of precipitation and frequency of occurrence of different duration periods.
These curves are generated from historical records of major precipitation events for a given region. The
standard project flood, used by the US Army Corps of Engineers, employs a similar approach.
44 of 54 05-03-2020, 19:51
The Probable Maximum Flood (or PMF), sometimes also referred to as the maximum possible flood, is
the theoretical upper limit that can be expected from the most severe combination of meteorological
and/or hydrologic conditions. It is calculated using the greatest depth accumulation of precipitation that is
theoretically possible for a given storm duration in a particular geographic region.
The PMF may also include structural failures, such as a dam or levee breaks, if applicable for the area. An
event such as this, with or without a storm, may produce the greatest magnitude flood for a region. This
level of flood is not generally used in a flood frequency analysis, since the probability and return period
45 of 54 05-03-2020, 19:51
value are unknown.
Another design event used in many locations is the Storm Transposition approach, in which an observed
storm is transposed spatially to another area of the same region.
46 of 54 05-03-2020, 19:51
Flood Frequency Exercise

In this exercise we will apply a design flood approach to a fictitious case and look at the effect of the
length of period of record on flood prediction.
You own a bicarbonate of soda warehouse near the river. Your business is protected for up to 900 flow
units. At 900 flow units there would be major damage to the merchandise causing the business to fizzle.
Based on a 30–year record, the annual series plot shows that 900 flow units has a _____ return period.
(Use the chart above to choose the best answer.)
a) 100–year
b) 50–year
c) 30–year
d) 25–year
You should have found that a flow of 900 has a 50–year return period. Thus, your 50–year design flood
has a flow of 900 flow units. Your business is designed to handle that much without damage.
47 of 54 05-03-2020, 19:51
Recently, more flow data was discovered that extends the period of record back another 30 years for a
total of 60 years. The newly discovered data indicate that higher flows with major flood events were
more common than the original 30–year dataset suggests. What would you expect the new graph of flow
versus return period to look like compared to the original? (Choose the most likely new plot from the
chart above.)
a) Right side is higher

b) Right side is the same
c) Right side is lower
The right side of the graph would go up, but the left side would be unchanged. This shows that the
occurrence of major floods is more frequent, and the magnitude of specific return period floods is
greater.
Use the the image below to answer the next two questions.
48 of 54 05-03-2020, 19:51
Looking at this graph that now shows the new and old data, please answer the following question:
With the longer period of data, the 50–year flood should now be associated with a flow magnitude of
about _____ flow units. (Choose the best answer.)
a) 2050
b) 900
c) 1130
The new 50–year flood has a magnitude of 1130 flow units.
Because your design flood is for a magnitude of 900 flow units, what is the return period of such a flow
with the new data? (Choose the best answer.)
49 of 54 05-03-2020, 19:51
a) about 10 years
b) 50 years
c) about 40 years
d) about 25 years
The return period of the flood associated with your design flood has changed from a recurrence of 50
years to one of about 24 or 25 years.
What does this mean for flood risk to your water–soluble business? (Choose all that apply.)
a) The business is still designed to handle up to the 50–year flood.

b) The business is only designed for a 25–year flood.
c) The chance of a damaging flood is greater than previously thought.
The correct answers are b and c.
Using the original data set, the design flood that the business was protected for had a return period of 50
years. The new data shows that the design flood actually may have a return period of 25 years. This
indicates twice the risk of damaging flooding in any given year. Thus, the business is likely to suffer
damages more frequently than originally thought.
50 of 54 05-03-2020, 19:51
Review Questions
The purpose of design storms and design floods is to _____. (Choose the best answer.)
a) develop "what if" scenarios for potential flood impacts

b) estimate the impact of extreme (probable maximum) events
c) estimate the impact of more common flood scenarios that are not extreme
d) all of the above
The correct answer is D, all of the above
A probable maximum precipitation value is always the highest precipitation ever observed for a given
basin. (True or False.)
a) True
b) False
In design events, a 10–year storm will produce a 10–year flood even though this is not always true in the
real world. (True or False.)
a) True
b) False
51 of 54 05-03-2020, 19:51
Summary
Overview of Flood Frequency Analysis
Flood frequency analysis provides information about the statistical potential for floods.
Return period expresses a time period (i.e., the 100–year flood) that is the statistical average time
duration between floods of a certain magnitude. The greater the return period, the larger the flood.
Exceedance probability or probability of occurrence for a given year is the inverse of the return
period. Therefore, the 100–year flood would have an exceedance probability of 1/100, or 1%.
Floods, however, do not occur at evenly spaced time intervals. Thus, a 100–year flood won't
necessarily occur just once in a 100–year period.
A given return period flood is not necessarily the result of a rainfall event with the same return
period. That is, a 25–year rainfall won't necessarily produce a 25–year flood.
There are two applications of flood frequency analyses:
1. Estimating the potential magnitude of a flood that can occur over a given time interval
2. Estimating the return period of a certain magnitude flood
The period of record should be at least half as long as the computed return period. That is, to
estimate a 500–year flood one should have at least 250 years of data.
Flood frequency statistics are more representative of potential floods when the period of record is
long enough to capture all types of events that are possible.
The probability of occurrence, or exceedance probability, of a flood over a given time period can be
computed with standardized formulas. This can answer a question such as, "What is the chance for
a 100–year flood over the next 30 years?"
The probability of an event not occurring, the non–exceedance probability, can be computed as
well.
Reliable flood frequency data require that floods are independent. That is, one flood doesn't directly
influence the next, and flow should return to baseflow between floods.
Reliable flood frequency data require that floods are homogeneous. That is, floods caused by dam
failures are not the same as floods caused by rainfall. Basin alterations, such as urbanization, also
violate the assumption of homogeneity in flood frequency analysis.
The annual series is used to compute flood frequency statistics using the peak flow in each calendar
year.
The instantaneous peak flow is the peak flow at any moment, but these measurements are not
commonly available in flood frequency analyses.
The annual maximum daily flow is often used in flood frequency analysis. It is the peak average
daily value over the course of a calendar year.
Annual maximum series data are ranked and then probabilities are computed with one of several
available formulas, such as the Weibull formula.
Return period data for the annual maximum series is computed by taking the inverse of the
probability.
Annual maximum series data are plotted as flow versus probability, or flow versus return period.
Annual maximum series data are not typically plotted on linear scales. One reason is that a linear
probability scale would imply that there are known probabilities of zero and 100 percent, which is
not the case.
The partial duration series is used to compute flood frequency statistics using all flow values above
a certain threshold and typically has more data points than the annual maximum series.
Compared to the annual maximum series, the partial duration series is better suited for smaller
flood events, especially those with short return periods.
The partial duration series graph may contain points with less than a 1–year return period.
52 of 54 05-03-2020, 19:51
Design Events
Design events are used to evaluate the risks from floods.

Design storm criteria contain a magnitude and duration. That is, a 25–year storm may be described
as 150 mm in 24 hours.
Design storms of a certain return period are assumed to produce a flood of similar stature. That is, a
25–year storm is assumed to produce a 25–year flood. We know that this can be a faulty
assumption.
Design storms are developed from statistical analysis of long–term precipitation records. A design
storm does not account for areal coverage or the spatial and temporal resolution of precipitation.
A probable maximum precipitation event and the resulting probable maximum flood (PMF) is a
theoretical estimate of the biggest event possible in a certain area. This value often reflects the
potential of floods from dam or levee failures.
The standard project flood is a design criteria used by the U.S. Army Corps of Engineers. It results
from a set of storm characteristics that can be reasonably expected in for a given area.
You have reached the end of this module. Please take a moment to provide us with your feedback by
completing a brief survey: Begin Survey
53 of 54 05-03-2020, 19:51
Copyright 2010, University Corporation for Atmospheric Research. All rights reserved. Legal notices
54 of 54 05-03-2020, 19:51

Flood Frequency Analysis: International Edition: Produced by The COMET Program Print Version

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Flood Frequency Analysis: International Edition: Produced by The COMET Program Print Version

Uploaded by

Copyright:

Available Formats

Flood Frequency Analysis: International Edition file:///H:/hydro/basic_int/flood_frequency/print.

Flood Frequency Analysis: International Edition

Introduction to Flood Frequency Analysis

1. To predict the possible flood magnitude over a certain time period

In this section you will learn to:

Define the meaning of return period (i.e., the 100–year flood).

Role in Flood Prediction Process

Time Distribution of Floods

Flood Frequency Misconceptions

Click to play animation.

Estimate of Flood Magnitude

Estimate of Return Period

a) designing transportation infrastructure such as bridges and culverts

The correct answer is d.

The 100–year flood is _____. (Choose all that apply.)

a) less frequent than the 500–year flood

The correct answers are b and d.

A 100–year rainstorm will always produce a 100–year flood. (True or False.)

The correct answer is b.

Statistical Representation of Floods

In this section you will learn to:

Explain how the period of record impacts flood frequency guidance.

Period of Record Issues

There are several different return periods traditionally used by hydrologists.

Confidence in Return Period Estimates

Exceedance probability = 1 – (1 – p)n

Example of Exceedance Probability

Calculation for Probability of 50–Year Flood Over 50–Year Period

Now we have (1–0.36), which is 0.64.

Calculation for Probability of 100–Year Flood Over 30–Year Period

n=30 and we see from the table, p=0.01 .

The 1–p is 0.99, and .9930 is 0.74.

Independent and Homogeneous Data

Impact of Basin Alterations

The correct answer is a.

The correct answer is b.

The correct answer is a.

a) Less than 0.01

The correct answer is c.

The correct answer is a.

The correct answer is d.

Analyzing Flood Data

In this section you will learn to:

Annual Maximum Daily Flow

The meaning of the term "peak flow" can vary in practice.

Annual Series Rank

Annual Series Probability

Note: Data above is abbreviated for ease of viewing.

Annual Series Display

Partial Duration Series

Partial Duration Series Exercise

To answer the following questions, refer to the table below.

The correct answer is c.

The correct answer is d.

The correct answer is a.

The correct answer is b.

a) Annual maximum series

The correct answer is b.

Preparing Partial Duration Series

Inverse Weibull Formula:

Partial Duration Display

a) probabilities can be computed for each data point

The correct answers are a and c.