Engineering Hydrology; Lecture note

CHAPTER FOUR

FREQUENCY ANALYSIS

4.1. General
Water resource systems must be planned for future events for which no exact time of occurrence
can be forecasted. Hence, the hydrologist must give a statement of the probability of the stream
flows (or other hydrologic factors) will equal or exceed (or be less than) a specified value. These
probabilities are important to the economic and social evaluation of a project. In most cases,
absolute control of the floods or droughts is impossible. Planning to control a flood of a specific
probability recognizes that a project will be overtaxed occasionally and damages will be
incurred. However, repair of the damages should be less costly in the long run than building
initially to protect against the worst possible event. The planning goal is not to eliminate all
floods but to reduce the frequency of flooding, and hence the resulting damages. If the socio-
economic analysis is to be correct, the probability of flooding must be eliminated accurately. For
major projects, the failure of which seriously threatens human life, a more extreme event, the
probable maximum flood, has become the standard for designing the spillway.

This chapter deals with techniques for defining probability from a given set of data and with
special methods employed for determining design flood for major hydraulic structures.

Frequency analysis is the hydrologic term used to describe the probability of occurrence of a
particular hydrologic event (e.g. rainfall, flood, drought, etc.). Therefore, basic knowledge about
probability (e.g. distribution functions) and statistics (e.g. measure of location, measure of
spread, measure of skewness, etc) is essential. Frequency analysis usually requires recorded
hydrological data. Hydrological data are recorded either as a continuous record (e.g. water level
or stage, rainfall, etc.) or in discrete series form (e.g. mean daily/monthly/annual flows or
rainfall, annual series, partial series, etc.).

For planning and designing of water resources development projects, the important parameters
are river discharges and related questions on the frequency & duration of normal flows (e.g. for
hydropower production or for water availability) and extreme flows (floods and droughts).

In many hydraulic-engineering applications such as those concerned with floods, the probability
of occurrence of a particular extreme rainfall, e.g. a 24-h maximum rainfall, will be of
importance. Such information is obtained by the frequency analysis of the point rainfall data.

Probability of Occurrence: The probability of occurrence (e.g. rainfall) whose magnitude is

equal to or in excess of a specified magnitude X is denoted by P.

Recurrence Interval: By recurrence interval of a given storm is meant the time interval during
which the given storm is likely to be equalled or exceeded. Thus, if we say that at a given station,
the maximum precipitation of 20 cm has got a recurrence interval of 5 years, we mean that on
this station, the chances of rainfall are such that once in five years, rain is likely to equal or
exceed 20 cm. This does not mean that a rain equal to 20 cm or more will occur after every five
years. It may occur twice in one set of five years and may not occur at all for consecutive eight

Page 1 of 24
 Engineering Hydrology; Lecture note

years or so. What it means is that during 100 years or so, such a rain will occur 20 times or so.
Therefore, the recurrence interval T, (also known as return period) is defined as T = 1 / P.

Time Series: The rainfall of a place is a random hydrologic process and the rainfall data at a
place when arranged in chronological order constitute a time series.

If the probability of an event occurring is P, the probability of the event not occurring in a given
year is q = (1-P).

The binomial distribution can be used to find the probability of occurrence of an event r times in
n successive years. Thus,

n
Pr,n = nCr P q = ----------- Pr qn-r
r n-r

(n-r) r

Where, Pr,n = probability of a random hydrologic event (rainfall) of given magnitude and
exceedence probability P occurring r times in n successive years. Thus, for example:

(a) Probability of an event of exceedence P occurring 2 times in n successive years is:

n
P2,n = ----------- P2 qn-2
(n-2) 2
(b) The probability of the event not occurring at all in n year is:
P0,n = qn = (1-P)n
(c) The probability of the event occurring at least once in n successive years (R):
R = 1 – qn = 1 – (1 – P)n
This probability is called risk and hence represented by R.

Example 4.1: Analysis of data on maximum one day rainfall depth at one station indicated that a
depth of 280 mm had a return period of 50 years. Determine the probability of one day rainfall
depth equal to or greater than 280 mm at that station occurring
(a) Once in 20 successive years
(b) Two times in 15 successive years
(c) At least once in 20 successive years
Solution: Here, P = 1 / 50 = 0.02
(a) n = 20, r = 1,
Then, P1,20 = 20 x 0.02 x (1 – 0.02)19
191
= 20 x 0.02 x (0.98)19 = 20 x 0.02 x 0.68123 = 0.272
(b) n = 15, r = 2

Then, P2,15 = 15 P2 q13

13 2

Page 2 of 24
 Engineering Hydrology; Lecture note

= 14 x 15 x 0.022 x 0.9813
2
= 14 x 15 x 0.0004 x0.769 = 0.0323
2
(c) P1 = 1 – (1 – P) n
= 1 – (1– 0.02)20 = 0.332
Example 4.2: A one day rainfall of 15.0 cm at a place X was found to have a return period of
100 years. Calculate the probability that a one-day rainfall of this or larger magnitude
(a) Will not occur at X during the next 50 years
(b) Will occur in the next year
Solution: P = 1 / 100 = 0.01
(a) n = 50 years, r = 0
P0,50 = qn = (1 – P)n = (1 – 0.01)50 = (0.99)50 = 0.605
(b) n = 1 year, r = 1, P = 0.01
P1,n = nP (1 – P)n-1
= 1 x 0.01 (1 – 0.01)1-1 = 0.01 (1 – 0.01)0 = 0.01
Example 4.3: On the basis of isopluvial maps the 50 years-24h maximum rainfall at Harar is
found to be 16.0 cm. Determine the probability of a 24h rainfall of magnitude equal to or
greater than 16.0 cm occurring at Harar. A map depicting maximum rainfall depth for different combinations of one
return period and one duration is called isopluvial map.
(a) Once in 10 successive years
(b) Two times in 10 successive years and
(c) At least once in 10 successive years
Solution: P = 1 / 50 = 0.02
(a) r = 1, n = 10
P1,n = n x P1 x qn-1 = n x P x (1- P)n-1
(n–1) x 1 (n–1) x1
= 10 x 0.02 x (1 – 0.02)9 = 10 x 0.02 x 0.989
9 x 1
= 10 x 0.02 x 0.83375 = 0.167
(b) P = 0.02, r = 2, n = 10

P2,10 = 10 x P2 x q10-2

(10–2) x 2
= 9 x10 x 0.022 x (1 – 0.02)8
2
= 9 x 10 x 0.00042 x 0.850763 = 0.0153
2
(c) P = 0.02, n = 10
R = 1 – qn = 1- (1 – P)n
= 1 – (1 – 0.02)10 = 1 – 0.9810 = 1 – 0.817072806 = 0.183

Page 3 of 24
 Engineering Hydrology; Lecture note

4.2. Flow Frequency

The question a planner or decision maker would ask a hydrologist concerning normal flows is
the length of time (duration) that a certain river flow is expected to be exceeded. An answer to
this question is provided by the flow duration curve (FDC) that is the relationship between any
given discharge and the percentage of time that the discharge is exceeded. Taking the n-years of
flow records from a river gauging stations, there are 365(6)n daily mean discharges. The
discharge is compiled, starting with the highest values. If N number of data points are used for
analysis, the plotting position of any discharge (or class value) Q is:

Pp = m / (N+1) (4.1)
Where, m = is the order number of the discharge (or class value)
Pp = percentage probability of the flow magnitude being equaled or exceeded.
The FDC only applies for the period for which it was derived. If this is a long period, say more
than 10 to 20 years, the FDC may be regarded as a probability curve or flow frequency curve,
which may be used to estimate the percentage of time that a specified discharge will be equaled
or exceeded in the future. An example is demonstrated in table 4.1 below.

Table 4.1: Flow records

Discharge %Exceeded or
(m3/s) Descending Equaled (m /
Order Rank (N+1))
(a) (b) (c) (d)
106.70 1200 1 8.33%
107.10 964.7 2 16.67%
148.20 497 3 25.00%
497.00 338.6 4 33.33%
1200.00 177.6 5 41.67%
964.70 148.2 6 50.00%
338.60 142.7 7 58.33%
177.60 141 8 66.67%
141.00 141 9 75.00%
141.00 126.6 10 83.33%
142.70 107.1 11 91.67%
126.60 106.7 12 100.00%

Page 4 of 24
 Engineering Hydrology; Lecture note

Figure 4.1: Flow Duration Curve

The shape of the flow-duration curve gives a good indication of a catchment’s characteristics
response to its average rainfall history. An initially steeply sloped curve results from a very
variable discharge, usually from small catchments with little storage where the stream flow
reflects directly the rainfall pattern. Flow duration curves that have very flat slope indicate little
variation in flow regime, the resultant of the damping effects of large storages.

4.3. Flood Probability

4.3.1. Selection of Data

If probability analysis is to provide reliable answers, it must start with a data series that is
relevant, adequate, and accurate. Relevance implies that the data must deal with the problem.
Most flood studies are concerned with peak flows, and the data series will consist of selected
observed peaks. However, if the problem is duration of flooding, e.g., for what periods of time a
highway adjacent to a stream is likely to be flooded, the data series should represent duration of
flows in excess of some critical value. If the problem is one of interior drainage of a leveled
area, the data required may consist of those flood volumes occurring when the main river is too
high to permit gravity drainage.

Adequacy refers primarily to length of record, but sparsity of data collecting stations is often a
problem. The observed record is merely a sample of the total population of floods that have
occurred and may occur again. If the sample is too small, the probabilities derived cannot be
expected to be reliable. Available stream flow records are too short to provide an answer to the
question: How long must a record be to define flood probabilities within acceptable tolerances?

Accuracy refers primarily to the problem of homogeneity. Most flow records are satisfactory in
terms of intrinsic accuracy, and if they are not, there is little that can be done with them. If the
reported flows are unreliable, they are not a satisfactory basis for frequency analysis. Even
though reported flows are accurate, they may be unsuitable for probability analysis if changes in
the catchment have caused a change in the hydrologic characteristics, i.e., if the record is not

Page 5 of 24
 Engineering Hydrology; Lecture note

internally homogenous. Dams, levees, diversions, urbanization, and other land use changes may
introduce inconsistencies. Such records should be adjusted before use to current conditions or to
natural conditions. There are two data series of floods:

(i) The annual series, and

(ii) The partial duration series.

The annual series constitutes the data series that the values of the single maximum
daily/monthly/annually discharge in each year of record so that the number of data values
equals the record length in years. For statistical purposes, it is necessary to ensure that the
selected peak discharges are independent of one another. This data series is necessary if the
analysis is concerned with probability less than 0.5. However, as the interest are limited to
relatively rare events, the analysis could have been carried out for a partial duration series to
have more frequent events. The partial duration series constitutes the data series with those
values that exceed some arbitrary level. All the peaks above a selected level of discharge (a
threshold) are included in the series and hence the series is often called the Peaks Over
Threshold (POT) series. There are generally more data values for analysis in this series than in
the annual series, but there is more chance of the peaks being related and the assumption of true
independence is less valid.

4.3.2. Plotting Positions

The purpose of the frequency of an annual series is to obtain a relation between the magnitude
of the event and its probability of exceedence. The probability analysis may be made either by
empirical or by analytical methods.

Empirical method: A simple empirical technique is to arrange the given annual extreme series
in descending order of magnitude and to assign an order (rank) number m. Thus, for the first
entry m=1, for the second entry m = 2 and so on till the last event for which m = N = number of
years of record. The probability P of an event equalled to or exceeded is given by the Weibull
formula:

P = (m)
N+1
The recurrence interval T = 1 / P = 1 = N+1
m / (N+1) m
The above equation is an empirical formula and there are several other such empirical formulae
available to calculate P.

Table 4.2 Plotting-position formulae

Method Formula
California m/N
Hazen (m – 0.5) / N
Weibull m / (N + 1)
Chegodayev (m – 0.3) / N + 0.4)
Blom (m – 0.44) / (N + 0.12)
Gringorten (m – 3/8) / (N +1/4)

Page 6 of 24
 Engineering Hydrology; Lecture note

The exceedence probability of the event obtained by the use of an empirical formula such as
above equation is called plotting position. Weibul formula is the most popular plotting position
formula.

Example 4.4: For a station A, the recorded annual 24-h maximum rainfalls are given below:
a) Estimate the annual 24-h maximum rainfall with a return period of 13 and 50 years.
b) What would be the probability of a rainfall of `magnitude equal to or exceeding 10
cm occurring in 24-h at station A?
Table: Annual maximum 24-h rainfall at station A
Year 50 51 52 53 54 55 56 57 58 59 60
Rainfall (cm) 13 12 7.6 14.3 16 9.6 8.0 12.5 11.2 8..9 8.9
Year 61 62 63 64 65 66 67 68 69 70 71
Rainfall (cm) 7.8 9.0 10.2 8.5 7.5 6.0 8.4 10.8 10.6 8.3 9.5
Solution:
(a) N = 22years
Rainfall Probability Return period, Rainfall Probability, Return
M m
(cm) P= m/N+1 T=1/P (years) (cm) P period
1 16.0 0.043 23.00 12 9.0 0.522 1.92
2 14.3 0.087 11.50 13 8.9 0.565 1.77
3 13.0 0.130 7.67 14 8.9 0.609 1.64
4 12.5 0.174 5.75 15 8.5 0.652 1.53
5 12.0 0.217 4.60 16 8.4 0.696 1.44
6 11.2 0.261 3.83 17 8.3 0.739 1.35
7 10.8 0.304 3.29 18 8.0 0.783 1.28
8 10.6 0.348 2.88 19 7.8 0.826 1.21
9 10.2 0.391 2.56 20 7.6 0.870 1.15
10 9.6 0.435 2.30 21 7.5 0.913 1.10
11 9.5 0.478 2.09 22 6.0 0.957 1.05
From the curve:
Return period, T years Rainfall magnitude, cm
13 14.60
50 18.40
(b) For rainfall 10 cm, T = 2.4 years and P = 0.417

Probability analysis seeks to define the flood flow with probability of p being equaled or exceed
in any year. Return period T is often used in lieu of probability to describe a design flood. Return
period and probability are reciprocals, i.e,

P = 1/T (4.2)

To plot a series of peak flows as a cumulative frequency curves it is necessary to decide on a

probability or return period to associate with each peak. There are various formulas for defining
this value as shown in table 4.2 above.

The probability of occurrence of the event r times in n successive years can be obtained from:

Page 7 of 24
 Engineering Hydrology; Lecture note

n!
Pm =❑nC r Pr q n−r= Pr q n−r (4.3)
( n−r ) ! r !
Where q = 1 - P.

Consider, for example, a list of flood magnitudes of a river arranged in descending order as
shown in Table 4.3. The length of record is 50 years.

Table 4.3: Calculation of Frequency T

Order No. m Flood magnitude Q (m3/s) T in years = 51/m

1 160 51.00
2 135 25.50
3 128 17.00
4 116 12.75
. . .
. . .
. . .
49 65 1.04
50 63 1.02
The last column shows the return period T of various flood magnitude, Q. A plot of Q Vs T
yields the probability distribution. For small return periods (i.e. for interpolation) or where
limited extrapolation is required, a simple best-fitting curve through plotted points can be used as
the probability distribution. A logarithmic scale for T is often advantageous. However, when
larger extrapolations of T are involved, theoretical probability distributions (e.g. Gumbel
extreme-value, Log-Pearson Type III, and log normal distributions) have to be used. In
frequency analysis of floods, the usual problem is to predict extreme flood events. Towards this,
specific extreme-value distributions are assumed and the required statistical parameters
calculated from the available data. Using these flood magnitudes for a specific return period is
estimated.

4.3.3. Statistical Analysis of Rainfall Data

1. Population and Sample: Observed values of x (variate) for a finite number of years is known
as “sample” of x. Say annual peaks or annual rainfall for 75 years gives the sample; on the
other hand, population consists of the values of annual flood peaks from time immemorial to
eternity. The “population parameters” can be estimated by means of parameters obtained from
the sample, known as “sample parameters”. Each phenomenon is characterized by a certain
value which vary in time and space. This characterization is called “variable” and its
particular value is a variate.
2. Central Tendency: Three types of parameters are generally used to represent measures of
central tendency.
(a) Expected Value: This value of random variable x is given by
∞
 = ∫ xf (x) dx where - ∞ < x < ∞
-∞
This population can be estimated by the sample parameters as:

Page 8 of 24
 Engineering Hydrology; Lecture note

(i) Arithmetic mean, ‾x = ∑x, for group data, ‾x = ∑ fx

n n
1/n
(ii) Geometric mean, ‾xg = (x1. x2. x3……………xn)

(iii) Harmonic mean, ‾xh = n

∑1/x
Where n = the size of the sample, say the number of years of annual flood peaks.
For any set of variables ‾x > ‾xg > ‾xh
In most cases the arithmetic mean gives the best estimate of the expected value i.e.  ≈‾x

(b) Median: The value of the variate such that half of the variates are below it and the
other half above it, is called the median of the series, i.e. it is the value of the variate
having a 50% cumulative frequency.

(c) Mode: The value of the variate having the highest frequency is called the mode. For
unimodal curves which are moderately skewed, the empirical relationship is:
Mean – Mode = 3 (Mean – Median)
3. Variability: The measures of variability or dispersion of a probability distribution curve are
given by the following parameters
(a) Mean deviation: The mean of the absolute deviations of values from their mean is
called mean deviation (M.D.).

M.D. = ∑ (x - ‾x) , for group data, M.D. = ∑ f (x - ‾x)

n n
(b) Standard deviation: It is the square root of the mean squared deviation of the
variables from their mean and the standard deviation for the population (σ p) is given
by:

σp = √∑( x - )2
n
and this is estimated from the standard deviation for the samples (σ) given by:

σ = √ ∑(x - ‾x) 2
(n – 1)

For group data, σ = √ ∑ f (x - ‾x) 2

(n – 1)
(c) Variance: The Square of the standard deviation is called variance i.e. given by σ p2 for the
population and σ2 for the sample.
(d) Coefficient of Variation: The standard deviation divided by the mean is called the
coefficient of variation (Cv) and is given by:

Page 9 of 24
 Engineering Hydrology; Lecture note

Cv = σp ≈ σ
 ‾x
(e) Skewness (asymmetry): The lack of symmetry of a distribution is called skewness or
asymmetry. The population skewness (α) is given by:

α = ∑( x - )3
n
Estimation from the sample skewness (a) is given by:
a = ∑(x - ‾x)3
(n – 1)

For group data, a = ∑ f (x - ‾x)3

(n – 1)
(f) Coefficient of Skewness:

(i) The degree of skewness of the distribution is wisely measured by the “coefficient of
skewness” (Cs) and is given by:

Cs = α ≈ a = ∑ f (x - ‾x)3 (CS ≈ Cv)

σp3 σ3 (n – 1) σ3
(ii) Another measure of skewness often used in practice is Pearson’s Skewness (S k) and is
given by:
Sk =  - Mode ≈ ‾x - Mode
σp σ
Or, Sk = 3 ( ‾x – Median)
σ
(g) Coefficient of Flood (Cf):
Cf = ‾x where A = Area of catchment in km2
A0.8/ 2.14
Example 4.5: The annual rainfall of a place for a period of ten years from 1961 to 1970 are
respectively as 30.3, 41.0, 33.5, 34.0, 33.3, 36.2, 33.6, 30.2, 35.5, and 36.3.
Determine the mean and median values of annual rainfall of the place.
Solution:
(i) Mean, ‾x = ∑ x = 30.3+41.0+33.5+34.0+33.3+36.2+33.6+30.2+35.5+36.3
n 10

= 343.9 = 34.39
10
(ii) Median: Arrange the samples in the ascending order;
30.2, 30.3, 33.3, 33.5, 33.6, 34.0, 35.5, 36.2, 36.3, 41.0
No. of items = 10 i.e. even

Page 10 of 24
 Engineering Hydrology; Lecture note

Median = 33.6 + 34.0 = 33.8

2
Note: If 11 years of record, say 1960 to 1970 had been given, the median would have been the
6th item (central value) when arranged in the ascending order.
Example 4.6: For the grouped data of flood in the river for 87 years (1885 – 1971), find the
mean, median and mode. Determine the standard deviation, coefficient of skewness
and the coefficient of variation when the class interval is fixed by exclusive method.
Class Interval Frequency
(cumecs)
0 -2 0
2–4 17
4–6 27
6–8 18
8 – 10 18
10 – 12 3
12 – 14 0
14 – 16 2
16 -18 1
18 – 20 1

Solution:

(i) Mean = ‾x = ∑ f x = 573 = 6.6 cumecs

n 87
n+1 - f
(ii) Median = Lmd + ( 2 ) x C.I
fmd

88 - 17
= 4 + (2 )x2 = 4 + (44 – 17) x 2 = 4 + 2 = 6 cumecs
27 27
(iii) Mode: The value which has maximum frequency (27) = 5 cumecs
(iv) Standard deviation
σ = √∑(x - ‾x) 2 = 858.1 = 3.16 cumecs
(n – 1) 86

(v) Coefficient of Skewness (CS) = ∑ f (x - ‾x) 3 = 4725.050 = 1.741

(n – 1) σ 3 2713.687

Pearson first coefficient, CS1 = ‾x – Mode = 6.6 – 5 = 1.6 = 0.507

σ 3.16 3.16

Pearson second coefficient, CS2 = 3 ( ‾x – Median) = 3 (6.6 – 6.0) = 3  0.6 = 0.57

Page 11 of 24
 Engineering Hydrology; Lecture note

σ 3.16 3.16
All the coefficients of skewness are positive and skew is to the right; if the coefficients were
negative, the skew would have been to the left.
(vi) Coefficient of Variation, CV = σ = 3.16 = 0.479 = 47.9%
‾x 6.6

C.I, cumecs X Freq, f f. x x - ‾x (x- ‾x)2

0–2 1 0 0 - 5.6 31.4
2–4 3 17 51 -3.6 13.0
4–6 5 27 135 -1.6 2.56
6–8 7 18 126 0.4 0.4
8 – 10 9 18 162 2.4 5.76
10 – 12 11 3 33 4.4 19.40
12 – 14 13 0 0 6.4 41.0
14 – 16 15 2 30 8.4 70.50
16 – 18 17 1 17 10.4 108.00
18 – 20 19 1 19 12.4 154.00
Σ= 87 Σ= 573 Σ = 858.1

f(x- ‾x)2 (x- ‾x)3 f(x- ‾x)3 C. F Probability = C.F/Σfx100 (%)

0 -176 0 87 100
221 -46.8 -796.00 87 100
69.2 -4.1 -110.00 70 80.5
2.9 0.064 1.15 43 49.5
103.8 13.82 249.00 25 28.8
58.2 85.30 255.90 7 8.05
0 262.60 0 4 4.6
141.0 593.00 1186.00 4 4.6
108.0 1123.00 1123.00 2 2.3
154.0 1910.00 1910.00 1 1.15
Σ= 4725.05

Mean, ‾x = Σ f.x / n = 573 / 87 = 6.6 cumecs

Standard deviation, σ = √ ∑(x - ‾x) 2 = 858.1 = 3.16 cumecs

(n – 1) 86

4.3.4. Theoretical Distributions of Floods

Statistical distributions are usually demonstrated by use of samples numbering in the thousands.
No such samples are available for stream flow and it is not possible to state with certainty that a
specific distribution applies to flood peaks. Numerous distributions have been suggested on the
basis of their ability to “fit” the plotted data from streams.

Page 12 of 24
 Engineering Hydrology; Lecture note

Ven Te Chow has shown that most frequency-distribution functions applicable in hydrologic
studies can be expressed by the following equation known as the general equation of hydrologic
frequency analysis:

x T = x́+ Kσ (4.4)

Where xT = value of the variate X of a random hydrologic series with a return period T, x́ = mean
of the variate, σ = standard deviation of the variate, K = frequency factor which depends upon
the return period, T and the assumed frequency distribution.

4.3.5. Extreme-Value Type I Distribution (Gumbel’s Method)

This extreme value distribution was introduced by Gumbel (1941) and is commonly known as
Gumbel's distribution. It is one of the most widely used probability-distribution functions for
extreme values in hydrologic and meteorological studies for prediction of flood peaks, maximum
rainfalls, and maximum wind speed, etc. Therefore, this extreme value theory of Gumbel is only
applicable to annual extremes. In contrast to the previous example, in the Gumbel method the
data are ranked in ascending order and it makes use of the probability of non-exceedence q=1-P
(the probability that the annual maximum flow is less than a certain magnitude). The return
period T is therefore given by T = 1 / P = 1 / (1-q).

Gumbel makes use of a reduced variate y as a function of q, which allows the plotting of the
distribution as a linear function between y and x (the maximum flow in this case). Gumbel also
defined a flood as the largest of the 365 daily flows and the annual series of flood flows
constitute a series of largest values of flow. According to his theory of extreme events, the
probability of occurrence of an event equal to or larger than a value x0 is
−y
P (X≥X0) = 1- e−e (4.5)
In which y is a dimensionless variable given by
y = α ( X−a )
a = x́−0.45005 σ x
α =1.2825/σ x (4.6)
1.2825 ( X− X́ )
Thus y = +0.577 (4.7)
σx
Where X́ =mean∧σ x =standard deviationof the variate X
In practice it is the value of X for a given P that is required as such Eq. (4.7) is transposed as
y =-ln (-ln(q)) = -ln (-ln(1-p)) (4.8)
Meaning that the probability of non-exceedence equals
−y
P(X ≤ x0) = q = e−e (4.9)
Noting that the return period T = 1/P and designating; yT = the value of y, common called the
reduced variate, for a given T

[ ( ( ))]
y T =− ln ln
T
T −1
(4.10)

Page 13 of 24
 Engineering Hydrology; Lecture note

[ ( T −1 ))]
y T =− 0.834 +2.303 log log( T

Now rearranging Eq. (4.7) the values of the variate X with a return period T is
X T =x́ + K σ x (4.11)
y −0.577
Where K= T (4.12)
1.2825
Note that Eq. (4.11) is of the same form as the general equation of hydrologic-frequency
analysis, Eq. (4.4). Further eqs (4.11) and (4.12) constitute the basic Gumbel's equations and are
applicable to an infinite sample size (i.e. N →∞).

Since practical annual data series of extreme events such as floods, maximum rainfall depths,
etc., all have finite lengths of record; Eq. (4.12) is modified to account for finite N as given
below for practical use.

4.3.6. Gumbel's Equation for Practical Use

Equation (4.11) giving the variate X with the return period T is used as
X T =x́ + K σ n−1 (4.13)
2
( x− x́ )
Where σ n−1 =standard deviation of the sample=
√
y T − ý n
∑ N −1
K = frequency factor expressed as K= (4.14)
Sn
In which y T =reduced variate , a function of T ∧is givenby

[ ( ( ))]
y T =− ln ln
T
T −1
(4.15)

[ ( T −1 ))]
Or y T =− 0.834 +2.303 log log( T

y T =reduced variate , a function of sample ¿¿ is given∈table 4.4 for N → ∞ , ý n →0.577

Sn=reduced standard deviation , a function of sample ¿ ¿ is given
¿ table 4.5 for N →∞ , S n → 1.2825

These equations are used under the following procedure to estimate the flood magnitude
corresponding to a given return period based on annual flood series.
1. Assemble the discharge data and note the sample size N. Here the annual flood value is the
variate X. Find x and σn-1 for the given data.
2. Using Tables 4.4 and 4.5 determine yn and Sn appropriate to given N
3. Find yT for a given T by Eq.(4.15).
4. Find K by Eq.(4.14).
5. Determine the required xT by Eq.(4.13).

To verify whether the given data follow the assumed Gumbel's distribution, the following
procedure may be adopted. The value of x T for some return periods T<N are calculated by using
Gumbel's formula and plotted as xT Vs T on a convenient paper such as a semi-log, log-log or

Page 14 of 24
 Engineering Hydrology; Lecture note

Gumbel probability paper. The use of Gumbel probability paper results in a straight line for x T
Vs T plot. Gumbel's distribution has the property which gives T = 2.33 years for the average of
the annual series when N is very large. Thus the value of a flood with T = 2.33 years is called
the mean annual flood. In graphical plots this gives a mandatory point through which the line
showing variation of xT with T must pass. For the given data, values of return periods (plotting
positions) for various recorded values, x of the variate are obtained by the relation T = (N+1)/m
and plotted on the graph described above. A good fit of observed data with the theoretical
variation line indicates the applicability of Gumbel's distribution to the given data series. By
extrapolation of the straight-line xT Vs T, values of xT> N can be determined easily.

The Gumbel (or extreme-value) probability paper is a paper that consists of an abscissa specially
marked for various convenient values of the return period T (or corresponding reduced variate y T
in arithmetic scale). The ordinate of a Gumbel paper represents xT (flood discharge, maximum
rainfall depth, etc.), which may have either arithmetic scale or logarithmic scale.

Table 4.4: Reduced mean yn in Gumbel's extreme value distribution, N = sample size
N 0 1 2 3 4 5 6 7 8 9
0.495 0.499 0.503 0.507 0.510 0.512 0.515 0.518 0.520 0.522
10
2 6 5 0 0 8 7 1 2 0
0.523 0.525 0.526 0.528 0.529 0.530 0.532 0.533 0.534 0.535
20
6 2 8 3 6 9 0 2 3 3
0.536 0.537 0.538 0.538 0.539 0.540 0.541 0.541 0.542 0.543
30
2 1 0 8 6 2 0 8 4 0
0.543 0.544 0.544 0.545 0.545 0.546 0.546 0.547 0.547 0.548
40
6 2 8 3 8 3 8 3 7 1
0.548 0.548 0.549 0.549 0.550 0.550 0.550 0.551 0.551 0.551
50
5 9 3 7 1 4 8 1 5 8
0.552 0.552 0.552 0.553 0.553 0.553 0.553 0.554 0.554 0.554
60
1 4 7 0 3 5 8 0 3 5
0.554 0.555 0.555 0.555 0.555 0.555 0.556 0.556 0.556 0.556
70
8 0 2 5 7 9 1 3 5 7
0.556 0.557 0.557 0.557 0.557 0.557 0.558 0.558 0.558 0.558
80
9 0 2 4 6 8 0 1 3 5
0.558 0.558 0.558 0.559 0.559 0.559 0.559 0.559 0.559 0.559
90
6 7 9 1 2 3 5 6 8 9
10 0.560
0 0

Table 4.5: Reduced standard deviation Sn in Gumbel's extreme value distribution, N = sample
size
N 0 1 2 3 4 5 6 7 8 9
0.949 0.967 0.983 0.997 1.009 1.020 1.031 1.041 1.049 1.056
10
6 6 3 1 5 6 6 1 3 5
Page 15 of 24
 Engineering Hydrology; Lecture note

1.062 1.069 1.075 1.081 1.086 1.091 1.096 1.100 1.104 1.108
20
8 6 4 1 4 5 1 4 7 6
1.112 1.115 1.119 1.122 1.125 1.128 1.131 1.133 1.136 1.138
30
4 9 3 6 5 5 3 9 3 8
1.141 1.143 1.145 1.148 1.149 1.151 1.153 1.155 1.157 1.159
40
3 6 8 0 9 9 8 7 4 0
1.160 1.162 1.163 1.165 1.166 1.168 1.169 1.170 1.172 1.173
50
7 3 8 8 7 1 6 8 1 4
1.174 1.175 1.177 1.178 1.179 1.180 1.181 1.182 1.183 1.184
60
7 9 0 2 3 3 4 4 4 4
1.185 1.186 1.187 1.188 1.189 1.189 1.190 1.191 1.192 1.193
70
4 3 3 1 0 8 6 5 3 0
1.193 1.194 1.195 1.195 1.196 1.197 1.198 1.198 1.199 1.200
80
8 5 3 9 7 3 0 7 4 1
1.200 1.201 1.202 1.202 1.203 1.203 1.204 1.204 1.205 1.206
90
7 3 0 6 2 8 4 9 5 0
10 1.206
0 5

4.3.7. Confidence Limits for the fitted data

Since the value of the variate for a given return period, T determined by Gumbel's method can
have errors due to the limited sample data used; an estimate of the confidence limits of the
estimate is desirable. The confidence interval indicates the limits about the calculated value
between which the true value can be said to lie with a specific probability based on sampling
errors only. For a confidence probability c, the confidence interval of the variate x T is bound by
value x1 and x2 given by
x1/2 = xT± f (c) Se (4.16)
Where f(c) = function of the confidence probability c determined by using the table of normal
variate as
C in percent 50 68 80 90 95 99
f(C) 0.674 1.00 1.282 1.645 1.96 2.58
σ
se = probable error =b n−1 N = sample size
√N
b=√ 1+ 1.3 K +1.1 K 2 (4.17)
K= frequency factor given by Eq. (4.14) σ n−1 =standard deviation of the sample

It is seen that for a given sample and T, 80% confidence limits are twice as large as the 50%
limits and 95% limits are thrice as large as 50% limits.

In addition to the analysis of maximum extreme events, there also is a need to analyze minimum
extreme events; e.g. the occurrence of droughts. The probability distribution of Gumbel,
similarly to the Gaussian probability distribution, does not have a lower limit; meaning that
negative values of events may occur. As rainfall or river flows do have a lower limit of zero,
neither the Gumbel nor Gaussian distribution is an appropriate tool to analyze minimum values.
Page 16 of 24
 Engineering Hydrology; Lecture note

Because the logarithmic function has a lower limit of zero, it is often useful to first transform the
series to its logarithmic value before applying the theory. Appropriate tools for analyzing
minimum flows or rainfall amounts are the Log-Normal, Log-Gumbel, or Log-Pearson
distributions.

4.3.8. Log-Pearson Type III Distribution

This distribution is widely used in USA. In this distribution the variate is first transformed into
logarithmic form (base 10) and the transformed data is then analyses. If X is the variate of a
random hydrologic series, then the series of Z variates where

Z = log x (4.18)
are first obtained. For this Z series, for any recurrence interval T, equation (4.4) gives
ZT = Z + σz Kz (4.19)
Where Kz = a frequency factor which is a function of recurrence interval T and the coefficient of
skew Cs,
2

σ z=standard deviationof the Z variate sample=

√ ∑ (Z− Ź ¿ ) and
N−1
∑ ( Z− Ź¿)3 ¿
¿

C s=coefficient of skew of variate Z=N (4.20)

( N −1 ) ( N −2 ) (σ Z )3
Where Ź = mean of the values N = sample size = number of years of record

The variations of Kz = f(Cs, T) is given in Table 4.6. After finding Z T by Eq.(4.19), the
corresponding value of xT is obtained by Eq.(4.18) as x T = antilog(zT). Sometimes, the coefficient
of skew Cs, is adjusted to account for the size of the sample by using the following relation
proposed by Hazen (1930)

^ s=C s 1+8.5
C ( N ) (5.21)

Where Cˆ s = adjusted coefficient of skew. However, the standard procedure for use of Log-
Pearson Type III distribution adopted by U.S. Water Resources Council does not include this
adjustment for skew. When the skew is zero, i.e. Cs = 0, the Log-Pearson Type III distribution
reduces to Log-normal distribution. The Log-normal distribution plots as a straight line on
logarithmic probability paper.
Table 4.6 Kz=F(Cs,T) for use in Log-Pearson Type III Distribution

Page 17 of 24
 Engineering Hydrology; Lecture note

The flood-frequency analysis described above is a direct means of estimating the desired flood
based upon the available flood-flow data of the catchment. The results of the frequency analysis
depend upon the length of data. The minimum number of years of record required to obtain
satisfactory estimates depends upon the variability of data and hence on the physical and
climatologically characteristics of the basin. Generally, a minimum of 30 years of data is
considered as essential. Smaller lengths of records are also used when it is unavoidable.
However, frequency analysis should not be adopted if the length of records is less than 10 years.
Flood-frequency studies are most reliable in climates that are uniform from year to year. In such
cases a relatively short record gives a reliable picture of the frequency distribution. With
increasing lengths of flood records, it affords a viable alternative method of flood-flow
estimation in most cases.

Page 18 of 24
 Engineering Hydrology; Lecture note

A final remark of caution should be made regarding to frequency analysis. None of the frequency
distribution functions have a real physical background. The only information having physical
meaning is the measurements themselves. Extrapolation beyond the period of observation is
dangerous. It requires a good engineer to judge the value of extrapolated events of high return
periods. A good impression of the relativity of frequency analysis can be acquired through the
comparison of result obtained from different statistical methods. Generally, they differ
considerably.
Example 4.7 -Annual maximum recorded floods in a certain river, for the period 1951 to 1977 is
given below. Verify whether the Gumbel extreme-value distribution fit the recorded values.
Estimate the flood discharge with return period of (i) 100 years and (ii) 150 years by graphical
extrapolation.
Solutions: The flood discharge values are arranged in descending order and the plotting position
return period TP for each discharge is obtained as
N +1 28
T p= =
m m
Where m = order number. The discharge magnitude Q can be plotted against the corresponding
TP on a Gumbel extreme probability paper.
Year Max. flood (m3/s Year Max. flood (m3/s Year Max. flood (m3/s
1951 2947 1960 4798 1969 6599
1952 3521 1961 4290 1970 3700
1953 2399 1962 4652 1971 4175
1954 4124 1963 5050 1972 2988
1955 3496 1964 6900 1973 2709
1956 2947 1965 4366 1974 3873
1957 5060 1966 3380 1975 4593
1958 4903 1967 7826 1976 6761
1959 3757 1968 3320 1977 1971
The statistics x and σn-1 for the series are next calculated and are shown in table below.
Order № (m) Flood discharge Tp years Order № (m) Flood discharge Tp years
3
x (m /s) x (m3/s)
1 7826 28.00 15 3873 1.87
2 6900 14.00 16 3757 1.75
3 6761 9.33 17 3700 1.65
4 6599 7.00 18 3521 1.56
5 5060 5.60 19 3496 1.47
6 5050 4.67 20 3380 1.40
7 4903 4.00 21 3320 1.33
8 4798 3.50 22 2988 1.27
9 4652 3.11 23 2947 -
10 4593 2.80 24 2947 1.17
11 4366 2.55 25 2709 1.12
12 4290 2.33 26 2399 1.08
13 4175 2.15 27 1971 1.04
14 4124 2.00
N = 27 years x́=4263 m3/s σ n−1 =1432.6 m3/s

Page 19 of 24
 Engineering Hydrology; Lecture note

Using these the discharge xT for some chosen return interval is calculated by using Gumbel's
formulae [Eqs.(4.15), (4.14) and (4.13)]. From Tables 4.4 and 4.5, for N = 27, y n = 0.5332 and Sn
= 1.1004.
Choosing T = 10 years, by Eq.(4.15), yT = -[ln(ln(10/9))] = 2.25037 and K = (2.25307-
0.5332)/1.1004 = 1.56 and xT = 4263 + (1.56*1432.6) = 6499m3/s.
Similarly, values of xT are calculated for two more T values as shown below.
T years XT [obtained by equation 3.13] (m3/s)
5.0 5522
10.0 6499
20.0 7436
When these values are plotted on Gumbel probability paper, it is seen that these points lie on a
straight line according to the property of the Gumbel's extreme probability paper. Then by
extrapolation of the theoretical xT Vs T relationship, from this plot, at T = 100 years, xT =
9600m3/s and at T = 150 years, x T = 10700m3/s. [By using Eq. (3.13) to (3.15), x100 = 9558m 3/s
and x150 = 10088m3/s.]
Example 4.8-Data covering a period of 92 years for a certain river yielded the mean and
standard deviation of the annual flood series as 6437 and 2951 m3/s respectively. Using Gumbel's
method, estimate the flood discharge with a return period of 500 years. What are the (a) 95% and
(b) 80% confidence limits for this estimate?

Solution: From Table 4.4 and 4.5 for N = 92 years, yn = 0.5589, and Sn = 1.2020. Then
y500 = -[ln((ln(500/499))] = 6.21361; K500 = (6.21361 - 0.5589)/1.2020 = 4.7044,
Hence, x500 = 6437 + 4.7044*2951 = 20320m3/s.
From equation (4.17), b= √ 1+1.3 ( 4.7044 ) +1.1 ¿ ¿
5.61∗2951
Se = probable error = =1726
√92
(a) For the 95% confidence probability f(c) = 1.96 and by Eq. (3.16) x1/2 =20320xT±
(1.96*1726), which results in x1 = 23703m3/s and x2 = 16937m3/s. Thus, the estimated discharge
of 20320m3/s has a 95% probability of lying between 23700 and 16940m3/s.
(b) For 80% confidence probability, f(c) = 1.282 and by Eq. (3.16) x1/2 =20320x T±
(1.282*1726), which results in x1 = 22533m3/s and x2 =18107m3/s. Thus, the estimated
discharge 20320 m3/s has an 80% probability of lying between 22533 and 18107m3/s.
For the data of Example 4.8, the values of xT for different values of T are calculated and can be
shown plotted on a Gumbel probability paper.

Page 20 of 24
 Engineering Hydrology; Lecture note

Figure 4.2: Confidence bands for Gumbel’s distribution: Example 4.8

Example 4.9- For the annual flood series data given in Example 4.7, estimate the flood discharge
for a return period of (a) 100 years (b) 200 years and (c) 1000 years by using Log-Pearson Type
III distribution.

Solution: The variate z = log x is first calculated for all the discharges in table below. Then the
statistics z , σz and Cs are calculated from table 4.6 to obtain
[ 27∗0.00303 ]
σ z=0.1427 Cs = =0.043
[ 26∗25∗( 0.1427 )3 ]
Ź=3.6071
The flood discharge for a given T is calculated as below. Here, values of Kz for given T and Cs =
0.043 are read from Table 4.6.

Page 21 of 24
 Engineering Hydrology; Lecture note

4.4. Risk, Reliability and Safety factor

Risk and Reliability: The designer of a hydraulic structure always faces a nagging doubt about
the risk of failure of his structure. This is because the estimation of the hydrologic design values
(such as the design flood discharge and the river stage during the design flood) involve a natural
or inbuilt uncertainty and as such a hydrological risk of failure. As an example, consider a weir
with an expected life of 50 years and designed for a flood magnitude of return period T=100
years. This weir may fail if a flood magnitude greater than the design flood occurs within the life
period (50 years) of the weir.
The probability of occurrence of an event (x≥xT) at least once over a period of n successive years
is called the risk, R. Thus the risk is given by R = 1 -(probability of non-occurrence of the event
x≥xT in n years)

(4.22)

It can be seen that the return period for which a structure should be designed depends upon the
acceptable level of risk. In practice, the acceptable risk is governed by economic and policy
considerations.

Page 22 of 24
 Engineering Hydrology; Lecture note

Safety Factor: In addition to the hydrologic uncertainty, as mentioned above, a water resource
development project will have many other uncertainties. These may arise out of structural,
constructional, operational and environmental causes as well as from non-technological
considerations such as economic, sociological and political causes. As such, any water resource
development project will have a safety factor for a given hydrological parameter M as defined
below.
Safety factor (for the parameter M) = (SF) m=

(4.24)

The parameter M includes such items as flood discharge magnitude, maximum river stage,
reservoir capacity and free board. The difference (Cam - Chm) is known as Safety Margin.

Example 4.10: A bridge has an expected life of 25 years and is designed for a flood magnitude
of return period 100 years. (a) What is the risk of this hydrological design? (b) If 10% risk is
acceptable, what return period will have to be adopted?

Solution:

(a) The risk R for n = 25 years and T = 100 years is:

Hence the inbuilt risk in this design is 22.2%.

say 240 years. Hence to get 10% acceptable risk, the bridge will have to be designed for a flood
of return period T = 240 years.

Exercise: Annual flood data of a certain river covering the period 1948 to 1979 yielded for the
annual flood discharges a mean of 29,600m3/s and a standard deviation of 14,860m3/s. for a
proposed bridge on this river near the gauging site it is decided to have an acceptable risk of 10%
in its expected life of 50 years.

(a) Estimate the flood discharge by Gumbel's method for use in the design of this
structure

Page 23 of 24
 Engineering Hydrology; Lecture note

(b) If the actual flood value adopted in the design is 125,000m3/s what are the safety
factor and safety margin relating to maximum flood discharge? (Answers (a) 105,000m3/s and
(b) (SF)flood = 1.19, Safety Margin for flood magnitude = 20,000m3/s)

Page 24 of 24