LECTURE NOTES - VI

WATER

RESOURCES

Prof. Dr. Atl BULU

Istanbul Technical University

College of Civil Engineering
Civil Engineering Department
Hydraulics Division

CHAPTER 6
FLOOD FLOWS AND THEIR ESTIMATION

6.1. DEFINITION AND CAUSES OF FLOODS

A river may get flooded due to the excessive rainfall or excessive melting of snow.
Whenever the water overflows the banks of the river, the river is said to be flooded.
However, on important rivers, arbitrary representative warning and danger levels are
generally assigned at important locations, and are called flood stages.
Apart from the overflow of rivers, the floods may be caused by the failure of some dam,
with a sudden release of huge amounts of water, causing considerable damage to life and
property.
6.2. DESIGN FLOOD
To avoid damage due to destruction of floods, it is necessary that the dam spillways have
sufficient capacity and afford adequate protection against scour at the toe; flood walls and
embankment should be sufficiently high as not to be easily topped, and the reservoirs
possess requisite storages in their design capacities.
Design flood is the maximum flood that any structure can easily pass. A design flood is
distinguished from the Standard Project Flood in the sense that the latter results from the
severe combination of meteorological and hydrological factors which may be considered
applicable in a particular region.
The Probable Maximum Flood is the extreme flood physically possible in a region as a
severest combination of meteorological and hydrological factors. The standard project
flood may be considered 50% of the probable maximum flood for the region.
6.3. FLOOD ESTIMATION
For estimation of floods, following methods are in use:
1.
2.
3.
4.

Rational method,
Empirical method,
Flood frequency studies,
Unit hydrograph method.

84

A.Bulu

6.3.1. Rational Method

It is based on the principle that if rain falls over an impervious area at a uniform rate, then
over a time period called time of concentration, the rate of runoff from the surface
becomes equal to the rate of rainfall. Mathematically,
Qm = CIA

(6.1)

in which,
Qm = Peak rate of runoff in m3/sec
A = Watershed area in m2
I = Rainfall intensity in m/sec for a duration equal to time of concentration,
C = Coefficient of runoff whose value depends upon the nature of watershed and the
rainfall intensity. Its values are given in Table (6.1).
The time of concentration tc may be obtained from an empirical equation,

t c = 0.0195L0.77 S 0.385

(6.2)

in which,
tc = Time of concentration in minutes,
L = Maximum length of travel of water in meters from the upper limit of the watershed to
the point where the maximum effect of flood is considered.
S = Slope of the catchment.
Table 6.1. Coefficient of runoff C for different surfaces.
(Bayazt, 2001)
Basin Type
Forests
Lawns
Sandy soil
Low permeability soil
Urban areas
Detached
Attached
Business and industrial districts
Light
Heavy
Roads

C
0.05-0.20
0.05-0.20
0.13-0.35
0.30-0.60
0.60-0.75
0.50-0.70
0.70-0.95
0.70-0.95

Note: Lower values ar efor flat, and higher values for step areas.

85

A.Bulu

The Rational method is suitable for design of storm drains, culverts, and other structures
conveying runoff from small areas generally up to 50 km2.
EXAMPLE 6.1: A catchment area of 1.2 km2 is drained by storm runoff which over a
duration of 45 minutes results in 3 cm of rainfall. The area has a time of concentration of
30 minutes and runoff coefficient of C = 0.30. Estimate the resultant maximum rate of
flow.
Solution:

Duration of rainfall = 45 minutes > time of concentration tc = 30 minutes

A = 1.2106 m2
3 10 2
I=
= 1.11 10 5 m sec
45 60
C = 0.30
Qm = CIA
Qm = 0.30 1.11 10 5 1.2 10 6
Qm = 4 m 3 sec

EXAMPLE 6.2: A watershed has a runoff coefficient of 0.20, area 1.5 km2 with the
general slope of 0.001 and maximum length of travel of overland flow of 1.25 km.
Information on the storm of 50 years return period is given as flows:
Duration
(minutes)
Rainfall
(mm)

15

30

45

60

80

40

60

75

100

120

Estimate the peak flow to be drained by a culvert for a 50-year storm.

Solution:

Time of concentration,
t c = 0.0195 L0.77 S 0.385
t c = 0.0195 1250 0.77 0.0010.385
t c = 67.5 min

86

A.Bulu

Maximum depth of rainfall for 67.5 min. duration is obtained from the data table,
120 100
7.5
20
= 107.5mm
= 100 +

107.5 10 3
Intensity of rainfall = I =
= 2.65 10 5 m sec
67.6 60
The peak flow,
Qm = CIA
Qm = 0.20 1.5 10 6 2.65 10 5
Qm = 7.95 m 3 sec

6.3.2. Empirical Methods

The empirical equations of peak flows are based on the characteristics of catchment; are,
rainfall, etc., which are peculiar to the region it is derived alone. Their use is, therefore,
restricted to the particular regions where they are initially developed.
6.3.3. Flood Frequency Analysis

In these methods, the predictions for the future floods are made on the basis of the
available records of the past floods. These methods can be safely used to determine the
maximum flood that is expected on a river with a given frequency, if sufficient past
records are available. Secondly, the predictions will be precise only if there has occurred
no appreciable change in the regime of the river during or after the period of records. For
the success of any probability method, sufficient past records must be made available.
If a flood of a given magnitude occurs with an average frequency of 100 years, then there
1
exists
100 = 1 per cent chance for this flood to occur, every year, and such a flood
100
is generally called T=100 year return period (recurrence interval) flood. Similarly, a
flood having a return period of 20 years has a chance to be seen every year,
1
100 = 5% . A flood having a return period of 100 years is generally misunderstood
20
as a flood that will occur after 100 years. But a flood with T = 100 years can occur at any
year with 1% chance or over 1000 years, it can be seen 10 times on average with 1%
chance at any time. (Bulu, 2004)

87

A.Bulu

The probability of occurrence of such a T = 100 year return period flood equaling or
exceeding in a given year, would be once in 100 years, 1/100=0.01. This probability of
occurrence or exceedance is generally represented by p, and would be equal to,
p=

1
, where T is the return period.
T

(6.3)

If the probability of occurrence of an event (such as precipitation or flood) is p, the

probability of its non-occurrence is q, would be equal to,
q = 1 p

(6.4)

The binomial distribution can be used to find the probability of the event occurring r
times in n successive years as;
pr .n = Crn p r q nr
p r ,n =

(6.5)

n!
p r q nr
(n r )! r!

Say for example,

a) When r =2, the probability of occurrence (exceedance) of an event twice in n
successive years would be;
p2,n =

n!
p 2 q n2
(n 2)!2!

(6.6)

b) When r =1, then, the probability of occurrence of an event once in n years would be;
p1,n =

n!
p1q n1 = npq n1
(n 1)!

(6.7)

Similarly when r =0, then

p0 , n =

n!
p 0 q n 0
(n 0)!0!

n! n
q
n!
p0 , n = q n

(6.8)

Hence, the probability of an event occurring non at all in successive years would be equal
to qn, which is equal to (1-p)n. Also, the probability of an event occurring at least once in
n successive years (R) would be,

88

A.Bulu

R = P(r 1) = 1 q n = 1 (1 p )

(6.9)

This probability is called risk and hence represented by R.

EXAMPLE 6.3: A flood of a certain magnitude has a return period of 25 years,

a) What is its probability of exceedance?

b) What is the probability that this flood may occur in the next 12 years?
Solution:

a) T = 25 years,
p = exceedance probability =

p=

1
T

1
= 0.04
25

b) The probability of non-occurrence of a flood in next n successive years is given by

equation,
p0,n = q n = (1 p )

= (1 0.04)

12

= 0.9612 = 0.613
Probability of occurrence of this flood at least once and more in the next 12 years, the
risk,
R = P(r 1) = 1 0.613
R = 0.387

EXAMPLE 6.4: The 50 year-24 hour maximum rainfall at a region is found to be 16 cm.
Determine the probability of 24 hr rainfall of magnitude equal to or greater than 16 cm
occurring at that region;

a) At least once in 10 successive years,

b) Two times in 10 successive years,
c) Once in 10 successive years.

89

A.Bulu

Solution:

Frequency (return period) of rainfall, T = 50 years,

p = Exceedance probability,
p=

1
1
=
= 0.02
T 50

a) Probability of equaling or exceeding at least once in 10 successive years, the risk,

R = P(r 1)

R = 1 (1 p )

R = 1 (1 0.02 )

10

R = 1 0.9810
R = 0.187

b) Probability of occurrence twice in 10 years,

n!
p 2 q n 2
(n 2)!2!
10!
(0.02)2 (0.98)8
=
8!2!
8!9 10
2
8
=
(0.02 ) (0.98)
8!2!
9 10
2
8
=
(0.02 ) (0.98)
2
= 0.0153

p2,n =
p2,10
p2,10
p2,10
p2,10

c) Probability of exceedance once in 10 years,

p1,n = n p q n1
p1,10 = 10 0.02 (0.98)

p1,10 = 0.167

90

A.Bulu

EXAMPLE 6.5: What return period you would adopt in the design of a culvert on a
drain if you are allowed to accept only 5% risk of flooding in the 25 years of expected
life of the culvert?
Solution: 5% risk means that there is a probability of 0.05 for the design flood to occur at
least once in successive 25 years. In other words, for 95% probability, the flood should
not occur.

R = 1 (1 p )

0.05 = 1 (1 p )

25

(1 p )25 = 1 0.05 = 0.95

1 p = 0.951 25 = 0.998
p = 1 0.998 = 0.002
1
1
T= =
= 500 years
p 0.002

6.3.3.1. Theoretical Probability Distribution Methods

For determining extreme flood events, specific extreme value distributions are assumed,
and the required statistical parameters are determined from the available data, from which
the flood magnitude Q for any specific return period T can be determined.
The basic statistical equation is given by,
X = X + K

(6.10)

where,
X = Any variable (discharge),
X = Mean of the variate,
= Standard deviation of the variate,
K = Frequency factor.
The Gumbel`s extreme value distribution can be well expressed by the above equation.
6.3.3.2. Gumbel Extreme Value Distribution

The probability of occurrence of an event equal to or larger than a value X0 is given by,
P( X X 0 ) = 1 e e

(6.11)

91

A.Bulu

where y = another dimensional variable given by,

y = (x )

(6.12)

where

1.2825

(6.13)

= X 0.45005
y=
y=

1.2825

1.2825

(X X + 0.4505 )
(6.14)

(X X ) + 0.577

In actual practice, it is the value of X for a given probability p, or return period T, is

required.
p = 1 e e

e e = 1 p

(6.16)

e y = Ln(1 p )
y = Ln[ Ln(1 p )]
Since p = 1/T,

yT = Ln Ln1
T

T 1
yT = Ln Ln

yT = Ln Ln

T 1

(6.17)

For a given return period,

yT =

1.2825

(X

X ) + 0.577

(6.18)

where XT = Value of X (flood discharge) for a return period T

92

A.Bulu

( yT

0.577 )

= XT X
1.2825
y 0.577
XT = X + T

1.2825
Since the basic statistical equation,

(6.19)

X T = X + K
Where,
K=

yT 0.577
(Frequency factor for Gumbel distribution)
1.2825

(6.20)

EXAMPLE 6.6: The mean and standard deviation of annual discharges of a river are
respectively 4580 m3/sec and 1560 m3/sec. What will be the flood discharges for 100 and
50 years of return period?
Solution: For T = 100 years,
T

yT = Ln Ln

T 1
100
yT = Ln Ln

99

yT = 4.60
yT 0.577
1.2825
4.60 0.577
= 3.137
K=
1.2825
K=

The flood discharge,

X 100 = X + K100
X 100 = 4580 + 3.137 1560
X 100 = 9474 m 3 sec
For T = 50 years,
50
yT = Ln Ln = 3.902
49
3.902 0.577
K=
= 2.593
1.2825
X 50 = 4580 + 2.593 1560
X 50 = 8624 m 3 sec

93

A.Bulu

EXAMPLE 6.7: For a river, the estimated flood peaks for two return periods by the use
of Gumbel distribution, are

T = 100 years, Q100 = 485 m3/sec

T = 50 years, Q50 = 445 m3/sec
What flood discharge in this river will have a return period of 1000 years?
Solution: Since,
X T = X + K
K=

yT 0.577
1.2825

y 0.577
3
X 100 = X + 100
= 485 m sec (1)
1
.
2825

y 0.577
3
X 50 = X + 50
= 445 m sec (2 )
1
.
2825

Subtracting (2) from (1), we get,

( y100 y50 )

1.2825

= 40 m 3 sec

Since,
T

yT = Ln Ln

T 1
100
y100 = Ln Ln
= 4.60
99

50
y50 = Ln Ln = 3.90
49
40 1.2825
=
y100 y50
40 1.2825
4.60 3.90
= 73 m 3 sec

94

A.Bulu

X = X T K100
y100 0.577
73
1.2825
4.60 0.577
X = 485
73
1.2825
X = 256 m 3 sec
X = X 100

For given T = 1000 years,

1000
y1000 = Ln Ln
= 6.907
999

6.908 0.577
= 4.936
K1000 =
1.2825
X 1000 = X + K1000
X 1000 = 256 + 4.936 73
X 1000 = 616 m 3 sec
When N is smaller, frequency factor K is modified as:

K=

yT y N
sN

(6.21)

Table 6.2. Expected means yN and standard deviations sN of Gumbel reduced variates

YN = Reduced mean, depending on N, values which are given in Table (6.2) maximum
value is 0.577 at N.
SN = Reduced standard deviation, depending on N values of which are given Table (6.2)
Maximum value is 1.2825 at N.

95

A.Bulu

EXAMPLE 6.8: Annual flood discharge records on a river have been collected for 17
years starting from 1975 to 1991, and the peak values of the floods observed during each
of these seventeen years are tabulated in the Table.

Estimate the magnitude of flood having return period equal to, a) 80 years, b) 400 years.
Q(m3/sec)
3000
4400
6000
3500
2900
4800
3900
3300
6700
5400
4300
3700
4200
9000
4000
3600
5100

Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991

Solution:
17

Mean peak discharge =

(X
N

2 =
Variance; 2 =

i =1

i =1

77800
= 4576 m3 sec
17

X)

N
2
i

2 X X i NX 2
+
N
N

N
X i2
2

2
=
(X )
N

96

A.Bulu

2
i

394 106
= 23.18 106
17
N
2 = 23.18 106 (4576 )2 = 2.24 106

For small samples, standard deviation is,

N
2
N 1

17
2.24 106
16
= 1543 m3 sec

For T = 80 years,
KT =

yT yn

yT = Ln Ln

T 1
80
y80 = Ln Ln = 4.376
79

y N for 17 years sample is from (Table 6.2) as equal to 0.5177. sN for 17 years sample is
found from Table (6.2) as equal to 1.0397.

K 80 =

y80 y n
sn

K 80 =

4.376 0.5177
= 3.711
1.0397

The expected flood discharge having a return period of 80 years is,

X 80 = X + K 80
X 80 = 4576 + 3.711 1543
X 80 = 10302 m 3 sec

Similarly, for T = 400 years,

97

A.Bulu

 400
y 400 = Ln Ln
= 5.99
399
y n = 0.5177, s n = 1.0397
K 400 =

5.99 0.5177
= 5.263
1.0397

The expected peak discharge having a return period of 400 years is,
X 400 = 4576 + 5.263 1543
X 400 = 12697 m 3 sec

In order to avoid large scale calculations while calculating frequency factor K values for
Gumbel distribution, ready-mate table has been prepared, which gives the values of K for
a given sample size N and return period T in Table (6.3).
Table 6.3. Frequency Factor K for Gumbel Distribution

98

A.Bulu

EXAMPLE 6.9: Annual flood discharges of a river for 57 years of observation has a
mean value of QP = 0.848105 m3/sec, and a standard deviation = 0.35105 m3/sec.
Determine the expected peak flows for floods having return periods of a) 50 years, b) 75
years, c) 100 years. Also determine the return period of a flood having a peak value equal
to 300000 m3/sec.
Solution:

The peak discharge,

QT = QP + KT
The values of frequency factor K for different return periods and a sample size of 57
years are interpolated from Table (6.3).
Return Period
50 years
75 years
100 years

Value of K
2.862
3.211
3.459

Peak discharge for 50 years of return period,

Q50 = 0.848 105 + 2.862 0.35 105
Q50 = 185000 m3 sec

Peak discharge for 75 years of return period,

Q75 = 0.848 105 + 3.211 0.35 105
Q75 = 197000 m3 sec

Peak discharge for 100 years of return period,

Q100 = 0.848 105 + 3.459 0.35 105
Q100 = 206000 m3 sec

Return period for 300000 m3/sec flood discharge,

99

A.Bulu

X T = X + KT
KT =

XT X

300000 84800
KT =
= 6.149
35000

KT =

yT yn
sn

For 57 years, yn = 0.5511, sn = 1.1708 are taken from the Table (6.2).
yT = y n + K T s n
yT = 0.5511 + 6.149 1.1708 = 7.75
y

p = 1 e e = 1 e e
T=

7.75

= 4.31 10 4

1
1
=
= 2320 years
p 4.31 10 4

6.3.4. Hydrograph Analysis

The study of floods and droughts require the determination of the hydrograph that shows
the variation of streamflow in time. It is necessary to consider the river basin as a system
that transforms the precipitation to the runoff in order to determine the relationship
between the hyetograph of the precipitation and the hydrograph of the streamflow.
(Bayazt, 2001).
6.3.4.1 Elements of a Hydrograph

Hydrograph shows the variation of the streamflow (discharge) in a stream cross-section

in time. The discharge, usually expressed in (m3/sec), is plotted on the vertical axis. Fig.
(6.1) shows a typical storm hydrograph. The shape of the hydrograph is related to the
characteristics of both basin and the rainfall. 3 elements can be distinguished on a
hydrograph.
a) Rising Curve (Limb): The discharge increases in time along the curve AB, which is
rather steep. Its shape depends on the rainfall characteristics (its variation in time and
space, its duration), previous conditions (soil moisture, vegetation cover), and basin
characteristics. It is usually concave because the contribution of the waters from the
upper parts of the basin increases in time.
b) Peak: Usually the hydrograph has a distinct peak (point B), where the discharge
reaches a maximum. The time difference between the centroid of the hyetograph and the
peak is called lag time.

100

A.Bulu

c) Recession Curve (Limp): The discharge decreases in time along the curve BD. This
curve is much flatter than the rising curve. Its shape is related to the basin characteristics.
Its upper part signifies the discharge of the water stored in the surface channels, and the
subsurface flow. Its lower part (CD) represents the base flow, and varies little from one
storm to another storm.

Figure 6.1. Hydrograph

6.3.4.2. Separation of Direct Runoff and Base Flow

It is possible to separate approximately the part of a hydrograph that represents the direct
runoff and the part corresponding to the base flow. Direct runoff terminates at a time
equal to the concentration time after the rainfall ends. In large basins, the number of days
N after the peak of hydrograph that the direct runoff terminates can be computed
approximately as,

N = 0.9 A 2

(6.22)

where A is the basin area in (km2). N increases with the basin area because recession is
slower for larger basins.

Figure 6.2. Separation of direct runoff and base flow

101

A.Bulu

Various assumptions can be made for the line that separates the direct runoff and base
flow. The part of the hydrograph prior to the rising curve can be extended as far as the
abscissa of the peak and joint to the point C. (ABC) curve. Recession curve can be
extended as far as the abscissa of the peak and joined to the point A (ADE). Or, recession
curve can be extended as far as the point where the recession curve can be extended as far
as the point where the recession curve changes its slope and joined to the point B (ABFE
curve). It is easiest to join the points A and C by a straight line. The total volume of base
flow will not change appreciably whatever assumption is made. (Bayazt, 2001)
6.3.4.3. Unit Hydrograph Theory

A river basin can be considered as a system that transforms the precipitation to runoff.
Laws of physics
Model
structure
i(t)

Q(t)
Model
parameters

Basin characteristics

Figure 6.3. Mathematical modeling of the basin system

This equation expresses that the basin transforms the hyetograph i(t) to the hydrograph
Q(t) at the basin outlet. Various losses having been separated, the excess precipitation
will be routed and dampened by storage to produce streamflow.

Unit hydrograph model is a black box model. The phenomena in the basin are not studied
in detail but the basin is considered as a box that transforms the excess rainfall to direct
runoff. The transform function f(.) of the black box is obtained from the precipitation and
runoff records of the basin.
The difficulty in setting up a black box model for the basin is related to the non-linearity
of the system that transforms the rainfall to runoff. Therefore, its behavior cannot be
described by a linear differential equation. Let Q1(t) be the runoff due to rainfall i1(t), and
Q2(t) be the runoff due to rainfall i2(t). Then, the runoff due to rainfall i1(t)+i2(t) cannot be

102

A.Bulu

obtained simply as Q1(t)+Q2(t). Such a system where the superposition principle is not
applicable is called a nonlinear system.
However, it can be assumed that the system that transforms the rainfall excess to direct
runoff is linear, where rainfall excess (effective rainfall) is obtained by subtracting the
interception, depression storage and infiltration from the rainfall and direct runoff is
obtained by subtracting the base flow from the total streamflow. This assumption
provides a sufficient approximation in practice.

Rainfall
excess

Direct
Runoff

Drainage
basin

Runoff

Rainfall
Infiltration and
Other losses

Base flow

Figure 6.3. System considered in the unit hydrograph model

The basin model that considers the stream basin as a linear system converting the rainfall
excess to direct runoff is known as the unit hydrograph model, first proposed in 1932 by
Sherman.
The unit hydrograph is defined as follows. It is the hydrograph of the direct runoff that is
produced by rainfall excess of unit (1 cm) depth of constant intensity distributed
uniformly over the basin

Figure 6.4. Elements of unit hydrograph

Following assumptions are made in the unit hydrograph theory:

1) Rainfall excess has constant intensity.
2) Rainfall excess is distributed over the basin uniformly.

103

A.Bulu

3) Duration of the direct runoff due to rainfall excess of certain duration (base length of
the hydrograph) is the same for rainfalls of any intensity.
4) Ordinates of direct runoffs of rainfalls of a given duration are proportional to the
rainfall excess depth.
5) The hydrograph of a rainfall of constant intensity with a given duration is the same
for each storm. This implies that the basin characteristics do not vary in time, which
is not always correct because the vegetation and some other basin characteristics may
change with the season, causing the unit hydrograph to be variable.

Although these assumptions are not strictly valid, results obtained by the unit hydrograph
theory are adequate in practical applications. This theory should not be applied in basins
smaller than 5 km2 and for runoff due to snowmelt.
6.3.4.3.1. Unit Hydrograph Derivation

Hydrograph of the runoff due to a storm, distributed uniformly over the basin, with
approximately constant intensity, of short duration, and separated from the other storms,
can be used to derive the unit hydrograph, as follows:

Figure 6.5. Derivation of unit hydrograph

a) Rainfall analysis: Basin average mass curve of the rainfall is obtained from the
records of the gages in the basin, and the hyetograph is derived from this curve.
b) Base flow is separated from the observed hydrograph, to obtain the direct runoff.
c) Total volume of direct is computed measuring the area below the direct runoff
hydrograph. It is then divided by the basin area to obtain the direct runoff RD.
d) All of the excess precipitation is converted to direct runoff. Therefore, excess
precipitation depth will be equal to RD by the principle of continuity. Ordinates of the
direct runoff hydrograph are divided by RD to obtain the ordinates of the unit
hydrograph, U.

104

A.Bulu

e) It is now required to determine the duration of the rainfall excess, t0, for the unit
hydrograph obtained as above. A horizontal line is drawn on the hyetograph such
that area above it is equal to the rainfall excess depth (it can be seen that the ordinate
of this line is the infiltration index). The time interval between points the points the
line intersects the hyetograph is the duration of the rainfall excess.
Once the unit hydrograph for a certain duration t0 is known, the unit hydrograph for any
other duration can be obtained as follows:
6.3.4.3.2. Superposition Principle

Figure 6.6. Derivation of the unit hydrograph for 2t0 from

the unit hydrograph for t0

This method is used if the desired duration is a multiple of t0. As an example, the unit
hydrograph for 2t0 is obtained by translating the unit hydrograph for t0 to the right by t0,
and summing up the two hydrographs. This gives the ordinates of the direct runoff
hydrograph caused by a storm of duration 2t0, and rainfall excess depth of 2 units. Unit
hydrograph for 2 t0 is obtained by dividing the ordinates by 2.

105

A.Bulu

6.3.4.3.3. S Hydrograph method

Figure 6.7. Derivation of the unit hydrograph for t0 from the S hydrograph

S hydrograph is defined as the hydrograph of an infinite rainfall of constant intensity i,

starting at t = 0.
S hydrograph of a basin is obtained by translating the unit hydrograph for t0 by t0 many
times, and summing up the ordinates of several of them. It suffices to superpose a finite
number of unit hydrographs because S hydrograph becomes horizontal (discharge
remains constant) after a while (equal to the base with of the unit hydrograph). The
number of unit hydrographs to be translated and summed up is equal to the ratio of the
base width of the unit hydrograph to its rainfall duration. Once S hydrograph is obtained,
unit hydrograph for any duration t0 is derived by translating the S hydrograph to the right
by t0 and taking the difference of the ordinates of the two S hydrographs.

106

A.Bulu

By superposition principle, the hydrograph of a rainfall excess of duration t0, and depth
it 0 = 1 t 0 is thus obtained. Division of the ordinates by it0 gives the unit hydrograph
t0
for t0.
The discharge at the outlet of the basin with area A (km2) for a given i0 (cm/hr)
continuous rainfall intensity can be computed by,

Q = Ai0 =

A 10 6 i0
= 2.78 Ai0 m 3 sec
100 3600

(6.23)

EXAMPLE 6.10: Let us have a 6-hour unit hydrograph and construct a S-hydrograph
from it.
Time

Ordinate of Ordinates of successive unit hydrograph

Unit Hydrog.
(hour)
(m3/sec)
(m3/sec)

Ordinate S
Hydrograph
(m3/sec)

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

00
03
06
09
12
15
18
21
24
27
30
33
36
39
42

0
9
20
35
49
43
35
28
22
17
12
9
6
3
0

0
9
20
35
49
43
35
28
22
17
12
9
6

0
9
20
35
49
43
35
28
22
17
12

0
9
20
35
49
43
35
28
22

0
9
20
35
49
43
35

0
9
20
35
49

0
9
20

0
9
20
44
69
87
104
115
126
132
138
141
144
144
144

Let us have a 6-hour unit hydrograph and construct a S-hydrograph from it. Here t0=6
hours. Column 1 of the Table shows the time intervals while column 2 shows the
ordinates of a 6-hour unit hydrograph for a drainage basin having area A = 311 km2.
Columns 3, 4, 5, 6, 7, 8 show the ordinates of successive unit hydrographs, each shifted

107

A.Bulu

by the unit time interval t0 = 6 hours. Column 9 shows the horizontal summation of the
ordinates of the original unit hydrograph (Column 2) and various successive unit
hydrographs. Evidently, Column 9 gives the ordinates of the S-hydrograph. The
summation is continued up to the time interval to reach the equilibrium rate of discharge.
For the present case, the equilibrium discharge is,

Q = 2.78 Ai0
A = 311km 2
i0 =

1(cm) 1
= (cm hr )
t 0 (hr ) 6

Q = 2.78 311

1
= 144 m 3 sec
6

This flow rate is reached at 36 hours which is base time of the unit hydrograph.

108

A.Bulu

EXAMPLE 6.11: For numerical illustration, let obtain unit hydrograph of 3-hour
duration from the hydrograph of 6-hour duration obtained at the preceding example.
Solution: Example has been solved in the following Table.

Ordinate of
Time Unit Hydrograph
(hour)
(1)
0
03
06
09
12
15
18
21
24
27
30
33
36
39

(m3/sec)
(2)
0
9
20
35
49
43
35
28
22
17
12
9
6
3

Ordinate
Of S-hydrograph
(m3/sec)
(3)
0
9
20
44
69
87
104
115
126
132
138
141
144
144

109

Ordinate
Of offset
S-hydrograph
(m3/sec)
(4)
0
9
20
44
69
87
104
115
126
132
138
141
144

R=y(6/3)
R=2y

(m3/sec)
(5)
0
9
11
24
25
18
17
11
11
6
6
3
3
0

(m3/sec)
(6)
0
18
22
48
50
36
34
22
22
12
12
6
6
0

A.Bulu

Column (2) of the Table gives the ordinates of the given hydrograph of 6 hours unit
duration. Column (3) gives the ordinates of S-hydrograph, derived from the given unit
hydrograph. This S-hydrograph represent i0=1/t0 =1/6 (cm/hour) constant excess rainfall
intensity. Column (4) gives the ordinates of the offset S-curve shifted by t0 = 3 hours.
Column (5) gives the difference y of the columns (3) and (4) of the two S-curves.
Column (6) gives the ordinates of the required unit hydrograph of 3 hours duration.
6.3.4.3.4. Computation of Runoff from Rainfall of Unit Hydrograph

The hydrograph corresponding to a given hyetograph can be obtained as follows using

the unit hydrograph of the basin of duration t0:

Figure 6.8. Computation of runoff from rainfall by

unit hydrograph

a) Rainfall excess hyetograph is obtained subtracting the infiltration curve from the
hyetograph. In practice, it is usually preferred to use a constant infiltration index. If
necessary, depression storage is estimated and subtracted.
a) Rainfall excess hyetograph is divided into parts, each of duration t0. Average rainfall
intensity ii is computed for each of these.
b) The ordinate of the hydrograph at time t is computed by the superposition principle.
Q(t ) = U [t (i 1)t 0 ]it 0

(6.24)

110

A.Bulu

U [t (i 1)t 0 ] is the ordinate of the unit hydrograph at time [t (i 1)t 0 ]. It is easier to

perform the computations graphically.
c) Base flow is added to the direct runoff computed.
6.3.4.3.5. Unit Hydrograph Applications

The hydrograph of a flow caused by a given rainfall can be determined by the unit
hydrograph. A design rainfall selected. Its return period may be 10, 20, .. years
depending on the problem. The hydrograph is then obtained using the unit hydrograph. In
major problems such as the design of the spillways of large dams, probable maximum
precipitation is taken as the design rainfall.
In basins where snowmelt has a contribution to the large floods, the hydrograph due to
the snowmelt is obtained using the maximum degree-day factors and the maximum
temperatures in the snow melt season. This hydrograph is then superposed to the
hydrograph due to the rainfall and the base flow to obtain the total flow hydrograph.
EXAMPLE 6.12: The hydrograph observed at the outlet of a basin area 162 km2 is
approximately triangular. It is assumed that the rainfall intensity is constant with 50
mm/hr and the infiltration index is 10 mm/hr. Derive the unit hydrograph of the basin.
Solution:
a) Hyetograph is drawn with constant rainfall intensity I = 50 mm/hr. Duration of the
rainfall will be estimated after the derivation of unit hydrograph.
b) The direct runoff and base flow must be separated. It will be assumed here that the
base flow is constant at 100 m3/sec from the observed flows. A triangle with peak
discharge 900 m3/sec and base width 120 min is obtained for direct flow hydrograph.
30min
t(min)

50(mm/hr)
=10(mm/hr)

I
1000

Q(m3/sec)
Direct Flow
Base Flow

100

t(min)
0

30

90

60

111

120

A.Bulu

c) The volume of direct runoff will be equal to the area of the direct runoff hydrograph.

900 120 60
= 3.24 10 6 (m 3 )
2
The direct runoff depth is found by dividing the volume by the basin area.

RD =

3.24 10 6
= 0.02m = 2cm
162 10 6

c) The ordinates of the direct runoff hydrograph are divided by 2 to obtain the unit
hydrograph because the unit hydrograph corresponds to 1 cm of direct runoff. The unit
hydrograph is a triangle of base width 120 min, with the peak discharge of 450 m3/sec
occurring at 30 minutes.
d) Duration of the unit hydrograph is determined as follows:
50 10 = 40(mm hr )

Rainfall excess intensity;

Rainfall excess depth is RD = 2 cm. therefore;

t0 =

20
= 0.50(hr ) = 30(min )
40

EXAMPLE 6.13: Derive the 60-min hydrograph of the basin derived in Example 6.12.
Q(m3/sec)
1cm

300

150

t(min
0

30

60

90

120

Q(m3/sec)
1cm
300

450

150
t(min)
0

30

60

90

112

120

150

A.Bulu

Q(m3/sec)
2cm

150
t(mi
0

60

Q(m3/sec)

90

15

120

1cm

75
t(min
0

30

60

90

120

150

Solution: The 30-min hydrograph obtained in the preceding example is translated by 30

minutes along the time axis and superposed with the original unit hydrograph. The
superposed hydrograph is for a direct runoff that results from a rainfall excess of 60
minutes duration and 1+1= 2 cm depth. Its ordinates are divided by 2 to obtain the 1 cm
unit hydrograph of 60 min. Its base is 150 min and peak of 375 m3/sec occurring at 60
minutes.
EXAMPLE 6.14: 20-min unit hydrograph of a basin of size 72 km2 is triangular with
peak discharge 2 cm/hr, lag time 10 minutes. The rainfall intensity of a storm in this basin
is 18 mm/hr in the first 20 minutes, 12 mm/hr in the second 20 minutes, and 6 mm/hr in
the third 20 minutes. Infiltration index of the basin is 6 mm/hr. Base flow is constant at
10 m3/sec. Determine the hydrograph of the flow for this rainfall storm.
Solution:

20

The peak discharge occurs at + 10 = 20 minutes, according to the definition of lag

2

time. The base width of the unit hydrograph can be found as follows, considering that its
direct runoff depth is 1 cm.
1
t b 2 = 1cm t b = 1hr = 60 min
2
It is thus determined that the 20-min unit hydrograph is a triangle of base width 60
minutes, with the peak discharge occurring at 20 minutes. The rainfall intensity for this
rain is,
113

A.Bulu

I t 0 = 1cm I =

1
1
60
=
=
= 3 cm hr
t 0 20 20
60
20mi

t(min

I(cm/h

Lag time = 10

Q(cm/h
2(cm/h

1(cm/h

t(min
2

40

Rainfall excess depths are,

t(min)
U(cm/hr)
R1U
R2U
Direct runoff (cm/hr)
Direct runoff (m3/sec)

0
0
0
0
0
0

20
2
0.8
0
0.8
160

40
1
0.4
0.4
0.8
160

60
0
0
0.2
0.2
40

80
0
0
0
0
0

Base flow (m3/sec)

Total flow (m3/sec)

0
0

10
170

10
170

10
50

10
10

R = t (I )

R1 =

20
(18 6) = 4mm = 0.4cm in the first 20 minutes.
60

R2 =

20
(12 6) = 2mm = 0.2cm in the second 20 minutes
60

R3 =

20
(6 6) = 0mm in the third 20 minutes
60

114

A.Bulu

In order to obtain the direct runoff hydrograph, the ordinates of the unit hydrograph are
multiplied by R1= 0.4 cm. The n the ordinates of the unit hydrograph translated to the
right by 20 minutes are multiplied by R2= 0.2 cm. These two hydrographs are combined
to obtain direct runoff hydrograph.
The direct runoff discharges in (cm/hr) can be converted to (m3/sec) by considering tat in
this basin;
1cm hr =

0.01
72 10 6 = 200 m 3 sec
3600

Total flows are found by adding the base flow of 10 (m3/sec) to the direct discharges.

Q(m3/sec)

170
50(m3/sec)

10
t(min)
20

40

60

80

EXAMPLE 6.15: The hydrograph observed at the outlet of a basin of are 90 km2 is
approximately triangular. It is assumed that the rainfall intensity is constant at 60 mm/hr
and the infiltration index is = 15 mm/hr. Derive the unit hydrograph of the basin.
Solution:

Volume of the direct flow is the area of the triangle of the direct flow hydrograph,

V=

150 60 600
= 2.7 10 6 m 3
2

Direct runoff depth = RD =

2.7 10 6
= 0.03m = 3cm
90 10 6

Rainfall excess intensity = 60 15 = 45 mm hr

115

A.Bulu

t
60

=15(mm/hr)

Q(m3/sec)

650
Direct flow

50
Base
t(min)
150mi
n

Duration of unit hydrograph = t 0 =

3
= 0.67hour = 40 min
(6.0 1.5)

The 40-minute unit hydrograph is,

t (hour)

I
Q(m3/sec)

t0=40min

200

t (min)
150min

116

A.Bulu