CHAPTER

3
Hyetograph and
Hydrograph Analysis

3.1 INTRODUCTION
Not all the rain that falls over a catchment contribute to the runoff volume at
the catchment outlet. A part of it is lost and the remainder of the rainfall,
termed excess rainfall, is the driving force causing runoff (or flow or discharge)
at the outlet of the catchment (or basin). A good understanding of these factors
that preclude rainfall from running off is necessary in order to fully use the
capabilities of observed hydrographs. Much of the rain falling during the first
part of a storm is stored on the vegetal cover as interception and in surface
depressions (puddles) as depression storage. As the rain continues, the soil
surface gets saturated and the water starts to run off towards the channel. Thus,
the part of storm rainfall which does not appear either as infiltration or as
surface runoff is known as surface retention. Interception due to vegetal cover
is relatively unimportant in moderate-to-severe floods, though it may be
considerable over a period of time.
Almost immediately after the beginning of rainfall, the small depressions
become full and overland flow begins. Some of the flows move towards larger
depressions, while the major overland flows contribute to the streams.
In$ltration is the phenomenon of water penetration from the ground surface
to the sub-surface layers. By far, infiltration is the major contributing factor
responsible for the water losses from rainfall which does not appear as runoff
in the stream. Rainwater infiltrates through the soil surface once it exceeds the
infiltration capacity of the soil. Thus, only excess rainfall will lead to runoff
Concepts related to precipitation,effective precipitation,relevant losses, runoff
and base-flow were briefly introduced in the previous chapter. Here, a detailed
80 Engineering Hydrology

coverage of hyetograph and hydrograph is considered with the main objective

of estimatingrunoff, giving equal emphasis to single-stormas well as multiple-
storm events.

3.2 HYETOGRAPHS
The graph that plots the rainfall over the basin along a time scale is called the
hyetograph. Although a number of techniques are available for separating the
losses from a rainfall hyetograph,the injltrution indices method is the simplest
and the most popular technique used for this purpose. In hydrological
calculations involving floods, it is convenient to use a constant value of
infiltration rate (including all losses from the total rainfall hyetograph) for the
duration of the storm.
The average infiltration rate is called inJiltration index, and two types of
indices are commonly used, they are: (i) &index and (ii) W-index.These indices
account for infiltration and other losses. It basically involves a trial and error
procedure, which does not differentiate between types of losses, e.g.,
interception, depression,and infiltration. Certain background on @-indexalong
with its computations has already been provided in the preceding chapter.
However, for clarifyingdistinctionbetween @-indexand W-index, both of these
are explained below.
3.2.1 @-Index
The @-indexor constant loss rate is defined as the rate of rainfall above which
the rainfall volume equals the runoff volume (Fig. 3.1). Mathematically, the
&index can be expressed as the average rainfall above which the rainfall volume
is equal to the direct runoff volume. The @-indexis derived from the rainfall

Y
,. I

4-index Total losses

*x
Time (hr)

Fig. 3.1 A generalized rainfall hyetograph showing

$-index in relation to total rainfall
 Hyetograph and Hydrograph Analysis 81

hyetograph which also helps to calculate the direct runoff volume. The
@-valueis found by treating it as a constant infiltration capacity. If the rainfall
intensity is less than @, then the infiltration rate is equal to the rainfall intensity.
However, if the rainfall intensity is greater than @, then the difference between
rainfall and infiltration in an interval of time represents the direct runoff volume.
The amount of rainfall in excess of the index is called excess rainfall. The
@-indexthus accounts for the abstraction of total losses and enables runoff
magnitudes to be estimated for a given rainfall hyetograph. The example given
below illustrates the calculation of @-index.
Mathematically, the @-indexcan be expressed as:
P-R
@ =1 (3.la)
1,

where, P = total storm precipitation (mm or cm)

R = total direct surface runoff (mm or cm)
t,= duration of the excess rainfall, ie., the total time in which the
total intensity is greater than @ (in hours), and
@ = uniform rate of infiltration ( d h r or cm/hr)
If antecedent precipitation conditions in the catchment is such that other losses
due to interception, depression, and evaporation, etc. are considerable in the
beginning of the storm, then these losses may have to be subtracted from the
rainfall, before applying the uniform loss rate method (@-index)for computing
the excess rainfall. Since infiltration accounts for majority of the water losses
in heavy storms, the @-indexmethod must be interpreted for all practical
purposes as a constant infiltration capacity method. It is a well-known fact
that the infiltration capacity decreases with time. Therefore, the @-indexmethod
is not that realistic due to varying infiltration capacity with time. It
underestimates the infiltration for early rainfall and overestimates it for later
rainfall. In addition, one does not know how to extrapolate the @-value
determined for one storm to another storm. In spite of these drawbacks, this
simple method is often used for computing the excess rainfall hyetograph.
The derivation of @-indexis simple, and the approach is described below with
the help of two examples.

Example 3.1 A storm with 10 cm precipitation produced a direct surface

runoff of 5.8 cm in the equivalent depth unit. The time distribution of the
storm is given in Table 3.l(a). Estimate the @-indexof the storm and the excess
rainfall hyetograph.

Table 3.l(a)
Time from start (hr) 1 2 3 4 5 6 7 8
Incremental rainfall in each hour (cm) 0.4 0.9 1.5 2.3 1.8 1.6 1.0 0.5
82 Engineering Hydrology

Solution
Compute total infiltration, i.e.,
Total infiltration = Total rainfall - Direct surface runoff
= 10 - 5.8 = 4.2 cm
Assume duration of the excess rainfall (t,) is equal to the total duration
of the storm for the first trial, i.e., t, = 8 hr
Compute the trial value of @ as:

Compare the above @ value with individual blocks of rainfall. The @

value makes the rainfall of the first hour and the eighth hour ineffective
as their magnitudes are less than 0.525 c d h r . Therefore, value oft, is
modified.
Assume t, = 6 hour in the second trial and modified value of total rainfall
for 6 hour duration will be 10 (0.4 + 0.5) = 9.1 cm
~

Compute a modified value of infiltration,

i.e., infiltration = 9.1 5.8 = 3.3 cm
~

Compute second trial value of @, i.e., @ = 3.3/6= 0.55 cmlhr.

This value of @ is satisfactory as it gives t, = 6 hr
Calculatethe excess rainfall, subtractingthe uniform loss fiom each block.
The excess rainfall hyetograph is given below:
Table 3.l(b) Calculation of excess rainfall
Time from start (hr) 1 2 3 4 5 6 7 8
Excess rainfall (cm) 0 0.35 0.95 1.75 1.25 1.05 0.45 0

Note that the total excess rainfall = 5.8 cm = direct surface runoff (cm)

Example 3.2 A storm with 12.0 cm precipitation produced a direct runoff of

6.8 cm. The time distribution of the storm is given in Table 3.2(a). Estimate
@-index.

Time from start (hr) 1 2 3 4 5 6 7 8

Incremental rainfall in each hour (cm) 0.56 0.95 1.9 2.8 2.0 1.8 1.2 0.61

Solution
Total infiltration = 12.0 - 6.8 = 5.2 cm
Assume t, = duration of excess rainfall = 8 hr for the first trial.
Then, @ = 5.218 = 0.65 c&
But, this value of @ makes the rainfalls of the first hour and the eighth hour
ineffective as their magnitude is less than 0.65 c&. Hence, the value oft, is
required to be modified.
 Hyetograph and Hydrograph Analysis 83

Assume t, = 6 hr for the second trial.

In this period:
Infiltration = 12.0 - (0.56 + 061 + 6.8) = 4.03
$ = 4.03/6 = 0.67 c d h r
This value of $ is satisfactory as it gives t, = 6 hr.
Table 3.2(b) Calculation of excess rainfall
Time from start (hr) 1 2 3 4 5 6 7 8
Excessrainfall(cm) 0 0.28 1.23 2.13 1.33 1.13 0.53 0

Total excess rainfall = 6.8 cm =total runoff.

3.2.2 W-Index
In an attempt to refine the $-index, the initial losses are separated from the
total abstractions and an average value of infiltration rate (called the W-index)
is calculated as given below:
W=(P-R-I,)/t, (3.lb)
where P = total storm precipitation (cm)
R = total storm runoff (cm)
I, = initial losses (cm)
t, = duration of the excess rainfall (in hours), i.e., the total time in
which the rainfall intensity is greater than inJiltration capacity
and W = average rate of infiltration (cm/hr)

Example 3.3 Determine $-index and W-index for a watershed SW-12 with
an area of 100 km2. A rainfall with the specifications given in the following
table occurred in the watershed.

0.300

The average flow measured at the outlet of the watershed was 100 m3/s.
Assume the retention to be 10% of rainfall.
Solution
Total rainfall P = 0.150 + 1.205 + 0.75 = 2.105 cm
Totalrunoff Q=(100m3/s) x(15Ox6Os)/(1OOx 106m2)=0.009m=0.9cm
Total loss = 2.105 - 0.9 = 1.205 cm
Total time T = 150/60 = 2.5 hr
84 Engineering Hydrology

The estimated value of 4 is = 1.205/2.5 = 0.482

Consider @initial = 0.7 cm /hr
Net rainfall is obtained as follows:
Table 3.3 Calculation of excess rainfall
Net rainfall in cm
[(I) - (2)] x (duration of rainfall in hr)
I 0.3 I 0.7 I 0
I 1.205 I 0.7 I 0.505
1.500 I 0.7 0.4

This shows that 4 = 0.7 c d h r is acceptable.

AS W=(P-R-I,)It,
2.105 - 0.9 - (0.2 x 2.105)
Tndex = = 0.52 cm/hr
1.5
In the calculation, 20% of rainfall is taken as initial losses.

3.3 HYDROGRAPHS
Design of structures, e.g., determining the height of a dam, spillway design,
design of bridge culverts, sizing the capacity of outlet works, etc., require
information about extreme magnitudes of floods. It requires quantitative
information on both peak flows and time distributionof runoff or hydrographs.
Thus, a hydrograph gives the temporal variation of flow at the outlet of the
catchment or basin.
The flow may be expressed in different units, e.g., in terms of discharge
[L3/T], or depth of flow per unit time [L/T]. The hydrograph yields time
distribution of runoff from the excess rainfall occurring over a catchment for
any given event. The runoff is generally measured at a gauging site located at
the outlet of a catchment. In India, systematic gauging records are maintained
by Central Water Commission and different state agencies for most of the
major and medium basins.
The unit hydrograph theory proposed by Sherman (1932) is primarily based
on principle of linearity and time and space invariance. For the gauged
catchments, unit hydrographs can be derived from the analysis of available
rainfall-runoff records. The procedures used to derive a unit hydrograph depend
on the type of storm, whether it is a single-period storm or a multi-period
storm. Thunderstorms, generally being intense and of short duration, are treated
as single-period storms. On the other hand, frontal storms of usually longer
duration are treated as multi-period storms.
A unit hydrograph converts the rainfall (excess rainfall) to runoff (direct
surface runoff). For its derivation, the duration of the unit hydrograph is
 Hyetograph and Hydrograph Analysis 85

generally taken as the duration of excess rainfall block. However, S-curve and
superimposition methods can be utilized for deriving a unit hydrograph of
desired duration. The former approach is more general than the latter, because
the former approach changes the duration of available unit hydrograph to any
duration length.
This chapter includes description of unit hydrograph theory, its assumptions,
and limitations; derivation of unit hydrograph and factors influencing its
derivation, instantaneous unit hydrograph (IUH), S-curve, change of unit
hydrograph duration, and derivation of average unit hydrograph.
The unit hydrograph derived from an event usually differs from another
unit hydrograph derived from a different event. There are two possible
approaches available for the derivation of the representative unit hydrograph
for a watershed. The first approach averages the unit hydrographs derived for
various events using conventionalarithmetic averaging. The second approach,
however, accumulates various events as a collective unit and then derives a
representative single unit hydrograph for the watershed.
3.3.1 Definition of Unit Hydrograph
A unit hydrograph (UH) is a hydrograph of direct surface runoff from a unit
excess rainfall occurring in unit time uniformly over the catchment area. The
excess rainfall (ER) excludes losses, i.e., hydrological abstractions,infiltration
losses from total rainfall; and unit excess rainfall volume equalling 1 mm may
be considered as a standard value since rainfall is usually measured in mm.
However, in certain cases, UH is derived for 1 cm ER. The selection of unit
time depends on the duration of storm and size of the catchment area. For
example, for small catchments, periods of 1 or 2 hours can be assumed; and
for larger catchments, 3 , 4 , 6, or even 12 hours can be adopted.
A unit hydrograph can be interpreted as a multiplier that converts excess
rainfall to direct surface runoff. The direct surface runoff (DSRO) is the
streamflow hydrograph excluding base-flow contribution. Since, a unit
hydrograph depicts the time distribution of flows, its multiplying effect varies
with time. In real-world application, the unit hydrograph is applied to each
block of excess rainfall; and the resulting hydrographs from each block are
added for computing direct surface runoff hydrographs, to which base-flows
are fwther added to obtain total flood hydrographs.
3.3.2 Unit Hydrograph Theory and Assumptions
The unit hydrograph theory can be described using the following underlying
assumptions:
Constant Base Length Assumption
The base of direct surface runoff (tB)corresponding to a rainfall amount of a
given duration (7') is constant for a catchment and does not depend on the total
86 Engineering Hydrology

runoff volume. For example, if T is the duration of excess rainfall, the base
length of the DSRO for all rainfall events of this duration is the same for all
events and is equal to tB,as shown in Fig. 3.1.
Proportional Ordinate Assumption
For storms of same duration, the ordinate of DSRO at a time, i.e., Q(t) is
proportional to the total volume of excess rainfall. Thus, Ql(t)= k PE1and
Q2(t)= k PE2,where Qi(t)is the ordinate of DSRO at any given time t for the ith
storm having an excess rainfall volume of PEi,and k is proportionality constant.
Similarly, the excess rainfall of another event, PE3,over a time period T is
equal to Pm T. Therefore, at a given time t, the following proportionalityholds.

Figure 3.2 shows the surface runoff hydrographs for three different excess
rainfall volumes of 10, 20, and 30 mm for the same unit period. The ratio of
qlO:q20:q30 = 10:20:30 = 1:2:3. Thus, the ordinates of the DSRO produced
by 60 mm excess rainfall will be 0.6 times the available 100 mm unit
hydrograph.

'\ '
/' ,/' ', '\. T T T

Fig. 3.2(a) Three different unit hydrographs with

the same base but varying peaks

Concurrent Flow Assumption

The hydrograph of surface runoff that results from a particular portion of storm
rainfall is not affected by the concurrent runoff resulting from other portions
of the storm. In other words, the total hydrograph of surface runoff is the surn
of the surface runoff produced by the individualportions of the excess rainfall.
Figure 3.3 shows two hydrographs: one resulting from excess rainfall PE1and
another from the excess rainfall PE2.The DSRO (Q,) from the first excess
rainfall (PE1)is not affected by the DSRO (Q2)from the second excess rainfall
(PE2).The total DSRO is, therefore, the surn of Q, and Q2.
 Hyetograph and Hydrograph Analysis 87

Q30
/-,

/' '\

Fig. 3.2(b) Surface runoff hydrographs for three different

excess rainfall volumes of same unit period

Fig. 3.3 The total hydrograph resultingfiom consecutive storms

UniformExcess Rainfall in Time Assumption

It is assumed that the excess rainfall is the result of constant intensity of excess
rainfall, which is equal to the excess rainfall divided by time duration T. The
assumption requires that: (i) the selected storm must be intense and of short
duration, and (ii) the resulting hydrograph be of short time base and single
peaked.
UniformExcess Rainfall in Space Assumption
It is assumed that the intensity of excess rainfall is uniform over the whole
catchment. For the validity of this assumption, the catchment considered should
88 Engineering Hydrology

be small. Therefore, the unit hydrograph theory is not applicable to large

watersheds (more than 5000 sq. km). It can, however, be applied in components
by subdividing the large watersheds.
Time Invariance Assumption
It is assumed that the unit hydrograph derived for a watershed does not vary
with time. This principle holds when: (i) the physical characteristics of the
watershed do not change with time and (ii) storm pattern and its movement do
not change with time.
3.3.3 Factors Affecting Unit Hydrograph Shape
The factors affecting the shape of unit hydrograph are the rainfall distribution
over the catchment and physiographic elements of the catchment, viz., shape,
slope, vegetation, soil type, etc.
Rainf a 11 Distribution
A variation in the areal pattern of rainfall, rainfall duration, and time intensity
pattern greatly affects the shape of the hydrograph. For example, a hydrograph
resulting from a rainfall concentrated in the lower part of a basin, i.e., nearer
to the outlet of the basin (gauging site) will have a rapid rise since most of the
rainfall volume will reach the outlet in a short time. This yields a sharp peak
and a rapid recession. On the other hand, rainfall concentrated in the upper
part of the same basin will yield a slow rising and receding hydrograph having
a broad peak. Therefore, it is natural that unit hydrographs developed from
rainfall of different areal distributions will exhibit differing shapes. This is
one reason why the unit hydrograph technique is restricted to small basins.
Small basins generally have a uniform structure throughout, which reduces
the difference between the shapes of two unit hydrographs of equal excess
rainfalls having same duration but different areal extensions.
For a given amount of runoff, the time base of the unit hydrograph increases
and the peak lowers as the duration of rainfall increases. This means that two
rainfall events having different intensities yield two different shapes of
hydrographs. On large basins, changes in storm intensity should last for several
hours to produce distinguishableeffects on the hydrograph. The storage capacity
of large basins tends to eliminate the effects of short time intensities and only
major variations in the time intensity pattern is reflected in the hydrograph.
On the other hand, on small basins, short bursts of excess rainfall lasting only
for a few minutes may yield clearly defined peaks in the hydrographs. Thus,
the effects of variation in the time intensity pattern can be lessened by selecting
a short computational interval so that the variation in patterns from one
computational interval to the next is not significant.
 Hyetograph and Hydrograph Analysis 89

Physiography of the Catchment

Physical characteristics of a watershed change with time due to man-made
interfaces, change in vegetative cover, and degradation of soil surface due to
changing physiographical effects. Due to this, the shape of the derived unit
hydrograph also changes. Due to high velocity of flow, steep catchment slopes
produce runoff peaks earlier than flatter slopes; consequently, giving rise to
earlier hydrograph peaks than that of flatter slopes. Seasonal and long-term
changes in vegetation or other causes, such as fire, also change the physical
characteristicsof the watershed. It resorts to developing a regional relationship
between unit hydrograph parameters and existing basin characteristics, for
deriving the unit hydrograph in the changed environment.
3.3.4 Base-flow Separation
The main objective while deriving a unit hydrograph (hereafter, we shall refer
it as UH) is to establish a relationship between surface flow hydrographs and
the effective rainfall. In other words, the response of the basin to any input
such as rainfall (in this case) is established using the available rainfall-runoff
(discharge) data. The surface flow hydrograph is obtained from the total storm
hydrograph by separating the quick-response flow from the slow-response
runoff. There are three methods available to separate the base flow. Some of
these methods have already been introduced in the previous chapter.
Method I
This is the simplest method and is generally practiced by field engineers. In
this method, the base-flow is separated by drawing a straight line from the
beginning of the surface runoff to a point on the recession limb representing
the end of direct runoff. In Fig. 3.4, point A represents the end of direct runoff,
identified by the sharp end of direct runoff at that point; and point B marks the
end of direct runoff. An empirical equation for the time interval of N days
from the peak to the point B is given by:
N = 0.83 (3-3)
where, A = drainage area in km2, and N in days
The value of N is approximate, and position of B should be decided by
considering the number of hydrographs.
Method 1 1
In this method, base-flow existing prior to the commencement of the surface
runoff is extended until it intersects with the ordinate drawn at the peak point
C. This point is joined to point B by a straight line. The segments AC and CB
demarcate the base-flow and surface runoff. This method is most widely used
in engineering practises.
Method 111
In this method, the base-flow recession curve, after depletion of flood water, is
extended backwards till it intersects with the ordinate at the point of inflection
90 Engineering Hydrology

Fig. 3.4(a) Schematic diagram showing various salient points

used for base-flow separation

(line EF in Fig. 3.4(a)). This method is realistic in situationswhere groundwater

contributions are significant and reach the stream quickly.
All the above methods of base-flow separation are arbitrary, and the selection
of any of these methods is dependant on local practise and successfulpredictions
achieved in the past.

Example 3.4 The ordinates of discharge for a typical storm of a catchment

having an area of 450 km2 are given below:

Table 3.4(a) Ordinates of discharge

Erne Discharge Erne Discharge Erne Discharge
(hrs) (rn3/s) (rn3/s) (m3/s)
0 12 36 90 72 27
6 15 42 75 78 22
12 25 48 63 84 15
18 75 54 45 90 13
24 120 60 35 96 13
30 110 66 30 102 13

Find ordinates of the direct surface runoff hydrograph using straight-line

technique for base-flow separation.
Solution
The computational steps involved are:
(i) Plot the recession limb of the discharge hydrograph on a semilog graph
paper, keeping time on arithmetic scale and discharge values on log scale.
 Hyetograph and Hydrograph Analysis 91

(ii) Locate the recession point in the above plot.

(iii) Draw a straight line from the rising point of the hydrograph to the recession
limb of the hydrograph for base-flow separation.
(iv) Subtract the base-flow ordinates from the corresponding discharge
hydrograph ordinates to estimate direct surface runoff hydrograph
ordinates.
Table 3.4(b) shows the above computations. Here the base-flow ordinates,
given in column (3) are subtracted from the discharge hydrograph ordinates,
given in column (2). The resulting hydrograph ordinates known as direct surface
runoff hydrograph ordinates are given in column (4).
It can be seen from the above plotted hydrograph (Fig. 3.4(b)) that the storm
hydrograph has a base-flow component. Using the straight line technique for
base-flow separation, we can find the time interval ( N days) from the peak to
the point B.

01 I

0 20 40 60 80 100 120
Time (hr)

Fig. 3.4(b) Baseflow separation

Using Eq. (3.3):

N = 0.83
= 2.8165 days = 67.6 hrs
N can be calculated by judgement.
Beginning of DRH t =0
End of DRH t=90hr
Peak t=24hr
Hence, N = 90 -24 = 66 hr
For convenience, N is taken as 66 hrs. A straight line is drawn between the
points A and B which demarcates the base-flow and surface runoff
92 Engineering Hydrology

Table 3.4(b) Estimation of direct surface runoff

7ime Dischavge Hydrograph Base--ow Direct Surface
(he) (m3 4 ) (m3 4 ) Runoff (m3/s)
(1) (2. (3) (4) = (2) - (3)
0 12 12 0
6 15 12 3
12 25 12 13
18 75 12 63
24 120 12 108
30 110 12.5 97.5
36 90 12.5 77.5
42 75 12.5 62.5
48 63 12.5 50.5
54 45 12.5 32.5
60 35 12.5 22.5
66 30 12.5 17.5
72 27 13 14
78 22 13 9
84 15.0 13 2
90 13 13 0
96 13 13 0
102 13 13 0

Example 3.5 The ordinates for discharge of a typical storm fiom a catchment
having an area of 50 km2 are given below. Estimate the direct surface runoff
and the volume of excess rainfall for the given event.
Table 3.5(a) Estimation of direct surface runoff
t (hrs.) Q (rn3/s) dQ/dt t (hrs.) Q (rn3/s) dQ/dt t (hrs.) t (hrs.) dQ/dt
(1) (2) (3) (1) (2) (3) (1) (2) (3)
0 61.67 24 378 -20 49 138.88 -13.34
1 63.67 2 25 356 -22 50 134.04 -12.92
2 64.5 ~~
0.83 26 341.75 ~~~~
-14.25 51 120.6 4.84
3 88.74 24.24
~~
27 344.42 ~~~~
2.67 52 122.1 -13.44
4 103.04 ~~
14.3 28 339.1 ~~~~
-5.32 53 99.01 1.5
5 146.04 ~~
43 29 296.8 ~~~~
42.3 54 98.11 -23.09
6 189.53 ~~
43.49 30 289.8 ~~~~
-7 55 96.55 -0.9
7 224.56 ~~
35.03 31 280.03 ~~~~
-9.77 56 95.63 -1.56
8 292.8 ~~
68.24 32 280.03 ~~~~
0 57 94.4 -0.92
9 389.3
-- 96.5 33 284.87 4.84 58 94.4 -1.23
(Contd.)
 Hyetograph and Hydrograph Analysis 93

(Contd.)
--
10 419.9 ~~
30.6 34 243.9 ~~~~
40.97 59 88.75 0
11 422.8 ~~
2.9 35 216.03 ~~~~
-27.87 60 88.7 -5.65
12 480.4 ~~
57.6 36 214 ~~~~
-2.03 61 89.91 -0.05
13 617.6 ~~
137.2 37 211.95 ~~~~
-2.05 62 89.91 1.21
14 641.87 ~~
24.27 38 210 ~~~~
-1.95 63 87.56 0
15 569 ~~
-72.87 39 198.17 ~~~~
-11.83 64 86.75 -2.35
16 514 ~~
-55 40 196.51 ~~~~
-1.66 65 85.0 -3.81
17 487 ~~
-27 41 194.6 ~~~~
-1.91 66 83.8 0.05
18 462.85 ~~
-24.15 42 190.3 ~~~~
4.3 67 82.7 0
19 450 ~~
-12.85 43 187.1 ~~~~
-3.2 68 82.7 -1.1
20 425.1 ~~
-24.9 44 177.11 ~~~~
-9.99 69 82.7 0
21 422.2 ~~
-2.9 45 171.74 ~~~~
-5.37 70 81.61 0
22 412 ~~
-10.2 46 165.4 ~~~~
-6.34 71 73.83 -1.09
23 398 ~~
-14 47 165.14 ~~~~
-0.26 72 71.83 -7.78
48 152.22 -12.92
Solution
The flood hydrograph for the above data is shown in Fig. 3.5. The peak of the
hydrograph is at t = 14 hours and the peak discharge (Q,) is equal to 641.87
m3/s.
Column (3) shows the variation of slope of the hydrograph at different time
ordinates. This is used to mark the point of inflection (Pi),
which marks the
change of hydrograph slope. dQldt changes from -24.9 to -2.9 at t = 20 hours.

700
.
G' 600
"8v 500
400
3
a
300
$ 200
b 100
0 I
1

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Time (hr)

Fig. 3.5 Flood hydrograph

This is taken as the point of inflection.

Using Eq. (3.3), N = 0.82 x 50°.2= 1.81 days
Therefore, point B (refer Fig. 3.5) is at t = 66 hours. The corresponding
discharge at point B is 83.5 m3/s (marked bold in table). From the data, it can
be seen that point A falls nearly at 3 hours. The ordinate of direct surface
runoff hydrograph using the straight-line technique is given in Table 3.5(b).
94 Engineering Hydrology

t Q DSRO t Q DSRO t Q DSRO

(hvs.) (m3/s) (m3/s) (hvs.) (m3/s) (m3/s) (hvs.) (m3/s) (m3/s)
~~~~~~~~~

0 61.67 - 24 378 294.2 49 138.88 68

~~~~~~~~~

~~~~~~~~~
1 I 63.67 I -
- I 25 I 356 I 272.2 I 50 I 134.04 I 55.08
~~~~~~~~~
2 I 64.5 I -
- I 26 I 341.75 I 258 I 51 I 120.6 I 50.24
3 I 88.74 I 4.94 I 27 I 344.42 I 260.6 I 52 I 122.1 I 36.8
I 103.04 I I I I I I 99.01 I
~~~~~~~~~

~~~~~~~~~
4 19.24 28 339.1 255.3 53 38.3
~~~~~~~~~
5 146.04 62.24 29 296.8 213 54 98.11 15.21
~~~~~~~~~
6 189.53 105.7 30 289.8 206 55 96.55 14.31
~~~~~~~~~
7 224.56 140.8 31 280.03 196.2 56 95.63 12.75
~~~~~~~~~
8 292.8 209 32 280.03 196.2 57 94.4 11.83
~~~~~~~~~
9 389.3 305.5 33 284.87 201.1 58 94.4 10.6
~~~~~~~~~
10 419.9 336.1 34 243.9 160.1 59 88.75 10.6
~~~~~~~~~
11 422.8 339 35 216.03 132.2 60 88.7 4.95
~~~~~~~~~
12 480.4 396.6 36 214 130.2 61 89.91 4.9
~~~~~~~~~
13 617.6 533.8 37 211.95 128.2 62 89.91 6.11
~~~~~~~~~
14 641.87 558.1 38 210 126.2 63 87.56 6.11
~~~~~~~~~
15 569 485.2 39 198.17 114.4 64 86.75 2.95
~~~~~~~~~
16 514 430.2 40 196.51 112.7 65 85.0 1.2
~~~~~~~~~
17 487 403.2 41 194.6 110.8 66 83.8 0
~~~~~~~~~
18 462.85 379.1 42 190.3 106.5 67 82.7
~~~~~~~~~
19 450 366.2 43 187.1 103.3 68 82.7
~~~~~~~~~
20 425.1 341.3 44 177.11 93.31 69 82.7
~~~~~~~~~
21 422.2 338.4 45 171.74 87.94 70 81.61
~~~~~~~~~
22 412 328.2 46 165.4 81.6 71 73.83
~~~~~~~~~
23 398 314.2 47 165.14 81.34 72 71.83
48 152.22 74.88

The summation of the DSRO is 10669 m3/s or 10669 x 3600 m3/hr.

In terms of depth, this is equal to (10669 x 3600)/(50 x lo6) = 0.000213 m
= 0.213 111111. This is the excess rainfall depth.

3.3.5 Features of UH and Distribution Graph

The distribution graph is a non-dimensional form of the unit hydrograph,
generally available in percent form. The sum of all percentage values should
be equal to 100. The main features of a unit hydrograph or a distribution graph
are:

It shows the time distribution of the discharge hydrograph of a watershed

produced by a uniform excess rainfall of given depth.
 Hyetograph and Hydrograph Analysis 95

0 It transforms the excess rainfall to time-varying discharge at the outlet

assuming a linear process.
0 It represents the watershed characteristicsby exhibiting the integrated effect
of the surface features on the routing of the excess rainfall through the
catchment.

It is worth mentioning that the derivation of a unit hydrograph usually

excludes records of moderate floods; and the principles of unit hydrograph
can be applied for estimating the design flood, supplementing missing flood
records, and short-term flood forecasting that is based on recorded rainfall.

3.4 DERIVATION OF U H
Unit hydrograph is extensivelyused in computation of surfacerunoffs resulting
due to storm events. Similarly, for known direct runoffs and storm characteristics
including magnitude and duration, unit hydrograph can be derived. In this
section, we discuss the procedures to derive the unit hydrograph.
3.4.1 Conventional Method using Single-Peak Storm
The unit hydrograph for a gauging station is derived from a single-periodstorm
as follows:

(i) Extract hydrographs for major floods from stream-flow records using
observed gauge-discharge or rating curves.
(ii) Separate base-flow from total hydrograph to produce direct surface runoff
hydrograph. Compute the surface runoff volume Q in mm. In practice,
floods of runoff volume greater than 20 mm are analysed. If records of
higher floods are not available, lower flow values can also be used, but
with caution.
(iii) Examine the rainfall records available for the rain gauge stations in the
catchment and compute average rainfall for the storm under investigation.
(iv) Derive the hydrograph and the total average rainfall for the time period
which corresponds to the desired unit period of the unit hydrograph.
(v) Determine the loss rate subtracting the direct surface runoff from the
total rainfall to estimate the excess rainfall in each of the unit periods.
(vi) Since the unit hydrograph is a multiplying factor to convert excess rainfall
to DSRO, the procedure for its derivation is as follows:
0 Plot the surface runoff hydrograph.

0 Determine the volume of excess rainfall, PE,in the single unit period.

(PEalso equals the volume of the direct surface runoff hydrograph

QJ .
0 Relate the value of PEto the unit volume of the unit hydrograph, QvH,
as demonstrated in the following example.
96 Engineering Hydrology

It is important to understand precisely the term “unit volume of the unit

hydrograph”. Normally, many textbooks use the word “unit hydrograph” to
imply that it is due to 1 cm effective rainfall. This may not always hold good.
Unit hydrograph may also imply that it is due to 1 rnm effective rainfall, or for
that matter, any other unit, say 1 feet, 1 inch, etc. In this book, we use unit
hydrograph in the sense that it is due to 1 rnm effective rainfall.
If 1 mrn effective rainfall occurs uniformly over a catchment for a unit time,
the volume of UH will be equal to (1 mm x area of catchment). If we divide
this volume by the catchment area A , this will be just equal to 1 rnm (the unit
used for effective or excess rainfall). Thus, the term “unit volume of unit
hydrograph” implies that the volume under the UH is equal to one unit used
for effective rainfall of one unit.

Example 3.6 For a given PE of 22.5 111111, obtain proportionality factor ( F )

to derive a unit hydrograph for 10 mm unit volume.
Solution
Give unit volume of unit hydrograph Q U H = 10 111111,the proportionality factor
F can be computed as:

Ordinates of the UH can be obtained by dividing the direct surface runoff

hydrographs by F. Please note that unit volume of such UH will be equal
10 mm.

Example 3.7 A thunderstorm of 6-hr duration with the excess rainfall of

154 mm produces the following surface runoff hydrograph:

Table 3.6(a) Data for Example 3.7

Date June 15 June 16 June 17
Time
(Hrs) 0600 1200 1800 M.N. 0600 1200 1800 M.N. 0600 1200 1800 M.N.
DSRO
(m3/s) 10 500 1600 3500 5200 3100 1500 650 250 0 0 0

Find the ordinates of a 6-hr UH of unit volume equal to 100 111111.

Solution
Since PE = 154 and QvH = 100 111111,the conversion factor F = 154l100
= 1.54.
The ordinates of unit hydrograph are computed dividing the direct surface
runoff hydrograph ordinates by the conversion factor. The calculation of 6-hr
unit hydrograph for the above storm is shown in Table 3.6(b).
 Hyetograph and Hydrograph Analysis 97

Table 3.6(b) Computation of 6-hr unit hydrograph

Date runoff 6-hv 100 mm unit hydvogvaph
(m3/s)

June 15
1800 1600 1039
2400 3500 2272
I 0600 I 5200 3377
1200 3100 2013
June 16
1500 974
I 2400 I 650 422
0600 250 162
June 17 1200 0 0
1800 0 0

3.4.2 UH Derivation from a Multi-period Storm

When a storm is of short duration and of fairly uniform intensity, the unit
hydrograph can be derived by simply applying the single period storm
technique. However, for a storm of long duration or a shorter one with variable
intensity, the storm should be treated as consisting of a series of storms and the
derivation of the unit hydrograph gets slightly more complicated. There are a
number of methods to derive unit hydrographs from multi-period storms, but
all are based on unit hydrograph theory as follows.
Let Pei= intensity of excess rainfall (mm/hr),PEi= volume of excess rainfall
(mm), Ui= ordinates of unit hydrograph (m3/sec), Qi= ordinates of direct
surface runoff hydrograph (m3/sec),and i = an integer value for an ordinate.
The underlying procedure of calculating the direct surface runoff hydrograph
from a multi-period storm is described below.
If Pei = rainfall rate ( d h r ) in period T, then PEi = volume of rainfall
= Peix T. For example, if Pei= 2 mm/hr and T = 2 hr, then PEi= 2 x 2 = 4 mm.
To determine the direct surface runoff hydrograph fiom 4 mm of rainfall that
occurred in 2 hrs, the 10 mm unit hydrograph of 2-hr duration having 7 unit
periods (time base = 7 x 2 = 14 hrs) is used. Thus,

(i) the base length of the direct surface runoff hydrograph for rainfalls in
excess of 2-hr duration is 14 hr
(ii) for a 10 mm unit hydrograph, the average runoff in the various periods
for the first rainfall in excess of 4 mm is computed using the proportionality
principle, as given below:
98 Engineering Hydrology

Q, = 0.4x U, Q3 = 0.4x U3 Qs = 0.4x Us Q7 = 0.4x U7

Q2 = 0.4X U2 Q4 = 0.4X U4 Q6 = 0.4X U6
(iii) If PE2= 4 d h r (2nd period of rainfall) or PE2 = 4 X 2 = 8 111111, the
ordinates for the second period of rainfall will be as follows:
Q, = 0.8 x U,, Q2 = 0.8 x U2, and so on.

It is to be noted that the excess rainfall in the second period is lagged by

T = 2 hr. Thus, Q , of the second period of rainfall is added to the second
period of runoff for first hydrograph ordinate. This principle continues for the
third and subsequent rainfall values. The equations for the combined surface
runoff can, therefore, be generalised as follows:
Qi =PEIui
Q2 = PEI u
2 + pE2 ui

Q3 3 +PE2 u
=PEIu 2 +Pm ui
Q4 = PEI u
4 + pE2 u3 +'E3 u2

Q5 = PEI u5 + pE2 u4 +'E3 u3

Qs = p ~ l u6 + pE2 us +'E3 u4

Q7 = pE1 u7 + pE2 u6 +'E3

Q,=O+PBU~+PE~U~
Q, =O+ O +PE3 U7
All the following procedures used for deriving unit hydrographs from multi-
period storms use the principles of proportionality and superposition.
Method of Single Division
This is basically a trial and error approach. The following procedure involves
solving each of the above equations relating to the unit hydrograph ordinates,
as shown in Fig. 3.6. For the first rainfall, both Q, and PEl are known; Q , from
the surface runoff hydrograph and PE1from the hyetograph excess rainfall.

Thus, Q, =PE1 X u, or 1=
u ~
Qi
pE 1

For the second rainfall, Q2 is known from the hydrograph; and PE1, U,, and
PB are already known.

Thus, the unit hydrograph ordinates (q.

) are computed.
 Hyetograph and Hydrograph Analysis 99

Fig. 3.6 Unit hydrograph derivation from a multiperiod storm

Since there are more equations than unknown values, the equations relating
Q, and Q, to u 6 and U,, respectively, may not be in balance. Therefore, it is
necessary to apply an ‘averaging’ procedure. It involves calculation of U,,
U2,..., U, from Q,, Q2,...,Q7 (forward run); and the calculation of U,, u 6 ,...,U1
from Q,, Q,, ...,Q, (backward run) as illustrated in Fig. 3.6 through a typical
example. The unit hydrograph ordinates are then derived by ‘averaging’ the
values of U,, U,, ...,U, computed from the two runs. The derived unit hydrograph
is then checked for producing a hydrograph that compares well with the
recorded hydrograph. This test is recommended in all unit hydrograph
derivations.
The direct surface runoff hydrograph computed using the above average
unit hydrograph does not match satisfactorily with the observed direct surface
runoff hydrograph. This deviation may be attributed to small errors in U,, U,,
u6, and U,, which are the leading and trailing ordinates, respectively; and
these values propagate in calculations of other ordinates. Computations rely
more on U , and U, in forward run and U, and U, in backward run. Furthermore,
it is possible that (i) the records of rainfall and runoff may not be representative
for the actual situation, (ii) the unit hydrograph theory may not be applicable
to the storm, and (iii) storms may not be uniformly distributed over the
catchment. These circumstances create difficulties in derivation of unit
hydrograph.
100 Engineering Hydrology

Least Squares Method

As is evident in single division method, there are more equationsthan unknown
values and the solution adopted depends on the selected set of equations. The
least squares method is a statistical curve fitting procedure which yields a ‘line
of best fit’.
The basic principle of the least squares method lies in the fact that the line
of best fit produces the best reproduction of the hydrograph, computed from
the application of the unit hydrograph to the excess rainfall, as shown in
Fig. 3.7. The developed unit hydrograph yields minimum deviation of the
ordinates of the reproduced hydrograph from the ordinates of the actual
hydrograph.

Fig. 3.7 Goodness-of-fit of observed and computed

hydrographs

As this method involves a number of tedious computations to minimize the

errors, it is not a practical proposition to carry out computations manually. The
readily available computer packages can be employed. This method is suitable
subject to the following criteria:

0 the hydrograph represents the basin and storm characteristics,

0 the rainfall and hydrograph ordinates are accurate, and
0 the unit hydrograph theory is applicable to the catchment.

If these criteria are not met, the ‘best fit’ unit hydrograph can assume any
shape and may yield even negative ordinates. Therefore, it is necessary to
stress that the derived unit hydrograph is more mathematical rather than just a
representativeof the characteristics of runoff peculiar to a particular catchment.
Kuchment (1967) suggested a procedure for deriving a smooth unit hydrograph
using a regularization factor determined by trial and error.
 Hyetograph and Hydrograph Analysis 101

Collins9Method
This is a trial and error approach. Assume that the unit hydrograph has four
ordinates, viz., U,, U2,U3, and U, correspondingto t values equal to T, 2T, 3T
and 4T; and there are three periods of time T of excess rainfall, viz., PE1, PE2,
and PE3with Pn representing the maximum rainfall volume. The surface runoff
ordinates are determined as described earlier in Section 3.2. It is to note that in
the previous example, the unit hydrograph had 7 ordinates; whereas in this
example, there are only 4 ordinates. It is not necessary that the unit hydrograph
ordinates are at the same time interval as the excess rainfall. The following
equations hold good for computations of Q, to Q4:
Qi = P E IUi
Q2 = PEI U2 + P E UI
~
Q3 = P E IU3 + P E U2
~ + P n UI
Q4 = P E IU4 + P E U
~3 + p m U2
Computationsfor other Q values can be done in a similar manner. For a gauged
catchment, a unit hydrograph can be derived using Collins’ method, according
to the following steps:

(i) Derive the flood hydrographs for major flood events from stream flow
records using rating curves.
(ii) Estimate the direct surface runoff by separating the base-flow, using a
suitable approach, from the total hydrograph.
(iii) Estimate the average rainfall hyetograph.
(iv) Compute the effective rainfall hydrograph by separating the losses from
the total rainfall hyetograph.
(v) Ignore the terms in the equation that contain the maximum rainfall [PE2,
termed PE
(vi) Assume a first trial unit hydrograph, i.e., a set of U-values, which appears
reasonable. To this end, use either constant value for unit hydrograph
ordinates or single division procedure for the first estimate of unit
hydrograph. The number of unit hydrograph ordinates can be determined
using the following relation:
Periods in unit hydrograph base = (Periods in direct surface runoff
base - number of periods of excess rainfall blocks) + 1
Arrange this set of values in Col. 3 of Table 3.7.
(vii) Apply the initial trial unit hydrograph to all periods of excess rainfall
except PE and determine the resulting hydrograph, as shown in
Columns 2 through 4 of Table 3.7.
(viii) Deduct the sum OfPEiUi from the actual direct surface runoff hydrograph
ordinates, and estimate the runoff from the maximum rainfall P!
as shown in Column 6 of Table 3.7. Retain only those values having the
term PE as the resultant DSRO is the response of PE
 Table 3.7 Collins' method
-
Time Excess Trial unit hydrograph (m3/s) Actual U Adjusted
-
Period rainfall UI u
2 u
3 u
4 DSRO (m3/s) U
(mm) (m3/s)

(1) (3) (4) (5)

0
1 pH pEIuI I I I
Qo
Qi
I
2 p H u2 Q2

3 pE2 Q3
4 Q4
5 Q5
6 Q6
I
 Hyetograph and Hydrograph Analysis 103

Divide the ordinates computed at step (viii) by PE and obtain another

estimate of the unit hydrograph ordinates, as shown in Column 7.
Compare the derived unit hydrograph ordinates (ui,)with the original
trial ordinates (UJ.
If the comparison is not satisfactory, take an average of the unit
hydrograph ordinates of the first trial for computing unit hydrograph
values for the second trial, as given below:
- MU, + NU,,
U=
M+N
where, A4 = total excess rainfall (= PE1+ Pn + ...) except PE(MAX), and
N = PE(MAX). Equation (3.4) ensures that the trial ordinates of the unit
hydrograph correspond to the unit runoff volume. Note that the unit runoff
volume can be any multiple of 1 mm. The conversion factor (6 is applied
as shown in Column 9 of Table 3.7. It is computed as follows:

where, X q is the sum of trial unit hydrograph ordinates, and X q Cis the
sum computed unit hydrograph ordinates falling between the limits of
PE influence.
Repeat step (vii) to (xi) until the calculated unit hydrograph agrees with
the trial unit hydrograph. The application is illustrated with the help of
an example.

Example 3.8 The excess rainfall and surface runoff ordinates for a storm of
a typical catchment of 1,70,000ha are given below. Derive a 6-hr 100 mm unit
hydrograph using Collins’ method.

Table 3.8 Data for Example 3.8

Erne Period PEi 4i
(rn3/s/6hr)
1 I 40 I 250 I

i6
7
~

I
I
1;: ~

I
I
2050
1050
4350
4150
2300
1070
,

I
I
8 I I 450 I
9 120
104 Engineering Hydrology

Solution
Tables 3.9 and 3.10 present computations for two trials of unit hydrograph
derivation using Collins’ method.
In Table 3.9, number of direct surface runoff periods = 9 (6-hr periods) and
rainfall periods = 3 (6-hr periods).
Therefore, the total period of UH = (9 - 3) + 1 = 7 (6-hr periods).
The first trial assumes a uniform unit hydrograph, determined as follows:
Calculate total surface runoff (from volume under hydrograph) or by
planimetering area. It is equal to 340 x lo6 m3.
Since the catchment area = 170,000 ha = 1700 x lo6 m2, the unit runoff of
1 mm corresponds to = 1700 x lo6 x lop3(= 1.7 x lo6) m3.
Therefore, unit runoff volume of 100 mm unit hydrograph will be equal to
170 x lo6 m3, and 340 x lo6 m3 of direct surface runoff (DSRO) volume will
correspond to 200 mm depth, which is equal to the sum of excess rainfall
depths (40 mm + 60 mm + 100 mm). Thus, the DSRO volume and the volume
of excess rainfall are perfectly matched.
Since a unit hydrograph of 100 mm is to be developed, it is necessary that
each value of excess rainfall be divided by 100. Thus, for PE1 = 40 mm, the
result is 0.4; for PE2= 100 mm, it is equal to 1.O; and for PE3 = 60 mm, it is 0.6.
It is to note that if the objective had been to develop a unit hydrograph of
1 mm, all values of PEwould have been unchanged.
For the first trial, assume a uniform rectangular-shaped unit hydrograph.
The magnitude of ordinates can be computed as Ui= 170 x 106/(42x 3600)
= 1125 m3/sec.Having computed the trial values of the 100mm unit hydrograph
ordinates, steps (vii) through (xii) can be followed.
The computations are shown in Table 3.9. In this table, Column 1 presents
the time steps and Column 2 presents the excess rainfall values which are
divided by 100 and the resulting values are shown in Column 3. Column 4
presents the trial unit hydrograph ordinates and computations of direct surface
runoff hydrograph from PE1 and PE2values of Column 3. The respective sum
of the ordinates which is the response of PE1 and PE2is shown in Column 5.
Column 6 presents the actual observed direct surface runoff and the difference
of Column 6 and Column 5. Column 7 lists the values of Column 6 divided by
PE which yields the ordinates of the 100 rnm unit hydrograph derived
from only PE
It is to note that the first negative value (Row 1 of Col. 7) is excluded
because the effect O f P E ( m )starts from second time-step onwards only. Since
these ordinates deviate far from the trial values, rainfall weighted ordinate
values are computed and presented in Column 8.
Since M = PE1 + PE2 = 0.4 + 0.6 = 1.0 and N = 1.00, the division of values
of Column 7 by 2 gives the values shown in Column 8. Here, it is necessary to
check the sum of trial unit hydrograph ordinates, to verify whether or not it
 Table 3.9 Example of UH derivation using Collins’ method (Trial 1)
Time Excess Excess Actual Adjusted
-
Period rainfall rainfall DSRO U
(mm) for (m3/s)
100 mm
UH
(2) (3) (6) (9) = 8/F
40 0.4 250 ~

100 1.oo 1050 82 1

60 0.60 2050 1025 975
4 4350 3225 2175 2069
5 I I I I 1125 4150 I 3025 I 2075 1974
6 I I I 1125 2300 I 1175 I 1150 1094
7 I 450 I 1125 1070 I -55 I 535 509
8 450 I -225 I 450 425
9 I I I 675 I 675 120 I ~ I
Total 200 I I 8273 7870
I I 7870
I F = 1.05
 Table 3.10 Example of unit hydrograph derivation using Collins’ method (Trial 2)
-
Eme Excess Excess Trial 2 unit hydrograph (m3/s) Actual U Adjusted
~

Period rainfall rainfall DSRO (in3/,) U

(mm) for (m3/s)
100 mm
UH
2 3 4 5 6 7 8 9 = 8/F

40 0.4 328.4 328.4 250 ~ ~ ~

100 1.OO ~

390 390.0 1050 660 74 1 729

60 0.60 429.6 ~ 827.6 1320.2 2050 729.8 852 838
4 I I I 11374.6 I 4350 I 2975.4 2522 2480
5 I I I 1241.41 ~ I 437.6 I I 11679.0 I 4150 I 2471.0 2223 2186
6 I I I 11388.0 I 2300 I 912.0 1003 986
7 I I I I I 656.4 I ~ I 171.2 I 827.6 I 1070 I 242.4 376 509
8 I I I I I I 305.4 I ~ I 305.4 I 450 I 144.6 286 425
9 1256.8 I 256.8 I 120 I
Total 200 8003 7870

F = 1.02
 Hyetograph and Hydrograph Analysis 107

equals that of Column 8 values. The sum of trial ordinates is 7875 whereas the
surn of values of Column 8 is 8273. Therefore, the conversion factor F is
computed as 1.05, and the ordinates of Column 8 are divided by 1.05, and the
resulting values are shown in Column 9. These values are the trial values for
the second trial, as shown in Table 3.10. Figure 3.8 represents the unit
hydrograph, derived using Collins’ method.

Time (hr)

Fig. 3.8 UHdenved using Collins’method

3.4.3 Selection of Unit Time Period of UH

The time period (7‘)should be determined using the following criteria:

(i) It should be short enough to define the hyetograph and the hydrograph
reasonably.
(ii) It should be a simple fraction of 24 hour. Use of time periods that are not
a whole fraction of 24 hours may not be used, for example, 5-hr,7-hr, etc.
(iii) Conventionally, a unit period of approximately 1/3 to 1/4 of the period of
rise of the hydrograph is used.
3.4.4 Instantaneous UH
Given the excess rainfall amount, if the period of a unit hydrograph is reduced,
peak flow magnitude of direct surface runoff hydrograph increases and the
base length decreases. If the period of excess rainfall reduces to zero and the
excess rainfall depth remains the same, the intensity of the excess rainfall
reaches infinity. This corresponds to the instantaneous application of a sheet
of water over the entire catchment area. This water drains off the basin through
gravity, as shown in the instantaneousunit hydrograph (IUH) in Fig. 3.9. Thus,
an IUH is purely a theoretical concept and represents the unit hydrograph
derived fi-om the unit excess rainfall volume that occurred instantaneously. An
IUH represents a characteristic of a catchment and is not affected by time
duration. It is a useful tool in regional unit hydrograph studies.
108 Engineering Hydrology

Fig. 3.9 Derivation of Instantaneous unit hydrograph

IUH is a single-peaked hydrograph with a finite base width and its properties
are as follows.

1. 0 I u(t) I a positive value, for t > 0

2. u(t) =o, for t I 0
3 . u(t) +o, ast+-
ca

4. I u(t) dt = unit depth over catchment

0

3.5 S-CURVE
An S-curve of T-hr unit hydrograph is the graph that results from 1/Tintensity
of excess rainfall which occurs for an infinite period, as shown in Fig. 3.10.
The S-curve may be derived as follows:

(i)Plot the unit hydrograph of unit volume at time t = 0.

(ii)Plot the unit hydrograph successively lagged by T-hr.
(iii)Add the ordinates of each unit hydrograph, taking the time lag into account.
(iv) At time tb (base length of the first unit hydrograph), the above sum of
ordinates reaches a constant value.
(v) Plot the sum of the ordinates as ordinate and time as abscissa. It is an
S-curve. Thus, the S-curve for the unit period T-hr is determined.
Sometimes, the S-curve fluctuates in the region approaching a constant
value. It is a problematic feature of the S-curve derivation. To avoid it, a
curve of best fit through the points is drawn, for using the values derived
fi-om the fitted curve.
(vi) The maximum ordinate of the S-curve corresponds to the equilibrium
discharge, and it is defined as:
 Hyetograph and Hydrograph Analysis 109

Unit excess rainfall 1 cm in T hr

Fig. 3.10 S-curve

where, A is the catchment area (sq. km), QUHis the unit volume (mm), and
T is the time duration of the unit hydrograph (hr). The derivation of
S-curve from a given unit hydrograph is explained using the following
example.

Example 3.9 For a typical catchment, the ordinates of 6-hr UH of 10 mm

volume are given below. If the equilibrium discharge is 16250 m3/s, compute
the S-curve ordinates.
Table 3.11 Data for Example 3.9
Date Eme interval 6hl: 10mm UH
(m3/s)
June 15 0600 0 0
0900 3 200
1200 6 500
1500 9 1000
1800 12 1600
2100 15 2400
2400 18 3500
(Contd.)
110 Engineering Hydrology

(Contd.)
June 16 0300 21 4200
0600 24 5200
0900 27 4400
1200 30 3100
1500 33 2300
1800 36 1500
2100 39 1000
2400 42 650
0300 45 400
0600 48 250
0900 51 150
1200 54 0

Solution
The steps followed with reference to Table 3.7 are:
(i) Enter the ordinates of 6-hr UH in Column 4.
(ii) Lag the 6-hr UH by 6-hr, 12-hr, etc. till 54 hours, which is the time base
of the UH; and enter these values in Columns 5 through 13.
(iii) Derive S-curve ordinates summing the values of Columns 5 through 13
at each time interval, as shown in Column 14.
(iv) Plot the S-curve obtained at step (iii).
(v) If the plotted S-curve shows hunting effect at the upper end, plot a best
fit curve through these points. Figure 3.11 shows the smoothened S-curve.
(vi) Read the discretized values of the S-curve, obtained at step (v), and enter
in the table, as in Column 15. This is the S-curve of intensity 10/6 &.
(vii) Multiply the ordinates of Column 15 by 6 to get the S-curve of intensity
10 d h r , shown in Column 16.

S-curve

Time (hr)

Fig. 3.1 1 Smoothed S-curve

 Hyetograph and Hydrograph Analysis 111

3.5.1 Relation between S-Curve and UH

There exists a simple relationship between the S-curve of a watershed and its
unit hydrograph, as shown in Fig. 3.12, where S, is the time-varying S-curve
for a continuous rain of i c& and S,, is the same curve shifted by Thr to the
right. The difference in the ordinates of the two S-curves (Sand S,,) represent
the ordinates of a hydrograph resulting due to excess rainfall depth of iT cm.
Therefore, T-hr and d-mm unit hydrograph ordinates (U,) can be derived as:
d
u,= -(st
iT
- st-,) (3.7)
If unit depth is equal to 1 mm and intensity (i) of excess rainfall is 1 mmh,
then
st - st-,
u,=
T

Fig. 3.12 S-curve and UH

3.5.2 Relation between S-curve and IUH

Since Ud(T, t) = [d(S,- &,)/(iT)]; as T approaches zero, Ul(T,t) approaches
Ul(O, t), and Ud(T,t) approaches (dS,/dt) (d/i). Here, Ul(O, t) and Ud(T,t)
represent 1-111111and d-mm T-hr unit hydrographs, respectively. Thus,

Or,alternatively,

(3.10)
112 Engineering Hydrology

Fig. 3.13 S-curve and IUH

If i = 1 and d = 1, the above equation reduces to

t
St= ju(t)dt (3.11)
0

3.5.3 Relation between UH and IUH

The relationship between rCJH [u(t)] and T-hr UH [ U(T, t)], both of the same
unit depth, can be derived using the two S-curves, as shown in Fig. 3.12.
Note: here, the subscript (i) of U is dropped for convenience.

(3.12)

t t-T
= j u ( t ) d t- u(t)dt
0 0

= tru(t)dt (3.13)
0
which equals the area between the two S-curves. The T-hr unit hydrograph
with a unit depth is given by:
t
j u(t>dt
U(T,t) = t-T (3.14)
T
Thus, a T-hr UH can be derived from an IUH. If the IUH is available, the
ordinate of the T-hr unit hydrograph at the end of the time unit T is equal to the
average ordinate of the instantaneous unit hydrograph over the T-hr period.
 Hyetograph and Hydrograph Analysis 113

3.6 CHANGE OF UNIT PERIOD OF UH

Having derived a unit hydrograph for a particular unit period (say, 6 hrs), one
may want to change the time period and derive a new hydrograph. The following
two methods are used for this purpose:

1. Superimposition method, and

2. S-curve method.
3.6.1 Superimposition Method
This method is suitable only when the new duration of the UH is an integer
multiple of the given unit duration. For example, the given duration is 6 hr,
and the hydrograph is to be converted to a 12-hr duration. In general, a unit
hydrograph of 2T-duration can be derived from a UH of T-duration as follows:

(9 Add the ordinates of the T-hr UH to the ordinates of an identical UH

lagged by T-hr.
(ii) Divide the ordinates of the resulting hydrograph of step (i) by 2 to obtain
a UH for unit duration of 2T. Figure 3.14 shows the conversion of a UH of
T-duration to 2T-duration.

Similarly, the UH of nT-duration, where n is an integer, can be derived by

successive lagging of the T-duration unit hydrograph n times and then dividing
the resulting hydrograph by n, as illustrated in the following example.

Example 3.10 The ordinates of 6-hr UH are given below. Derive a 12-hr
UH of the same unit volume as of the 6-hr UH.

Table 3.12 Data for Example 3.10

Erne 6-hr UH Erne 6-hr UH
(hr) (rn3/s) (rn3/s)
0 0 30 3100
3 200 33 2300
6 500 36 1500
9 1000 39 1000
12 1600 42 650
15 2400 45 400
18 3500 48 250
21 4200 51 150
24 5200 54 0
27 4400
114 Engineering Hydrology

b lcm lcm

m,
\
\
\
\
\
\
\; ,T-hr UH

\
\

0 T 2T 3T 4T 5T 6T 7T 8T 9T
Time (hr)

A 2T-hrUH
Ordinate = (A+B)/2

Time (hr)

Fig. 3.14 Change of unit period of UH

Solution
With reference to Table 3.13, the computational steps are:

(i) Enter the ordinates of 6-hr UH in Column 2.

(ii) Lag the 6-hr UH by 6-hr and enter these values in Column 3.
(iii) Add the respective ordinates of Columns 2 and 3. The resulting hydrograph
ordinates are shown in Column 4.
 Hyetograph and Hydrograph Analysis 115

(iv) Divide the ordinates of the hydrograph derived from step (iii) by 2 to
compute UH of 12-hr duration. The ordinates of this UH are given in
Column 5.

Table 3.13 Derivation of 12-hr UH from 6-hr UH

Time 6-hr UH UH Superimposed 12-hr UH
(hrs) ordinates ordinates lagged by hydrograph (m3/s)
(m3/s) (m3/s) 6-hr
(m3/s)
0 1 0 1 - I 0 1 0
3 I 200 I -
I 200 I 100
6 I 500 I 0 I 500 I 250
9 I 1000 I 200 I 1200 I 600
12 I 1600 I 500 I 2100 I 1050
15 I 2400 I 1000 I 3400 I 1700
18 I 3500 I 1600 I 5100 I 2550
21 I 4200 I 2400 I 6600 I 3300
24 I 5200 I 3500 I 8700 I 4350
27 I 4400 I 4200 I 8600 I 4300
30 I 3100 I 5200 I 8300 I 4150
33 2300 4400 6700 3350
36 1500 3100 4600 2300
39 1000 2300 3300 1650
42 I 650 I 1500 I 2150 I 1075
45 I 400 I 1000 I 1400 I 700

250 250
150 150
60

3.6.2 S-Curve method

This is a more general method than the method of superimposition.After having
derived the S-curve of unit (or some other) intensity (for example, 1 d h r )
from T-duration unit hydrograph, the unit hydrograph of T-duration can be
obtained as follows:

(i) Shift the S-curve by T hrs to obtain another S-curve, as shown in

Fig. 3.15.
(ii) Subtract another S-curve ordinate from the original S-curve.
(iii) The difference between the two S-curves represent the unit hydrograph
for time T with a unit volume equal to iT = T rnrn (as i = 1 d h r ) .
116 Engineering Hydrology

Fig. 3.15 Changing UH duration using S-curve method

(iv) To derive a 1 mm UH of duration T divide the difference of step (iii)

by T.

lhe S-curve lhe S-curve

(hr) hydrograph (hr) hydrograph
ordinates ordinates
(rn3/s) (rn3/s)
0 I 0 I 30 I 82800
3 I 1200 I 33 I 88200
6 I 3000 I 36 I 91800
9 I 7200 I 39 I 94200
12 I 12600 I 42 I 95400
15 I 2 1600 I 45 I 96600
18 33600 48 97200
21 47400 51 97380
24 I 62400 I 54 I 97500
27
I 74400
I 57 I 97500
 Hyetograph and Hydrograph Analysis 117

Solution
With reference to Table 3.15,the computational steps are as follows:
Enter the S-curve ordinates in Column 2. This S-curve is derived from
6-hrUH used in Example 3.4 of S-curve derivation.
Shift the S-curve by 3-hrand enter these values in Column 3.
Subtractthe two S-curves [Column (2) minus Column (3)] and enter these
values in Column (4).This represents the UH of 3-hrduration with a unit
volume equal to 3 cm.
Compute the 3-hr UH of unit volume equal to 1 cm by dividing the
hydrograph obtained from step (iii) by 3. The ordinates of the resulting
UH are given in Column 5.
Table 3.15 Derivation of 3-hrUH from 6-hr UH (S-curve method)
The S-curve ordinates of S-curve lag-ged 3-hr UH 3-hr UH
(hrs) excess rainfall by 3 hrs with 3-cm depth of 3-ern depth
intensity 1 crn/hr (rn3/s) (rn3/s) (rn3/s)
(rn3/s)

(1) (2) (3) (4) = (2) ~ (3) (5) = (4)/3

0 0 ~

0 0
3 1200 0 1200 400
6 3000 1200 1800 600
9 7200 3000 4200 1400
12 12600 7200 5400 1800
15 2 1600 12600 9000 3000
18 33600 2 1600 12000 4000
21 47400 33600 13800 4600
24 62400 47400 15000 5000
27 74400 62400 12000 4000
30 82800 74400 8400 2800
33 88200 82800 5400 1800
36 91800 88200 3600 1200
39 94200 91800 2400 800
42 95400 94200 1200 400
45 96600 95400 1200 400
48 97200 96600 600 200
51 97380 97200 180 60
54 97500 97300 120 40
57 97500 97500 0 0
118 Engineering Hydrology

-
+ S-curve +S-curve lagged by 3-hr

3-hrUH volume 3 cm -3-hr UH volume 1 cm

120000
?
m

-5
E
100000
80000
60000
0
.sn 40000
20000
0
0 10 20 30 40 50 60
Time (hr)

Fig. 3.16 UH derivation

3.7 DERIVATION OF AN AVERAGE UH

In practice, unit hydrographs derived from different rainfall-runoff events of a
particular catchment differ significantly from one another. These unit
hydrographs are averaged for computing a basin-representative average unit
hydrograph.
3.7.1 Graphical Method
The graphical method (Wilson, 1974) of deriving an average UH can be applied
by the following steps:

(i) Plot the unit hydrographs derived from a number of individual events on
a single plotting paper.
(ii) Mark a point such that it corresponds to the average peak discharge
ordinates and average time to peak, as shown in Fig. 3.17.
(iii) Draw an average unit hydrograph such that it passes through the average
peak point, has unit volume, and generally conforms to the characteristic
shapes of the individual unit hydrographs. Some trials for adjusting the
ordinates of the derived unit hydrograph may be required for preserving
the volume.
3.7.2 Statistical Method
The statistical approach characterizes unit hydrographs by their shapes, and
the shapes are described by volume, mean, coefficient of variation, skewness,
and peakedness. These five parameters are termed as shape factors. It is to
note that the volume and mean, however, represent the scale and location
characteristics of a distribution (or unit hydrograph), respectively. The steps
for using shape factors for determiningan average unit hydrograph are described
below.
 Hyetograph and Hydrograph Analysis 119

-Average time to peak

PI-t

Time (hr) +

Fig. 3.17 Averaging UH by the graphical method

(i) Calculate the values of shape factors for each derived unit hydrograph
and averagethem. Averaging based on median values minimizes the effect
of outliers.
(ii) The values of shape factors of each unit hydrograph are investigated and
that unit hydrograph is selected the parameters of which closely match
with the average value of the parameters.
(iii) The unit hydrograph derived fiom step (ii) represents the average unit
hydrograph for the catchment.
Further details on statistical methods are given in Chapter 4.

3.8 CONCEPTUAL MODELS

The conceptual models occupy an intermediate position between the fully
physical approach and empirical black-box analysis. Such models are
formulated on the basis of a relatively small number of components, each of
which is a simplified representationof one process element in the system being
modeled. The purpose of this breakdown is primarily to enable the runoff
from catchments to be estimated using standard parameters together with actual
data. A detailed discussion of two popular conceptualmodels, i.e., Nash model
and Clark's model are discussed below.
3.8.1 Nash Model
The derivation of the Nash KJH is described using the unit impulse function
theory. Let I, Q, and S represent the rainfall, runoff (discharge), and storage
capacity of a basin at any arbitrary time t; then the following water balance
equation holds:
120 Engineering Hydrology

I - Q = dS/dt
Since for any time-invariant system S = KQ (refer Chapter 5),
dS/dt = K(dQ/dt)
Or I = Q + K(dQ/dt)

Or dQ/dt + Q/K = I/K

By multiplying the factor e@lK)
to both sides of the above equation, we have:

&(QetfK)= (I/K)etfK
dt
This means:
Qet'K = (I/K)letiKdt
Or Qet'lK= +A
where A is the integration constant which can be derived as follows:
At t = 0, Q = 0, therefore A = -I(e)(-"fl = I
Or Q = 1(1 e-"">
-

F o r I = 1, Q = 1 -ePtfK
This equation gives the unit impulse, i.e., the output for given unit input
( I = 1). The Q in the equation represents the ordinates of the S-hydrograph
when there is a continuous input of I = 1.
Therefore, the IUHcan be computed by deriving the equation as:
-tlK
q(t) = dQ(t)/dt = -e--

Fig. 3.18 Concept of Nash Model

If there are n reservoirs in series, and there is an input of unit rainfall occurring
within a small time, i.e., At + 0, the output forms the ordinates q(t) of an IUH.
-e-tfK
For the first reservoir, the output ql(t) = - K . This forms the input to the

second reservoir, therefore, the output fi-om second reservoir is:

q2(t)= (-e-t'K/K>(I - ept'/K>
 Hyetograph and Hydrograph Analysis 121

In a similar way, the output from the nth reservoir is derived in form of a gamma
distribution function as follows:

This is the Nash model. Table of r(n)is given in Appendix-I. Explicit estimation
of Nash parameters is also explained in Chapter 4.
Computational steps for deriving V H by Nash Model
The following steps are involved for deriving the unit hydrograph by the Nash
Model using the rainfall-runoff data of a particular storm:

(i) Obtain mean rainfall values at each computational interval taking the
weighted mean of the observed values at different stations.
(ii) Estimate direct surface runoff separating the base-flow from the discharge
hydrograph using one of the base-flow separation techniques.
(iii) Estimate the excess rainfall hyetograph separating the loss from total
rainfall hyetograph.
(iv) Estimate the first and second moment of effective rainfall hyetograph
about the origin.
(v) Estimate the first and second moment of direct surface runoff hydrograph
about the origin.
(vi) Find out the parameters n and K using the values of moments obtained
from step (iv) and (v).
(vii) Estimate the unit hydrograph of duration T hours using Nash model.

Example 3.12 The ordinate of excess rainfall hyetograph and direct surface
runoff hydrograph for a storm of typical catchment of the size 1700 sq. km.
are given below. Find the Nash model parameters.

Table 3.16 Data for Example 3.12

lhe Excess rainfall Direct surface
(hrs) (mm) runof (m3/sec)
I 0 0
1 6 I 40.209 I 250 I
I 12 I 100.209 I 1050 I
I 18 I 60.209 I 2050 I

I
I
I
24
30
I 0
0
I 4350
4150

120
I
I
122 Engineering Hydrology

Solution
(i) First and second moment of effective rainfall:

- (40.209 x 3) + (100.209 x 9) + (60.209 x 15)

40.209 + 100.209 + 60.209

M
C xiti'
&ffx= +
C xi
i=l
- [40.209 x (3)2] + [100.209 x (9)2] + [60.209 x (15)2]
40.209 + 100.209 + 60.239

(ii) First and second moment of direct surface runoff

$T+T+1 N

2
ti. C (T + T+Jtir
-
- i=l
lMfY = i=l N

$*
i=l
c (T +
i=l
T+1>

250 (250 + 1050) (1050+ 2050) 15

-x3+
+

=
2 2 2

+ (2050 +2 4350) 21 (4350+4150)

2
+ x27

+ (4150 +2 2300) 33 (2300 2+ 1070) 39

+

+ (10702+ 450) 45 + (450+120)

2
x 51 + 120 x 57
2
~

250 (250 + 1050) (1050 + 2050) (2050 + 4350)

, M j . = -+
+ +

2 2 2 2

+ (4350+4150)
2
+ (4150+2300)
2
+ (2300+ 1070)
2
 Hyetograph and Hydrograph Analysis 123

(450 + 120) (120)

+ (1070+
2
450) +

2
+- 2
-
435720 = 27.595 hr
-~
15790
To compute second moment, ti is to be replaced by square of ti in the
above expression, i.e.,
250 (250 + 1050) 92+ (1050 + 2050) 152
~ x 32+
2 2 2

+ (2050+2 4350) 212+ (4350+2 4150) 272+ (4150+2 2300) 332

120 2
+-x57
2

(iii) Solve the following equations to get the parameters n and K.

nK = ,M’y - 2Mfx
n(n + 1)K2+ 2nK,Mfx = 2M’y ,Mfx-

where ,Mfx = 9.598 hour and 2Mfx= 109.785 h o d

,M’y = 27.395 hour, and 2M’y = 852.572 hour2
.. nK = 27.595 9.598 = 17.997
-

and n (n+l) K2 + 2 n K x 9.598 = 852.572 -109.785

n (n+l) K2 +19.196 n K = 742.787
( T Z K ) ~ + T Z19.196nK=742.787
K~+
(17.997)2+ 17.997 K + (19.196 x 17.997) = 742.787
323.892 + 17.997 K + 345.470 = 742.787
17.997 K = 73.425
K = 4.08

(iv) Compute the IUH using n and K. The procedure to convert IUH into
S-curve or unit hydrograph is already explained. The procedures used to
compute moments in case of rainfall and runoff are illustrative only and
can be used interchangeably. Please check if estimates of n and K will be
sensitive to the use of a particular procedure of computing moment and
if so, up to what extent?
124 Engineering Hydrology

3.8.2 Clark’s Method

Similar to Nash model discussed in previous section, Clark’s method routs the
time-area histogram to derive an IUH. The time area diagram is developed for
an instantaneousexcess rainfall over a catchment and represents the relationship
between the areas that contributes to the runoff at the outlet versus the time of
travel. It is assumed that the excess rainfall first undergoes pure translation by
a travel time-area histogram, and the attenuation is attained by routing the
results of the above through a linear reservoir at the outlet. The details are
discussed below.
Time-AreaCurve
The time of concentration is the time taken for a droplet of water to travel
from the upper boundary of a catchment to the outlet point. This is generally
computed using empirical formulae which relate the time of concentration to
geomorphological parameters of the catchment. In gauged areas, the time
interval between the end of the excess rainfall and the point of inflection of the
resulting surface runoff provides a good way of estimating t, from known
rainfall-runoff data. In simple terms, the total catchment area drains into the
outlet in t, hours.
First, the catchment is drawn using a toposheet map, and then the points on
the map having equal time of travel, (say, t,, hr where t,, < t,), are considered
and located on a map of the catchment. The t,, is calculated using the length of
stream, length of centroid of the area from the outlet, average slope, and the
area (from empirical equations). Subsequently,a line is joined to all such points
having same tlo to get an isochrones (or runoff isochrones). These are shown
in an example below.
To make this procedure simple, one can start with selection of the longest
water course. Then, its profile plotted as elevation versus distance from the
outlet; the distance is then divided into Nparts and the elevations ofthe subparts
measured on the profile transferred to the contour map of the catchment.
The areas between two isochrones denoted as A , , A,, ..., A,,, are used to
construct a travel time-area histogram. If an excess rainfall of 1 cm occurs
instantaneouslyand uniformly over the catchment area, this time-area histogram
represents the sequence in which the volume of rainfall will be moved out of
the catchment and arrive at the outlet. For example, a sub-area of A , km2
represents a volume of A , km2 x 1 cm = A , x lo4 (m3) moving out in time
Atlc = hours. Clark routed this time-area diagram through a linear reservoir
N
which he assumed to be hypothetically available at the outlet to provide the
requisite attenuation.
Routing
The linear reservoir at the outlet is assumed to be described by S = KQ, where
K is the storage time constant. The continuity equation can be written as:
 Hyetograph and Hydrograph Analysis 125

Or K = -Q,/(dQ/dt)I (3.15)
where, suffix ‘i ’ refers to the point of inflection, and K can be estimated from
a known surface runoff hydrograph of the catchment. The K can also be
estimated from the data on the recession limb of the hydrograph. Knowing
K of the linear reservoir, the inflows at various times are routed by the
Muskingum method. The Muskingum equation can be written as:
S = K[xI+ (1 - x)Q] (for details, refer Chapter 5) (3.16)
where, x is known as the weighting factor. For a linear reservoir, value of x = 0.
The inflow rate between an inter-isochrones area of A,, km2 with a time
interval At,, hr is:

I = A, x lo4 = 2.78 ~
Alc (m3/s)
3600(Atc) At1c
The Muskingum routing equation can be written as:
Q2 = CoI2 + CiIi + C2Q1 (3.17)
where C, = (0.5 Atc)@+ 0.5 Atc)
C, = (0.5 Atc)@+ 0.5 Atc)
C2 = ( K - 0.5 Atc)@+ 0.5 Atc)
i.e., C, = C,. Also, since the inflows are derived from the histogram, I , = I 2 for
each interval. Thus, Eq. (3.17) becomes:
Q2 = 2CJi + C2Q1 (3.18)
Routing of the time-area histogram by Eq. (3.18) gives the ordinates of IUH
for the catchment. Using this KJH, any other D-hr UH can be derived.
The computational steps involved in Clark’s model are as follows:
Make first estimate of Clark‘s model parameters, t, and R, from the excess
rainfall hyetograph and direct surface runoff hydrograph.
Construct the time-area curve, taking the t, value obtained from step (i),
using the procedure described.
Measure the area between each pair of isochrones using planimeter.
Plot the curve of time versus cumulative area. Note that the abscissa is
expressed in percent oft,. Tabulate increments between points that are at
computational interval At apart.
Route the inflow using Eqs. (3.17) and (3.18) to get KJH ordinates.
126 Engineering Hydrology

(vi) Compute the unit hydrograph of the excess rainfall duration using the
given equation:
UHi = l / n [0.5Uip,+U,p,+l +........ Uip, + 0.5 Ui]
where, n = DIAt, and D is duration of the UH.

Example 3.13 Clark’s model parameters, t, and R, derived fi-om a short-

duration storm excess rainfall hyetograph and direct swEace runoff hydrograph,
are 8 and 7.5 hours respectively for a typical catchment of area 250 h2. The
ordinates of time-area diagram are:
Time 1 2 3 4 5 6 7 8
(W
Area 10 23 39 43 42 40 35 18
(h2)

Derive a 2-hr UH using Clark’s model.

Solution
Computational steps are: (Ref. Table 3.17)

(9 Convert the units of inflow using the equation: Ii = Kai I At

where ai = the ordinates of time-area curve in m2
Ii = the ordinates of the same time-area curve in m3/s
K = conversion factor
At = computational interval (hours; 1 hour in this example)
If the volume under the time-area curve is taken equivalent to 1 cm excess
rainfall, then the conversion factor K would be:

Therefore, the above equation may be written as:

Ii = 2.778 aill = 2.778 ai
(ii) Compute the routing coefficent (c> using the following equation:
c= R +At0.5At -- 7.5 + 0.5 =0.125

:. 1 - C = 1 - 0.125 = 0.875
(iii) Route the time-area curve obtained from step (i); (see Columns 3,4 and
5)
u i = c I i + ( l - c ) ui-1
Ui = 0.125 Ii + 0.875 U,. - 1
where U, = the ithordinate of lLTH in m3/s.
 Hyetograph and Hydrograph Analysis 127

(iv) Compute 2-hr unit hydrograph (see Column 6).

Hi = l / n [0.5Uip,+Uipn+l+. ... ....Uipl +0.5 U , ]
For this example:
D = 2 hours
t = 1 hour

1
1
.*. UHi= - (0.5 Uip, + U,.-, + 0.5Ui]
2
See Table 3.17 for the detailed computation.

Table 3.17 Clark's model computation

Erne Time area 0.125 I, = 2.78 0.875 X Co@) Col (3) + (4) 2-hr UH
(hd diagram x 0.125 x Col(2) (in3/,) IUH ordinates ord;
(kmz, (m3/s) (m3/s) (m'/s)
(1) (2) (5)
0 0 0 I 0 0
1 10 I 3.5 I 3.5 0.875
2 8.0 I 3.1 I 11.1 4.525
I
3 39 I 13.5 I 9.7 I 23.2 12.225
4 43 I 14.9 I 20.3 I 35.2 23.175
5 42 I 14.6 I 30.8 I 45.4 34.750
6 40 13.9 39.6 53.5 44.875
7 35 12.1 46.8 58.9 52.825
8 18 I 6.2 I 51.4 I 57.6 57.225
9 0 1 0 I 50.5 I 50.5 56.150
10 0 I 44.1 I 44.1 50.675
11 0 I 38.6 I 38.6 44.325
12 0 1 0 I 33.8 I 33.8 38.775
13 0 I 29.6 I 29.6 33.950
14 0 I 25.9 I 25.9 29.725
15 0 1 0 I 22.7 I 22.7 26.025

SUMMARY

In this chapter, an attempt has been made to introduce the concept of hydrograph
analyses. Understanding of different terms like Unit Hydrograph, lLTH and
S-curve is very relevant in this context. While analysis of single storm runoff
is relatively simpler than that of complex storm runoff, one is often confronted
128 Engineering Hydrology

with analysis of complex storms. From the analysis of single runoff storm
hydrograph, based on application of Z-transform, it is also possible to identify
hyetograph and unit hydrograph. However, such an approach has not been
included here. A catchment normally has different orders of streams, and for a
given channel network, geomorphological instantaneous unit hydrograph
(GKJH) can be derived using known catchment characteristics. One can also
use synthetic approaches to develop unit hydrograph.

SOLVED EXAMPLES

Example 3.14 In a catchment, the following 1-hrUH due to 1 mm effective

rainfall is known:
Time (hr) 0 1 2 3 4
QUH (m3/s) 0 3 2 1 0

(a) Estimate the direct runoff if the effective rainfall during 1-hr is 5 mm.
(b) Estimate the direct runoff if the effective rainfall is 5 mm during the first
hour and 10 mm during the second.
Solution
(a) UH ordinates need to be multiplied by 5 because effective rainfall is 5 mm.
Erne UH Qdimct

)
(in '/s, (in '/s,

15
4 0 0

(b) We need to obtain two DRHs corresponding to 5 mm and 10 mm effecvtive

rainfall. Columns 3 and 4 show these at 1-hrlag.
nine UH Qdir-ectl Qdirect2 Qdirect total
(m3/s) (in 3 4 (m3/s) (m3/s)
0 0 0 0
1 3 15 0 15
2 2 10 30 40
3 1 5 20 25
4 0 0 10 10
5 0 0 0
 Hyetograph and Hydrograph Analysis 129

Example 3.15 During and after a rainfall event, the following discharge
was measured in a river. Plot the hydrograph both in linear/linear and linear/
log scale.
(a) What was the base flow if it can be assumed to be constant?
(b) What was the base flow if it can be assumed to follow the recession curve?
Tid (hr) 0 1 2 3 4 5
Qmi (m3/s> 0.1 0.1 2.50 1.30 0.25 0.1

Solution
(a) Base flow constant = 0.1 m3/s

The direct runoff ends when the flow follows the recession curve. This point
can be found when the flow follows a straight line in a lin/log plot. Thus, the
base flow increases from ( t = 2, Q = 0.1) to ( t = 4, Q = 0.25), after the point
Q = Qbasc
Example 3.16 A catchment in Egypt experiences a long period without rain.
The discharge in the river which drains the catchment is 100 m3/s 10 days
after the long period without rain, and 50 m3/s after another 30 days. What
flow can be expected to occur on day 120 if there is no rainfall during this
period?
Solution
The recession can normally be assumed to follow an exponential function
given by:
Q = Q, epkt
Data given:
t = 10 days, Q = 100 m3/s
t = 40 days, Q = 50 m3/s
Applying the above function, we have:
100 = Qo epkl0
50 = Q e+40
0
130 Engineering Hydrology

Therefore, 100/50 = e +40k

30k
2=e
k = 0.023 1
For t = 120 days:
- AOk + 120k = ,80k =6 347
Q40/Q120 -

Q120= 50/6.347 = 7.9 m3/s

Answer Flow after 120 days would be 7.9 m3/s.

Example 3.17 Given below are the discharge values for a 30-min UH.
(a) What is the size of the catchment area?
(b) Calculate and plot the hydrograph for a 60 minute rain with an effective
intensity of 10 Illlll/hT.
Time
(min) 0 15 30 45 60 75 90 105 120 135 150
Discharge
(m3/s) 0 4.5 10 12.5 11 9 6.5 4 2.5 1 0

Solution
Eme Discharge
(rnin) (rn '/s)
0 0 0
15 4.5 4050
30 10 9000
45 12.5 11250
60 11 9900
75 9 8100
90 6.5 5850
105 4 3600
120 2.5 2250
135 1 900
150 0 0
Sum = 54900

(a) Total volume of runoff is given above.

Totally, 1 rnrn effective rain has fallen over the catchment.
Thus, the area can be calculated by 54900/0.001= 5.49 x lo7m2= 54.9 h2
(b) The given hydrograph has a duration of 30 min and is valid for an intensity
of 2 InmAlr.
Add two hydrographs multiplied by 5 with a 30 min lag between the
hydrographs.
 Hyetograph and Hydrograph Analysis 131

Rme UH 2 mmhv UHxS UHxS Sum

(min) (m3 4 (m3/s) (m3/s) (m3 4
0 1 0 I 0 I I
15 I 4.5 I 22.5 I I 22.5

45 I 12.5 I 62.5 I 22.5 I 85

60 I 11 I 55 I 50 I 105
75 I 9 I 45 I 62.5 I 107.5
90 I 6.5 I 32.5 I 55 I 87.5
105 I 4 I 20 I 45 I 65
120 I 2.5 I 12.5 I 32.5 I 45
135 I 1 1 5 I 20 I 25
150 I 0 I 0 I 12.5 I 12.5
165 I I 1 5 I 5
180 0 0

Example 3.18 A catchment received a continuous rainfall for 3 hours. The

rain is distributed in time with 17 mm during the first hour, 22 mm during the
second hour, and 14 mm during the final hour.
#-index is 12 mtnlhr and the measured hydrograph is given below.
(a) Find the l-hr UH. (Assume the base flow to be constant.)
(b) What is the time base (duration) and the lag-time between the maximum
rainfall and the maximum flow of the hydrograph?
Time
(W 0 1 2 3 4 5 6 7 8 9 1 0 1 1
Discharge
(m3/s) 0 1 5.5 12.4 15.4 13 9.6 6.2 3.05 0.9 0.1 0

Solution
The measured runoff can be described by the sum of three unit hydrographs
scaled to the rainfall intensity @.
Subtract the @-indexfrom the rainfall intensities to get the effective rainfall.
(a) For the first hour,
effective rainfall = 17 mrn - 12 mm = 5 rnm
For the second hour,
effective rainfall = 22 mrn - 12 mm = 10 rnm
For the third hour,
effecitve rainfall = 14 mrn - 12 mm = 2 rnm
(b) Columns (2), ( 3 ) and (4) are the multiplied values of UH ordinates with
the effective rainfalls of 5 mm, 10 mrn and 20 rnm respectively. The values
are placed at 1-hr lag.
132 Engineering Hydrology

Column (5) will be the sum of the values in columns (2), ( 3 ) , and (4).
This implies
5UHO=O
5UH1+ 10 UHO = 1
5 UH2 + 10 UH1= 2 UHO = 5.5
and so on.
The solution of these equations is given in the following table.
l3me 5xUH 1OxUH 2 xUH Sum
(hr) (m3 4 (m3/s) (m3/s) (m3 4
0 5 x UHO 0
1 5xUH1 10 x UHO 1
2 5xUH2 10 x UH1 2 x UHO 5.5
3 5xUH3 10 x UH2 2 x UH1 12.4
4 5xUH4 10 x UH3 2 x UH2 15.4
5 5xUH5 10 x UH4 2 x UH3 13
6 5xUH6 10 x UH5 2 x UH4 9.6
7 5xUH7 10 x UH6 2 x UH5 6.2
8 5xUH8 10 x UH7 2 x UH6 3.05
9 5xUH9 10 x UH8 2 x UH7 0.9
10 5xUH10 10xUH9 2xUH8 0.1
11 5xUHll 10 x UHlO 2xUH9 0

I I Solving for UH,

I uH0 I 015 I 0 1
I uH1 I 115 I 0.2 1
uH2 (5.5 - 10 X 0.2 - 0)/5 0.7
uH3 (12.4 - 10 X 0.7 - 2 X 0.2)/5 1
uH4 (15.4 - 10 X 1 - 2 X 0.7)/5 0.8
uH5 (13 - 10 X 0.8 - 2 X 1)/5 0.6
uH6 (9.6 - 10 X 0.6 - 2 X 0.8)/5 0.4
uH7 (6.2 - 10 X 0.4 - 2 X 0.6)/5 0.2
uH8 (3.05 - 10 X 0.2 - 2 X 0.4)/5 0.05
uH9 (0.9 - 10 X 0.05 - 2 X 0.2)/5 0
uHl0 (0.1 - 10 X 0.2 X 0.05)/5 0
 Hyetograph and Hydrograph Analysis 133

EXERCISES

3.1 Explain UH theory with its basic assumptions and applications.

3.2 Describe the governing factors affecting the shape of the unit hydrograph.
How is the UH different fi-om the distribution graph?
3.3 Describe the methods of UH derivation from rainfall-runoff data of an
event.
3.4 In Question 3, how is the conventional method advantageous over other
methods? If the available rainfall-runoff data correspond to multi-period
storms, which methods are suitable for deriving a UH of the desired
duration?
3.5 Describe Collins’ method. State the reasons for the occurrence of
oscillating ordinates in the recession limb of the derived UH. What are
the procedures followed to do away with the oscillating ordinates?
3.6 Describe the methods of single division and least squares for deriving a
UH and discuss their advantages and disadvantages.
3.7 What are the unit period and volume of the UH? State the theory for
changing the unit volume of the UH of a specific duration.
3.8 Describe IUH and S-curve hydrograph and their relationship. How are
these related with the UH?
3.9 What is superimposition method and how is it advantageous over the
S-curve method?
3.10 Discuss various methods of averaging unit hydrographs.
3.11 A thunderstorm of 1-hr duration with 20 mm excess rainfall volume
resulted in the following direct surface runoff (DSRO) hydrograph at
the outlet of the catchment. Find out the UH of 1 mm excess rainfall
volume and 1-hr duration.
Erne DSRO Erne DSRO
(k) (rn 3 4 (k) (rn3/s)
0 0 12 91.6
1 61.0 13 61.4
2 3 14.6 14 40.6
3 561.0 15 26.4
4 673.4 16 17.0
5 645.0 17 11.0
6 584.2 18 7.0
(Contd.)
 ;
134 Engineering Hydrology

10
11
475.0
365.4
269.8
192.8
134.4
22 1.o

3.12 The excess rainfall hyetograph and the direct surface runoff hydrograph
ordinates for an event of Godavary basin Sub-zone 3f (bridge no. 807/1)
are given below. The catchment area is 824 sq. km. Determine 1-hrUH
of 1 mm volume using Collins’ method.
Excess Rainfall DSRO 7ime DSRO
(mm) (in3/,) (W (m'/s)
0 3.58 0 12 91.6
1 4.07 61.0 13 61.4
2 2.54 3 14.6 14 40.6
3 ~

561.0 15 26.4
4 ~

673.4 16 17.0
5 ~

645.0 17 11.0
6 ~

584.2 18 7.0
I ~

475.0 19 4.4
8 ~

365.4 20 2.8
9 ~

269.8 21 1.6
~

192.8 22 1.o
11 ~ 134.4

3.13 The ordinates of a 2-hr UH for a catchment (area = 824 sq. km) are given
below. Determine 1-hr UH using S-curve method.
llme UH coordinates llme UH coordinates
0%) (m'/s) 0%) (m'/s)
1 1.53 13 3.83
2 9.39 14 2.55
3 21.89 15 1.68
4 30.86 16 1.09
5 33.46 17 0.70
6 3 1.23 18 0.45
7 26.48 19 0.29
8 21.01 20 0.18
9 15.88 21 0.11
10 11.57 22 0.07
11 8.18 23 0.03
12 5.65
 Hyetograph and Hydrograph Analysis 135

3.14 The ordinates of 1-hr UH for a catchment (area = 824 sq. km)are given
below.
Rme UH coovdinates Rme UH coovdinates
(hv) (m’/.) hv (m’/.)
1 3.05 12 4.58
2 15.73 13 3.07
3 28.05 14 2.03
4 33.67 15 1.32
5 32.25 16 0.85
6 29.21 17 0.55
7 23.75 18 0.35
8 18.27 19 0.22
9 13.49 20 0.14
10 9.64 21 0.08
11 6.72 22 0.05

(a) Determine the 5’-curve hydrograph for the excess rainfall of unit
intensity (1 d h r ) .
(b) Find out the ordinates of 2-hr UH (1 mm) using 5’-curve and
superimposition methods.
(c) Discuss the advantages of using superimposition method over the
5’-curve method for the given problem. State the circumstancesunder
which the superimposition method fails.
3.15 A catchment (area = 300 km2) receives an uniform rainfall of 30 mm in
one day. During thenext few days, the discharge is observed in the river
that drains the catchment. The readings are given in the following table.

T (days) 0 1 2 3 4 5
Q (m3/s) 2 5 10 8 4 2

The base-flow is assumed constant Qba,, = 2 m3/s during the discharge

period.
(a) How much was the direct runoff during the period (answer in m3/s)?
(b) How much was the total direct runoff during the period (answer in
mm)?
(c) How much was the total losses for the rainfall (answer in mm)?
(d) What was the maximum discharge in the water course during the
period?
Answer
(a) 3, 8, 6 and 2 m3/s or an average of 4.75 m3/s
= = = =
(b) 4.75 4 24 3600/300 lo6 m = 5.5 mm
(c) 24.5 mm (d) 10 m3/s
136 Engineering Hydrology

OBJECTIVE QUESTIONS

1. The inflection point on the recession side of the hydrograph indicates the
end of
(a) Base flow (b) Direct runoff
(c) Overland flow (d) Rainfall
2. The concept of unit hydrograph was first introduced by
(a) Dalton (b) Sherman
(c) Horton (d) Thiessen
3. UH is the graphical relation between the time distributions of
(a) Total rainfall and total runoff
(b) Total rainfall and direct runoff
(c) Effective rainfall and total runoff
(d) Effective rainfall and direct runoff
4. The word “unit” in unit hydrograph refers to the
(a) Unit depth of runoff
(b) Unit duration of the storm
(c) Unit base period of the hydrograph
(d) Unit area of the basin
5. The range of the area of the basin where the unit hydrograph is applicable
is
(a) 200 hectar-5000 km2 (b) 100 hectar-10000 km2
(c) 100 km2-5000 km2 (d) 250 km2-7500 km2
6. The basic principles of unit hydrograph theory are
(a) Principle of superposition (b) Principle of time invariance
(c) Both (a) and (b) (d) None of these
7. The S-curve hydrograph is the summation of the
(a) Unit hydrograph (b) Total runoff hydrograph
(c) Effective rainfall hyetograph (d) Base-flow recession curve
8. The S-curve hydrograph is used to
(a) Estimate the peak flood flow of a basin resulting from a given storm
(b) Develop synthetic unit hydrograph
(c) Convert the UH of any given duration into a UH of any other desired
duration
(d) Derive the UH from complex storms
9. The lag-time of the basin is the time interval between the
(a) Centroid of the rainfall diagram and the peak of the hydrograph
(b) Beginning and end of direct runoff
 Hyetograph and Hydrograph Analysis 137

(c) Beginning and end of effective rainfall

(d) Inflection points on the rising and recession limbs of the hydrograph
10. The direct runoff hydrograph of a basin can be approximated as a triangle
with a base period of 90 hr and a peak flow of 230 m3/sec during the lSth
hour. If the area of the basin is 1863 km2, what is the depth of runoff
indicated by the hydrograph?
(a) 1 cm (b) 2 cm (c) 5 cm (d) lOcm
11. The 6-hr UH of a basin can be approximated as a triangle with a base
period of 40 hr and peak ordinate of 150 m3/s. Then the area of the basin
is
(a) 2160km2 (b) 1080km2 (c) 540km2 (d) 1280km2
12. The peak discharge in 2-hr and 4-hr UHs of a basin occurs at t , and t2.
What can you infer from this information?
( 4 t,= t2 (b) t,> t2 (c) t,< t2 (d) t,<=t2
13. The 4-hr and 8-hr UH of a basin having the same base period is plotted.
Which of the following statement is true?
(a) The peak discharge at 4thhr is greater than Sth hr
(b) The peak discharge at Sthhr is greater than 4thhr
(c) Both the peak occurs at the same time
(d) None of the above
14. A 6-hr storm produced rainfall intensities of 7, 16,20, 14,8, and 5 mm/hr
in successive one hour intervals over a basin of 1000 sq km. The resulting
runoff is observed to be 2600 ha. m. Determine @-indexfor the basin.
(a) 8mm/hr (b) 9mm/hr (c) lOmm/hr (d) 11 mm/hr
15. In an 8-hr storm, the total depth of precipitation is 80 mm. Total storm
runoff is 55 mm. Losses due to depression storage and interception is
5mm. Find out the W-index.
(a) 2.0mm/hr (b) 2.5 mm/hr (c) 3.0mm/hr (d) 4.0 mm/hr

REFERENCES

Chow, V.T., 1964, Runoff Section 14, Handbook of Applied Hydrology,

McGraw-Hill Book Co., New York.
Clark, C.O., 1945, ‘Storage and the Unit Hydrograph’, Trans. Am. SOC. Civil
Engg., 110, pp. 1419-144.
Collins W.T., 1939, “Runoff distribution graphs from precipitation occurring
in more than one time unit”, Civil Engineering., 9(9), pp. 559-561.
Kuchment, L.S., 1967, “Solution of inverse problems for linear flow models”,
Soviet Hydrology, Selected Papers, Vol. 2, p. 94.
138 Engineering Hydrology

Larsson, R., 2000, CalculationExercises in Hydrology,Lund University, Lund,

Sweden.
Malm, J., M. Persson, and R. Larsson, 2005, Exercises in Water Resources,
Lund University, Lund, Sweden.
Nash, J.E., 1959, “Synthetic determination of unit hydrograph parameters”.
J. Geophys., 64( l), pp. 111-1 15
Nash, J.E., 1960 ‘A unit hydrograph study with particular reference to British
catchments’, Proc. Inst. Civil Engg, London, 17, pp. 249-282.
Ojha, C.S.P., K.K. Singh, and D.V.S. Verma, 1999, “Single Storm Runoff
Analysis using Z-Transform”, J. ASCE Hydrologic Engineering, 4( l),
pp. 80-82.
Singh, V.P., 1991, Elementary Hydrology, Prentice Hall, Englewood Cliffs,
New Jersey.
U.S. Army Corps of Engineers, 1959, “Flood Hydrograph analysis and
applications”., Engineering Manual, pp. 1110-1603, US Army,
Washington D.C.
Wilson, E.M., 1969, Engineering Hydrology, McMillan, London.