Unit Hydrograph Method

The unit hydrograph is defined as a hydrograph resulting from one unit depth of effective
rainfall applied uniformly over the catchment at a uniform rate during a specified period of
time. The unit depth may be taken as 1 mm or 1 cm. The duration is taken as T generally in
hours: T-hour unit hydrograph, indicated by u(T,t) or TUH. Hence the rain intensity
becomes 1/T mm/hr.

The unit hydrograph was first introduced by Sherman in 1932 as unit-graph, where “unit”
referred originally to unit of time; later “unit” was interpreted as a unit depth of rainfall. The
unit hydrograph is a simple linear model. It can be used to derive the hydrograph from any
excess rainfall applying the principles of linearity: proportionality and superposition. The
following basic assumptions are inherent in this model (Chow et.al., 1988):

1. The excess rainfall has a constant intensity within the effective duration.
2. The excess rainfall is uniformly distributed throughout the whole catchment area.
3. The base time of the direct runoff resulting from an excess rainfall of given duration is
constant, irrespective of the rain intensity.
4. The ordinates of all hydrographs of a common time base (or the direct runoff
hydrographs produced by effective rainfalls of the same duration) are directly
proportional to the total amount of direct runoff represented by each hydrograph.
5. For a given catchment, the hydrograph resulting from a given excess rainfall reflects the
unchanging characteristics of the catchment.

Assumptions 1 and 2 are incorporated in the definition of unit hydrograph itself. Assumption
4 represents the linearity principle (See Fig. 3.6), whereas assumptions 3 and 5 imply that the
shape of the runoff hydrograph remains the same irrespective of whatever time it may occur
as long as the duration of rainfall is the same, i.e. the principle of invariance.

1
 Figure 3.6 Linearity (proportionality and superposition) principle

Limitations: in literature very different ranges of catchment areas for which the method may
be applicable are given. However the method should not be used for catchment areas greater
than 5000 (km)2. Assumption 1 requires that storms of short duration are selected, since they
are likely to produce intense and nearly constant excess rainfall rate, yielding a well-defined
single-peaked hydrograph of short time base. Assumption 2 requires that the area is not
taken too large since else the assumption of uniform distribution throughout the basin will be
violated. To overcome this problem the total catchment area may be subdivided into smaller
areas. Assumption 3 on the constant time base depends on the method of baseflow
separation. Theoretically it should depend on the rainfall intensity too. Assumption 4, the
principle of linearity is basically not valid for overland flow or channel flow. Unit
hydrographs derived from low intensity rainfall will generally have a lower peak than those
derived from large storms. If such types of non-linearity exists the unit hydrograph theory
should be applied with caution or better not be applied at all for design purposes.

2
Assumption 5, the principle of invariance, implies that the hydrograph from a given excess
rain would be the same irrespective of the season. From overland flow theory we know that
the hydraulic roughness plays a role in shaping the hydrograph. Soil cover, vegetation, will
affect the hydraulic roughness and so a season dependency is present.

Note that the unit hydrograph cannot be applied when a major portion of storm precipitation
is in the form of snow.

Derivation of the unit hydrograph from single storms: The derivation of the unit
hydrograph of a catchment from single storms goes along the following lines:

1. The rainfall records are scanned to find a storm of desired duration that gave a fairly
uniform distribution in time and space. The hyetograph of this storm is constructed using
a convenient uniform interval.
2. The base flow is separated from the hydrograph using one of the methods presented in
the preceding section.
3. The surface runoff volume is determined as a depth of flow by numerical integration:

(3.30)

Where: d = depth of surface runoff in mm

t = uniform time interval in hours at which the ordinates of the
surface runoff are measured
Q = Sum of all ordinates of the surface runoff hydrograph in m3/s
A = catchment area in km2.

4. The ordinates of the surface runoff hydrograph are divided by the runoff depth d to get
the ordinates of the unit hydrograph.
5. The duration of excess rain producing the unit hydrograph is determined by applying the
- index method to the hyetograph such that the excess rainfall is exactly d cm. The
duration of the excess rainfall gives the value for T (in hours) in the TUH or u (T,t).
The procedure is shown in the following example.

Example 3.6: given is the following cumulative rainfall and total runoff from a catchment of
size 104 km2:
Date Time hr Cum. Runoff Date Time hr Cum. Rain Runoff
rain mm m3/s mm m3/s
10/6 2 0 16.5 11/6 2 76.4
4 15 15.0 4 65.0
6 30 14.2 6 55.2
8 105 158.5 8 46.7
10 190 260.0 10 39.6

3
 12 200 286.0 12 34.0
14 221.0 14 28.3
16 186.5 16 22.7
18 157.0 18 21.5
20 133.0 20 20.6
22 113.0 22 19.9
24 93.4 24 19.2

Required: the derivation of the unit hydrograph of appropriate duration

Solution: from the rainfall record the hyetograph is constructed at two hours intervals:

Date 10/6 hrs 2-4 4-6 6-8 8-10 10-12 12-14

Rain intensity (mm/hr) 7.5 7.5 37.5 42.5 5 0

The surface runoff clearly starts at the end of the 6 th hour on day 10/6 and it is rather
arbitrarily chosen to end at 16th hour on 11/6. The baseflow is separated by a straight line.
Then the following unit hydrograph derivation follows.

Date Time hr Qtotal m3/s Qbase m3/s Qsurface m3/s TUH Time after
m3/s/mm start hr
1 2 3 4 5 6 7
10/6 6 14.2 14.2 0 0 0
8 158.5 14.7 143.8 1.252 2
10 260.0 15.2 244.8 2.132 4
12 286.0 15.7 270.3 2.354 6
14 221.0 16.2 204.8 1.784 8
16 186.5 16.7 169.8 1.479 10
18 157.0 17.2 139.8 1.218 12
20 133.0 17.7 115.3 1.004 14
22 113.0 18.2 94.8 0.826 16
24 93.4 18.7 74.7 0.651 18
11/6 2 76.4 19.2 57.2 0.498 20
4 65.0 19.7 45.3 0.395 22
6 55.2 20.2 35.0 0.305 24
8 46.7 20.7 26.0 0.226 26
10 39.6 21.2 18.4 0.160 28
12 34.0 21.7 12.3 0.107 30
14 28.3 22.2 6.1 0.053 32
16 22.7 22.7 0 0 34
Q = 1658.4

The surface runoff depth becomes with equation (3.30): d = 3.6 x 2. 1658.4/104 = 114.8 mm.

4
Hence all ordinates of Qsurface are divided by 114.8 to get the TUH ordinates.

The computation of T follows from the hyetograph. The total rainfall was 200mm fallen in
10 hours time. So, the first estimate of the loss rate is (200-114.8)/10 = 8.52 mm/hr. From
the hyetograph it is observed that this intensity is only exceeded in the third and fourth time
step. The total rainfall in those two intervals is 160 mm. Hence the -index becomes (160-
114.8)/4 = 11.3 mm/hr. The excess rainfall intensity over the third and fourth interval
becomes 26.2 mm/hr and 31.2 mm/hr, respectively. These are nearly equal intensities over a
period of 4 hours, so the period T of the TUH becomes 4 hrs.

If this procedure is repeated for different storms the shape of the TUH may be slightly
different; e.g. the peak values (UP) as well as the time to peak (TP) may differ. The advocated
procedure to get an overall acceptable TUH is then to take the average value of the TUH
peaks (UP,av = (UP,1+UP,2+….+UP,n)/n) occurring at the average time to peak (T P,av = (TP,1 +
TP,2+…..+TP,n)/n) and draw the rest of the hydrograph such that it resembles the general shape
of the individuals, see Figure 3.5 below. One should not take the average of the ordinates of
the hydrographs at corresponding times as this will lead to an underestimation of the peak
value; the latter is an important value for design.

Figure 3.5 Derivation of TUH from many storms

5
Changing the duration of the unit hydrograph: there are two simple procedures to change
the duration of the unit hydrograph: (i) by superposition from u (T1, t) to u (T2,t), where T2 =
n*T1, with n an integer > 1, hence only enlargements to some multiple of T 1 are possible, (ii)
by S-curve technique from u(T1,t) to u (T2,t), where T2 = *T1, with  a real > 0, hence T2
may be larger but also smaller than T 1. It is remembered that a unit hydrograph refers to unit
depth of rainfall excess (e.g in mm), so if the duration is T hours then the excess rainfall
intensity is 1/T mm/hr!!!

Enlargement of T by superposition: say, it is required to derive from u(T 1,t) a unit

hydrograph u(T2,t) with T2 = 3T1. The first one refers to a rainfall intensity 1/T 1 to give unit
depth, whereas u (T2,t) should refer to an intensity 1/T 2 to give unit depth. The u(T 2,t) is
obtained by superposition of three u(T1,t) shifted T1 hours apart. By adding the ordinates of
u(T1,t) , u(T1,t-T1) and u(T1,t-2T1) at the corresponding times the resulting hydrograph Q(t)
will refer to an effective rainfall of: 3T 1*1/T1 = 3 mm. Hence to get u(T 2,t) all Q(t) ordinates
have to be multiplied by: i 2/i1 = (1/T2)/(1/T1) = T1/T2 = 1/3; to let it refer to a unit depth of
rainfall (see Fig.3.6).

Figure 3.6 Adjustment of rainfall intensities when enlarging T

Example 3.6: the above outlined procedure is worked out in the example below where from
a one-hour unit hydrograph a three-hour unit hydrograph has to be derived (see Figure 3.7)
and the table below. The ordinates of the one-hour unit hydrograph are presented in column
2, whereas in the columns 3 and 4 the ordinates of the one-hour unit hydrograph are
presented with respectively 1 hour and 2 hours offset. Then for the corresponding times the
ordinates are summed, column 5, and finally divided by 3 to get the three-hour unit
hydrograph. What is the time base or base length Tb of the three-hour unit hydrograph?

6
Example 3.6: the above outlined procedure is worked out in the example below where from
a one-hour unit hydrograph a three-hour unit hydrograph has to be derived (see Figure 3.7)
and the table below. The ordinates of the one-hour unit hydrograph are presented in column
2, whereas in the columns 3 and 4 the ordinates of the one-hour unit hydrograph are
presented with respectively 1 hour and 2 hours offset. Then for the corresponding times the
ordinates are summed, column 5, and finally divided by 3 to get the three-hour unit
hydrograph. What is the time base or base length Tb of the three-hour unit hydrograph.

Table (for example 3.6) Conversion of u (1,t) to u (3,t)

Time Ordinates of 1-hour 1-hour UH-s displaced by Sum of 3-hour UH

hours UH (/s)/mm 1h 2h ordinates ordinates (x1/3)
(1) (2) (3) (4) (5) (6)
0 0 0 0
1 50 0 50 17
2 140 50 0 190 63
3 320 140 50 510 170
4 750 320 140 1210 403
5 1100 750 320 2170 723
6 1050 1100 750 2900 967
7 950 1050 1100 3100 1033
8 830 950 1050 2830 943
9 700 830 950 2480 827
10 600 700 830 2130 710
11 440 600 700 1740 580
12 330 440 600 1370 457
13 250 330 440 1020 340
14 185 250 330 765 255
15 140 185 250 575 192
16 103 140 185 428 143
17 77 103 140 320 107
18 58 77 103 238 79
19 43 58 77 178 59
20 33 43 58 134 45
21 24 33 43 100 33
22 13 24 33 75 25
23 0 13 24 55 18
24 0 13 31 10
25 13 4

7
S-curve: the S-curve is the hydrograph of runoff of continuous rainfall of intensity i e = 1/T1
(e.g. mm/hr). To derive the S-curve assume a T-hour unit hydrograph with non-zero
ordinates:
u1, u2, u3, …., un. The base length Tb of this hydrograph is (n+1)T. The S-curve is obtained
by superposition of n T-hour unit hydrographs as shown in Fig.3.8. The maximum is reached
after n time periods of T hours. This maximum is equal to Q e, i.e. the equilibrium discharge
= ie.A: (3.31)
Where: Qe = equilibrium discharge in m3/s
T1 = unit storm duration in hours
A = catchment area in (km)2

Figure 3.8 S-curve derivation

The S-curve is computed using the following scheme:

S1 = u1
S2 = u1 + u2 = u2 + S1
S3 = u1 + u2 + u3 = u3 + S2
..
.
Sn = u1 + u2 + u3 + … + un = un + Sn-1

So, generally:

Si = ui + Si-1, for i = 1,…, n

Si = Si-1, for i > n

8
The T2-hour unit hydrograph is obtained from the difference between two S-curves distanced
T2 hours apart, corrected for the effective rainfall intensity as follows. Since the S-curve
refers to continuous rain of 1/T1 mm/hr, the difference between the S-curves T2 hours apart
refers to a rain intensity i1 = 1/T1, which multiplied with the duration T2 will create a depth
1. A rainfall with duration T2 requires an intensity i2 = 1/T2 to give unit depth. Hence, the
S-curve difference has to be multiplied with the ratio i2/i1 = (1/T2)/(1/T1) = T1/T2 to get a unit
rain depth in T2 hours. Hence u(T2,t) follows from:

(3.32)
Note that the base length follows from: Tb2 = Tb1 – T1 + T2. The procedure is shown in Fig.
3.9

Figure 3.9 u (T2,t) from S(T1,t)-curve differences

Example 3.7: in the table below an example is worked out by using the S-curve technique to
derive a 12-hour unit hydrograph from a 4-hour unit hydrograph. Note that the offset
between the S-curves is 12 hours. The difference between the two curves is presented in
column 6 and the u(12,t) in column 7: u(12,t) = (4/12) * column 6.

Table (for example 3.7) u(12,t) from S(4,t)-curve difference

Time Ordinates of S-curve S-curve ordinate S-curve lagged S Col.6 * = u(12,t)
(h) u(4,t) addition (Col.2 + Col.3) by 12 h (m3/s) (Col. 4 –
(m3/s)/mm (m3/s) (m3/s) Col. 5) ordinates (m3/s)/mm

(1) 2 3 4 5 6 7
0 0 - 0 - 0 0
4 20 0 20 - 20 6.7
8 80 20 100 - 100 33.3
12 130 100 230 0 230 76.7
16 150 230 380 20 360 120.0
20 130 380 510 100 410 136.7
24 90 510 600 230 370 123.3
28 52 600 652 380 272 90.7
32 27 652 679 650 169 56.3
36 15 679 694 600 94 31.3
40 5 694 699 652 47 15.7
44 0 699 699 679 20 6.7
48 699 699 694 5 1.7
54 699 699 0 0

Example 3.8: using the 4 -hour unit hydrograph from the table above the ordinates of the 2-
hour unit hydrograph are computed in the table below with the help of the S-curve technique.
First by some interpolation u(4,t)- values are estimated at 2 hours intervals, column 2. Next
the S-curve computation technique is applied to the columns 2, 3 and 4. Note that the
elements of column 4 are brought foreword 4 hours (and not 2 hours) in accordance with S-

9
curve computational algorithm presented before. Next the S-curve difference is computed
from the difference between column 4 and 5, where column 5 is offset by 2 hours. The
difference between the two curves is presented in column 6 and u(2,t) in column 7: u(2,t) =
(4/2) * column 6.
Table (for example 3.8) u(2,t) from S(4,t)-curve difference
Time Ordinate of S-curve S-curve ordinate S-curve lagged S U(2,t) ordinate
u(4,t) addition Col. (2) + (3) by 2 h Col. (4) – Col. (6) x
(m3/s) (5)
(h) (m3/s) 3
(m /s) (m3/s) (m3/s)
(m3/s)
Col-1 Col-2 Col-3 Col-4 Col-5 Col-6 Col-7
0 0 0 0 0
2 8 8 0 8 16
4 20 0 20 8 12 24
6 43 8 51 20 31 62
8 80 20 100 51 49 98
10 110 51 161 100 61 122
12 130 100 230 161 69 138
14 146 161 307 230 77 154
16 150 230 380 307 73 146
18 142 307 449 380 69 138
20 130 380 510 449 61 122
22 112 449 561 510 51 102
24 90 510 600 561 39 78
26 70 561 631 600 31 62
28 52 600 652 631 21 42
30 38 631 669 652 17 34
32 27 652 679 669 10 20
34 20 669 689 679 10 (20)15
36 15 679 694 689 5 (10)10
38 10 689 699 694 5 (10)6
40 5 694 699 699 (0) (0)3
42 2 699 701 699 (2) (4)0
44 0 699 699 701 (-2) (-4)0
Final adjusted values are given in Col. 7. Unadjusted values are given in parentheses.

Note: The duration of a unit hydrograph depends on the size of the catchment. It is recommended that the
duration T should not exceed 1/5 to 1/3 of the basin lag or lag time! For a catchment area of 250 (km) 2, T=6
hours is satisfactory.

Instantaneous Unit Hydrograph (IUH)

The main advantage of IUH is that it is independent of the duration of excess rainfall
hyetograph and thus has one parameter less than a TUH. This fact and the definition of IUH
make it eminently suitable for theoretical analysis of rainfall excess-runoff relationship of a
catchment. For a given catchment, IUH being independent of rainfall characteristics, is
indicative of the catchment storage characteristics.

10
Derivation of IUH: the relationship between the S-curve and the IUH is derived as follows.
1
Let the S-curve derived from a TUH be S 1. In this the intensity of rainfall excess, i e =
T
mm/h. Let S2 be another S-curve of intensity ie mm/h. If S2 is separated (or lagged) from S1
by a time interval dt and the ordinates are subtracted, a Direct Runoff Hydrograph (DRH)
dt
due to a rainfall excess of duration dt and magnitude i edt = is obtained. A unit
T
hydrograph of dt hours is obtained from this by dividing the above DRH by i edt. Thus the dt-
 S2  S1 
h UH will have ordinates equal to   . As dt is made smaller and smaller, i.e. as dt 
 ie dt 
0, an IUH results. Thus for an IUH, the ordinate at any time t is
 S2  S1  1 dS
u(t) = dtlim   (3.33)
 0  i . dt  ie dt
e

If ie = 1, then where S' represents a S-curve of intensity 1mm/h. Thus the ordinate
of an IUH at any time t is the slope of the S-curve of intensity 1mm/h (i.e S-curve derived
from u(1,t).

Derivation of T-hour Unit Hydrograph from IUH from the above equation, dS'= u(t) dt

and Integrating between two points 1 and 2:

If u(t) is essentially linear within the range 1-2, then for small values of t = (t2 - t1), by
taking 

But = ordinate of a UH of duration T1 = (t2 - t1). Thus, in general terms, for small

values of T1, the ordinates of a T1 -hour unit hydrograph are obtained by the equation

(3.34)

Thus if two IUHs are lagged by T 1 - hour where T1 is small and their corresponding ordinates
are summed up and divided by two, the resulting hydrograph will be a T 1 -hour UH. After
obtaining the ordinates of T-hour UH from the above equations, the ordinates of any T 2 -hour
UH can be obtained by the superposition method or S-curve method. From accuracy
considerations, unless the limbs of IUH can be approximated as linear, it is desirable to
confine T1 to a value of 1-hour or less.
Example 3.9: the coordinates of the IUH of a catchment are given below. Derive the DRH
for this catchment due to a storm of duration 4-hours and having a rainfall excess of 5cm.

11
Time (hrs) 0 1 2 3 4 5 6 7 8 9 10 11 12
IUH ordinates 0 8 35 50 47 40 31 23 15 10 6 3 0
U(t) m3/s

Solution: the calculations are performed in table below.

1- First, the ordinates of 1-h UH are derived as follows: col. 2 = ordinates of given IUH =
1
u(t), col. 3 = ordinates of IUH lagged by 1-h, and col.4 = (col. 2 + col. 3) = ordinates of
2
1-h UH.

2- Using the 1-hour UH, the S-curve is obtained and lagging it by 4-hours the ordinates of
4-h UH are obtained. col. 5 = S-curve additions, col. 6 = (col.4 + col.5) = S-curve
ordinates, col. 7 = col.6 lagged by 4-hours = S-curve ordinates lagged by 4-h, col. 8 =
(col. 6 - col. 7) = ordinates of a DRH due to 4cm of excess rainfall in 4-hours, and col. 9
=
col.8  ordinates of 4-hour UH.
4

3- The required DRH ordinates due to 5.0 cm excess rainfall in 4-hours are obtained by
multiplying the ordinates of 4-h UH by 5.0; col. 10 = (col. 9)x 5.0 = ordinates of required
DRH.

[Note: Calculation of 4-hour UH directly by using T 1 = 4h will lead to errors as the

assumption of linearity of u(t) during T1 may not be satisfied.]

12
Time u(t) u(t) lagged ordinates of S-curve S-curve S-curve DRH of 4cm Ordinates of DRH due to
by 1-h u(1,t) additio ordinate lagged by 4- in 4-h u(4,t) 5cm ER in 4-h
[(2)+(3)]/2 n h [6] - [7] [8]/4 [9]*5
(h) (m3/s) (m3/s) (m3/s) (m3/s) (m3/s) (m3/s) (m3/s)
(m3/s) (m3/s)
1 2 3 4 5 6 7 8 9 10
0 0 - 0 - 0 0 0 0 0
1 8 0 4 0 4.0 0 4.0 1.0 5.0
2 35 8 21.5 4.0 25.5 0 25.5 6.38 31.9
3 50 35 42.5 25.5 68.0 0 68.0 17.0 85.0
4 47 50 48.5 68.0 116.5 0 116.5 29.13 145.7
5 40 47 43.5 116.5 160.0 4.0 156.0 39.0 195.0
6 31 40 35.5 160.0 195.5 25.5 170.0 42.5 212.5
7 23 31 27.0 195.5 222.5 68.0 154.0 38.6 191.1
8 15 23 19.0 222.5 241.5 116.5 125.0 31.3 156.3
9 10 15 12.5 241.5 254.0 160.0 94.0 23.5 117.5
10 6 10 8.0 254.0 262.0 195.5 66.5 16.6 93.1
11 3 6 4.5 262.0 266.5 222.5 44.0 11.0 55.0
12 0 3 1.5 266.5 268.0 241.5 26.5 6.6 33.1
13 0 0 268.0 268.0 254.0 14.0 3.5 17.5
14 268.0 268.0 262.0 6.0 1.5 7.5
15 268.0 268.0 266.5 1.5 0.4 1.9
16 268.0 268.0 268.0 0 0 0