Methods for Developing IDF Curves

Intensity Duration and Frequency (IDF) curves have been found to be very helpful in the
estimation of design flows for the design of hydraulic structures like culverts and sewers. IDF
curves show the intensity of rainfall for a given duration and expected frequency. From the
curves, intensities of given duration at required return periods are are obtained and used in
estimating peak flows.
a) Watkins and Fiddes
The procedure adopted was as follows:
(i) Rainfall data for the selected urban centres was obtained from the Meteorology
Department. These location of urban centres are shown in figure 1. The collection of
data involved extraction of rainfall depths together with duration of occurrence. The
maximum rainfall depth in every month was considered from which the maximum in
every year selected. Rainfall depths, with duration between 15 and 210 minutes were
selected

(ii) For each town, daily maximum rainfall values together with the duration for the available
period, were read from the cards and ranked from 1 to N ( the number of years of
record) in a decreasing order.

(iii) The corresponding return period was estimated for each data set, using Weibull’s
plotting position formula1:

T = (N+1) (2)
M
where M is the event rank number (1, 2, …, N).

(iv) The maximum rainfall depths were plotted against return periods on a semi-log paper
and a line of best fit plotted through the points.

(v) The correlation coefficient between the maximum daily rainfall and log of return period
was determined.

(vi) From the graphs, maximum rainfall depths at desired return periods were read off from
the graphs.

(vii) The desired sequence of duration ( e.g. 10, 15, 30 minutes) was chosen for each data
set and used to calculate the rainfall intensity using the basic mathematical form of the
intensity-duration-frequency curve represented by the rectangular hyperbola:

I = a (3)
(t+b)n
where I is the intensity (mm/hr), t the duration (hours) and a, b, n are
coefficients/constants developed for each IDF curve.

(viii) The intensities were plotted against duration for each desired return period to give the
IDF curves.
The following procedure was adopted for the determination of coefficients a, b and n in the
formulae used.

(i) A value of b from other studies in the region; (for East Africa, b = 1/3) was applied.

(ii) The effective duration of storms in the area, teff was estimated. This is the length of time
over which 60% of the days rainfall occurs. This can be done by either judgement,
discussion with local experts or a review of local storm data. The latter option was used
in this study.

(iii) The value of n from the original Watkins and Fiddes equation was calculated:

ln(14.4)
n = teff . (4)
ln b+24
b+teff

(iv) From the graphs in figure 2, the maximum daily rainfall for a particular return period R T24
can be determined. This is divided by 24 to determine the daily rainfall intensity for a
particular return period iT24:

iT24 = RT24 (5)

24

(v) The value of aT for each return period was calculated using the formula:

aT = iT24 x (b+ 24)n (6)

The value of aT and other constants b and n were applied in equation (3) to determine
the intensity in mm/hr for a particular duration and return period.

(b) Bells Method

Bells method computes the intensity as a function of duration and the hourly
precipitation for any given return period by the relation:

iTt = (0.54t 0.24 - 0.50) PT60 (7)

t

where iTt = the rainfall intensity (in mm/hour) of return period T and duration t minutes and P T60
is the 60 minute of return period T. It can be applied to storms of less than two hours duration
where hourly totals are available.

The following procedure was adopted.

a) Based on the data available for each station the hourly intensities were calculated and
were ranked according to Weibull formula.
b) A semilog plot of return period verses hourly intensities in mm/hr was made.

c) The correlation coefficient between the log of return period and the hourly intensities was
determined.

d) The 60-minute intensities for given return periods were calculated from the graph.

e) The intensities for other duration were applied to Bells formula to obtain a set of IDF
Curves.

In a study (Rugumayo et al, 2002) on data for different stations in Uganda observes that there is
a good correlation between log of return period and the daily maximum rainfall, which is easier
to measure accurately, whereas a poor correlation between the log of return period and hourly
rainfall intensities on which, Bells method is based. In the absence of data, only estimates of
hourly rainfall can be made from daily rainfall. Furthermore, the original Bells equation was
based on North American catchments, which have less thunderstorm days than tropical
thunderstorms. The approach of Watkins and Fiddes method is based on the determination of
constants a, b and n in the basic equation describing IDF curves in which, intensity is a function
of duration at any given return period, T.
Developing IDF Curves:

Example

Thus;
a
itT 
t  b n
where itT is the intensity(in mm/hr) of t hours duration with a return period of T years. Watkins
and Fiddes suggest the following procedure for the determination of constants:
(i) Start with a value of b from other studies in the region (b=1/3 for East Africa)
(ii) Estimate the effective duration of the area, teff.
(iii) Given teff compute the value of n from the original Watkins and Fiddes equation as
n  I 14.4n teff  Inb  24 b  teff 
Finally the value of a can be computed from the equation a T  i24 xb  24 where i 24
T n T
(iv) is
the rainfall intensity (mm/hr) of duration 24 hours with return period T years. It can be
estimated as i24 T
 R24
T T
24 where R 24 is the maximum 24-hour rainfall with a T-year return
period.
Using the daily maximum rainfall and corresponding duration of Kabale develop the IDF curves
for the area.

Year Maximum Daily Rainfall Duration (hours)

1967 12.3 1.1
1968 12.2 0.4
1969 25.6 1.5
1970 39.1 3.5
1971 233 0.9
1972 55.1 2.3
1973 44.2 1.5
1975 26.7 1.5
1976 11.3 0.7

Take b to be 1/3 and a review of local data shows that teff (60% of the day rainfall) for Kabale
can be taken as 1.5 hours. (Hint: First rank the maximum daily rainfall and determine the return
period (or frequency) T from T  N  1 r where N is the number of years of record and I is the
rank).

Solution
a) Calculation of the coefficient n:
B=1/3(value for East Africa)
teff= 1.5 hours

Therefore
n  In14.4 1.5 In1 3  24 1 3  1.5
b) Calculation of coefficient a
(i) Calculate the return periods for Maximum daily rainfall
Year Maximum daily Duration Rank r Return period
rainfall (mm) (hours) T  N  1 r
1972 55.1 2.3 1 10.0
1973 44.2 1.5 2 5.0
1970 39.1 3.5 3 3.3
1975 26.7 1.5 4 2.5
1969 25.6 1.5 5 2.0
1971 23.3 0.9 6 1.7
1967 12.3 1.1 7 1.4
1968 12.2 0.4 8 1.3
1976 11.3 0.7 9 1.1

(ii) Plot rainfall depths against return period

Daily Maximum rainfall Vs Return Period

140
120
Daily Maximum
Rainfall(mm)

100
80
60 y = 4.9092x + 12.319
40
20
0
1 3 5 7 9 11 13 15 17 19
Return Period(hours)

(iii) Read off from the line, at desired return periods depths (Max. 24 hour rainfall) and
calculate Max.24 –hour intensity and the coefficient a:
T
For T=2 years, Max. 24 hour rainfall R24 =23.8mm
T
i24  R24
T
/ 24  28.8 / 24  0.99

Hence a T  i24 xb  24  0.99x1 / 3  24  13.16

T n 0.81

Return Period Max.24-hour Rainfall (mm) Max. 24 hour aT

T (years) Intensity
(mm/hr)
1 9.12 0.38 5.04
2 23.80 0.99 13.16
5 43.20 1.80 23.88
10 57.87 2.41 32.00
25 77.28 3.22 42.72
50 91.95 3.83 50.84
100 106.63 4.44 58.95
c) Calculation of intensity and plotting of IDF curves
Calculate the intensity for each desired duration, with the given return period T years.

For T=2 years, a=13.16

Using t=10 minutes
Intensity I  a t  bn  13.6 10 / 60  1 / 30.81
=23.06mm/hr

Return 1 2 5 10 25 50 100
Period T
(years)
Duration
t (min)
10 8.84 23.06 41.87 56.10 77.54 89.13 103.35
20 7.00 18.27 33.17 44.44 60.54 7.060 81.87
30 5.84 15.25 27.68 37.09 49.97 58.93 68.33
45 4.72 12.33 22.38 29.99 39.88 47.64 55.25
50 4.45 11.61 21.08 28.24 37.42 44.87 52.03
60 3.99 10.42 18.92 25.35 33.36 40.27 46.70
75 3.47 9.07 16.46 22.05 28.77 35.04 40.63
90 3.08 8.05 14.62 19.58 25.37 31.11 36.08
120 2.54 6.62 12.02 16.11 20.61 25.59 29.68

IDF Curves