EECQ 4141: Irrigation and Drainage Engineering I

Irrigation water requirement:

Rainfall
Precipitation is water originally contained by the air and reaching the earth surface in the form of rain,
drizzle, snow, hail fog condensation, or frost. In areas with irrigated agriculture, rainfall is the main
component of precipitation.
Rainfall distribution is highly variable in both space and time. Over large areas covering countries and
continents the times of year when rainfall is received, and the amount received varies. This forms the
basis of classification of climates over the global ranging from deserts through semi-arid to humid
climate (See the following table).
Table. Major climatic zones
Climate Annual rainfall Wet period
(mm) (months)
Desert <100 0-1
Arid 100-400 1-3
Semi-arid 400-600 3-4
Sub-humid 600-1200 4-6
Moist sub-humid 1200-1500 6-9
Humid >1500 9-12

Irrigation is needed when the amount of precipitation going into the root zone is not enough to meet
the water requirements of crops. The ‘wet period’ in in the previous tables shows the months of the
year when irrigation will not be needed in each of the climatic zones.
The total rainfall received in a given time period at a location varies from year to year and therefore,
rather than using mean rainfall data a dependable level of rainfall should be selected. For instance, the
dependable rainfall may be the rainfall, which can be expected in 8 out of 10 years (80%). In order to
estimate such dependable rainfall levels, a statistical analysis needs to be made from long-term rainfall
records. Dependable rainfall can only be determined through calculation of the rainfall probability or
frequency. The frequency distribution provides the most complete description of the nature of the
variation in rainfall records. Methods of computing rainfall probabilities are not given in this course
but the student is requested to consult a statistical hand book for hydrology.
For preliminary planning purposes, monthly rainfall data can be used. When designing irrigation
systems and irrigation schedules however, monthly intervals are too long, as they may hide dry spells

1
 EECQ 4141: Irrigation and Drainage Engineering I

of one or two weeks, which can be highly critical for the development and yield of some crops. Decade
(10-day) or weekly rainfall data should therefore be used.

Definitions
Historical or actual rainfall data
The historical rainfall data is the actual recorded rainfall during a specified period. Real time irrigation
scheduling requires this data in order to determine the timing of the next irrigation. The data is also
useful for evaluating purposes.
Average or normal rainfall
Average or normal rainfall is the arithmetic mean derived from a record of several years of historical
rainfall data. Although the monthly or annual average rainfall is commonly available, the average
rainfall appears only as information to indicate the ‘normal’ amount of rainfall one can expect. When
compared with actual amount of rainfall, it gives an idea of the departure of the actual amount from
the normal value.
Dependable rainfall
Dependable rainfall is defined as the rainfall, which can be, expected a set number of years out of a
total number of years. For instance, the dependable rainfall may be the rainfall, which can be expected
in 9 years out of each 10 years (90%). The percentage (90%) gives the probability, that the rainfall will
be obtained or exceeded i.e. the probability that the actual rainfall will be equal to or higher than the
dependable rainfall. One year out of 10, the rainfall amount will be smaller.
The determination of the probability level is related to the risk, one wants to accept. In the case of
expensive structures such as bridges or dams and intakes in rivers one may want to restrict the risk,
that the rainfall (causing the flood discharge) will exceed a certain value, to once in 50 or once in 100
years. The corresponding probabilities of exceedance here are 2% and 1% respectively.
For agriculture the risks involved are the reduction in or the loss of the yield once in so many years.
The selected dependable level of rainfall is the depth of rainfall that can be expected 3 out of 4 years
or 4 out of 5 years. The probabilities of exceedance are respectively 75% and 80%. This ‘minimum’
rainfall is used as a design norm for the dimensioning of irrigation system as well as for water
management. A higher level of dependable rainfall (say 9 out of 10 years) may need to be selected
during the period that crops are germinating or are most sensitive to water stress and yields are
severely affected.
Dry, normal and wet conditions
For management and planning purposes the information on the amount of rainfall which one can
expect in a specified period under dry, normal and wet conditions is important. The period under
consideration may be any period such as a week, a decade, month or year.
A period is dry if the rainfall received during that period will be exceeded 4 out of 5 years, i.e. having
a probability of exceedance of 80%. This information is of importance to ascertain if rainfall is sufficient
to support rainfed agriculture or if supplementary irrigation is needed. When calculating the capacity

2
 EECQ 4141: Irrigation and Drainage Engineering I

of storage reservoirs and of main canals, the rainfall amount with a 80% probability of exceedance is
generally used.
The rainfall in a period is normal, if the rainfall received during that period will be exceeded 1 out of 2
years, i.e. each other year. The probability of exceedance is equal to 50% and the rainfall corresponds
to the average or mean rainfall.
A period is wet if the rainfall received during that period is only exceeded 1 out of 5 years, i.e. having
a probability of exceedance of 20%.
The three values are useful for the programming of irrigation supply and simulation of irrigation
management conditions. The rainfall amount which one can expect with 20, 50 and 80% probability in
each of the decades of rainy period of Moi University Main Campus are given as an example in the
following figure.

Fig. Mean monthly reference evapotranspiration and dependable rainfall amounts for each of the
months of the rainy period in Moi University Main Campus with a 20, 50 and 80% probability
of exceedance representing a wet, normal and dry month
Effective rainfall
The effective rainfall is defined as that part of rainfall which is effectively used by the crop after rainfall
losses due to surface run off and deep percolation have been accounted for. The effective rainfall is
the rainfall ultimately used to determine the irrigation requirement of the concerned crop.

Frequency analysis
Dependable rainfall can only be determined through calculation of the rainfall probability or
frequency. This is normally done through the use of statistical methods. The problem with rainfall
however is that it seldom satisfies the statistical theories on which these methods are based. For
instance (extremely) heavy rainfalls and (extremely) low rainfalls may have different statistical
probabilities than ‘normal’ rainfalls. Also, in mountainous areas local climatic differences are great and
data may show large departure from the mean. Most methods give similar results near the middle
(around 50%) but different results for low and high values (extremes).

3
 EECQ 4141: Irrigation and Drainage Engineering I

Another problem arises from the fact that the frequency distribution obtained is considered
representative for frequencies in the future and is therefore being used to predict the probability of
certain rainfall. However, the frequency distribution has been obtained from series of rainfall data over
limited period and this time period does not contain all (present and future) rainfall depths.
Now, if all rainfall depths form the population, the frequency distribution based on the sample data
available forms only an estimate of the probability distribution belonging to the population. The longer
the period of observation the more similar the frequency distribution will be to the probability
distribution (assuming that the climate will not change in the future). A period of 30 years and over
normally is thought to be satisfactory.
The procedure of frequency analysis based on depth ranking and assuming a normal distribution is as
follows:
Step 1: Rank the (𝑛) data (𝑃𝑖 ) in a descending order, the highest value first and the lowest value last.
Step 2: Attach a serial rank number (𝑟) to each value (𝑃𝑖 ) with 𝑟 = 1 for the highest value (𝑃1 ) and 𝑟 = 𝑛
for the lowest value (𝑃𝑛 ).
Step 3: Calculate the frequency of exceedance 𝐹(𝑃 > 𝑃) as:
Method Frequency of exceedance
r
C alifornia
n
r − 0.5
H azen
n
r
W eibull
n+1
r − 0.44
G ringorten
n + 0.12
The frequency of exceedance corresponds with the plotting position on the probability scale of the
probability paper. The Weibull and Gringorten plotting positions are theoretically better sound. All
four relationships give similar values near the centre of distribution, but may vary considerable at the
extreme ends.
Step 4: Plot the data on normal probability paper
After ranking the data and calculating the frequency of exceedance, the calculated frequencies of
exceedance are plotted on normal probability paper. If the plotted data fall in a reasonable alignment,
it can be assumed that the data can be approximated by the assumed normal distribution. Since it
would be rare for a set of data to plot exactly on a line, a decision must be made as to whether or not
the deviations from the line are random deviations or represent true deviations indicating the data
does not follow the given probability distribution.
After fitting a straight line through the points, the magnitude of rainfall corresponding to various
probabilities is derived from the probability plot.

4
 EECQ 4141: Irrigation and Drainage Engineering I

If the points in the probability plot do not fall in a reasonable alignment, it often means that the data
is not distributed as the selected normal distribution. To convert the distribution to normality,
transformation can be attempted by plotting on the normal probability paper, the square root or
logarithm of the same rainfall data. The transformed data might be closer to the normal distribution.

Effective rainfall
Infiltration and soil water storage
Effective rainfall in its simplest sense means “useful” or “utilisable” rainfall. Rainfall is not necessarily
useful or desirable at the time, rate or amount in which it is received. Some of it may be wasted while
some may even be destructive.
Rainfall may evaporate in the atmosphere, be intercepted by vegetation or it may reach the soil, where
it either runs off over the surface or infiltrates into the soil. Infiltrated rainfall may join the ground
water and be lost as groundwater flow. Thus, a part of the rainfall will be lost through surface run-off
and percolation below the root zone and will not be available for evapotranspiration.
Run-off will depend on the intensity and duration of rainfall, on its timing (related to earlier rainfall),
on soil characteristics (texture, structure), on the slope of the terrain and on the vegetation cover.
Deep percolation will depend on initial soil moisture content, soil water holding capacity (soil type)
and depth of the rooting zone (crop type).
Effective rainfall may be defined as the fraction of rainfall, stored in the root zone and used by plant-
soil system for evapotranspiration.
This is purely technical definition: for a scheme manager rainfall will only be effective when one or
more irrigation gifts can be omitted. Also, rainfall, which falls on fallow land, will only be considered
effective in as far as it will be stored in the soil for use by the next crop.
For a farmer, rainfall which leaves his field as run-off is not effective, although it may be effective to
his neighbour receiving this runoff or to farmers further away who may re-use the water for their
irrigation scheme by deriving it from the river.
Furthermore, the decrease in radiation and temperature and thus in evapotranspiration equally might
be considered a useful effect of rainfall.
Thus, it is always important to consider in which setting and from whose point of view the effectiveness
of rainfall is considered. For agricultural purposes rainfall may be considered effective in as far as it
satisfies water needs for land preparation, consumption by crops, salt leaching requirements and/or
percolation needs (as for rice or fishponds).
Estimation procedure
Although run-off and deep percolation can be measured it is not practical to do this for all
meteorological stations or scheme areas as these measures are complicated, time consuming and
expensive.

5
 EECQ 4141: Irrigation and Drainage Engineering I

A number of empirical formulae have been developed for the estimation of effective rainfall. Most of
these formulas take into account total precipitation and consumptive use or (potential)
evapotranspiration.
Fixed percentage of rainfall
The effective rainfall (𝑃𝑒𝑓𝑓 ) is calculated as a fixed percentage (a) of total rainfall (𝑃𝑡𝑜𝑡 )
𝑃𝑒𝑓𝑓 = 𝑎. 𝑃𝑡𝑜𝑡 𝑓𝑜𝑟 𝑎 = 0.7 𝑡𝑜 0.9
This method is rather crude, and results will not be very satisfactory, as -among others- the influence
of the depth of rainfall is disregarded.
USDA SCS method
The Soil Conservation Service of the U.S. Department of Agriculture (USDA SCS) has developed a
procedure for estimating effective rainfall as a function of the evapotranspiration/precipitation ratio
by processing precipitation and soil moisture data for a period of 50 years for 22 experimental stations
representing different climatic and soil moisture data. The results are given in the following table.
Table. Average monthly effective rainfall as related to Average Monthly 𝐸𝑇𝑐𝑟𝑜𝑝 and Mean Monthly
Rainfall (USDA SCS, 1969)

The average monthly effective rainfall is based on a net water storage depth of 75 mm. Correction
factors to be used for different net water depth storage are: