Keynote Lectures

Drainage in arid regions

J.H. Boumans
Euroconsult, Arnhem, The Netherlands

1 Introduction
The need for drainage to control the watertable and soil salinity in irrigated arid land
is now generally accepted. 30 years ago this was not the case. In 1955 for instance
people in Iraq still doubted whether drainage was the right answer to combat saliniza-
tion and to reclaim saline land. It was for that reason among others, that in 1955
the Dujailah Drainage Experiments in Iraq were started (ILRI 1963).
When discussing drainage, distinction has to be made between drainage of ground-
water and drainage of surface water. Drainage of irrigated arid land refers primarily
to groundwater drainage (needed for salt control) but surface drainage may also be
required to remove excess rain or excess irrigation water, especially for soils with low
infiltration rates. Surface drainage is further needed if rice is grown. Shallow surface
drains can also serve to leach out salts by surface as well as subsurface flow. For
instance, with ditches of only 0.4 m depth the farmers in Egypt succeed in leaching
their new salty land. If rice is grown in rotation with dry foot crops, such as on the
old land of the Egyptian Nile Delta, subsurface drainage, required for groundwater
control of the non-rice crops, implies the risk of excessive percolation losses in periods
that rice is grown. The solution worked out in Egypt is to grow rice in drainage units
in which the watertable can be backed up by means of a closing device at the drainage
outlet.
Groundwater drainage can, apart from with horizontal drains, be achieved with
pumped vertical wells. Vertical drainage requires the presence of a permeable aquifer.
Vertical drainage may be an attractive solution, especially if the pumped drainage
water is of good quality, so it may be re-used for irrigation. In that case the required
pumping for drainage should be integrated into a regional programme for optimal
management of surface waters for both irrigation and drainage purposes.
In arid lowland areas groundwater is often saline and as a rule salinity increases
with depth. Deep vertical drainage wells will produce highly saline water unfit for
re-use and difficult to dispose of. In that situation a solution called ‘skimming drain-
age’ may be considered. Skimming drainage intends to drain the least salty upper
water only, and can be achieved with shallow low capacity wells but much better with
horizontal drains.
There are three successive phases involved in implementing a drainage scheme:
1. Surveys and studies;
2. Design;
3. Construction.
Construction, which in fact is very important and is probably the most critical phase

22 .
 Keynote Lectures

for the success of the drainage operation, will be discussed in a separate paper.
This paper deals mainly with survey and design aspects of horizontal pipe drainage
of irrigated, non-rice cropped arid land. After a comparison of the drainage in arid
and humid climates and a brief review of the past and present drainage activities in
arid regions, some selected items regarding technological developments, lay-out op-
tions, design criteria and monitoring are discussed.

2 Drainage for salinity control

Groundwater drainage in irrigated arid and semi-arid regions is in many aspects differ-
ent from drainage in humid areas. The aim of drainage in humid temperate areas
is to control soil water conditions for better aeration, workability, and temperature
regime. The primary objective of draining irrigated arid land is to control soil salinity.
This may be a subject of debate, but is easily understood if we consider the hypoth-
etical situation of irrigation in an arid region, where neither the irrigation water nor
the groundwater contains any salt. In that case drainage is not needed to leach and
evacuate salts. Nor it is needed to avoid waterlogging as groundwater depth and soil
water conditions can be adequately controlled by proper irrigation. A watertable too
near to the surface can be corrected by reducing the irrigation amount or frequency,
after which, in a dry hot climate the watertable will drop rapidly due to evapotranspira-
tion. This hypothetical situation corresponds with the dry season irrigation in regions
with a tropical monsoon climate. Because of excess rain in the wet season, neither
soil nor groundwater are saline. In those areas, groundwater drainage is practically
not existent. There is apparently no need for it.
Apart from this fundamental difference in the drainage objective between irrigated
arid and rainfed humid areas, there are many other aspects in which drainage of arid
land differs from that of humid land. Arid regions are by definition characterized
by the absence of an active natural drainage system. For irrigation projects situated
in river valleys the river represents the main water supply for irrigation and domestic '
uses, and should not be contaminated with salty drainage water. Therefore, drainage
of irrigated land involves not only the installation of field and collector drains, but
also the construction of a drainage infrastructure.
This may eventually imply the need for constructing an extensive and costly main
outfall system to carry the drainage water from the irrigated area to the sea. Examples
of such man-made main drain networks are found in Egypt, Iraq and Pakistan. In
Egypt an enormous network of main and secondary drains has been constructed
throughout the delta to transport drainage water. In Iraq the main outfall system
from north of Baghdad to the Arabian Gulf has been designed and its lower section
is under construction. In Pakistan the construction of the Left Bank Outfall Drain
in the Indus Valley has been started. In Haryana State (India) where, as a result of
increased irrigation water supplies, waterlogging and soil salinization are becoming
a threat for agricultural production, the disposal of the saline drainage water is a not
yet solved problem. A main outfall would require 450 km of drain and 17 pumping
stations with a total lift of 100 m, and is not considered feasible. As long as there

23
Keynote Lectures

is no major drainage outfall system, temporary solutions are often applied, such as
disposal into evaporation ponds or, if drainage water is not too saline, into the river
or other irrigation systems.
Field drainage conditions in arid regions are also different from those in humid
regions. Drains have to be situated at a greater depth for salinity control and are .
usually spaced much wider apart (30 to 150 m) than in humid areas (8 to 30 m). Deeper
drains and wider spacing imply that soil investigations to determine the hydrological
parameters needed for design have to be carried out to a considerably greater depth
in arid regions than in humid regions. This requires application of more complicated
and costly investigation techniques and equipment.
Finally, the construction of a drainage infrastructure in arid and humid regions
is also different. The construction of a deep, widely spaced drainage network in arid
land requires larger pipes and heavier installation equipment and is more difficult
than that of a shallow, narrow spaced drainage system in humid areas. Trenchers
used in arid regions are in the power range of 200-300 kW compared to 100-200 kW
in humid regions.
Although the drainage objectives, approach, and construction in arid and humid
regions are different, the flow of groundwater to the drains is identical in both cases.
The same drainage formulae are used for the design of drainage schemes and the drain-
age requirements are formulated in similar terms as for drainage in humid temperate
regions.

3 Drainage activities in arid regions

The need for drainage in arid and semi-arid regions is directly linked to irrigation.
Archeological studies in Iraq (NN 1958) have shown that as early as 2400 B.C. water-
logging and salinization were causing yield reduction and were reasons to abandon
irrigated land in the Mesopotamian Plain. Drainage of excess surface water to depres-
sions and marshes was already applied at that time, but no indications have been
found of drainage with a view to lower the groundwater table and to leach out salts.
Still farmers succeeded at that time to keep salts low and to cultivate the land for
many centuries in succession. This was probably achieved by means of an adapted
farming system which, as described by Russel (NN 1958) consisted of long fallow
periods (summer-winter-summer) between two winter crops. During such fallow peri-
ods the watertable dropped to 2 m and deeper by transpiration of deep rooting weeds
such as shok (Proposis Stephaniana) and camel thorn (Alhagi Maurorum). This made
it possible to leach the salts accumulating in the surface layers to the subsoil at the
start of the next irrigated winter crop season. This practice resembles in some way
the farming method known as Niren system, applied in Iraq until recently.
The modern history of land drainage for irrigated land started in the USA around
the end of the last century. In 1886 Hilgard (Luthin 1957) had already noted the need
for drainage in the San Joaquin Valley of California. One of the first drainage schemes
for the control of watertable and soil salinity was probably constructed in the Pecos
Valley Irrigation Scheme (New Mexico), where in 1918, 1800 m of deep open drains

24
 Keynote Lectures

and 150 m of covered drains were completed (Euroconsult 1985). In 1950 large areas
of irrigated land in California and other states had already been pipe drained. Also
in sugar estates in Central and South America drainage, including pipe drainage to
control the watertable was applied long before 1950. Undoubtedly drainage for water-
table control was also applied before 1950 in the irrigated arid regions of Australia
and South Africa.
In Egypt drainage problems developed after the introduction of perennial irrigation
at the beginning of this century. As early as in 1892the British engineer Scott Mongrieff
(Wilcocks 1913) stressed the need for drainage which should receive priority above
extension of water supply to new areas. In 1952 a reported 50 O00 feddan were pipe
drained in Lower Egypt.
A general feature of the drainage before, say, 1945 was its empirical approach.
Drains were designed and constructed according to local experience, and later on in-
tensified, deepened or in another way adjusted whenever this appeared necessary. After
1945 drainage and land reclamation received a scientific base. In 1940 Hooghoudt
published his well-known analytical approach to the flow of groundwater to drains,
following which many other researchers turned their attention to this field. They con-
firmed, improved and extended Hooghoudt’s work, and drainage formulae for steady
and non-steady flow and for complicated multi-layered aquifer systems were devel-
oped. In 1957 the well-known handbook ‘Drainage of Agricultural Lands’ (Luthin
1957) was published and contained many contributions to drainage theory. In the
same period in different parts of the world research on the drainage and reclamation
of saline land was conducted. In 1948 Reeve et al. published a bulletin on reclamation
of saline alkali soils by leaching and in 1954 the US Salinity Laboratory Handbook
60 (USDA 1954) was published on the diagnosis and improvement of saline and alkali
soils. Basic and operational research on the reclamation and drainage of saline and
alkaline land was carried out in The Netherlands in connection with the reclamation
of the Zuiderzee, and the floods of 1945 and 1953 which inundated large parts of
the low-lying land with sea water. In 1955 operational research started in Dujailah
and other parts of Iraq to study the drainage and reclamation of extremely saline
and alkaline land (ILRI 1963).
After 1950 the application of pipe drainage in the arid regions rapidly expanded
in North Africa and the Middle East. In Egypt, as a result of the law of 1965, the
Government made itself responsible for the implementation of field drainage. A long-
term implementation programme, the largest in the world, was started and is still in
operation. In Tunisia the first pipe drains were installed about 1958 in the Medjerda
Valley. In Iraq all new or reconstructed irrigation schemes since 1950 have been pro-
vided with pipe drainage. The total area drained, still zero in 1950, now covers several
hundred thousands of hectares. The first pipe drainage in Morocco dates from about
1970, when the Gharb Valley Irrigation Project was implemented in the Sebou Basin.
In other Middle Eastern countries like Syria, Israel, Lebanon, Jordan and Iran, drain-
age of irrigated land was mainly introduced after 1965.
A different kind of development took place in Pakistan. Because of increased water
supplies for irrigation, waterlogging and salinization affected in creasingly more agri-
cultural land and after 1950 became a major danger to agriculture in the Indus Basin.

25
Keynote Lectures

In the early sixties with the assistance of United Nations and World Bank, large-scale
studies were undertaken and a number of SCARP's (Salinity Control and Reclamation
Projects) identified. Tubewell drainage appeared to be the best solution in most areas
due to a highly permeable aquifer and groundwater of good quality. The extent and
degree of the problem and the remedial measures in the Indus Basin were further
investigated during 1964 - 1966. The first SCARP with tubewell drainage was imple-
mented in this period, followed by many other tubewell SCARP'S. The first pipe drain-
age scheme - the East Khairpur Tile Drainage Project - started in 1978 in an area
unsuitable for tubewell drainage. A second pipe scheme is now under construction
and a third is in the preparation phase.
During the last decennia the drainage activities in the Middle East were concentrated
in the Nile Delta of Egypt, the Mesopotamian Plain of Iraq and the Indus Valley
in Pakistan, all areas with comparable climate, land slope and irrigation water quality.
The method of drainage, however, varies greatly in the three countries due to differ-
ences in soil conditions and cropping systems.
In the Nile Delta of Egypt the soils are heavy and poorly permeable and the aquifer
conditions are not suitable for vertical drainage. The land is practically permanently
irrigated without fallow periods, and rice is part of the crop rotation, with the result
that there is no serious salinization problem. The pipe drains in Egypt are therefore
installed at a relatively shallow depth (1.2 to 1.5 m) and, as compared to other regions,
narrowly spaced (30 to 60 m).
The Iraqi soils of medium to heavy texture offer better permeability, but aquifer
conditions are not suitable for well drainage, this is because the groundwater is saline.
Winter is the main cropping season and summer irrigation is limited which results
in a great part of the land being fallow in summer, and subject to salinization by
the capillary rise of saline groundwater. The pipe drains in Iraq are therefore installed
much deeper (1.7 to 2.2 m) and with wider spacings (60 to 150 m) than in Egypt.
In Pakistan soils are medium textured, well-permeable and underlain by a highly
permeable aquifer with often good-quality water. Cropping is semi-intensive (60 to
70% cultivation in summer and winter). Conditions in many areas are favourable for
tubewell drainage, and this is largely applied. As in Iraq the pipe drains are installed
deep with even larger spacings. Owing to the highly permeable aquifer, the natural
drainage conditions as a result of different topographies determine the need and degree
of (additional) drainage, more so than in Iraq and Egypt.
To end this review it may be concluded that since 1950 when, except for USA, drain-
age of irrigated land was practically unknown, much has changed. The need for drain-
age of irrigated land is generally accepted these days and large areas have been pro-
vided with drainage facilities. Future prospects include the continuation of activities
such as those in Egypt, Iraq and Pakistan and the expected start of new large scale
drainage programmes in countries like Turkey, Iran, India and China and in South
America.

26
 Keynote Lectures

4 Developments in design and construction technologies

Present principles and concepts for draining irrigated land are the same as 25 years
ago, but design methods and construction techniques have developed as a result of
the advanced techniques which were made available to the design and the construction
engineer. In this respect particular mention should be made of the use of the computer
for design work and the introduction of the flexible corrugated drainage pipe.

4.1 Computer Aided Design

In 1960 computers were scarce and mainly used for administrative purposes or scientif-
ic research. At that time computers were permanently installed in a central office and
only operated by specialized staff. Use of computers for drainage design had no prior-
ity and moreover was impractical as data had to be processed far away from the project
area. This lead to delays and made it impossible for the engineer to intervene during
the data processing stage. Although in the 70’s computers increased in number and
availability, there was still too great a distance both literally and figuratively, between
the engineer in the field and the computer in the main office. After 1980 the situation
rapidly changed. Programmable pocket calculators had been introduced, and portable
microcomputers became available as well as user-friendly software which opened the
way for every drainage technician to make use of these powerful instruments in a
direct interactive way. The use of computers for design rapidly grew. Computers were
used for field surveys, data processing, groundwater modelling, drain spacing calcula-
tions, and for detailed design of drainage networks including the preparation of maps,
drawings and cost estimates. The computer made it possible to improve the speed
and quality of the drainage studies and designs. Groundwater flow analyses and calcu-
lations of the salt-water balance, which had previously been considered too complicat-
ed and too costly for manual execution, now became possible. Two examples of com-
puter applications in drainage follow:
1. The aquifer of alluvial plains generally consists of stratified sediments of different
hydraulic conductivity. For drain spacing calculations the layered aquifer is usually
simplified, as shown in Figure 1, to one homogeneous layer (Hooghoudt model)
or sometimes to two layers (Ernst model), despite the fact that in 1971 Toksöz
and Kirkham (T-K) had already presented a correct solution for multi-layered soils.
The T-K solution, however, has found little application. It was unsuitable for man-
ual calculations and even the proposed graphic solution was too difficult. With
a computer, however, application of the T-K formula is very easy and there is no
reason at present not to apply the T-K approach for spacing calculations in strati-
fied soils. This is particularly true for drainage of irrigated land, where because
of large spacings, the groundwater flow to the drain penetrates deeply into the
stratified aquifer. The error which can result by using the one-layer model for a
three-layer aquifer is presented in Figures 1 and 2. In the one-layer approach the
second stratum with a permeability of only 1/10 of the layer above is taken as
the impervious barrier, and a drain spacing of 38 m is calculated for the hydraulic

27
Keynote Leclures

ONE LAYER MODEL TWO LAYER MODEL THREE LAYER MODEL

(HOOGHOUDT) (ERNST) (TOKSÖZ-KIRKHAM)
k S = 3 8 m d k ~ = 3 m 9- j ( kLS=68 4 -m

1 1 1 1 1 1 1
q=O.O03 mld
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
q=0.003 mld q=O.O03 mld
0.8
7F
D,LI m K,=0.5 mld

D314m K3=2 mld

I

Figures I and 2 One, two and three layer steady flow drainage models

head and discharge rate assumed. If calculated for a three-layer profile with the
T-K formula the result would be 68 m, almost twice as much as the one-layer model;
2. The second example refers to the question whether and to what extent the designed
system is able to reduce and control soil salinity. It is possible to calculate the leach-
ing requirement with one of the available formulae and to check whether this requir-
ement is met during the irrigation period. This, however, is a rather arbitrary meth-
od which does not take into account that the salt regime in the soil is a dynamic
process and that a lack of leaching in one period may be compensated by an excess
in other periods. A better approach is to analyse the salt balance in the soil through-
out the entire cropping cycle, which may be one or more years. The computer is
an excellent tool for these salt simulation studies. It analyses and processes rapidly
the complex water and salt movements in the soil. An example of a salt balance
analysis for a two-year crop rotation is presented in Figure 3.
These examples concern relatively simple computer applications. There are many
other possibilities with various degrees of complexity. Much basic research has already
been done in this field and is or soon will be available for practical application. At
the International Seminar on Land Drainage held in Helsinki in 1986, I 1 of the total
36 presented papers dealt with computer applications for drainage research and design.
Promising simulation models of the saturated soil water conditions throughout the
cropping cycle and depending on rainfall, evapotranspiration and drainage conditions,
are already available.
Developments in computer application in drainage will continue and proceed rapid-
ly. After a number of years the approach, the techniques, and the procedures of drain-
age design may be quite different to those of today. But common sense and good
field experience will remain essential tools for the new computerized designer as much
as they were for the drainage engineer in the past.

28
Keynote Lectures

5 Complete underground systems

The singular system (Figure 4-A) is the traditional method of draining flat land in
humid areas. The system consists of pipe drains which discharge directly into an open
collector ditch. The length of the pipe drain is usually not more than 200-300 m, as
it can then conveniently be cleaned with flushing apparatus. The main advantages
of this system are its simple lay-out and construction, the visibility of its performance
at outflow pipes, and its easy maintenance. However, the great length of open collector
drains implies fragmentation and loss of valuable land for agriculture. This also leads
to high excavation and especially maintenance costs and additional construction costs
for crossings with irrigation canals and roads.
The singular lay-out also initially served as a model for the drainage in irrigated
arid and semi-arid regions. But in view of the risk of salinization under these condi-
tions, the drains have to be installed at a much greater depth, and disadvantages count
much more than for shallow drainage in humid areas. This is shown in Table 1 for
the excavation costs and loss of land of required collector ditches. It holds even more
for the maintenance costs which also increase disproportionately with depth. More-
over, instability and collapse of the side slopes of deep ditches makes proper mainte-
nance a permanent and often almost unsolvable problem.
As open drains may need crossings with roads and irrigation canals, there is a strong
argument to design the drainage system in arid areas in such a way as to minimize
the length of open collector drains. This can be achieved in two ways: by using the
composite lay-out (Figure 4-B), in which the collector ditch is replaced by a buried
collector pipe, or by using the extended lateral lay-out (Figure 4-C), in which there
are very long laterals discharging directly into the main system.

Table 1 Excavation costs for and land loss due to drain ditches *

Depth of ditch (m) Relative excavation cost Land loss (%)

1.6 1O0 2.1

2.1 161 3.5
2.6 235 4.3
3.1 324 5.2

* Drains with bed width 1.0 m, side slopes of gradient 3h:2v, a 2 m berm, a maximum spoil height of
2 m, and spaced at 500 m intervals.

A possible next step is the pumped composite lay-out shown in Figure 4-D. In this
system which has been applied in Khairpur (Pakistan), for example, the collector pipes
discharge into a sump equipped with a pump to lift the drainage water into a shallow
disposal drain. In this concept not even a deep main or outfall drain are required,
but the price of installation and operation of the pumping stations will have to be
paid.
In the past developments towards large, composite drainage networks have been
slowed down by the fact that construction of deep composite pipe networks with clay

30
 Keynote Lectures

A B C
SINGULAR SYSTEM COMPOSITE SYSTEM EXTENDED SINGULAR SYSTEM
LATER[: OF 250 m ~~~ LATERALS OF 250 m LATERALS OF 1000 m

_II__II_

D
PUMPED COMPOSITE SYSTEM

ZE
-_ ----- ---
Figure 4 Lay-out alternatives
-------
-
:-

---LI-,,
-
main and collector drain ditch
shallow disposal drain
collector pipe with junction box
lateral pipe drain
provision for maintenance
outflow structure
sump with pumping station

or concrete pipes was difficult and hence costly. However, as mentioned above, con-
struction problems could be solved to a great extent due to the introduction of corru-
gated drainage pipe, which also reduces the risk of silting up and need for maintenance.
We therefore expect that in irrigated arid regions the construction of extended under-
ground pipe drainage networks to minimize the need for deep, impractical, open drains
will find more and more support and application in coming years. The final target,
and one which is believed attainable in the future, is an almost fully underground
pipe drainage network. With the aid of reliable and payable mineral or synthetic filters
this network should be practically maintenance free.

6 Discussion of criteria
6.1 Formulation of requirements

The required performance of a groundwater drainage system can be defined by the

combination of the minimum groundwater depth to be maintained in the critical peri-
ods, and the amount of excess groundwater to be drained during those periods.
This requirement formulation is very appropriate and commonly applied to drainage
in relation to irrigation and salt control. The quantity to be drained is predictable
and to a certain extent controllable, and is related to irrigation losses, seepage supplies
and leaching requirements. The minimum groundwater depth to be maintained is relat-
ed to avoiding damage by waterlogging and by capillary salinization. Waterlogging
and capillary salinization, however, refer to different critical periods of the cropping
cycle, namely the irrigation and the fallow periods. This implies that there are two

31
 Keynote Lectures

I independent depth criteria, one related to waterlogging in the irrigation periods, and
one related to capillary salinization in fallow periods. The situation is demonstrated
in Figure 5 with the annual movement of the watertable in an irrigated field with
winter cropping and summer fallow.
The watertable reaches its maximum level during the irrigation period. It is in that
period that waterlogging has to be avoided. No capillary salinization then occurs;
on the’contrary, salts are leached owing to the net downward groundwater flux of
percolation losses from irrigation. The watertable is at its lowest level in the non-
irrigated period, thus no risk for waterloggïng but upward capillary transport of
groundwater causes salinization of the upper soil layers. The capillary transport, ini-
tially fast, rapidly slows down and finally stops due to the accompanying drop in
the watertable by drainage and evapotranspiration, unless the watertable in the fallow
land is recharged by lateral inflow of seepage water from neighbouring higher areas
or adjoining irrigated land, both common phenomena. In that case the role of drainage
in the fallow period is to intercept the lateral seepage inflow, and to keep the watertable
at such a depth that the capillary rise is considered negligible or acceptable. This depth
is called ‘critical depth’.
As a result the drainage criterion for irrigated land is indicated by two requirements
which need to be satisfied. The drainage should be able:
1. To discharge the peak drainable surplus during the irrigation season,maintaining
the watertable at the minimum required depth to avoid waterlogging;
2. To discharge the drainable surplus during the fallow season, the: seepage supply,
maintaining with the watertable at the depth needed to avoid capillary salinization.
It should be noted that in fallow periods and even during each interval between irriga-
tions, salinization of surface layers by a redistribution of the salts in the soil profile
cannot be avoided, whatever the depth and intensity of the drainage system might
be. These salts are leached prior to sowing at the start of the irrigation season and
further with each following irrigation.

drainable surplus
rlonrh
__r... (mm)
wheat I fallow I beneem cotton
I
20 20
30
40 - I
’’
,

60 -
example of possible actual
fluctuations of the watertable +
short term fluctuation of the
tertable as corresponding
! b r, I
80
drainable-surplus-recharge
1O0
al fluctuation of average
120 of the watertable

140

160

Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr Mey Jun Jul Aug Sep

I Figure 5 Groundwater fluctuations in a drained field

 Keynote Lectures

6.2 Depth t o watertable requirements

Waterlogging should be avoided in the irrigation season which is usually specified

as the watertable being kept below the root zone of the crop. In steady state drainage
calculations the depth requirement for field crops is usually applied in the range of
0.8 to 1.2 m. The FAO publication, 1980, on drainage design factors recommends
values of 1 .O to 1.2 m for steady state and 0.9 m for non-steady state drainage calcula-
tions depending on the soil. The non-steady state requirement of 0.9 m refers to the
so-called minimum depth of the watertable after irrigation. So-called, because in non-
steady state calculations the rise of the watertable after irrigation is based on recharge
equal to the net drainable surplus, whereas the actual recharge and thus the rise may
be much more, due to the fact that the actual groundwater recharge mostly includes
water which is later consumed by the crop.
As for the depth requirement, it is noted that the relation between depth of the
watertable and crop development is much more complicated than can be expressed
by a simple rule, as the watertable should be kept below the root zone. Furthermore,
the rooting depth is a purely relative notion, as crops may adapt their root system
to the prevailing soil and water conditions. It might therefore be more practical to
consider the depth criteria as empirically established values which, although not scien-
tifically proven, have shown to be satisfactory for practical design purposes. It is
doubtful whether better depth criteria can be established with field experiments. The
empirical criteria could be tested with computerized models (growth simulation mo-
dels).
With the choice of the waterlogging-depth criterion, the choice between the steady
or non-steady state approach also arises. The following remarks result from this
choice:
1. As mentioned the actual watertable fluctuations will differ from the calculated ones,
which do not take into account the fact that after irrigation the watertable rises
more than that due to the recharge, equal to the drainable surplus (irrigation minus
evapotranspiration), and that the following drop of the watertable is not only due
to drainage, but also to evapotranspiration. The actual minimum depth of the wa-
tertable will thus be less than the one calculated;
2. There is no evidence that the minimum depth in the irrigation interval is a better
yardstick for evaluating the relation between watertable and crop development than
the average interval depth, independent of the manner by which the minimum is
to be calculated. Analyses of the watertable variation within the irrigation interval
are thus not required, and there is thus no reason for applying the non-steady state
approach for that purpose;
3. The non-steady state or dynamic equilibrium analyses, however, are useful for si-
mulating the seasonal, annual or multi-annual variations of the depth of the water-
table. This is particularly true for crop rotations which include fallow periods (see
Figure 5). This Figure shows watertable conditions and the critical periods for wa-
terlogging and salinization throughout the year;
4. Further, the non-steady state analyses take account of the possible reducing effect
of water storage on the peak discharge requirements;

33
Keynote Lectures

5. The non-steady state approach is thus preferable in situations with significant sea-
sonal watertable fluctuations.
The watertable should be kept below the ‘critical depth’ in the fallow season which
in general terms is defined as the depth at which the upward capillary flow becomes
negligibly small, although there is no precise definition of what rate is to be considered
negligible. The critical depth is related to the type of soil and is usually taken in the
range 1.4 m (for soils with fine and coarse textures) to 1.7 m (for soil with a medium
texture) (FAO 1980). The critical depth is an essential design parameter which deter-
mines the minimum depth at which drains are to be installed in irrigated land. Some
additional remarks on proper application of this design parameter are relevant here:
1. The critical depth is only relevant as a drainage criterion in situations with seepage.
Without seepage the watertable will automatically drop in the fallow period to
below the critical depth (even without drains) as a result of any evapotranspiration
and natural drainage. Seepage to fallow land is, however, a common phenomenon,
as discussed before.
2. In the case of intensive cropping and more or less continuous irrigated land, the
critical depth also has no practical meaning. This explains that in the Egyptian
Nile delta, with a more than 200% cropping intensity, drains are installed at relative-
ly shallow depths of 1.2 to 1.5 m, compared to depths of 1.7 to 2.2 m applied in
Iraq and Pakistan;
3. The critical depth has been related to type, or better, the capillary properties of
the soil. It should, however, also be related to the length of the fallow period and
the evapotranspiration in that period. If the given values of 1.4 to 1.7 m refer to
long summer fallows in extreme hot climates (Iraq and parts of India and Pakistan),
then smaller values could be applied for winter fallows, short summer fallows and
summer fallows in less extreme climates (Tunisia, Morocco, Algeria). It is therefore
suggested that the critical depth design parameter must be redefined in terms of
the depth of the watertable where the capillary soil water transport to the surface,
totalized over the entire fallow period, should not exceed a certain established limit.
Apart from soil type, the totalized capillary water transport is also related to the
length of the fallow period and its climate, for which the totalized potential evapora-
tion of the fallow period might serve as a yardstick.

6.3 The required discharge rate

The required discharge rate, also called the drainage coefficient, is the rate required
to discharge the drainable surplus in the critical periods. The drainable surplus in
the irrigation season is composed of leaching water, irrigation losses, excess rainwater,
canal seepage, and seepage from neighbouring areas. In the fallow season the drainable
surplus consists of seepage water only. The drainable surplus is usually determined
by means of soil water balance studies. The different components of the drainable
surplus will be examined.
The irrigation component of the surplus consists of percolation losses which are
related to the irrigation method and to the efficiency of the field water management.

34
 Keynote Lectures

With localized irrigation methods such as drip irrigation and for sprinkling systems
such as central pivot and linear move, practically no irrigation water need to be lost
to the subsoil. In those cases the irrigation component of the drainable surplus may
- be small and the drainage coefficient not determined. With gravity irrigation, however,
percolation losses generally constitute the major part of the drainable surplus. The
percolation losses are usually assumed to be a constant percentage of the irrigation
supply throughout the entire growing season. It is doubted whether a constant percen-
tage is realistic, whether the percentage of losses is higher in the initial growing stage
when water requirements are low (crops are shallow rooted and cover only part of
the soil surface) than in the full development stage with maximum water demands,
full crop coverage, and deep rooting system. In fact no evidence could be found which
justifies the constant percentage assumption. On the contrary, the available informa-
tion points to a decreasing percentage of losses during the growing season.
Figure 6 shows the field water balance of a cropped field in Iraq taken from data
of the Dujailah Drainage Experiments. The area was underlain at a depth of 4 m
by an impervious barrier, so the drain discharge could be taken equal to the percolation
losses. Figure 6 also shows that the percentage of losses for winter crops are higher
than for summer crops and that, as expected, this percentage decreases during the
growing period. In the peak summer period the losses are 8% only.
Other indications to support our theory can be observed in many irrigation schemes
in arid and semi-arid countries. With constant proportional losses, the drains in those
schemes should have an increasing discharge during the growing season, and a maxi-
mum flow in periods of peak irrigation demand. In reality, however, it is observed

irrigation t r a i n
fallow green gram , (mm)
barley
I
irrigation
c 50
rainfall
-c
1O0
70 r
150

50

40

30

20

10

O
1 L 1 . k
* r
Keynote Lectures

that drain discharges do not increase. On the contrary, drains often have minimum
flow in periods of peak irrigation demand.
Application of a constant percentage for field percolation losses overestimates the
drainable surplus as well as the irrigation requirement. It is therefore, suggested to
apply a variable decreasing percentage for field losses during the irrigation period.
A rather conservative example of variable field efficiencies is given in Table 2. The
effect of decreasing losses on the calculated drainage requirement in Table 2 is evident;
the peak (July) drainable surplus, which would be 71 "/month if constant losses
are assumed, decreases to 29 "/month for variable losses.

Table 2 Drainable surplus in relation to application efficiency

Feb* Mar Apr May Jun Jul Aug Sep Total

Net irrigation requirements 80 13 46 108 138 165 126 20 696

Variant I: Constant applic.

efficiency 0.7 0.7 0.7 0.7 0.7 0.7 0.7
Field requirements 80 19 66 154 197 236 180 29 961
Drainage 6 20 46 59 71 51 9 265

Variant 11: Variable applic.

efficiency 0.5 0.7 0.8 0.8 0.85 0.85 0.7
Field requirements 80 26 66 135 173 194 148 29 851
Drainage 13 20 27 35 29 22 9 155

* pre-irrigation

The leaching component or the leaching requirement (LR) is usually defined as the
leaching needed to control the salinity of the soil at a specified tolerated maximum
level, expressed as a percentage of the irrigation supply. Several formulae for calcula-
ting the LR have been developed and used since the LR concept was introduced in
Handbook 60 (USDA 1954 No 60), (FAO 1976 No 29), (ILRI 1963 No 1 1 ) .
Leaching occurs by rain or irrigation water percolating through the soil to the wa-
tertable. If the percolation from losses is not sufficient to control the salt, additional
irrigation is needed for leaching. Only in that case does the leaching requirement affect
the required discharge rate. This might occur in case of drip and sprinkler irrigation.
The relation between irrigation, drainage and soil salinity is more complicated than
expressed by the LR formula. It is not necessary that the LR be satisfied for each
irrigation interval individually, because excess leaching in one period may compensate
for shortages in other periods. Leaching operations may therefore be scheduled outside
the irrigation season when water is available and the drains are not heavily charged.
As stated previously, a study of the salt and water regime in the soil in relation to
irrigation and drainage by means of computer simulation models provides the best
method of evaluating the drainage requirements for leaching. Simple models are al-
ready available. Development of more precise complicated models is only a question
of time and priority.

36
 Keynote Lectures

The seepage component is the most difficult one of the series of drainable surplus.
It includes canal seepage, seepage from irrigated to fallow land, and groundwater mo-
vement at greater depth and over larger distances from high land to areas of depression.
The latter is responsible for the differences in natural drainage and seepage conditions.
Canal seepage is a local problem. It may be accounted for by designing seepage
interceptor drains, and it may also be accounted for in the drainage coefficient by
assuming an uniform seepage rate over a strip of land on both sides of the canal.
Both approaches are not very accurate but considered satisfactory, in view of the low
accuracy with which the canal seepage losses are usually known. To estimate canal
losses and their distribution, flow analyses with two-dimensional numerical computer
models are nowadays applied.
Seepage from irrigated to fallow land is very common and is the main cause of
salinization in fallow land. As is the case with canal seepage, quantification of this
seepage is difficult. It is usually taken as an estimated uniform rate for the area con-
cerned.
In areas underlain by a highly transmissible aquifer, the natural drainage and seep-
age conditions resulting from lateral groundwater movements may have a great effect
on the drainable surplus and on the discharge criterion for design. Areas with natural
drainage have a small drainage coefficient or require no drainage at all. Areas with
seepage have an increased surplus and are moreover subject to salinization if not pro-
perly drained. For a quantitative analysis of the natural drainage and seepage condi-
tions in areas where these are considered of importance, groundwater modelling is
the only proper solution, especially as drainage/seepage conditions will vary through-
out the year. Groundwater modelling, however, is still quite a complicated and costly
exercise, and is preferably done as a part of an integrated regional groundwater man-
agemen t study.
An example of groundwater modelling for assessment of drainage coefficients is
the 250000 ha SCARP VI project (Figure 7), in the central part of the Indus Basin
in Pakistan. This project area is underlain with a highly permeable aquifer. The lateral
groundwater movement provides for natural drainage in some parts of the project
area, and causes serious drainage needs in other parts. A computer model study, aimed
at optimal integrated ground and surface water management, and the determination
of the areas in need of drainage and the regional variation in drainable surplus, was
developed and calibrated with historical watertable records. Some results of the model
study are given in Figure 7 (WAPDA 1981).

6.4 The drain depth

The drain depth refers to the installation depth of underground drains or to the depth
to the watertable in open drainage ditches. The drain depth is directly related to the
drainage criterion. The drains should at least be installed below the required level
of the groundwater table during the irrigation season or below the critical depth if
the criterion of the fallow season is to be met. Drains are usually installed deeper
than the required minimum. The optimum drain depth is to be found as a compromise

37
Keynote Lectures

POLYGON NETWORK OF SCARP V I AREA FOR GROUNDWATER MODEL

____-_
i* boundary project area

DRAINABLE SURPLUS IN SALINE GROUNDWATER AREA DETERMINED BY MODEL STUDY

drainable surplus
(mm/davl
-0-1
1- 1-2 20 30 40 50 km

watertable In fresh water zone

controlled by skimming irrigation Welk

Figure 7 Groundwater modelling for determining drainable surplus

between the advantages and disadvantages of a greater depth.

Advantages of deep drains are greater hydraulic head and more storage capacity
in the soil, both resulting in larger spacings and thus less length of drain per unit
area. Another advantage is that deep drains will discharge continuously, and thus
may have less risk of silting up.
Disadvantages of deep drainage are increased technical problems in proper installa-
tion, increased costs of installation and materials per unit length of lateral, and possible
costs for deepening the main outfall system and pumping. The installation depth of
lateral drains has been steadily increasing since 1960, and installation of deep drainage

38
 Keynote Lectures

has become better and cheaper. There is an economic optimum drain depth, corres-
ponding to minimal costs per unit area. This depth has recently been analysed (Bou-
mans and Smedema 1986), and corresponds rather well with the depth ranges actually
applied in humid as well as arid regions. We expect that the existing trend towards
deeper drainage will continue, because further improvements in the construction tech-
nology will reduce the disadvantages of deeper installation.

Importance of monitoring
It will have become clear that there are still incomplete answers and thus further re-
search is required into fundamental questions such as relation of drainage with crop
yield, choice of design criteria, salt regime and drainage. Research is still necessary
for testing and improvement of current survey and design techniques, drainage mater-
ials, construction equipment and installation methods. Research is carried out in the
laboratory, on experimental fields and with simulation computer models. Another
very valuable research instrument is monitoring the performances of the drainage
scheme in operation. Monitoring is a direct way of measuring the effect of the imple-
mented works on crops, salinity and groundwater: pre and post-project conditions
as well as actual and forecasted performances can be compared; operation and main-
tenance aspects studied; costs and benefits evaluated; and design and construction
shortcomings identified.
Unfortunately monitoring has not generally been applied to date. Reliable monitor-
ing of drainage schemes is almost non-existent. Monitoring should therefore be pro-
moted in the interest of operational and fundamental drainage research, and also for
proper operation and maintenance of the drainage works. The drainage and project
authorities responsible, have to be convinced of the importance of monitoring. Moni-
toring should already receive proper attention in the study and design phase of a drain-
age project. A detailed monitoring programme, and the requirements of staff, equip-
ment and funds, should be part of the project planning documents. The drainage design
should take monitoring into account by providing adequate facilities for data collec-
tion.
Monitoring is considered costly because of the staff required for field work and
data processing. However, monitoring nowadays can make use of modern technology
and this is very advantageous. Not only the staff requirements for data collection
can be greatly reduced but also better data-processing techniques can be applied with
less staff. Of the many possible applications of modern technology for monitoring
of drainage schemes the following are mentioned:
- The use of remote-sensing images for recording changes in soil water and salt condi-
tions and developments in land use and crop growth;
- The use of automatic electronic recorders for all kinds of data collection. These
recorders are preferably fitted with facilities to transmit recorded data directly to
the computer processors and eliminate the need for manual input. A good example
of such equipment is the ‘Preslog’, a small battery powered instrument which is
very easy to install and can register and store in a memory block (Eprom) hourly
I
39
Keynote Lectures

records of waterlevels or rainfall for one year. The Eprom can be read by a computer
which also is used for data processing and filing;
- The use of computers for data filing, data processing and reporting. The great ad-
vantage of computers compared to traditional data processing by hand is obvious.
Monitoring as well as processing and evaluation of data requires highly qualified staff.
It may therefore be practical to carry out monitoring in collaboration with a specialized
drainage research institute which can advise on the kind and method of data collection
and on the equipment required, and can participate in data processing. The different
objectives of monitoring may then be served in the best way. A good example of such
a specialized institute is the ‘Drainage Research Institute (DRI)’ in Egypt which is
an agency of the Ministry of Irrigation.

8 I Final remarks and conclusions

In this paper we have discussed the horizontal drainage of irrigated arid lands which
is different from the drainage in humid temperate areas in many ways. Developments
with respect to design, construction, and research have been discussed. Present design
practices, criteria and standards have been critically examined.
The major conclusions are summarized below in the form of statements presented
for further discussion.
1. Drainage or better subsurface drainage for irrigated arid land is very different from
subsurface drainage in humid areas, with respect to the objectives, requirements
and criteria, field investigations, design options, construction practices, equipment
and materials. It is questionable whether the difference between drainage in humid
and arid regions is sufficiently reflected in the structure and programme of the
International Post-Graduate Course on Land Drainage;
2. Computer drainage applications have been started and are expected to increasingly
contribute to drainage research and to the introduction of improved and automated
design procedures for the drainage of irrigated land. Particularly simulation models
for studying the salt and water regimes as related to soil, irrigation and drainage
will be important in these respects. Simulation models have the advantage that
they can be adjusted to local conditions, since they can be calibrated with actual
field data;
3. Monitoring is a valuable tool for fundamental and operational drainage research,
and for controlling the proper functioning of the system. Monitoring should be
included in the design and operation of any drainage scheme. Monitoring can make
use of modern electronic equipment. Monitoring should, if possible, be coordinated
by a national drainage research organization;
4. With respect to lay-out and construction of drainage systems, the projected devel-
opment is an ongoing trend towards extended underground piped drainage net-
works in order to minimize length of deep open drains. The final target being an
almost fully underground system and practically maintenance free;
5. The drainage requirements for the irrigation and fallow seasons should be clearly
distinguished. The irrigation season requirements must be related to waterlogging

40
 Keynote Lectures

and leaching, and the fallow season requirements related to intercepting seepage
and maintaining the watertable at the critical depth. The critical depth must be
defined in terms of the depth for which the totalized upward capillary soil water
flux during the fallow period does not exceed a certain norm value;
6. The common practice by which field percolation losses and, consequently the drain-
able surplus are taken as proportional to the irrigation supply may lead to an overes-
timation of the drainage discharge requirement. The percentage of losses is general-
ly higher in the initial growing stage than in the period that crops reach their full
development. Application of decreasing losses (or increased field efficiencies) dur-
ing the growing season is therefore recommended in water balance studies for as-
sessing drainable surplusses;
7. The relation between leaching requirement and drainage can best be analysed by
means of salt regime simulation models;
8. In schemes with a highly permeable aquifer, groundwater modelling may be needed
for a proper assessment of both the natural drainage conditions and the spatial
variations in drainage requirements and drainable surplusses;
9. Drain installation depth has the tendency to increase as a result of improved con-
struction technology.

References
Boumans, J.H. and L.K. Smedema, 1986. Derivation of cost-minimizing Depth for Lateral Pipe Drains.
Agricultural Water Management, 12 (1986)41-51.
Ernst, L.F. 1986. Computation of groundwater flow between parallel open conduits (in Dutch with English
summary). Comm. Hydr. Res. TNO Proc.: 48-68. The Hague.
Euroconsult, 1985. Drainage of gypsiferous soils in the Carlsbad Irrigation District. New Mexico, USA.
Report of field visit by W.J. van Diest.
FAO 1976. Water quality for agriculture. FAO Irr. and Drain Paper 29. R.S. Ayers and D.W. Westcot,
Rome.
FAO 1980. Drainage design factors. FAO Irr. and Drain Paper 38, Rome.
Hooghoudt, S.B. 1940. Dewatering through parallel drains (in Dutch). Versl. Landb. Ond. 4651 5~707.
Alg. Landsdrukkerij, Den Haag.
Helsinki 1986. Proceeding of International Seminar on Land Drainage, J. Saavalainen, P. Vakkilainen,
editors. Helsinki University of Technology, Dep. Civ. Eng.
ILRI, 1963. Reclamation ofsalt affected soils in Iraq. P.J. Dieleman edit. Publication 1 1 , ILRI Wageningen.
Luthin, J.N. ed. 1957. Drainage of Agricultural Lands. American Soc. of Agr. Madison, Wisconsin.
NN, 1958. Niyala Basin Archeological Project. Salinity and Irrigation Agriculture in antiquity. Report
on essential results June 1, 1957 to June I, 1958.
NESPAK/ILACO, 198 I . Panjnad-Abbasia Salinity Control and Reclamation Project (SCARP VI): Final
Plan Report for the WAPDA of Pakistan.
Reeve, R.C., L.E. Allison and D.F. Peterson, 1948. Reclamation of Saline Alkali soils by leaching. Delta
Area, Utah. Utah Agr. Exp. Sta. Bul. 335.
Toksöz, S and O. Kirkham, 1971. Steady drainage of layered soils I and 11. Irr. and Drain of the Div.
Am. Soc-Civ. Eng. 7985: IRRpapers 7985: 1-17; 7986: 19-37.
U.S.D.A. 1954. Diagnosis and improvement of saline and alkali soil. Handbook 60.
U.N. 1985. Studies for the use of saline water in command'areas of irrigation projects. Haryana, India.
Wilcocks, W. and J.I. Craig, 1983. Egyptian Irrigation. 3rd ed. London.
WAPDA/ILACO 1985. East Khairpur Tile Drainage Project. Project Completion Report.

41