17.

5 Examples of Agricultural Drainage Criteria

17.5.1 Rain-Fed Lands in a Temperate Humid Zone

How drainage criteria are used for the design of drainage systems in rain-fed lands
in temperate humid zones will be exemplified with design particulars for field and
collector drainage systems in The Netherlands.

Example 17.1 Field Drainage Systems in The Netherlands

The water balance of field drainage systems in many parts of The 'Netherlands can
be written in the simple form, neglecting groundwater flow components
Dr = q,At = P - E - AW (1 7.2)
where
Dr = drainage (mm)
q, = drainage rate (mm/d)
At = period (d)
P = precipitation (mm)
E = evapotranspiration (mm)
AW = water storage (mm)

Figure 17.31 presents the monthly balances of rainfall (P) and evapotranspiration (E)
in The Netherlands. It shows that, in summer, the evapotranspiration exceeds rainfall
and AW = P-E, so that, according to Equation 17.2, no drainage is required (qd
= O), except in areas with a strong upward seepage of groundwater. In winter, the
rainfall exceeds the evapotranspiration plus the storage by about 180 mm, which, for
4 months, gives an average drainage rate q, = 1.5 mm/d. Crop-response functions
have indicated that, in winter, an average depth of the watertable of 0.9 m below
the soil surface is amply sufficient. This represents a long-term, steady-state
agricultural criterion for subsurface drainage systems. Assuming a drain depth of 1.O
m, we find the average hydraulic head to be h = 1.0 - 0.9 = 0.1 m. Defining the
drainage intensity ratio as q,/h (d-I), we find qd/h = 0.001 5/0.1 = 0.015 d-l.
In The Netherlands, when the depth of the watertable midway between the drains
is 0.5 m, subsurface field drainage criteria are expressed as a normative discharge
(q, = 7 mm/d). This normative discharge and reference level of the watertable are
exceeded only once a year on average, so we are dealing with a short-term, unsteady-
state situation. The drainage intensity ratio becomes qd/h = 0.007/(1.0 - 0.5) = 0.014
day-', which is only slightly less than the ratio q/h = 0.015 found for the steady-state
situation.
The q,/h ratio for the steady state is very sensitive to changes in drain depth; for
example, if we take a drain depth of 1.1 m instead of 1 .O m, the q,/h ratio becomes
0.0075 instead of 0.015. Therefore, and because the agricultural effects of drainage
are usually more responsive to average long-term water levels than to short-term
extreme water levels, the qd/h ratio should not be employed as a drainage criterion
outside The Netherlands or without extensive empirical evidence.
For situations in which the incoming or outgoing groundwater flows cannot be
ignored, Van Someren (1 958) used the observed watertable depths to derive normative

670
 r = rair
I

J F M A M J J A S O N D
summer season deficit A W < O
winter season drainable surplus
winter season excess replenishing
soil watdr A W > O

Figure 17.31 Monthly values of rainfall, evapotranspiration, storage, and drainage surplus in The
Netherlands

discharges and reference levels of the watertable for a subsurface field drainage system
(Table 17.5). The underlying principle is that shallow watertables indicate net
groundwater inflow and upward seepage, and deep watertables indicate net
groundwater outflow and natural drainage. The table shows that the drainage rates,
qd, diminish as the observed watertables are deeper (i.e. as the upward seepage reduces
and the natural drainage increases). Since this is not generally true in other parts of
the world, Table 17.5 is not directly applicable outside The Netherlands.
With the established intensity ratio’s for subsurface drainage, either for long-term
steady-state or short-term unsteady-state conditions, one can proceed to design the
subsurface field drainage systems. The steady-state conditions permit the use of steady-
state drainage equations. Since the data of Table 17.5 have already taken the storage
into account (they are specified in terms of normative drain discharge exceeded on
an average of only once a year, which equals the corresponding recharge, less storage),
here too steady-state drainage equations can be used.

Example 17.2 Collector Drains in The Netherlqds

In The Netherlands, we recognize two criteria for water levels in open collector drains
(Figure 17.32): a high water-level criterion (HW) and a normal water-level criterion
(NW). The HW criterion specifies that the water level in the collector may exceed
a level of 0.5 m below the soil surface only 1 day a year. The NW criterion specifies

67 1
Table 17.5 Normative extreme discharge (qd) and corresponding watertable depth (i)for subsurface field
drainage systems in The Netherlands, by type of land use. The q/h ratios (d-I) (where the height
h = 1.1 -j, for a drain depth of 1.1 m) are indicated between brackets (Van Someren 1958)

Reference level observed to be Normative discharge

exceeded by the watertable only once ("W
a year
Grassland Arable land Orchards
(m below soil surface) j = 0.3 m j = 0.5 m j = 0.7 m

O 7 (0.009) 7 (0.012) 7 (0,018)

o. 1 7 (0.009) 7 (0.012) 7 (0.018)
0.2 3 (0.004) 5 (0.008) 6 (0.015)
0.3 O 3 (0.005) 5 (0.013)
0.4 O 4 (0.010)
0.5 3 (0.008)
0.6 2 (0.005)
0.7 1 (0.003)
2 0.8 O

that the water level in the collector may exceed the outfall level of the laterals (i.e.
1.0 to 1.1 m below soil surface) no more than 15 days a year. For collectors serving
small areas, the second criterion is the most critical and will therefore be adopted
for the design, whereas for collectors serving large areas, the first criterion is the
appropriate one.
According to Blaauw (1961), the collector discharge (SIS)that is exceeded 15 days
a year is about half the discharge (ql) that is exceeded only 1 day a year (qI5= o.5ql).
In general, he found for the discharge that is exceeded in x days a year
qx = ql(1 - 0.44 log x) mm/d
The extreme discharge, q , , is found from the empirical relationship
q1 = 8.64 B (0.53 - 0.05 log A) mm/d
where A is the area (ha) served by the collector, and B is a factor depending on the
area's hydrological conditions. The value of B is usually between 2 and 3, depending
on the soil type, kind of cropping'system, and intensity of the field drainage system,
but when upward seepage or natural drainage occurs, the factor B may go up to 4
or down to 1 , respectively.

. . . . . . . .

. . . . .
. . . . . .

Figure 17.32 High water-level (HW) and normal water-level (NW) criteria used in The Netherlands for
the design of collectors

672
1 With the water-level criteria and the corresponding discharges thus determined, we
can proceed with the design of the capacity and dimensions of the collector system,
~

using Manning's steady-state formula (Chapter 19), because the dynamic storage of
1, water in the collector system is small compared with recharge and discharge (Section
17.3.4).

~ 17.5.2 Irrigated Lands in Arid and Semi-Arid Regions

How subsurface drainage criteria are used in arid zones and how the corresponding
~
water balances are applied will be illustrated with one case from Egypt and two from
, Peru.

Example 17.3 Egypt

For Egypt's Nile Delta, the agricultural drainage criterion reads: 'The seasonal average
depth of the watertable midway between the drains should be 1 .O m.' (Safwat Abdel-
Dayem and Ritzema 1990). Although there are indications that this depth could be
somewhat less (Figures 17.19 and 17.20), the value 1.0 m was adopted for safety
reasons. On the other hand, it would be inefficient to lower the average watertable
to more than 1.2 m, because this would increase the deep percolation losses and reduce
the irrigation efficiency (Oosterbaan and Abu Senna 1990).
Starting from the overall water balance given in Chapter 16 (Equation 16.8), we
may ignore rainfall, surface evaporation, and surface runoff, and add a drainage term
to obtain
qd = qsi - E + qgi - qgo (17.3)
where
qd = drainage rate (mm/d)
qsi = surface irrigation (mm/d)
E = evapotranspiration rate ("id) '

q, = groundwater inflow (mm/d)

q,, = groundwater outflow (mm/d)
' The continuous irrigation throughout the year and the steady-state long-term
agricultural criterion for subsurface drainage in Egypt permits us to neglect also the
change in storage.
In many parts of the Nile Delta, it has been observed that there is natural drainage
of groundwater to an underlying deep aquifer (so that qgo> qgi).Hence, we can expect
the value of qd to be relatively small.
Safwat Abdel-Dayem and Ritzema (1990) reported on measurements of drain
discharge and found an average rate of qd = 0.6 mm/d. This discharge includes the
discharge from subsurface-drained rice fields, which is in fact not desired (Qorani
et al. 1990). When the drain discharge rates were determined per crop, these rates
were found to be distributed as shown in Table 17.6. From that table, we can conclude
that, if the drainage from the rice fields could be restricted, a design discharge rate
of q = 0.4 mm/d (corresponding to the average discharge rate of the maize fields)
would be amply sufficient for the design. The value of qd can be so low because it

673
Table 17.6 Average drain discharge in Egypt's Nile Delta per season and per crop (Safwat Abdel-Dayem
and Ritzema 1990)

Season Winter Summer

Crop Berseem Wheat Cotton Maize Rice
~ ~~

Drain discharge (mm/d) 0.2 o. 1 o. 1 0.4 1.3

is only supplementary to the natural drainage.

With the steady-state agricultural criterion for a subsurface field drainage system
(i.e. the seasonal average depth of the watertable midway between the drains equals
1.0 m) and the corresponding design discharge rate (i.e. qd = 0.4 mm/d), we can
proceed with the design of the field drainage system, using steady-state equations.
In a pilot area in the Nile Delta, it was found that the rate of natural drainage to
the underground (qg0-qgi)amounted to 0.5 mm/d (Oosterbaan and Abu Senna 1990).
The total water flow through the profile for this area would amount to 0.9 mm/d, the
artificial drainage contributing 0.4 mm/d and the natural drainage 0.5 mm/d.
The irrigation rate causing this drainage flow amply satisfies the leaching
requirement, as is shown in Figure 17.33. Before the drainage system was installed,
the area had a slight salinity problem, because a small percentage of salinity data
were higher than the critical value EC, = 5 dS/m and the corresponding yields were
lower than average. After the drainage system had been installed, all soil salinity data
showed an EC, below 2 dS/m, a very safe value, and the corresponding crop yields
are independent of soil salinity. No additional amount of leaching water therefore
need be included in the design discharge. We can also note from Figure 17.33 that
the average crop yield (5 t/ha) after drainage is higher than the average yield of the
data with EC, < 5 dS/m before. Apparently, by reducing the soil salinity and lowering
the watertable, drainage hascontributed to thegeneral yield improvement. In addition,
improved agricultural practices upon the introduction of drainage have had a further
positive effect on crop yields.

8
crop: wheal
year: 78/79
no of samples: 40
- .._.-
4 .
-
crop: wheal
year: 85/86
no o1 samples: 45
::.
6 ' . . a

::..:.
t.:..
-
--..-
..
-0.

O
I
!
..'
- \

674
 The subsurface drainage systems of the Nile Delta consist of piped field drains and
piped collector drains. The discharge capacity and the required diameter of the
collectors should not be based on the average discharge rate, but on a more extreme
and less frequent rate. This is because the collector system has a small buffer capacity
and it has to function properly during the relatively short periods of peak discharge,
otherwise the field drainage system fails. For the design of the collector system, Safwat
Abdel-Dayem and Ritzema (1990) proposed to use the discharge rate from maize fields
that is exceeded only 10% of the time. This rate was found to be 1.2 mm/d. Such
a design discharge would also provide a certain safety margin because it occurs only
infrequently.
1 With the technical criterion: 'The collector pipe is just filled to the top at the design
discharge', the design procedure of the collector drains can start, based on Manning's
steady-state formula, even though the design discharge rate is essentially unsteady.

Example 17.4 Coastal Peru

The first Peruvian example concerns an area in the coastal delta of a river that
originates in the Andean mountain range. The coastal area is arid, and agriculture
is totally dependent on irrigation from rivers descending from the Andes, where
rainfall does occur. The irrigation in the river valleys is accompanied by considerable
percolation losses. In the underlying deep and permeable aquifers, the percolation
losses are transported towards the coast. A salt water wedge intruding from the ocean
and a decreasing land slope towards the coast forces the aquifer water to flow upwards,
and the watertable becomes shallow (Figure 17.34). The continuous upward seepage
of groundwater feeds capillary rise into the unsaturated zone. The subsequent
evaporation causes salts to accumulate in the topsoil. For these two reasons, irrigation
and agriculture can only be practised in seepage zones when a subsurface drainage
system is installed.
The area in the delta has light-textured soils and it had to be prepared for irrigated
sugarcane (Suclla Flores 1972). This cane has a growing season of 14 to 16 months,
with irrigation for a period of 10 to 12 months (the vegetative period), followed by
an unirrigated period of 4 to 6 months (the ripening or drying period), during which
the cane augments its sugar content. The average depth of the watertable in the

upward see
irrigation, percolation ,

drainage

.....

ocean

. \ . .
h.....
.'saline groundurax.0 . . , .- . _ . _ V I ~ L K I W I
. . . . . . . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. . . . . . . . . . . . . .

Figure 17.34 Cross-sectional sketch of the geohydrological situation in Coastal Peru (Example 17.4)

675
irrigation season is permitted to be 0.8 m (such a value is also found from Figure
17.7, which refers to sugarcane in Australia), but during the ripening period the average
depth should be more than 1.3 m; otherwise the crop uses too much of the capillary
rise and the ripening does not proceed well. There are therefore two agricultural criteria
for the subsurface drainage system, and the system has to satisfy both. The slight
resalinization of the soil during the ripening period is not a problem, because, with
the first consecutive irrigations, the accumulated salts will be removed again quickly.
The rate of upward seepage from the deep aquifer (called qglv)can be estimated
from the equilibrium depth of the watertable before irrigation and drainage systems
were introduced. In that situation, the topsoil was dry (pF = 4.0) and the seepage
rate equalled the steady rate of capillary rise from the saturated zone (G), which also
equalled the rate of evapotranspiration (qglv= G = E). Under such conditions, the
rate of capillary rise can be found from the steady-state relationship between depth
of watertable, hydraulic properties of the soil, and soil-water content (Chapter 11).
An example is shown in Figure 17.35. If the average depth of the watertable before
drainage was 0.8 m, the estimated rate of capillary rise from the saturated zone was
2.0 mm/d, which gives us the value of the average seepage rate qglv.
In the seasonal water balance of the soil profile, we may ignore the storage term,
and we get
qd = - + qgiv (1 7.4)
where
qd = drainage rate (mm/d)
R = percolation rate (mm/d)
G = capillary rise (mm/d)
qglv= upward seepage (mm/d)

The irrigation system is designed to apply 2400 mm/yr (i.e. during the vegetative
period), of which 800 mm/yr is assumed to be lost as deep percolation. The average

capillary rise
in mmldav

Figure 17.35 The relationship between depth of the watertable and rate of capillary rise in Example 17.4,
under the conditions that the watertable depth is steady, the rates of upward seepage and
capillary rise are equal, there is no irrigation or rainfall, and the topsoil is dry (pF = 4)

676
 Table 17.7 Estimate of the drain discharge from the components of the water bakdnce for irrigated
sugarcane in Coastal Peru (Example 17.4)
~~ ~ ~ ~~

Area Seepage rate Percolation rate Capillary rise Drain discharge

Irrigation season
A 2.0 2.2 O 4.2
B 3 .O 2.2 O 5.2
C 1 .o 2.2 O 3.2
Ripening season
A 2.0 O 0.5 1.5
B 3.0 O 0.5 2.5
C 1.o O 0.5 0.5

percolation rate thus equals R = 800 / 365 = 2.2 mm/d, and the capillary rise G
is nil. Hence, the average drain discharge during the irrigation season can be estimated
+
from Equation 17.4 as qd = 2.2 2.0 = 4.2 mm/d (see Table 17.7). During the ripening
period, there is no percolation ( R = O), but some capillary rise will take place as
the soil becomes dry; it is estimated at G = 0.5 mm/d (Figure 17.35). The drain
discharge is now estimated from Equation 17.4: qd = 2.0 - 0.5 = 1.5 mm/d.
The capillary rise (G = 0.5 mm/d or 90 mm over 180 days) will cause some salinity
build-up in the rootzone, but the amount of percolation of 800 mm/yr is amply
sufficient to cover the leaching requirement, even when its irregular spatial distribution
in the field is taken into account.
To satisfy the agricultural criterion for the ripening period, the depth of the drains
u)
(8) should be greater than the watertable depth of 1.3 m; say 1.5 m. The available
hydraulic head (h) during the irrigation period is h = g - j = 1.5 - 0.8 = 0.7 m,
and, during the ripening period, it is 1.5 - 1.3 = 0.2 m.
The required drain spacings for the irrigation and ripening periods can now be
calculated with the equations given in Chapter 8. The drain spacing adopted should
be the one that satisfies both drainage criteria. It should also be possible to vary the
drain depth (say from 1.5 to I .7 m) so that an optimum combination of drain depth
and drain spacing can be found. Table 17.8 presents an example of the result of
calculations for areas with different seepage rates. The table shows that the required
drain spacings are wider as the drain depth increases and the seepage diminishes. In
Area B, which has the highest seepage rate, the ripening period appears to be critical
for drainage design, because this period requires the smaller drain spacings. In Area
C, which has the lowest seepage rate, the vegetative period (corresponding to the
irrigation season) is critical. In Area A, the seasonal influence on the required drain
spacing depends on the drain depth. The possible combinations are therefore:
- Area A: 1.5 m depth with 77 m spacing, and 1.7 m depth with 120 m spacing,
determined by, respectively, the ripening period and the vegetative period (irrigation
season);

I 677
 Table 17.8 Calculation of drain depth and spacing in Coastal Peru (Example 17.4)

Area Drain Depth of the Hydraulic Drain Calculated drain

depth watertable head discharge* spacing**
g j h=g-j qd L
(m) (m> (m> (“/d> (m>
Irrigation season
A 1.5 0.8 0.7 4.2 96
B 1.5 0.8 0.7 5.2 81
C 1.5 O.8 O. 7 3.2 120
Ripening season
A. 1.5 1.3 0.2 1.5 77
B I .5 1.3 0.2 2.5 51
C 1.5 1.3 0.2 0.5 196
Irrigation season
A 1.7 0.8 0.9 4.2 120
B 1.7 0.8 0.9 5.2 100
C 1.7 0.8 0.9 3.2 150
Ripening season
A 1.7 1.3 0.4 1.5 140
B 1.7 1.3 0.4 2.5 91
C 1.7 1.3 0.4 0.5 355

* From Table 17.7

**
Calculation based on the method presented in Figure 8.4 (Chapter 8) using a
hydraulic conductivity K = 1.0 m/d, a drain radius r = 0.1 m, and a depth of the
impermeable layer D = m m

- Area B: 1.5 m depth with 51 m spacing, and 1.7 m depth with 91 m spacing, both
determined by the ripening period;
- Area C: 1.5 m depth with 120 m spacing, and 1.7 m depth with 150 m spacing,
. both determined by the vegetative period.

In view of the difficulty of installing drains below the watertable, it is a sound technical
practice to place the drains as shallowly as possible (i.e. at 1.5 m depth).

NOTE
The requirement of a fairly deep drain depth in this example is dictated more by the
specific crop requirements during the ripening period than by the need for salinity
control, which is automatically fulfilled by the percolation losses. Under most
agricultural conditions, drain depths can be shallower than 1.5 m, which often
enhances ease of installation and reduces installation cost per m length of drain. This
offsets the disadvantage of needing more drains per ha than with deeper drains.

678
,
Example 17.5 Northern Peru
Figure 17.36 shows a cross-section through sloping agricultural land in Northern Peru.
The land is arid and is equipped with an irrigation system. The land had to be
abandoned, however, owing to problems of waterlogging and salinity. The soil is
sandy, but at some depth the presence of a compact clay layer was noted. At the
downslope end of the land, this clay layer rises to the soil surface, but farther upslope
it is deeper. Here, massive irrigation occurred and the resultant percolation losses
I
continued downslope as groundwater flow.
I
Since the slope (s) of the watertable at the right-hand side of the figure equals the
slope of the interface of the clay layer, which is about 1 % (s = O.Ol), and the hydraulic
1 conductivity of the sandy soil could be estimated at K = 2 m/d, the amount of
horizontal groundwater flow per metre width through the sandy layer could be
calculated, with Darcy, as K.D.s = 2 x 2 x 0.01 = 0.04 m2/d, where D is the level
of the watertable above the interface. Over a length of 1000 m (Figure 17.36), this
means a horizontal groundwater inflow rate, q,,,, of 0.04 mm/d.
Since the land considered is no longer irrigated, the climate is dry, and the watertable
1 remains shallow, we can conclude that the continuous capillary rise from the
watertable and the subsequent evapotranspiration of the weeds and shrubs is fed by
1 an inflow of groundwater. According to the physical principles of steady-state
capillary rise, its rate can be estimated from the depth of the watertable (Chapter
II 11). Thus the rate of capillary rise could be estimated at G = 3 mm/d. Hence, the
upward seepage of groundwater, qglv,also equals 3 mm/d.
The value of q,, is almost 100 times greater than the value of qglh.This leads to
the conclusion that the clay layer has sufficient permeability to permit the passage
of the seepage flow. Hence, the greater part of the groundwater seeping up into the
land originates from a great depth, and there must be a deep and permeable aquifer
below the clay layer. We can therefore conclude that, somewhere downslope of the

Figure 17.36 Cross-sectional sketch of the hydrological situation in Northern Peru (Example 17.5)

679
land, there must be an underground barrier to the flow of groundwater which forces
this flow upwards.
The technical solution to the problem of waterlogging would be to install a regular
subsurface field drainage system or to introduce pumped wells, employing the usual
agricultural criteria for subsurface field drainage systems (Section 17.3.7), while
ensuring that irrigation can be effectively resumed. For example, we could use the
agricultural criterion that the average depth of the watertable during the irrigation
season should be 0.8 m, which satisfies the requirements of most crops. We can find
the corresponding design discharge from a water balance, taking into account the
percolation stemming from the irrigation and the upward seepage of the groundwater,
both taken as an average rate during the season considered.
The example of Northern Peru shows that the mere horizontal flow of groundwater
contributes little to the subsurface drainage problem, but that the main causes must
be sought in vertical recharges from percolation and/or upward seepage. Intercepting
the almost horizontal flow, qgih,would therefore not alleviate the problem. The reasons
are two-fold:
- If the impermeable layer is shallow, an interceptor drain would catch 100% of the
horizontal groundwater flow, but the amount of flow is so small that it cannot
create an extensive problem of waterlogging, so the interceptor drain is hardly
needed;
- If the impermeable layer is deep, there is an aquifer with a high transmissivity which
can cause waterlogging over an extensive area, but an interceptor drain would catch
only a very small fraction of the groundwater flow and would not significantly solve
the extensive waterlogging problem.

17.5.3 Irrigated Lands in Sub-Humid Zones

Sub-humid zones are often characterized by a rainy season with high rainfalls (say
more than 100 mm per month, and with extreme rainfalls up to 100 mm per day),
followed by a dry season. The rainy season may coincide with a cool winter period
(e.g. as in North-West Africa), or with a hot summer period (e.g. the monsoon in
South-East Asia and West Africa, south of the Sahara). However, also in tropical
areas without distinct winter or summer seasons, there may be pronounced rainy and
dry seasons (e.g. East Africa).
In the sub-humid zones, irrigation is often practised during the dry season, but also
during the rainy season if the rainfall is erratic. When drainage problems occur, salinity
problems are often also apparent. The drainage systems to solve these problems should
be clearly distinguished in surface drainage systems for the rainy season, subsurface
drainage systems for the dry (irrigated) season, and perhaps combined surface and
subsurface drainage systems for the rainy season. The drainage criteria have to reflect
this differentiation. In addition, a thorough study is required to check whether the
drainage problem is entirely the result of local rainfall or of incoming groundwater,
or whether inundations from side slopes, rivers, lakes, or seas are the main cause.
Where such inundations occur, a drainage system should not be implemented without
a flood control system, and perhaps this alone will be sufficient to relieve the
waterlogging.

680
 In the following, an example will be given of the development of criteria for subsurface
field drainage systems by pipes in North-West India, which has a monsoon climate.
Unfortunately, the practice of combined surface and subsurface drainage systems in
sub-humid zones is not well developed, so that we have little experience on drainage
criteria for combined systems to draw from. In irrigated lands of the sub-humid zones,
drainage systems are often lacking or, if present, are either solely surface or solely
subsurface systems. When there is only surface drainage, salinity problems are not
counteracted, and when there is only subsurface drainage, the surface drainage
problems either persist or are tackled with an excessively expensive subsurface system
geared to cope with very high discharges.

Example 17.6 North- West India

Rao et al. (1990) describe the results obtained with subsurface drainage by pipes in
an experimental area in North-West India. The area was waterlogged during the
monsoon period and was very saline. Pipe drainage systems were installed at a depth
of 1.75 m and with spacings of 25, 50, and 75 m. The average drain discharge was,
respectively, 2.7, 1.1, and 0.9 mm/d during the irrigation season from October to
February. This reveals that the discharge of the drainage system with 25 m spacing
was high, that more irrigation water was applied there, and that the irrigation was
less efficient.
After drain installation, the area’s annual rainfall of about 700 mm, occurring
mainly in the months of July to September (the monsoon season), desalinized the
soil. The rainwater was conserved in the field by bunds, so surface drainage was
impeded and infiltration was enhanced. Table 17.9 shows the measured soil salinities.
The initial soil salinity corresponded to an electric conductivity of a saturated paste
(EC,) of about 50 dS/m in the surface layer of 0.20 m, and about 20 dS/m in the
deeper layers down to 1.2 m. Within 4 months, the soil salinities had come down
to levels that were generally below 10 dS/m. The reduction in soil salinity as well as
the yield increase of the crops was faster with the 25 m spacing than with the larger
spacings. After a period of three years, however, significant differences were no longer
observed.
The faster reclamation with the 25 m spacing was achieved at the cost of a much
more expensive drainage system and of less efficient irrigation in the post-monsoon season.

Table 17.9 Soil salinity (EC, in dS/m) in the Sampla pilot area before (June 1984) and at the end of the
first monsoon season after drain installation (October 1984) (Rao et al. 1990)

Depth of soil Drain spacing (m)

layer
25 50 75

June Oct. June Oct. June Oct.

I
(m)
o - 0.2 50.7 5.3 50.7 8.1 46.1 8.3
0.2 - 0.4 23.6 4.0 19.4 4.7 26.4 9.1
0.4 - 0.6 19.4 3.7 15.8 7.9 13.4 9.0
0.6 - 0.9 17.0 4.8 16.8 11.1 11.1 9.4
0.9 - 1.2 12.2 7.6 15.5 14.3 12.6 10.2

68 1
With the 25, 50, and 75 m spacings, the watertable rose above 1.0 m for about 85,
90, and 108 days, respectively, during the 5-year period from 1984 to 1988. Yet, during
the monsoon season, the time-averaged depth of the watertable remained well below
0.8 m with all spacings. This suggests that the spacings can be fairly wide ( > 75 m)
and/or that the drain depths can be considerably reduced.
The average discharge rates during the monsoon season (i.e. from July to September)
for the 25, 50, and 75 m spacings were, respectively, 8. I , 2.2, and I . 1 mm/d. The rate
for the 25 m spacing is very high, and is difficult to explain. It is much higher than
the leaching requirement. In such a situation, one ought to consider a combination
of surface and subsurface drainage systems to relax the subsurface drainage
requirements, or one ought to examine whether water conservation could be improved.
The last objective could be achieved by restricting the drain outflow (Qorani et al.
1990), but also by reducing drain depth and increasing the spacing (Oosterbaan and
Abu Senna 1990).
The evacuation of the salty drainage water in the dry season is not desirable because
it would contaminate the river water below the outlet. It was found that the drainage
water can be re-used for irrigation in the dry season when the salt concentration of
the drainage water is reduced from the usual 12 kg/m3 to 6 kg/m3 by mixing it with
fresh irrigation water (Sharma et al. 1990). With such a mix, the crop production
is hardly affected, provided that the resulting accumulation of salts in the soil is
removed by drainage during the rainy season. Evacuating the salty drainage water
in the rainy season is not harmful owing to the high river discharges so that the
contamination of the river water is negligible. Instead of pumping the drainage water
for irrigation, one can also refrain from pumping, letting the crops use groundwater
directly (Ra0 et al. 1992), thereby saving irrigation water.
Suitable drainage criteria appear to be the following:
- During the monsoon season, the average depth of the watertable should be 0.8 m

to ensure sufficient dryness of the soil;

- During the dry season, the average depth of the watertable should be 0.5 m to ensure

an efficient irrigation and to provide an opportunity for the plants to use

groundwater by capillary rise, yet providing sufficient soil aeration.

With these criteria, an adequate salt balance of the soil is guaranteed and
environmental requirements are met.
The design discharge during the monsoon season follows from the average excess
rainfall in that period. During the dry season, the water balance will show that the
design discharge is nil, so that no drainage is required. The required depth of the
watertable is brought about naturally.

17.5.4 Rain-Fed Lands in Tropical Humid Zones

The humid tropics are characterized by long-lasting rainy seasons (more than 8
months) with an annual rainfall exceeding 2000 mm. Waterlogging occurs frequently
in the flat areas. As in the sub-humid zones, one has to assess the extent to which
inundations from rivers, lakes, or seas contribute to the waterlogging. When the
inundations have a strong influence, no attempt should be made to implement a

682
drainage system without a flood-control scheme. Further, investigations ought to be
made to check whether an adjustment of the cropping system would be sufficient to
eliminate the drainage problem. If a drainage system is still found to be necessary,
a surface drainage system is usually the appropriate choice, because subsurface
drainage systems in the humid tropics are often prohibitively expensive as they would
have to be designed for very high discharge capacities and would need very narrow
spacings. Only when the soil's hydraulic conductivity is very high could the spacing
be wide enough to be practically feasible.
In the following paragraphs, an example will show how the discharge capacity was
determined for the collectors serving a surface drainage system in a coastal plain in
Guyana. Another example will demonstrate the effects of subsurface drainage systems
on agriculture in a coastal plain of Kalimantan, Indonesia.

Example 17.7 Guyana

This example concerns the collectors for surface drainage systems in sugarcane
plantations in the coastal region of Guyana (Naraine 1990).
According to Equation 16.4 (Chapter '1 6), the surface water balance, for a period
of one day, reads
D,, P - I - E, + D,i - AWs (17.5)
where
D,, = runoff depth (mm)
P = precipitation (mm)
I = infiltration (mm)
E, = evaporation from the surface (mm)
DSi = surface inflow depth (mm)
AW, = change in storage of surface water (mm)

In this example, the term D,, can be set equal to zero. Because we consider a short
period with intensive rainfall, the term E,, can also be neglected. Thus Equation 17.5
can be reduced to
D,, = P - I - AW,
The Curve Number Method (Chapter 4) uses this balance (Equation 4.2) to calculate
the runoff. This will also be done here.
Table 17.1O shows data on the cumulative 5-day rainfall with a I0-year return period
and the resulting cumulative surface runoff D, calculated with the Curve Number
method, using a Curve Number value of 40. This empirical method takes into account
the storage AWs and infiltration I in the sugarcane fields, but not the dynamic storage
in the fields that is needed to induce the discharge, as will be explained below. Table
17.1O also shows the daily surface runoff Dl and the surface runoff rate q,, as a time
average of the cumulative surface runoff: q,, = D,/t, where t is the time (days). Note
that D, = CD, and qso= ZD,/t.
The design discharge of the main drainage system can be chosen as the maximum
value of the average surface runoff rate: qdeslgn = qso(mex) = 35 mm/d. It occurs after
3 days, which is the critical period because, with shorter or longer durations, the qso
values are less than 35 mm/d.

683
Table 17.10 Example of a rainfall-runoff relationship with a return period of 10 years in the case study
of Guyana, using the Curve Number method with a Curve Number value C N = 40

~~
Surface runoff
~

Duration Cumulative Cumulative Daily Average surface

t rainfall DC Di runoff rate
P 9so = Dc/t
(d) (") ("> (") ("/a
1 2 3 4 5
1 150 14 14 14
2 250 59 45 29
3 325 104 45 35
4 360 128 24 32
5 375 138 10 28

The cumulative surface runoff (Dc,Column 3 in Table 17.10) is plotted in Figure 17.37
against the time. It shows a curve with an S-shape. The slope of the tangent line from
the origin to this curve indicates the required discharge capacity of the collectors,
with a return period of 10 years (qdesign
= 35 mm/d).
The S-shape of the runoff curve, which is initially quite flat, shows that the drainage
system cannot immediately function at its maximum capacity: there is a delay in the
functioning and a necessary dynamic storage. The daily dynamic storage can be found
from
AWi = Di - qso (17.6)
Table 17.11 shows the development of AWi and cumulative dynamic storage AWc =

cumulative
runoff I discharge
in mm
200

tan a = 35 m m l d

150 ~ -__ -

1 O0

50

O 1 2 3 4 5 6
time in days
-a
. cumulative runoff fromthe field (Dc, Table 17.10)
O----. cumulativedischargeiromthefield (Oc'. Table 17.11)
t j c u m u l a t i v e live storagefieldanddrain(AWc.Table17.11)
- - tangent line from origin
Figure 17.37 Runoff and discharge versus time in the example of Guyana

684
Table 17. I 1 Daily and cumulative dynamic storage and discharge derived from Table 17. IO.

Time Storage Discharge

Daily Cumulative Cumulative Daily
AWi AWC D,* D;
(d) (mm) (") (") (")
1 O O 14 14
2 16 16 43 29
3 10 26 78 35
4 -6 20 108 30
5 -18 2 136 28

CAW, with time. Further, it shows the cumulative discharge = D, - AWc and
the daily discharge DT = Di - AWi. Note that = CD:.
It can be seen from Table 17.11 that the daily storage AW! is positive up to the
critical time t = 3 days, after which it becomes negative. The cumulative storage AW,
= CAW, therefore increases up to t = 3 days, and afterwards decreases. The table
also shows that the maximum daily discharge (DT = 35 mm/d) occurs during the
3rd day and it equals the design discharge qdesign determined from the tangent line
in Figure 17.37 and from q,,,,,,, in Table 17.1O.
Naraine (1990) plotted the yield versus the number of high-water days (NHW),
defined as the number of days per season during which the water level in the collectors
exceeded a level corresponding to a depth of 0.9 m below the soil surface (Figure
17.38). The figure shows that there is a tendency towards decreasing crop yields when
the NHW value is greater than about 7. Therefore NHW = 7 can be taken as a design
criterion for the collector drainage system.

The above analysis shows that the design of the collector drainage system can be based
on criteria that use the same principles as described for collector drainage systems
in The Netherlands (Section 17.5.1); only the quantitative values need to be adjusted:
- There should be a high water-level criterion (HW) specifying the water level in the
drain that may be exceeded only once in I O years. (In the example of Guyana,
however, this level has not yet been determined.) The corresponding discharge is
35 mm/d;
- There should be a normal water-level criterion (NW) specifying that the water level
in the drain may be shallower than 90 cm below soil surface only for 7 days per
season. (The corresponding discharge in the example of Guyana has yet to be
defined).

Despite the relative shortcomings in the example of Guyana, the analysis permitted
Naraine to distinguish the well-drained and the poorly-drained plantations and to
recommend criteria for improved drainage systems and to calculate a benefit/cost
ratio.

685
yield of sugarcane
intlha
125

íío

95

80

65

50

NHW

Figure 17.38 Crop yield versus number of days (NHW) with a high water level (above 90 cm below soil
surface) in the collector system in the example of Guyana (Naraine 1990)

Example 17.8 Indonesia

The coastal area of Southern Kalimantan, Indonesia, is characterized by the presence
of large, deep rivers, between which flat marine and alluvial soils have formed. The
soils often contain large amounts of organic matter and/or large amounts of acidifying
sulphuric material.
The climate is characterized by an annual rainfall of about 2800 mm, of which
roughly 2000 mm evaporates. The excess rainfall, therefore, is about P - E = 800
mm/yr. Despite the high excess rainfall, few inundations from the rivers occur owing
to their enormous hydraulic transport capacity. Inundations are only apparent near
the sea shore and stem from oceanic tides.
Long ago, the inhabitants dug canals from the riversides into the interior of the
land. These hand-dug canals are 5 to 10 km long and are spaced at 300 to 500 m.
They have an important drainage function as they evacuate the main part of the excess
rainfall; they are also used for transport by boat.
A research project in the region has established that the hydraulic conductivity of
the soils is extremely high (AARD and LAWOO 1992). Over a depth of D = 2 m
or more, the highly cracked soils have a hydraulic conductivity of K = 100 to 300
m/d. The soils’ hydraulic transmissivity values therefore range between K D = 300
and 800 m2/d.During the months with the highest rainfalls (November to May), the
excess rainfall P - E (equalling the net recharge Rd)can be estimated at 700 mm to
800 mm, giving an average of about Rd = 3 mm/d. From Equation 17.1 (Section
17.3.4), setting Ah = O, we find that the average drain discharge qd also equals 3
mm/d. Using a canal spacing of L = 500 m and a transmissivity value K D = 500

686
 m2/d, we can calculate the hydraulic head h, using Hooghoudt’s drainage formula
(Chapter 8) and taking q = q,/lOOO m/d, as

h = - q Lz - 0.003 (500)z = 0.2m

8KD- 8x500
Since the water level in the canals has an average depth, g, of about 0.5 m below
the soil surface, the average depth of the watertable, j, is found at j = g - h = 0.5
- 0.2 = 0.3 m below the soil surface. For short periods with high intensity rainfalls,
the watertable may rise close to the soil surface, so that the hydraulic head equals
h = 0.5 m. The discharge rate of the canals then becomes
I

8KDh - 8 x 500 x 0.5 = o,oo8 mld

q = F - 50O2
This is a high value, and many farmers in the region have observed that it is difficult
to maintain a permanent water layer on their rice fields, and that high water levels
in their fields after intensive rainfall drop in a matter of 2 or 3 days. Therefore, the
drains are equipped with check gates.
It is not yet possible to decide whether the region is excessively drained by the
traditional hand-dug canals or not. To evaluate the agricultural drainage criteria, it
would be necessary to take into account the drainage requirements of crops other
than rice, the extent to which rice fields and fields with other crops are contingent,
and the possibility that occasionally deeper watertables may have a favourable effect
on the soil structure or the quality of the soil’s organic matter. There are indications
that maintaining the watertable at a modest depth below the soil surface during a
period with non-rice crops, has a positive effect on the soil’s fertility and acidity
(AARD & LAWOO 1992).

References
A A R D & LAWOO 1992. Acid sulphate soils in the humid tropics : water management and soil fertility.
AARD, Jakarta, 283 p.
Advisory Panel on Land Drainage in Egypt 1982. The influence of the watertable depth, soil salinity, and
groundwater salinity on the crop production of cotton. Drainage Research Institute, Cairo.
Baily, A.D. 1979. Benefits from drainage on clay soils. Paper No. 79-2549, ASAE winter meeting, American
Society of Agricultural Engineers, St. Joseph.
Berryman C. 1975. Improved production from drained grassland. Techn. Bulletin 75/7, Field Drainage
Experimental Unit, Cambridge.
Blaauw, H. 1961. De berekeningen van waterlopen en kunstwerken. Cultuurtechn. Tijdschr., I , 3, pp.
79-100.
Bolton, E.F., J.W. Aylersworth and F.R. Hore 1970. Nutrient loss through tile drains under three cropping
systems and two fertility levels on a Brookston clay soil. Canadian Journal of Soil Science, SO, 3, pp. 275-279.
De Jong, M.H. 1979. Drainage of structured clay soils. In: J. Wesseling (ed.), Proceedings of the
International Drainage Workshop. ILRI Publication 25, Wageningen, pp. 268-280.
Doty, C.W., S.T. Currin and R.E. McLin 1975. Controlled subsurface drainage for southern coastal plains.
Journal of Soil and Water Conservation, 30,2, pp. 82-84.
Eriksson J. 1979. Soil functions and drainage. In: J. Wesseling (ed.), Proceedings of the International
Drainage Workshop. ILRI Publication 25, Wageningen, pp. 180-212.
Euroconsult 1976. Technical and economic evaluation of different drainage techniques. Euroconsult,
Arnhem.

687
FDEU 1972. Annual Report. Field Drainage Experimental Unit, Ministry of Agriculture, Cambridge.
Feddes, R.A. 1971. Water, heat and crop growth. Thesis, Agricultural University, Wageningen, 184 p.
Feddes, R.A. and A.L.M. van Wijk 1977. An integrated model approach to the effect ofwater management
on crop yields. Agricultural Water Management, I , I , pp. 3-20.
Found, W.C., A.R. Hill and E.C. Spence 1976. Economic and environmental impacts of agricultural land
drainage in Ontario. Journal of Soil and Water Conservation, 3 I , pp. 20-23.
Harris, C.R., H.T. Erickson, N.K. Ellis and J.E. Larson 1962. Water level control in organicsoil, as related
to subsidence rate, crop yield and response to nitrogen. Soil Science, 94, pp. 158-161.
Irwin, R.W. 1981. On farm drainage benefit. Omaf Res. Progr. 39, Proj. A-39-742, School of Engineering,
University of Guelph, Canada.
Jaynes, D.B., R.C. Rice and R.S. Bowman 1988. Independent calibration of a mechanistic-stochastic model
for field-scale solute transport under flood irrigation. Soil Science Society of America Journal, 52, pp.
1541-1456.
Lenselink, K.J. 1972. Drainage requirements for banana in the young coastal plain (in Dutch, title translated
by author). De Surinaamse Landbouw, Vol. 20, pp. 22-36.
Minessy, F.A., M.A. Barakat and E.M. EI Azab 1971. Effect of some soil properties on root and top
growth and mineral content of Washington Navel orange and Balady mandarin. Plant and Soil, 34,
pp. 1-15.
Naraine, D.S. 1990. The development of drainage criteria and their application to sugar cane cultivation
in Guyana. MSc. thesis, International Institute of Hydraulic Engineering, Delft.
Nijland, H.J. and S. EI Guindy 1984. Crop yields, soil salinity and watertable depth in the Nile Delta.
In: ILRI Annual Report 1983, Wageningen, pp. 19-29.
Nijland, H.J. and S. EI Guindy 1986. Crop production and topsoil/surface water salinity in farmers’ irrigated
rice fields, the Nile Delta. In: K.V.H. Smith and D.W. Rycroft (eds.), Hydraulic Design in Water
Resources Engineering : Land Drainage. Proceedings of the 2nd International Conference, Springer,
Berlin, pp. 383-392.
Nolte, B.H., N.R. Fausey and R.W. Skaggs 1982. Time available for field work in Ohio. Summer Meeting,
Paper No. 82-2076, American Society Agricultural Engineers, St. Joseph.
Oosterbaan, R.J. 1982. Crop yields, soil salinity and watertable depth in Pakistan. In: ILRI Annual Report
1981, Wageningen, pp. 50-54.
Oosterbaan, R.J. 1988. Agricultural criteria for subsurface drainage: a systems analysis. Agricultural Water
Management, 14, pp. 79-88.
Oosterbaan, R.J. 1992. Agricultural land drainage : a wider application through caution and estraint. In:
ILRI Annual Report 1991, Wageningen, pp. 21-36.
Oosterbaan, R.J., H.A. Gunneweg and A. Huizing 1987. Water control for rice cultivalion in small valleys
of West Africa. In: ILRI Annual Report 1986, Wageningen, pp. 30-49.
Oosterbaan, R.J. and M. Abu Senna 1990. Using Saltmod to predict drainage and salinity in the Nile
Delta. In: ILRI Annual Report 1989, Wageningen, pp. 63-75.
Oosterbaan, R.J., D.P. Sharma, K.N. Singh and K.V.G.K. Rao 1990. Crop production and soil salinity:
evaluation of field data from India by segmented linear regression. In: Symposium on land drainage
for salinity control in arid and semi-arid regions, Vol. 111. Drainage Research Institute, Cairo, pp. 373-383.
Qorani, M., M.S. Abdel Dayem and R.J. Oosterbaan 1990. Evaluation of restricted subsurface drainage
in rice fields. In: Symposium on land drainage for salinity control in arid and semi-arid regions, Vol.
3. Drainage Research Institute, Cairo, pp. 415-423.
Rao, K.V.G.K., P.S. Kumbhare, S.K. Kamra and R.J. Oosterbaan 1990. Reclamation ofwaterlogged saline
alluvial soils in India by subsurface drainage. In: Symposium on land drainage for salinity control in
arid and semi-arid regions, Vol. 2. Drainage Research Institute, Cairo, pp. 17-25.
Rao, K.V.G.K., D.P. Sharma and R.J. Oosterbaan 1992. Subirrigation by groundwater management with
controlled subsurface drainage in semi-arid areas. International Conference on Supplementary Irrigation
and Drought Management, Bari, Italy.
Rudd, A.V. and C.W Chardon 1977. The effects of drainage on cane yields as measured by watertable
height in the Machnade Mill area. In: Proceedings of the 44th Conference of the Queensland Society
of Sugar Cane Technology, Australia.
Safwat Abdel-Dayem and H.P. Ritzema 1990. Verification of drainage design criteria in the Nile Delta,
Egypt. Irrigation and Drainage Systems, 4,2, pp. 1 17- I3 I .
Schothorst, C.J. 1978. Het zakkingsproces bij ontwatering van de westelijke veenweide gronden.
Landbouwkundig Tijdschrift, 6, pp. 167-175.

688
Schwab, G.O., G.S. Taylor, J.L. Foussand Ehud Stibble 1966.Crop response from tile and surface drainage.
Soil Science Society of America Proceedings, 30, pp. 634-637.
Shalhevet, J. and P.J. Zwerman 1962. Nitrogen response of corn under variable conditions of drainage.
Soil Science, 93, pp. 172-182.
Sharma, D.P., K.N. Singh, K.V.G.K. Rao and P.S. Kumbhare 1990. Re-use of saline drainage water for
irrigation in a sandy loam soil. In: Symposium on land drainage for salinity control in arid and semi-arid
regions, Vol. 3. Drainage Research Institute, Cairo, pp. 304-312.
Sieben, W.H. 1965. Het verband tussen ontwatering en opbrengst bij de jonge zavel gronden in de
Noordoostpolder. Van Zee tot Land, 40, Tjeenk Willink, Zwolle, 117 p.
Skaggs, R.W. 1980. Combination surface-subsurface drainage systems for humid region soils. Journal
Irrigation and Drainage Division, Proceedings American Society of Civil Engineers, 106, IR4, pp.
265-283.
Suclla Flores, J.A. 1972. Diseño de la infraestructura de recupjxación en zonas afectadas con problemas
de drenaje en el Anexo Chiquitoy - C.A.P. Cartavio, Trujillo - Peru. Tesis, Programa Académica de
lngenería Agricola, Universidad Nacional Agraria ‘La Molina’, Lima.
Todd, D.K. 1980. Groundwater hydrology, 2nd. ed. Wiley, New York, 535 p.
Trafford, B.D. 1972. The evidence in literature for increased yields due to drainage. Technical Bulletin
72/5, Field Drainage Experimental Unit, Cambridge.
Van Alphen, J.G. and F. de los Rios Romero 1971. Gypsiferous soils. ILRI Bulletin 12, Wageningen,
44 p.
Van Hoorn, J.W. 1958. Results of a groundwater level experimental field with arable crops on clay soils.
Netherlands Journal of Agricultural Science, 6, pp. 1-10,
Van Hoorn, J.W. 1979. The effect of capillary flow on salinization and the concept of critical depth for
determining drain depth. In: J. Wesseling (ed.), Proceedings of the International Drainage Workshop,
ILRI Publication 25, Wageningen, pp. 686-700.
Van Someren, C.L. 1958. Het drainage-onderzoek in Nederland. Cultuurtechnische Dienst, Utrecht, 5 1 p.
Wesseling, J. 1974. Crop growth and wet soils. In: J. van Schilfgaarde (ed.), Drainage for agriculture.
Agronomy 17, American Society of Agronomy, Madison, pp. 7-38.
Wind, G.P. and J. Buitendijk 1979. Simulation over 35 years of the moisture content of a topsoil with
an electric analog. In: J. Wesseling (ed.), Proceedings of the International Drainage Workshop. ILRI
Publication 25, Wageningen, pp. 214-219.
Yamada, N. 1965. Some problems of irrigation and drainage in rice culture. International Rice Commission
Newsletter, 14,3, pp. 13-28.
Young, T.C. and J.T. Ligon 1972. Watertable and soil moisture probabilities with tile drainage.
Transactions of the American Society of Agricultural Engineers, 15,3, pp. 448-45 I .

689