Contribution to irrigation from shallow water table
under field conditions
C. Babajimopoulos1, A. Panoras2, H. Georgousis1, G. Arampatzis2,
E. Hatzigiannakis2, D. Papamichail1
1
Laboratory of General & Agricultural Hydraulics & Land Reclamation, Faculty of Agriculture
Aristotle University of Thessaloniki,
2
Land Reclamation Institute, N.AG.RE.F.
SUMMARY
The mathematical model SWBACROS was applied to estimate the contribution of a
shallow groundwater to the water needs of a maize crop. The model was applied with
the top and boundary conditions defined by the observed irrigation/rainfall events and
the observed water table depth. The simulated water contents of the root zone were
very close to the observed values. Furthermore the model was applied with an
assumed free drainage bottom boundary condition. The difference of the computed
water content profiles under the presence and absence of the water table gave a very
good estimate of the capillary rise. It was found that under the specific field
conditions about 3.6 mm/day of the water in the root zone originated from the shallow
water table, which amounts to about 26% of the water, which was transpired by the
maize crop.
Keywords: shallow water table, irrigation, capillary rise, SWBACROS
1. INTRODUCTION
World population today is about 6.5 billion and it is estimated that it will be increased
to 9.1 billion by the year 2050 (UN, 2004). Irrigation supplies approximately 40 % of
the world foodstuffs on less than 18% of the arable land and has a significant future
role in meeting the projected world food demand (Ayars et al., 2006). It is estimated
that irrigation consumes more than 80% of the good quality water. This makes it the
greatest user of water, very far from its other competitors, namely urban, industrial
and environmental use. It is more than certain that competition among agricultural,
urban, industrial and environmental needs will be even more intense in the near
future. Any effort towards improving irrigation efficiency is worthwhile because it
can lead to saving large quantities of good quality water.
Shallow ground water table exists in many areas of the world. This shallow ground
water can be used by plants either by using drainage water for irrigation or through
in–situ use. Saline drainage ground water has been studied extensively as a
supplemental source of irrigation water (Rhoades et al., 1989; Ayars et al., 1993;
Ayars et al., 2006). In–situ use of ground water by crops is a more complicated matter
than irrigating with drainage ground water. It depends on several factors such as depth
to the water table, hydraulic properties of the soil, stage of the crop growth, ground
water quality etc. Quantification of the water taken by the roots from the shallow
water table is of great significance and has been a topic of extensive research in the
last few decades.
Wallender et al. (1979) found that 60 % of the evapotranspiration (ET) of a cotton
crop was extracted from a 6 dS m-1 shallow water table.
Ayars and Schoneman (1986) found that capillary rise of water of ECe = 10 dS m-1
from a water table of 1.7 – 2.1 m deep contributed to up to 37% of evapotranspiration
(ET) of a cotton crop.
Prathapar and Qureshi (1998) observed that under shallow water table conditions
irrigation can be reduced up to 80 % without affecting crop yield and increasing soil
salinization.
Soppe and Ayars (2003) by using weighing lysimeters maintained a saline (14 dS m -1)
water table at 1.5 m depth and found that ground water contributed of up to 40% of
daily water used by safflower crop. On a seasonal basis 25% of the total crop water
use originated from the ground water. The largest contribution occurred at the end of
the growing season when roots were fully developed. The applied irrigation in the
presence of a water table was 46 % less than irrigation applied to the crop without a
water table.
Kahlown et al. (2005) investigated the effect of shallow water tables on crop water
requirements by using 18 large size drainage type concrete lysimeters. They found
2
that when a water table was kept at a depth of 0.5m, wheat met its entire water
requirement from the ground water. Sunflower required only 20 % of its total need
from irrigation. However maize and sorghum were found to be waterlogging sensitive
crops whose yield were reduced with higher water table.
Most of the previous research is conducted by using weighing and drainage
lysimeters. The method is very accurate but has serious limitations because of the
cost of construction, operation and maintenance of the lysimeters. A common
characteristic of most of this research is that it is focused on the existing conditions of
the experiment and the conclusions are difficult to be extended to other situations.
Thessaloniki plain, Greece, covers an area of about 100000 hectares. The plain is
cultivated with cotton (34%), maize (12%), rice (17%), sugarbeets (6%), alfalfa (3%),
orchards (20%) and some other crops in smaller extend. Bad irrigation management,
non functional drainage systems and also seepage from the rice fields, which are
covered with water for long periods of time, result to high water table during summer
in many parts of the plain. Salts accumulate in the root zone but are leached in winter
when the water table deepens.
In some areas of the plain the depth to the water table during the cultivation period
can be as low as 40–50 cm. Capillary rise is very large and contributes to crop water
requirements significantly. Even though farmers reduce irrigation taking advantage of
this shallow water table, contribution of this to transpiration has never been
quantified. It is obvious that if it is managed correctly, groundwater can contribute
significantly to crop water needs and therefore reduce applied irrigation.
The objective of this paper is to estimate ground water contribution to transpiration.
Towards this goal, the mathematical model SWBACROS (Babajimopoulos et al.,
1995) is used. The model is used with the top and bottom boundary conditions
defined by the observed irrigation/rainfall events and water table depths, respectively.
The simulated water contents are very close to the observed values. Furthermore a
free drainage bottom boundary condition is assumed. The difference of the computed
water content profiles under the presence and the absence of the water table is an
estimation of the water used by the crop because of the capillary rise. It is found that
about 26% of transpiration of a maize crop originates from the shallow water table.
3
2. MATERIALS AND METHODS
2.1 The Mathematical model
The mathematical model is based on the equation which describes the unsaturated,
transient, water flow in a heterogeneous soil under the presence of a crop:
C  
h  
  h 

K   
 1   S  h 

t  z 
  z 
(1)
where C(θ)=  θ/  h is the specific moisture capacity function (1/L), h is the pressure
head (L) which is negative in unsaturated soil, K(θ) is the unsaturated hydraulic
conductivity (L/T), θ is the volumetric soil water content (L3/L3), z is the vertical
dimension directed positive downward (L), t is the time (T) and S is the root water
uptake (1/T).
A detailed description of the model is given by Babajimopoulos et al. (1995). It is
referred here that the unsaturated hydraulic conductivity and the specific moisture
capacity functions are computed as in Van Genuchten (1978,1980). The sink term is
computed as in Belmans et al. (1983) by:
S h   hSmax
(2)
where a( h ) is a dimensionless prescribed function of pressure head and Smax is the
maximal possible water extraction by roots. Figure 1 is a graphical representation of
eqn (2) with the x axis representing absolute values of the pressure head. Water
uptake is maximal between h1 and h2 and varies linearly between 0 and h1 and
between h2 and h3. Actual transpiration is computed by integration of eqn (2) with
respect to depth.
Figure 1: General shape of the sink term S (Feddes et al., 1978)
The potential transpiration rate, E, is computed by
4
E  ETP  EV
(3)
where ETP is the potential evapotranspiration rate computed by the modified Penman
method (Doorenbos and Pruitt, 1984) and EV is the potential evaporation rate from
soil surface computed as in Al–Khafaf et al. (1978) by
EV  ETP exp  0.623LAI 
(4)
where LAI is the Leaf Area Index function.
If irrigation/precipitation is less than 10 mm/day (Belmans et al., 1983) then
evaporation from soil surface is computed as in Al–Khafaf et al. (1978) by
EVA  t 0 .6   t 1
0 .6
(5)
where σ is a soil dependent parameter and t is the time in days after the dry period
starts. When evaporation is given in mm/day then according to Al–Khafaf et al.(1978)
and Ritchie (1972) σ varies between 3.34 and 5.8.
Equation (1) is solved by the Douglas–Jones predictor–corrector method (Douglas
and Jones, 1963, Babajimopoulos, 1991, Babajimopoulos, 2000).
2.2 Area of study and field data
The model was applied in a 6 acre experimental field situated in the area of the
Thessaloniki plain (φ: 40 39΄ 13΄΄ , λ: 22 46΄ 01΄΄ WGS84 projection) and
cultivated with maize. The experimental field was irrigated by surface irrigation
methods. The water table of 0.93 dS m-1 was observed at an average depth of 0.6m
below soil surface and its variations were recorded by two piezometers. Figure 2
shows the variation of water table depth with time.
0.0
0.1
Water table depth (cm)
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
22/06/04 02/07/04 12/07/04 22/07/04 01/08/04 11/08/04 21/08/04 31/08/04 10/09/04
Date
Figure 2: Variation of water table depth with time
5
The soil retention curve was estimated by pressure cell and pressure plate methods
(Childs, 1969). The coefficients a and n of the Van Genuchten model were
determined by non linear regression of the soil water retention data. The results of the
mechanical composition and the values of the parameters describing the hydraulic
properties of the soil (measured or computed) are given in Table 1.
Table 1: Mechanical composition and hydraulic parameters of the soil
Van Genuchten retention parameters
Depth Sand
Silt
Clay
USDA
θs
θr
a
Ks
(cm)
(%)
(%)
(%)
texture
n
(cm3 cm-3) (cm3 cm-3) (m-1)
(m/day)
0 – 30
33.2
47.0
19.8
L
0.452
0.000
2.497
1.089
0.5
30 – 70
32.4
53.5
14.1
SiL
0.517
0.160
0.349
1.736
0.003
The values of the parameters h1, h2 and h3 of the function S(h) were obtained from
Al–Khafaf et al. (1978) and are the following: h1 = -100 cm, h2=-600 cm, h3=-15000
cm. Α value of σ equal to 4 (Al–Khafaf et al., 1978) was chosen in eqn (5) for the
calculation of evaporation under non potential conditions. Root depth was measured
by isolating the root zone and carefully washing it. The Leaf Area Index (LAI)
function with respect to calendar day was measured by a Delta-T Devices SunScan
Canopy Analysis System. An equation of the form (Dale et al., 1980):
A
Amax
1  a exp[b(t  to )]
(6)
was used to describe LAI and rooting depth (RD), where t is calendar day. The
parameters appearing in eq. (6) are defined in Table 2.
Table 2: Values of parameters of eq. (6)
A(LAI or RD)
LAI
Amax
α
b
to
RD
111<=t<=219
219<t<=255
4.7
14971.6
0.18076
111
4.7
0.242·10-3
0.2752
219
50
223331.2
0.156892
105
Climatic data were obtained from an adjacent weather station. In addition to rainfall,
a total of 73.9 mm of water were applied by the farmer during six irrigation events to
cover crop water needs. Disturbed soil samples were used to measure gravimetrically
the soil moisture content twice a week. Undisturbed soil samples were used to
determine the saturated hydraulic conductivity of each layer using a falling head
permeameter.
6
3. RESULTS
The mathematical model was applied to predict the moisture content of the field under
observed irrigation/rainfall conditions and observed water table depth. The simulation
period started on 15 June 2004 and ended on 11 September 2004. The simulation
ended on that particular day because after that an extreme rainfall event (82.8mm)
flooded the experimental field for 7 consecutive days.
Figure 3 shows the moisture content of the top zone of the soil predicted by the model
and the observed moisture content. The average mean error was 0.53·10-2 cm3/cm3
and the coefficient of variation (CV) was 6.7%. In the same figure, the predicted
moisture content under an assumed free drainage bottom boundary condition is also
shown.
16/06/04
01/07/04
16/07/04
01/08/04
16/08/04
01/09/04
20
0.6
30
0.4
0.2
0.0
Legend
Irrigation + water table
Irrigation, no water table
Θobs
5.0
4.0
3.0
2.0
1.0
0.0
16/06/04
01/07/04
16/07/04
01/08/04
16/08/04
Groundwater
contribution (mm)
moisture content (cm3 . cm-3)
10
irrigation and
precipitation (mm)
0
01/09/04
Date
Figure 3: Observed moisture content, simulated moisture content with the existence and absence of a
water table and groundwater contribution to crop needs.
The difference between the two simulated curves of this figure is due to the capillary
rise because of the presence of the high water table. Groundwater contribution to the
moisture content of the root zone for the simulated period is also shown in Figure 3.
For the period from July 1st to September 11th, 2004 which was simulated, this
contribution varies between 1.68 and 4.83 mm with an average value of 3.6 mm day-1.
This sums up to about 282.2 mm. When this amount is added to the applied irrigation
of 73.9 mm this gives a sum of 356.1 mm.
It is pointed out that potential
evapotranspiration for the period of 1st July to 11 September 2004 as it is computed
by modified Penman method is 344mm.
7
Rooting depth below surface and groundwater contribution are shown in Figure 4.
Groundwater contribution is higher when roots are fully developed. It is worth
mentioning in this figure that a large amount of the water originated either from
irrigation or from ground water is added to the top zone when the crop is not fully
developed yet. Therefore this amount is not contributing significantly to the
transpiration of the crop. That is obvious in Figure 5 where the computed transpiration
under the presence and the absence of a water table is shown. It appears that about
26% of the transpired water is due to the shallow water table. That seems to be in
contrast with the fact that the water of the root zone for the simulated period was
356.1mm while the estimated evapotranspiration was 344mm. Actually it is not
because a large amount out of the 356.1 mm was added to the top zone before the
time the crop was able to use it.
16/06/04
01/07/04
16/07/04
01/08/04
16/08/04
01/09/04
10
0.0
30
0.2
0.4
0.6
0.8
1
5.0
4.0
3.0
2.0
1.0
0.0
16/06/04
01/07/04
16/07/04
01/08/04
16/08/04
Groundwater
contribution (mm)
Rooting depth (cm)
20
irrigation and
precipitation (mm)
0
01/09/04
Date
Figure 4: Variation of rooting depth and groundwater contribution with time
16/06/04
01/07/04
16/07/04
01/08/04
16/08/04
01/09/04
0
Transpiration (mm)
20
Legend
Potential transpiration
Actual transpiration (irrig and water table)
Actual transpiration (irrig, no water table)
400
irrigation and
precipitation (mm)
10
30
300
200
100
0
16/06/04
01/07/04
16/07/04
01/08/04
16/08/04
01/09/04
Date
Figure 5: Cumulative potential and actual transpiration under the existence and absence of a water
table.
8
It is apparent that groundwater contribution, as it is estimated in this work is restricted
to the specific field conditions. Even though our experimental field represents a
typical soil type in the plain of Thessaloniki, application of the model to the rest soil
types of the plain will lead to a very good estimate of the groundwater contribution
under different field conditions. This will certainly lead to a much better management
of irrigation in the plain of Thessaloniki.
4. CONCLUSIONS
Crop water use from shallow water tables is affected by several factors such as depth
to water table, rooting depth, ground water quality, crop salt tolerance, soil type,
irrigation frequency and application depth. Because of this complexity it is impossible
to conduct experiments that cover all the factors at once and the research is focused
on only a single component, i.e. water use relative to the water table depth, or ground
water quality or soil type (Ayars et. al., 2006). The method which was used in this
paper is relatively easy, does not have the limitations of the weighing lysimeters (cost
of construction, operation, maintenance) and can study several factors at the same
time e.g. depth to the water table, crop growth stage, soil type.
It was found that under the specific field conditions groundwater contribution to the
root zone was about 3.6 mm/day. This amounts to about 26% of the transpired water
for the period of 1st July to 11 September 2004. The model can be applied to the
prevailing soil types in the plain of Thessaloniki. This will lead to a very good
estimate of the groundwater contribution to crop needs and therefore to a significant
improvement of the irrigation efficiency.
REFERENCES
Al–Khafaf, S., Wierenga, P.J., Williams, B.C., 1978. Evaporative flux from irrigated
cotton as related to leaf area index, soil water and evaporative demand, Agronomy
Journal, 70, 912–917.
Ayars, J.E., Schoneman, R.A., 1986. Use of saline water from a shallow water table
by cotton. Trans. ASAE, 29, 1674–1678.
Ayars, J.E., Hutmacher R.B., Schoneman, R.A., Vail, S.S., Pflaum, T., 1993. Long
term use of saline water for irrigation. Irrig. Sci., 14, 27–34.
9
Ayars, J.E., Christen, E.W., Soppe R.W., Mayers, W.S., 2006. The resource potential
of in–situ shallow ground water use in irrigated agriculture: a review. Irrig. Sci.,
24(3), 147–160.
Babajimopoulos, C., 1991. A Douglas–Jones Predictor–Corrector method in
simulating one–dimensional unsaturated flow in soil. Ground Water, 29(2) 271–
286.
Babajimopoulos, C., Budina, A., Kalfountzos, D., 1995. SWBACROS: A model for
the estimation of the water balance of a cropped soil. Environmental Software, 10,
3, 211–220.
Babajimopoulos, C., 2000. Revisiting the Douglas–Jones method for modeling
unsaturated flow in cultivated soil. Environmental Modeling & Software, 15, 303–
312.
Belmans, C., Wesseling, J.G., Feddes, R.A., 1983. Simulation mode of the water
balance of a cropped soil. SWATRE, J. of Hydrology, 63, 271–286.
Childs, E.C., 1969. The physical basis of soil water phenomena. Wiley Interscience,
London, 493 pp.
Dale R.F., Coelho D.T., Gallo K.P, 1980. Prediction of daily green leaf area index for
corn. Agronomy Journal 72: 999–1005.
Doorenbos, J. and Pruitt, W.O., 1984.
Guidelines for predicting crop water
requirements, FAO Irrigation and Drainage Paper 24.
Douglas, J.J. and Jones, B.F., 1963. On predictor–corrector method for non linear
parabolic differential equations, J. SIAM, 11, 195–204.
Feddes, R.A., Kowalik, P.J., Zaradny, H., 1978. Simulation of field water use and
crop yield. Simulation Monograph, Pudoc, Wageningen.
Kahlown, M.A., Ashraf, M., Zia–ul–Haq, 2005. Effect of shallow groundwater table
on crop water requirements and crop yields. Agric. Water Manage., 76(1), 24–35.
Prathapar, S.A., Qureshi, S., 1998. Modelling the effects of deficit irrigation on soil
salinity, depth to water table and transpiration in semi–arid zones with monsoonal
rains. Water Resour. Dev., 15, 1/2, 141–159.
Rhoades, J.D., Bingham, FT., Letey, J., Hoffman, G. J., Dedrick, A.R., Pinter, P.J.,
Replogle, J.A., 1989. Use of saline drainage water for irrigation: imperial valley
study. Agric. Water Manag. 16, 25–36.
Ritchie, J.T., 1972. Model for predicting evaporation from a row crop with
incomplete cover, Water Resources Research, 8, 1204–1213.
10
Soppe, R.W.O., Ayars, J.E., 2003. Characterizing groundwater use by safflower
using lysimeters. Agric. Water Manage., 60, 59–71.
UN, 2004. The world population prospects: The 2004 revision population data base.
Department of Economics and Social Affairs. http://esa.un.org/unpp/
Van Genuchten, M.Th., 1978, Calculating the unsaturated hydraulic conductivity with
a new, closed–form analytical model,
Research Report 78–WR–08, Water
Resour.Program, Dep. of Civil Engin., Princeton Univ., Princeton, N.J., 1978.
Van Genuchten, M.Th., 1980. A closed–form equation for predicting the hydraulic
conductivity of unsaturated soils, Soil Sci. Soc. Am. J., 44, 892–898.
Wallender, W.W., Grimes, D.W., Henderson, D.W., Stromberg, S.K., 1979.
Estimating the contribution of a perched water table to the seasonal
evapotranspiration of cotton. Agron. J., 71, 1056–1060.
11