The Effect of Irrigation Schedules on Water Table Depth and Root Zone Soil Moisture

The Effect of Irrigation Schedules on Water Table Depth
and Root Zone Soil Moisture
1H. T. Nguyen and 1J. P. Walker
1Department of Civil and Environmental Engineering, The University of Melbourne,
E-Mail: thuhien@civenv.unimelb.edu.au
Keywords: Irrigation schedules; water table depth; soil moisture; root zone
EXTENDED ABSTRACT
for the simulations are typical of semi-arid south-
eastern Australia. Therefore, the results should
Indiscriminate use of irrigation water, particularly
provide at least a qualitative indication of relative
in existing areas of shallow water table, can result
effects of different irrigation scenarios on water
in further water table rise leading to water logging
table depth and root zone soil moisture.
and secondary salinity problems. Hence, it is
essential that irrigators have a clear understanding
This study shows that the time interval between
of how their often ad-hoc irrigation scheduling
flood irrigation events has a more significant
practices impact on both the local water table
impact on the depth to water table than the duration
level and on-farm soil moisture content, which
of inundation. In order to control or limit future
influences crop yield, a primary motivator of
water table rise, the interval between irrigation
irrigators.
events should be sufficiently far apart; at least 14
days in our situation. This is almost a 50% increase
We have studied the impact of irrigation
in the time between irrigation events as compared
scheduling on both water table rise and root zone
to typical practice being 4 hours every 10 days.
soil moisture content in a desk-top study. A
Moreover, an inundation period of 2 hours was
Richards’ equation based soil moisture model has
found to be sufficient to mitigate any undue water
been used to study the effect of flood irrigation
stress on the crops. This is a further water saving
frequency and duration of inundation on the water
with a 50% decrease in the inundation duration.
table depth and root zone soil moisture content.
Hence a 2 hour flood irrigation event once every
While the study was not intended to represent a
14 days during the irrigation season was found to
specific study site, the results should be
be more sustainable than the current practice.
applicable to typical flood irrigation regions in
semi arid regions having a shallow water table
This study indicates that in addition to improved
depth, such as that in south-eastern Australia.
irrigation techniques, the key to avoiding water
table rise is improved efficiency in scheduling
Using a series of simulations, we explored the
irrigation to meet as precisely as possible the water
effect of altering time between flood irrigation
needs of the crop, rather than applying irrigation
events from 5 to 20 days, and duration of flood
water in a more ad-hoc approach.
irrigation events from 1 to 6 hours. The initial
water table level, soil type and climatic data used

1286

---

shallow saline water table in south-eastern
1. INTRODUCTION
Australia by using the TOPOG-Dynamic
simulation model; and Silberstein et al. (2002)
Irrigation water is used to maximise crop yield by
have used a steady state hydrological model to
minimising water stress in the root zone. However,
study the occurrence of seasonal water logging
this is often done in an ad-hoc manner.
across a 639 ha catchment in south-western
Indiscriminate use of irrigation water has led to
Australia. It was found that there are some
problems of rising water tables causing widespread
solutions could reduce both rates of groundwater
land degradation (Schofield et al., 1989; Anderson
recharge and the area of salinised land such as a
et al., 1993). Thus in areas where the water table is
plantation rotation (Silberstein et al., 1999), the
rather shallow (less than 2m), the most significant
establishment of deep-rooted trees (Silberstein et
problem facing irrigators has become not how
al. 2002, Pavelic et al., 1997) or use of shallow
much water is available or used, but the long term
groundwater pumps (Prathapar et al., 1996).
impact this has on the agricultural productivity of
However, the cost of revegetation for reclaiming
the area, and the environmental impact in general.
salinised land is high.
Recent estimates indicate that one-half of the
existing irrigation areas around the world have
While there are several studies that look at
shallow water tables, and require careful irrigation
solutions to salinity and water-logging problems
management practices to prevent water-logging
associated with shallow water tables, no studies
and secondary salinisation (Pratharpar et al.,
addressing the impact of timing between irrigation
1996).
events and length of inundation on water table
levels have been found. The identification of
The predominant cause of waterlogging and
appropriate irrigation schedules to limit, and even
salinisation is an increase in recharge to the water
reverse, the rise in water table level can have a
table, which occurs when excess water infiltrates
significant impact on land and water management
past the root zone of the plant. The high rate of
of shallow water table areas. While ceasing or
water table recharge is exacerbated when irrigation
significantly limiting irrigation will no doubt be
application rates exceed the consumptive use of
the best environmental solution, this would in most
plants. Smith (1998) suggests that the irrigation
instances render the land agriculturally unviable.
application efficiency for Australia is likely to be
As such, any recommendation on irrigation
only 60 percent, with flood irrigation as low as 40
scheduling must also address the likely impact on
percent.
root zone soil moisture content and its effect on
crop yield.
High water table levels have the same effect as
water-logging, with the added problem of salinity
In this research, a soil moisture model has been
when the ground water is saline. Water-logging
used to simulate the depth to water table and
and salinity are a potentially serious problems for
average root zone soil moisture content in response
the agricultural industry, because of the significant
to different irrigation schedules for a typical flood
negative impact on crop yield and long-term
irrigation district in south-eastern Australia. The
impact on agricultural productivity; they can
changes in water table depth and associated root
reduce the potential yield by as much as 30-80
zone soil moisture content have been analysed for
percent for many crops and pastures in the greater
a range of irrigation scenarios, including no
than 400 mm rainfall zone (McFarlane and
irrigation, and an appropriate schedule
Williamson, 2002).
recommended.
Several researchers have studied solutions related
2. MODELS
to waterlogging and secondary salinisation
problems in shallow water table areas. For
In order to study the response of root zone soil
example, a Soil, WAter and Groundwater
moisture content and water table depth to the
SIMulation model (SWAGSIM) has been
interval between and duration of flood irrigation
developed by Prathapar et al. (1996) to facilitate
events, the movement of soil moisture through the
evaluation of shallow water table management
unsaturated zone and recharge to a local water
options in south-eastern Australia; Pavelic et al.
table must be simulated. This is because the
(1997) have used an integrated modeling approach
logistical constraints and potential environmental
to explore a range of land management options to
impacts associated with field-based studies, and
control salinity in a 105-km2 site on a coastal plain
time requirements to undertake such studies,
in southern Australia; Silberstein et al. (1999) have
would be untenable.
studied the growth and hydrologic impact of a
small 21-year old plantation growing over a
1287

---

The simplest approach to water balance modeling
one dimensional soil moisture model. This
is to use a lumped soil moisture model of the
assumption is appropriate because the topography
unsaturated zone. Current practice of simulating
of flood irrigation areas is typically quite flat and
the unsaturated zone is mostly based upon this
flood irrigation is applied uniformly across
type of approach (Sarma and Mani, 1992; Rogério
reasonably large expanses of land. The second
and Chandra, 1996). In principle, this method
assumption is that there is no water flow across the
involves a simple book keeping of various mass
bottom boundary of the soil column. As we
balance components of the unsaturated zone
modeled to a depth of 4 m, and the initial water
(infiltration, evapotranspiration, change in water
table depth was taken to be at 1 m depth (typical of
storage and recharge to the water table). Usually
many irrigation districts in south-eastern
only the water movement in the root zone is
Australia), this assumption was not considered to
modeled, with the rest of the unsaturated zone
have any major impact on our results. The final
assumed to be at field capacity. The infiltration is
assumption was uniform soil properties throughout
determined after deducting the runoff from rainfall
the soil column. The major impact of this
and applied irrigation. The excess infiltration
assumption on the results would be through
(above field capacity and after accounting for
changes in hydraulic conductivity and soil porosity
actual evapotranspiration) is assumed to be
with depth. However, the deep soils in south-
available as recharge.
eastern Australia and the shallow water table depth
means that the active zone is limited to the top 1 m
An alternate approach to simulation of soil
or so, a zone of slowly varying soil type. Hence
moisture in the unsaturated zone and recharge to
this assumption is not likely to have a significant
the water table is based on the theory of Philip and
impact on the results provided the dominant soil
de Vries (1957). With the recent advances in
type has been correctly identified. Moreover, given
computing power this has become one of the most
that this study is largely synthetic in nature, correct
widely used modeling tools to date (Milly, 1982;
specification of soil type is not a major
Silberstein et al, 1999; Dam and Feddes, 2000).
consideration.
Under the assumptions of negligible vapour flux
and isothermal conditions, the unsaturated flow
3. NUMERICAL EXPERIMENTS
equations of Philip and de Vries (1957) simplify to
the well-known Richards’ equation:
A set of numerical experiments have been
undertaken to explore the effect of irrigation
??
? K
?
frequency and duration on both the depth to water

= ?? ? ?? + K? ?



(1)
?t
table and soil moisture content in the root zone.
? C?
?
While this study is not intended to represent a
specific study site, we have used typical soil
where ? is volumetric soil moisture content, C? is
properties, water table depth, meteorological data
soil capillary capacity factor ( C
and irrigation application for a typical irrigation
? = ?
? / ?
? ), ? is
district of northern Victoria (south-eastern
the matric potential and K? is the isothermal
Australia), in order to represent some realism.
moisture conductivity.
The topography in the region is very gently
This study used an implicit finite difference
sloping to level, with approximately 60 percent of
approximation to the one-dimensional ?-based
the agricultural land in the district irrigated
Richards’ equation to simulate soil moisture
(mainly pasture for dairying) and the remainder
content and depth to the water table at the point of
used for dryland grazing. The irrigation season is
application. The input data of the model include i)
from 15 August to 15 May, and the typical
soil properties such as volumetric porosity,
irrigation method is flood irrigation with

saturated hydraulic conductivity, saturated matric
4 hours of inundation every 10 days. The water
potential, residual volumetric soil moisture
table depth in the region has been rising at a rate of
fraction, Brooks and Corey parameters for the
around 0.2m/year over the past 20 years and is
soil-water characteristic curve and unsaturated
currently between 0 and 3m below the soil surface,
hydraulic conductivity relationship, ii) with a typical value of around 1m (Department of
meteorology data such as precipitation, potential
Natural Resources and Environment, 2002).
(or actual) evapo-transpiration, and iii) timing and
duration of irrigation events.
The dominant crops in this region are perennial
pasture (lucerne) and annual pasture (white clover
Three key simplifying assumptions have been
and ryegrass) which have an active root depth of
made in the application of this model. First, we
approximately 0.5m, while the main soil types are
have assumed there are no lateral flows by using a
loam and sandy loam. Because of a lack in specific
1288

---

data on soil hydrologic properties and the generic
4. RESULTS AND DISCUSSION
nature of this study, we adopted the typical Brooks
and Corey parameters for a loam soil from Rawls
The simulation model has been used to predict the
et al (1982); volumetric porosity 0.463, saturated
sensitivity of water table depth and root zone (top
hydraulic conductivity 13.2 mm/h, saturated matric
0.5 m) soil moisture content for the case of no
potential 40.12 cm and residual volumetric soil
irrigation and the 15 irrigation scenarios described
moisture 0.027.
in Table 1. Comparison is made with results from
the typical (4 hours every 10 days) irrigation
Mean annual rainfall and class A pan evaporation
schedule as the control. Average monthly water
for the region is 460 mm and 1600 mm
table depth and root zone soil moisture content
respectively. The maximum mean monthly rainfall
have been used in the analysis as we are more
occurs in May (48 mm), and the minimum occurs
interested in the longer term impacts than short
in February (25 mm), while the maximum mean
term fluctuations due to rainfall and irrigation
monthly pan evapotranspiration occurs in January
events. However, as crop response to soil moisture
(264 mm) and the minimum occurs in June (35
content may be more sensitive to short periods of
mm). Mean summer maximum and minimum
low soil moisture content, we include an
temperatures are 32oC and 15oC respectively, and
assessment of minimum monthly soil moisture
mean winter maximum and minimum temperatures
content in our analysis.
are 14oC and 3.2oC respectively (Bureau of
Meteorology, 1988). Meteorological data used in
Table 1 summarises the water table depth results
our simulations were obtained from a local climate
for the second year of simulation for each scenario
station for the year 1996. The simulation model
tested while Table 2 summarises the minimum
used Penman-Monteith potential evapo-
monthly root zone soil moisture results in order to
transpiration, factored by a soil moisture stress
check the possibilities of crop water stress.

index, in order to estimate actual
Figures 1 and 2 show the average monthly water
evapotranspiration.
table depth and root zone soil moisture content
respectively for some of the key scenarios tested.
Based on the current water table situation of this
area, the initial water table depth was assumed
The purpose of irrigation is to supply water to
equal to 1 m. The initial soil moisture content of
eliminate crop water stress in the root zone and
the top layer was set to 30 % v/v and increased
maximise crop yield. The crop water stress point is
linearly with depth to a saturated soil moisture
the limit of soil moisture content at which the
content at the water table. Using the data described
water in the soil ceases to become readily available
above and an irrigation interval of 10 days with a 4
to the roots for photosynthesis. If the amount of
hour inundation period, the simulated seasonal
soil moisture content in the root zone falls below
variation in water table depth was found to have
this point then the crop growth will be reduced due
good agreement with the observed water table
to lack of freely available water.
fluctuation at a bore in the region.
In the case of no irrigation, the water table depth
In this study, soil moisture profiles and water table
draws down significantly during the first few
depth have been simulated for 2 years under a
months and becomes stable at a depth of around
range of irrigation scheduling options, with the
3 m during the second year. However, the soil
same meteorological data used for both years of
moisture content in the root zone is below (about
simulation. This allowed for spin-up effects
10%) the crop water stress limit of about 30%
associated with the assumed initial conditions to be
(Wood et al., 2002). This reinforces the point that
mitigated. The simulations assumed that
this area requires regular irrigation for agricultural
availability of irrigation water was unlimited.
production.
Several irrigation scenarios were established, with
irrigation frequency ranging from 5 to 20 days and
The results in Table 1 indicate that the depth to
inundation ranging from 1 to 6 hours including the
water table varies considerably for different
typical irrigation of frequency and inundation of
irrigation intervals. It can be seen that the shorter
10 days and 4 hours respectively. A simulation
the irrigation interval the greater the rise in water
scenario with no irrigation was also made. No
table level. However, even with a short irrigation
irrigation water was applied from 15 May until 15
duration (1 hour) the water table still continues to
August in any of the simulations.
rise above that for the control for intervals of 5 and
7 days. This indicates that these intervals are not
appropriate in term of control of a rise in the water
table.
1289

---

Table 1: Monthly average water table depth (cm) of different irrigation scenarios
Irrigation
Jan. Feb. Mar. April
May June
July Aug.
Sep. Oct. Nov. Dec. Average
Scenarios
2h/5days
46.4
43.5 41.2 33.4 40.6 48.5 28.9
30.7 33.0 39.9 45.3 45.3
39.7
1h/5days 45.4 43.5 41.5 33.6 37.2 43.5 28.9
32.1 33.3 40.2 45.9 46.0
39.3
2h/7days
52.2
47.5 42.9 34.7 42.2 47.0 28.9
32.9 37.4 44.3 47.2 49.8
42.2
1h/7days
59.3
50.0 42.8 34.8 49.7 61.6 28.9
33.1 37.3 44.8 48.3 52.8
45.3
6h/10days
63.3
58.5 48.3 39.4 41.1 48.5 28.9
32.8 39.3 46.7 55.5 54.9
46.4
4h/10days*
64.1
59.3 49.7 39.8 41.3 48.5 28.9
33.0 39.5 47.1 55.7 55.9
46.9
2h/10days
61.6
60.1 50.7 40.3 41.4 48.5 28.9
33.2 40.6 47.9 56.1 56.8
47.2
1h/10days
76.6
66.9 50.6 40.5 41.4 48.5 28.9
33.3 40.8 52.6 57.9 61.8
50.0
6h/14days
70.4
65.8 50.4 41.8 54.8 61.0 28.9
33.0 46.9 51.3 66.3 63.8
52.9
4h/14days
70.5
66.5 50.9 42.3 55.8 61.7 28.9
33.1 47.0 51.8 66.1 64.4
53.3
2h/14days
71.9
68.8 51.3 42.7 56.5 61.7 28.9
33.2 47.4 53.9 68.2 67.4
54.3
1h/14days
114.9 110.9 97.8 69.7 56.7 61.6 28.9
33.3 47.0 56.0 78.6 91.4
70.6
6h/20days
84.4
82.6 67.0 57.0 50.9 64.0 28.9
33.9 40.1 56.7 73.1 80.9
60.0
4h/20days
85.1
82.4 69.6 57.3 51.1 64.0 28.9
33.9 40.3 57.3 73.5 82.0
60.4
1h/20days 111.7
121.2
124.3 99.1 72.4 82.5 39.7 34.0
41.5 59.8 85.5 111.5 81.9
No irrigation 310.0 305.5 321.5 328.8 329.4 330.0 327.2 339.9 353.2 344.6 349.7 343.0
331.9
Note: * indicates typical irrigation schedule; bold numbers indicate the maximum and minimum values.

Table 2: Monthly minimum soil moisture content (% v/v) in the rootzone (top 0.5m) of
different irrigation scenarios
Irrigation
Jan. Feb. Mar. April
May June July Aug. Sep. Oct. Nov. Dec.
Scenarios
2h/5days 40.00 39.70 40.30 41.70 40.10 39.80 43.10 42.20 41.90 40.50 40.00 36.10
1h/5days 39.17 39.62 40.28 41.69 40.92 40.50 43.13 42.03 41.84 40.44 39.97 39.56
2h/7days 38.70 38.20 40.00 41.40 40.50 40.10 43.10 42.00 41.60 39.20 38.40 38.40
1h/7days 36.92 38.19 39.99 41.34 39.03 38.67 43.13 42.03 41.57 39.15 38.33 38.16
6h/10days 36.00 36.30 38.90 39.60 40.00 39.80 43.10 42.00 40.30 37.80 37.50 37.10
4h/10days* 35.90 36.20 38.70 39.50 40.00 39.80 43.10 42.00 40.20 37.60 37.50 37.00
2h/10days 35.80 36.20 38.60 39.50 39.90 39.80 43.10 42.00 40.20 37.60 37.40 36.90
1h/10days 33.80 34.70 38.60 39.40 39.90 39.80 43.10 42.00 40.10 37.20 36.20 36.90
6h/14days 33.80 34.20 37.20 39.70 39.10 38.80 43.10 42.00 39.60 35.50 36.10 35.00
4h/14days 34.80 34.10 37.20 39.60 39.10 38.70 43.10 42.00 39.60 35.40 36.10 34.90
2h/14days 33.20 34.00 37.20 39.60 39.00 38.70 43.10 42.00 39.60 35.40 35.50 34.80
1h/14days
28.42 28.29 31.45 35.49 39.03 38.67 43.13 42.03 39.58 35.32 34.57 31.79
6h/20days
28.79
32.92 33.24 37.35 38.64 38.56 43.13 42.03 39.95 34.22 32.69 32.22
4h/20days
29.55
32.84 33.14 37.27 38.64 38.56 43.13 42.03 39.95 34.13 32.61 32.13
1h/20days
27.11 25.95 25.22 32.93 34.43 36.79 40.53 42.03 39.91 33.74 32.23 27.26
No irrigation
9.95 10.08 9.97 9.87 10.01 9.93 9.80 9.91 9.88 9.66 9.82 9.68
Note: * indicates typical irrigation schedule; bold numbers indicate the soil moisture content below the crop
water stress limit.
1290

---

Figure 1: Monthly average water table depth for some typical irrigation scenarios.



Figure 2: Monthly average volumetric soil moisture content in the root zone (top 0.5m) for some typical
irrigation scenarios. The horizontal line shows the soil moisture limit in order to mitigate crop stress.
Although the average monthly soil moisture
For the purpose of crop water supply to eliminate
content in the root zone of every irrigation scenario
the crop water stress in the root zone, and control
is greater than the crop water stress level in Figure
or limit rise in the water table, the results show that
2, Table 2 shows the minimum soil moisture for
the most appropriate irrigation schedule in the
the month does not satisfy the crop water
region is one irrigation every 14 days for 2 hours.
requirements for some irrigation schedules. These
include the scenario of 14 day interval and 1 hour
5. CONLUSIONS
duration and all scenarios of 20 days interval.
Although this latter interval can lower the water
A Richards’ equation based soil moisture model
table level considerably (see Table 1), this interval
has been used to study the effect of frequency and
is too long even with the case of 6 hours irrigation
duration of flood irrigation on the water table and
duration, because there is a period in which the soil
root zone soil moisture content. While the study
moisture content in the root zone falls below the
was not intended to represent a specific study site,
limit of crop water stress. For all the other
the results should be applicable to typical flood
irrigation intervals with irrigation durations from 2
irrigation regions of south eastern Australia and
to 6 hours, the soil moisture contents in the root
provide a qualitative indication of relative effects
zone satisfy the plant requirements.
of different irrigation scenarios on water table
depth and root zone soil moisture.
1291

---

The study shows that the time interval between
Agricultural Water Management, 35(1-2), 75-
flood irrigation events has a more significant
93.
impact on the depth to water table than the
Philip, J.R and D.A. de Vries (1957), Moisture
duration of inundation. In order to control or limit
movement in porous materials under
future water table rise, the interval between
temperature gradients, Eos Trans. AGU, 38(2),
irrigation events should be sufficiently far apart
222-232.
and the inundation duration should be decreased;
14 days and 2hrs respectively in this application.
Prathapar, S. A., W. S. Meyer, M. A. Bailey and
D.C. Poulton (1996), A soil water and
The key to avoiding water table rise is improved
groundwater SIMulation model: SWAGSIM,
Environmental Software
efficiency in both irrigation techniques and
, 11(3), 151-158.
scheduling to meet as precisely as possible the
Rawls, W.J, D.L. Brakensiek and K.E. Saxton
needs of the crop. This can be improved by
(1982), Estimation of soil water properties.
frequently monitoring soil moisture content, and/or
Transactions of the ASAE 25, 1316-1328.
applying numerical models to predict soil moisture
Rogério, T.F. and A.M. Chandra (1996),
content with observed meteorology data, soil, crop
Simulation of soil moisture profiles for wheat
and irrigation information.
in Brazil, Agricultural Water Management,
31(1), 35-39.
6. ACKNOWLEDGMENTS
Sarma, P.B.S. and A. Mani (1992), A simple soil
The authors wish to thank Cressida Savige for her
Water balance model, Journal of Hydrology of
Indian Hydrology Association,
insight on typical irrigation practices and
No. 3-4, 13-20.
environmental conditions in south-eastern
Schofield, N.J., I.C. Loh, P.R. Scott, J.R. Bartle,
Australia.
P. Ritson, R.W. Bell, H. Borg, B. Anson and,
R. Moore, (1989), Vegetation strategies to
7. REFERENCES
reduce stream salinity of water resource
catchments in south-west Western Australia,
Anderson, J., R. Britten, and J. Francis (1993),
Water Authority of Western Australia, Report
Dryland Salinity. 1: The Causes, Dept.
No. WS33, 81 pp.
Conservation and Land Management, NSW,
Silberstein, R. P., R. A. Vertessy, J. Morris and P.
and Dept. of Water Resour., NSW.
M. Feikema (1999), Modelling the effects of
Bureau of Meteorology (1988), Climatic Atlas of
soil moisture and solute conditions on long-
Australia, Bureau of Meteorology, Australia.
term tree growth and water use: a case study
from the Shepparton irrigation area, Australia,
Dam, J.C. and R.A. Feddes (2000), Numerical
Agricultural Water Management, 39(2-3),
simulation of infiltration, evaporation and
283-315.
shallow groundwater levels with the Richards
equation, Journal of Hydrology, 233(1-4), 72-
Silberstein, R. P., G. A. Bartle, R.B. Salama, , T.
85.
J. Hatton, P. Reggiani, G. Hodgson, , D. R.
Williamson and, P. Lambert (2002),
Department of Natural Resources and
Mechanisms and control of water logging and
Environment, (2002), Nanneella and District
Local Area Plan. Victoria, Australia
groundwater flow in the ‘Ucarro’ sub-
.
catchment, Agricultural Water Management
McFarlane, D.J. and Williamson D.R. (2002), An
53(1-3), 227-257.
overview of water logging and salinity in
Smith, D.I. (1998), Water in Australia – Resources
southwestern Australia as related to the
and Management, Oxford University Press,
`Ucarro' experimental catchment, Agricultural
Water Management,
South Melbourne, Australia.
53(1-3), 5-29.
Wood, M., H. Malano and H. Tural, (2002), Real-
Milly, P.C.D. (1982), Moisture and heat transport
time monitoring and control of on-farm
in hysteretic, inhomogeneous porous media: A
surface irrigation systems - Final report,
matric head based formulation and numerical
Department of Civil and Environmental
model, Water Resources Research, 18(3),
Engineering, The University of Melbourne,
489-498.
Melbourne, Victoria.
Pavelic, P., P.J. Dillon, K.A. Narayan, T.N.
Herrmann and S.R. Barnett (1997), Integrated
groundwater flow and agronomic modelling
for management of dryland salinity of a
coastal plain in southern Australia,
1292

---