Modification of the FAO-56 Spreadsheet Program for Scheduling Supplemental Irrigation of Winter Crops in a Mediterranean Climate

Abstract

Population growth and urbanization are increasing demands on limited renewable water resources in the Mediterranean region. Irrigation is a major water user, and so there has been increased effort to improve its efficiency. Using supplemental irrigation to increase and stabilize the yield of rain-fed crops is potentially an efficient use of water, but scheduling irrigation is difficult because of the unpredictability of the weather. The objectives were: (1) to develop a simple irrigation decision support tool based of the irrigation scheduling spreadsheet program presented in FAO Irrigation and Drainage Paper 56 (Allen et al., 1998), but with modifications that allow its use in supplemental irrigation; and (2) to evaluate the effect of uncertainties in the input parameters using a 27-year daily climate record for northern Syria. Modifications to the FAO model were incorporated that allow infiltrated rainfall to be stored within the potential root zone so that it can be accessed by the crop later in the season when the root depth has increased. The modified model was tested using a 4-year data set on supplemental irrigation of wheat at Tel Hadya in which a neutron probe was used to measure soil water content in 15 cm increments within the soil profile. The modified model predicted the depth of water within a 1.2-m root zone with a mean absolute error of 23 mm compared to the measured values. Applying the irrigation schedule developed by the model for each year of the climate record and a specified set of conditions to a range of conditions typical for the local area reduced the ratio of actual crop ET to non-stressed crop ET by a maximum of 0.03 at most, from 0.93 to 0.90. This model has potential for use as an irrigation decision support tool at the farm level and also at the level of strategic planning on irrigation water use. © 2008 American Society of Agricultural and Biological Engineers.

Full text

Applied Engineering in Agriculture
Vol. 24(2): 203‐214 E 2008 American Society of Agricultural and Biological Engineers ISSN 0883-8542 203
MODIFICATION OF THE FAO‐56 SPREADSHEET PROGRAM
FOR SCHEDULING SUPPLEMENTAL IRRIGATION OF
WINTER CROPS IN A MEDITERRANEAN CLIMATE
I. R. McCann, A. Bruggeman, T. Y. Oweis, M. Pala
ABSTRACT. Population growth and urbanization are increasing demands on limited renewable water resources in the
Mediterranean region. Irrigation is a major water user, and so there has been increased effort to improve its efficiency. Using
supplemental irrigation to increase and stabilize the yield of rain‐fed crops is potentially an efficient use of water, but
scheduling irrigation is difficult because of the unpredictability of the weather. The objectives were: (1) to develop a simple
irrigation decision support tool based on the irrigation scheduling spreadsheet program presented in FAO Irrigation and
Drainage Paper 56 (Allen et al., 1998), but with modifications that allow its use in supplemental irrigation; and (2) to evaluate
the effect of uncertainties in the input parameters using a 27‐year daily climate record for northern Syria. Modifications to
the FAO model were incorporated that allow infiltrated rainfall to be stored within the potential root zone so that it can be
accessed by the crop later in the season when the root depth has increased. The modified model was tested using a 4‐year
data set on supplemental irrigation of wheat at Tel Hadya in which a neutron probe was used to measure soil water content
in 15 cm increments within the soil profile. The modified model predicted the depth of water within a 1.2‐m root zone with
a mean absolute error of 23 mm compared to the measured values. Applying the irrigation schedule developed by the model
for each year of the climate record and a specified set of conditions to a range of conditions typical for the local area reduced
the ratio of actual crop ET to non‐stressed crop ET by a maximum of 0.03 at most, from 0.93 to 0.90. This model has potential
for use as an irrigation decision support tool at the farm level and also at the level of strategic planning on irrigation water
use.
Keywords. Rain, Evapotranspiration, Weather, Soil water, Irrigation scheduling.
he Mediterranean region, which includes countries
from southern Europe, North Africa, and the
Middle East, experiences a climate characterized
by cool temperatures, low reference
evapotranspiration (ET0), and rainfall during the winter; and
high temperatures, high ET0, and little or no rainfall during
the summer. Annual rainfall is significantly lower than
annual ET0. Cropping systems encompass the entire range
from rain‐fed to fully irrigated. Irrigation enables more
profitable crop production, but uses large amounts of water.
Increasing population, urbanization, and industrialization
within the region is putting pressure on the often limited
water resources. Because irrigation consumes a large portion
of the regions total water use (Allen et al, 1998) and because
water will be increasingly taken away from agriculture for
use in other sectors, any reduction in irrigation water use
Submitted for review in May 2007 as manuscript number SW 7030;
approved for publication by the Soil & Water Division of ASABE in
November 2007.
The authors are Ian R. McCann, ASABE Member Engineer,
Assistant Professor, Department of Bioresources Engineering, Research
and Education Center, University of Delaware, Georgetown, Delaware;
Adriana Bruggeman, ASABE Member Engineer, Agricultural
Hydrologist, Theib Y. Oweis, ASABE Member Engineer, Director
Integrated Water and Land Management Program, and Mustafa Pala,
Agronomist (retired), International Centre for Agricultural Research in Dry
Areas (ICARDA), Aleppo, Syria. Corresponding author: Ian R. McCann,
Research and Education Center, University of Delaware, 16483 County
Seat Highway, Georgetown, DE 19947; phone: 302‐856‐7303; fax:
302‐856‐1845; e‐mail: mccann@udel.edu.
through improvements in efficiency will help to maintain
production levels with less water.
An example of a Mediterranean climate, which shows
cumulative rainfall and crop evapotranspiration (ETc) for
wheat for Tel Hadya, Syria is presented in figure 1. The
cumulative daily crop evapotranspiration deficit remains
negligible from November (the beginning of the growing
season) until March, but after March ETc increases rapidly
and rainfall decreases, resulting in a rapid increase in the
deficit. One of the most efficient uses of water is to
supplement winter rainfall in areas such as this where
rain‐fed production is possible but where yields are limited
by water stress in the latter part of the growing season. For
example, the increase in yield made possible by
supplemental irrigation (SI) of wheat in Syria can result in
irrigation water productivity (increase in wheat yield per unit
volume of irrigation water) of more than 2 kg/m‐3 (Oweis and
Hachum, 2006). Deficit SI can result in even higher irrigation
water productivity, although at the expense of lower yield
(Zhang and Oweis, 1999).
Although well‐managed irrigation has national and
regional benefits from a water productivity viewpoint,
farmers worldwide tend to apply excess water in order to
eliminate the risk of yield losses from applying too little. This
is particularly so where the incremental cost of applying
irrigation water is relatively low compared with the
economic return realized from increased yield. Surface water
supply systems often provide water on a fixed schedule at low
cost, and groundwater is often available for the cost of
pumping it without regard to its sustainable use. Other factors
T

204 APPLIED ENGINEERING IN AGRICULTURE
may also impact irrigation decisions, such as equipment and
labor availability.
In Syria, wheat production currently accounts for more
than half of the irrigated land and more than 30% of the
non‐irrigated land (MAAR, 2006). The fraction of the total
wheat production area that is irrigated has increased
substantially, from 14% in 1981 to 45% in 2005. However,
during the same period the total land irrigated by
groundwater more than tripled, from 250,000 to 860,000 ha
(MAAR, 2006), and this has led to excessive drawdown in
some aquifers. The MAAR has adopted a plan for reducing
water use by irrigated agriculture that provides economic
incentives and technical assistance to farmers to help them
convert from surface irrigation to potentially more efficient
sprinkler and drip irrigation (NAPC, 2003). Schweers et al.
(2004) found that farmers in the Khanasser Valley in northern
Syria who applied supplemental irrigation to wheat by
surface irrigation applied more than double the required
water needs. However, regardless of the type of irrigation, it
is often the on‐farm management and irrigation decisions that
determine whether water is used efficiently.
In the case of supplemental irrigation of winter crops in
Mediterranean countries, the soil profile, particularly near
the surface, is very dry in the fall. When the crop is planted,
typically in November, the first rains have wetted the surface
to allow germination. A relatively light irrigation at planting
is an option farmers consider if necessary to ensure adequate
crop establishment. There are usually no more irrigations
until the spring. Rainfall during the winter is generally
sufficient to meet crop water requirements and provide some
recharge to the soil profile for later use by the crop. However,
the depth of wetting may not be sufficient to allow root
development at the same rate as under full irrigation. During
dry years or when rain is poorly distributed, stress occurs
during winter and early spring. There are also practical
constraints that might not apply to a fully irrigated production
system, such as that farmers will not usually apply frequent
light irrigations because of the labor required to do so.
Instead, once the weather starts warming up and rainfall
events become less frequent, farmers will apply a few large
irrigations sufficient to refill the root zone.
Figure 1. Cumulative daily precipitation, wheat evapotranspiration and
evapotranspiration deficit during the growing season at Tel Hadya, Syria.
The bars show the value at the end of each month averaged over the 27
seasons from 1979/80 to 2005/06 The error bars show the corresponding
minimum and maximum values.
Making regular direct in‐field measurements of soil water
content (SWC) to schedule irrigation is usually too laborious,
time consuming, difficult, or expensive for individual
farmers. A more feasible approach is to estimate SWC based
on parameters that can be more easily measured. A large
effort has been made to model the various processes that
determine SWC and movement. Models of SWC range from
the simple to the complex. Bastiaanssen et al. (2007) provide
a comprehensive review and reference source for many of the
models developed for irrigation and drainage during the last
quarter century. The simplest models are based on the water
balance approach, in which irrigation, rainfall, and
sometimes upward movement from below the root zone all
increase SWC while evaporation, transpiration, and drainage
decrease it. More complex models take account of water
fluxes within the soil profile, but they often require more
parameters and knowledge on the part of the user. The
simpler a model is in terms of the required data and
parameters, the easier it is to use. However, if the model is too
simplistic in terms of physical processes and/or requires site
specific empirical coefficients to perform satisfactorily, it
may be of limited value as a practical tool.
The internet has enabled easy access to up‐to‐date weather
information and evapotranspiration estimates from
automated climate stations, making the use of
evapotranspiration‐driven water balance models for
irrigation scheduling more generally applicable. Examples
of online networks include CIMIS (California Irrigation
Management Information System, www.cimis.water .ca.
gov); AGRIMET (www.usbr.gov/pn/agrimet/) in the
northwest states; and AEMN (Automated Environmental
Monitoring Network, www.griffin.uga.edu/aemn) in
Georgia, all from the United States, and www.agric.wa.
gov.au in Western Australia. Climate station networks may
also include irrigation scheduling programs that can directly
use the climate data, such as WISE (Washington Irrigation
Scheduling Expert, www.sis.prosser.wsu.edu/wise.htm);
AZSCHED (Arizona Irrigation Scheduling System, www.ag.
arizona.edu/crops/irrigation/azsched/azsched.htm) and
Wateright (www.wateright.org). Another approach is to use
a spreadsheet, such as Kansched and KISCORN from Kansas
(www.oznet.ksu.edu/mil/ToolKit.htm). But it should be
noted that these programs are generally set up to schedule full
irrigation. However, in Syria, as in most other countries in the
Mediterranean region, there is currently no system and little
services to assist farmers with their day to day irrigation
decisions. There is a clear need for simple and robust methods
that can be used by agricultural research and extension
personnel to provide farmers with real‐time irrigation
scheduling advice and to help make sound water allocation
decisions.
FAO Irrigation and Drainage Paper 56 (Allen et al., 1998)
is a standard reference for crop evapotranspiration. It
provides a comprehensive description of the widely accepted
Penman‐Monteith method for estimating ET0 from data on
air temperature, humidity, wind speed, and solar radiation;
and procedures for computing ETc under standard and
non‐standard (stressed) conditions. This publication also
includes a spreadsheet program (available from the
University of Idaho web site, www.kimberly.uidaho.edu/
water/fao56/index.html) for irrigation scheduling under
standard conditions. In this spreadsheet program the root
zone is treated as a single layer from which water is depleted

205Vol. 24(2): 203‐214
by the crop. The root zone increases linearly to its maximum
value as a function of time, while water depletion from the
root zone determines how much irrigation is needed to refill
it to field capacity. The spreadsheet was developed for full
irrigation and cannot be directly used for supplemental
irrigation because of some of the explicit and implicit
assumptions. One of these assumptions is that the complete
profile is at field capacity, except for a user‐specified deficit
in the surface layer due to evaporation. Thus, the moisture
content in the root zone increases when it expands downward
due to root growth, and any drainage from the current root
zone leaches directly from the profile and is therefore not
available for later use.
A multi‐layer model can better account for the movement
of water and the growth of roots within the profile, but it will
be more complex. But even some of the more advanced
irrigation scheduling and soil water balance models that have
been used for supplemental irrigation of wheat (e.g. Oweis et
al., 2003) do not include a dynamic root growth function. If
sufficient data are available, crop models (e.g. Pala et al.,
1996) could also be used to provide additional management
and production information. We sought a compromise
between the potentially higher accuracy but greater
complexity of multi‐layer and crop models, and the less
realistic but greater simplicity of a single layer model such as
the FAO‐56 spreadsheet‐based model. Our goal was to
provide agricultural research and extension personnel in the
Mediterranean region with a robust and user‐friendly tool to
help farmers make daily irrigation scheduling decisions and
to aid with the development of water management and
allocation strategies.
Our specific objectives were to build upon the FAO‐56
spreadsheet program to (i) modify and evaluate it for
scheduling of supplemental irrigation in a rain‐fed
Mediterranean environment; (ii) evaluate the effect of
uncertainties in the model's input parameters on its
usefulness as an irrigation and water management tool in the
region.
MATERIALS AND METHODS
SOIL WATER BALANCE MODEL
The FAO‐56 spreadsheet program estimates actual crop
water use (ETa ) from the reference evapotranspiration (ETo),
using a crop coefficient (Kc); a soil evaporation coefficient
(Ke) to account for evaporation directly from a moist soil
surface; and a stress coefficient (Ks) to account for reduced
crop water use when SWC is limiting. The model uses the
dual crop coefficient method that separates evaporation from
transpiration. This process can be expressed as:
ETa = (Ks Kcb + Ke) ETo (1)
where ETa is the actual crop evapotranspiration (mm); ETo is
the reference evapotranspiration (mm); Ks is a stress
coefficient with values from 0 (fully stressed) to 1 (not
stressed); Kcb is the basal crop coefficient; and Ke is the soil
evaporation coefficient. Ke is a function of the evaporation
reduction coefficient (Kr), the maximum and basal crop
coefficient, and the exposed and wetted soil fraction (Allen
et al., 1998).
Irrigation is automatically initiated when soil water in the
root zone falls below the readily available water level
(RAW). The readily available water is indicated by p, the
threshold at which stress starts to occur, expressed as a
fraction of the total available water (TAW), the water held
between field capacity (FC) and wilting point (WP). Hence
RAW = p TAW, and under well‐managed full irrigation SWC
should be maintained above the stress level and the value of
Ks should never decrease below its maximum value of 1.0.
However, when SWC does fall below p, Ks decreases linearly
towards its minimum value of 0.0 at WP. Similarly, the soil
evaporation reduction coefficient (Kr) is at its maximum
value of 1.0 until the readily evaporable water (REW) has
evaporated, and then decreases linearly to 0.0 as evaporation
approaches the soil's total evaporable water (TEW).
To model rain‐fed and supplemental irrigation production
systems, we modified the FAO‐56 spreadsheet program by
dividing the potential root zone into two dynamic layers
(fig. 2). Layer 1 extends from the surface down to the current
root depth, with the remainder of the potential root zone
being layer 2.
Roots are allowed to grow downwards only when the
SWC in either layer 1 or layer 2 is above a user‐specified
threshold value, defined as a percentage of available water
content. When water content is above the threshold, the daily
increase in root depth is a constant, subject to the defined
maximum root depth and the user‐specified period of root
growth.
The initial depth of layer 1 should be sufficiently large (at
least equal to or larger than the depth of the soil from which
water evaporates) to be able to store some water before it
drains into the usually much dryer layer 2. Because roots start
growing from the actual planting depth, layer 1 only starts
expanding after the roots have reached the bottom of the
initial value of layer 1. This process is expressed as follows
DRZi = RG for SWC1i‐1 ≥ (q × TAW1 i‐1)
or [SWC2 i‐1 ≥ (q × TAW2 i‐1) and
L1i‐1 < RZmax and TPL ≤ i ≤ TRG] (2a)
DRZi = 0 if all above conditions are not met (2b)
where DRZ i is the change in root depth on day i (mm); RG
is the vertical root growth rate (mm/d); SWC1i‐1 and
SWC2i‐1 are the water contents in layers 1 (root zone) and 2,
respectively, on day i‐1 (mm); q is the threshold for root
growth, expressed as a fraction; TAW1 i‐1 and TAW2 i‐1 are
the total available water contents in layers 1 and 2,
respectively (mm) on day i‐1; L1i‐1 is the depth of layer 1 on
day i (mm); RZmax is the maximum potential root zone depth
(mm); and TPL and TRG are the planting date and the date that
root growth ceases, respectively. Thus, the root zone is
computed as:
RZi = RZi‐1 + DRZi (3)
where RZi and RZi‐1 are the root zone depths on day i and day
i‐1 (mm). The depth of layer 1 is given by:
L1i = RZi for RZi ≥ L1PL (4a)
L1i = L1PL for RZi < L1PL (4b)
where L1i is the depth of layer 1 on day i (mm); and L1PL is
the initial depth of layer 1 at planting (mm).

206 APPLIED ENGINEERING IN AGRICULTURE
Figure 2. Division of the root zone into two layers under rain‐fed
conditions.
This process represents a simplified model of the growth
of roots following downward water movement through
cracks and macropores. In an earlier version of the model
(Bruggeman et al., 2005), the root depth was constrained to
the zone that had been fully recharged to FC. However, the
simulated root depths were often smaller than the root depths
ascertained from the pattern of soil water extraction shown
by the measurements of soil water with depth.
Rainfall and irrigation replenish layer 1, while
evaporation, transpiration, and drainage deplete it.
Following a rainfall or irrigation event large enough to cause
layer 1 to exceed field capacity, the excess will drain into
layer 2 and be stored there, thereby increasing layer 2 water
content. If the drainage from layer 1 into layer 2 causes layer
2 to exceed field capacity, the excess will drain out of layer
2 and the profile. The total water content in the profile is the
sum of the water content of layers 1 and 2. If SWC permits
sufficient root growth, layer 2 will eventually decrease to
zero with layer 1 occupying the entire profile. This process
is expressed by:
SWC1i = SWC1i‐1 + PPTi + IRRi ‐ ETi
‐ DRN1i + (DRZi × q2i‐1) (5)
DRN1i = SWC1i ‐ (L1i × FC)
for SWC1i > (L1i × FC) (6a)
DRN1i = 0 for SWC1i v (L1i × FC) (6b)
SWC2i = SWC2i‐1 + DRN1i
‐ (DRZi × q2i‐1) ‐ DRN2i (7)
DRN2i = SWC2i ‐ (L2i × FC)
for SWC2i > (L2i × FC) (8a)
DRN2i = 0 for SWC2i v (L2i × FC) (8b)
where PPTi, IRR i, and ETi are rainfall, irrigation, and
evapotranspiration on day i (mm) and SWC1i‐1 and SWC2i‐1
are the water contents of layer 1 and layer 2, respectively, on
the previous day (day i‐1) (mm); DRZi is the change in layer
1 on day i resulting from root growth (mm); q2 i‐1 is the water
content of layer 2 on day i‐1 (volumetric fraction); DRN1i
and DRN2i are drainage from layers 1 and 2, respectively, on
day i (mm), L1i and L2i are the depths of layer 1 and 2 on day
i (mm); and FC is the field capacity of the soil (volumetric
fraction).
MODEL EVA L U AT I O N USING FIELD TRIALS
To evaluate the modified spreadsheet program, a data set
from a multi‐year experiment on supplemental irrigation of
durum wheat (Triticum turgidum L.) was used (Oweis et al.,
1999). The experiment was conducted at ICARDA's main
research station at Tel Hadya, near Aleppo in northern Syria
(36.01° N, 36.93° E; 284 m above sea level) where the soil
is a deep red clay, classified as a Calcixerollic Xerochrept,
and representative of soils in the region (Ryan et al., 1997).
The data set includes the four consecutive cropping seasons
1992/1993, 1993/94, 1994/95 and 1995/96. In all years,
wheat was planted in early November and received no
irrigation until spring (March or April). The wheat was
adequately fertilized with 40 to 50 kg P/ha and 100 kg N/ha.
Daily climate data from an automatic station at this location
were used to compute ET0. The crop was irrigated at a rate
sufficient to replenish the root zone to field capacity based on
a p value of 50%.
The values we used for the variables required by the model
came from Allen et al. (1998) (see table 1). In all years, the
development stage began around 1 January and continued to
approximately 1 April, followed by the mid stage which
lasted until 15 May, and the end stage which lasted until the
end of the crop season on 25 May. The total length of the
growing season was 200 days.
Soil water content was measured during the season in
15‐cm increments down to at least 1.50 m using a neutron
probe, while SWC in the top 15‐cm was determined using
cored soil samples. Soil water content below 1.2 m did not
change much over the season, indicating that this would be
a reasonable estimate of maximum effective rooting depth or
depth of water infiltration. The field capacity (FC) of this soil
is in the range 36 to 40% while the wilting point (WP) is 20
to 24%. Analysis of the measured volumetric soil water data
indicated that a value of 38% for FC and 22% for WP would
best represent the soil at the field sites used during this study.
The value of TEW for the top soil was computed from the
FC and WP, as specified by Allen et al. (1998), considering
an effective depth of the soil evaporation layer of 20 cm.
Although the depth to which evaporation takes place in these
cracking clay soils is deeper, soil moisture observations
indicate that evaporation during the majority of the crop
season is restricted to the top 15 cm. The depth of the initial
root zone should be equal or larger than the soil evaporation
layer. If the depth of the initial root zone is too small the
amount of water it can hold will also be small and rapidly
depleted by evaporation. It will also drain after a relatively
small rainfall and this drainage into the correspondingly
larger and usually much dryer layer 2 will not have much
effect in raising its water content. If the depth of the initial
root zone is too large the average water contents will be too
low and the roots won't grow.
Barraclough and Leigh (1984) found root growth rates for
adequately irrigated winter wheat in the UK of 6 mm/d when
planted in October and 12 mm/d when planted in September
and related the differences to temperature. They quoted
similar wheat root growth rates of 5 to 6 mm/d in winter
(Gregory et al., 1978; Ellis and Barnes, 1980) and 18 mm/d
in spring (Gregory et al., 1978). Sato et al. (2006) monitored

207Vol. 24(2): 203‐214
Table 1. Model input parameter values.
Symbol Parameter Value Units
PD Planting date 5‐10 Nov
Kcb ini Crop coefficient in initial stage 0.15
Kcb mid Crop coefficient at mid stage 1.10
Kcb end Crop coefficient at end stage 0.40
L ini Duration of initial stage 57 d
L dev Duration of development stage 90 d
L mid Duration of mid stage 44 d
L late Duration of late stage 10 d
FC Field capacity 38 % by volume
WP Wilting point 22 % by volume
p Water depletion at which crop
stress starts
50 % (of TAW)
REW Readily evaporable water 8 mm
TEW Total evaporable water 54 mm
RZ ini Initial root zone 0.20 m
RZ max Maximum potential root zone 1.20 m
RG Root growth rate 11 mm/d
q Threshold for root growth 50 % (of TAW)
Ht max Maximum crop height 0.75 m
root growth of bread wheat planted on 23 December in Tel
Hadya and found that, under non‐water limiting conditions,
roots had reached the 45‐ to 60‐cm layer after 56 days,
indicating an approximate root growth of 10 mm/d. Also for
wheat at Tel‐Hadya, but in a much drier year with a 50‐mm
irrigation applied at planting, Izzi et al. (2007) measured an
average downward root growth of 7 mm/d during the first 74
days after emergence. Although roots are generally
considered to reach their maximum depth at anthesis, Brown
et al. (1987) and Gregory et al. (1992) found that the roots of
barley and wheat still increased in weight between anthesis
and maturity. Considering the above, we used a vertical root
growth rate of 11 mm/d, with roots growing from 5‐cm depth
at planting down to a maximum potential effective root depth
of 120 cm. We used a SWC value of 50% of available water
as the threshold above which root growth can proceed. This
threshold corresponds to the same SWC as specified by the
stress threshold. Root growth was assumed to cease towards
the end of April, halfway between anthesis and maturity.
The measured values of SWC were screened to identify
measurement errors. There were a total of 71 measurement
days during the 4‐year period. Average daily ET between
adjacent measurement dates was calculated as the residual of
the soil water balance, including irrigation, rainfall, and the
change in SWC measured to the depth of the neutron probe
access tube. The calculated daily ET on 5 of the 71 dates was
found to be an unrealistic 200% greater or less than ET0, as
estimated from climate data, and so these dates were
eliminated. (three in 1992/1993 and two in 1995/96).
To evaluate the model, we used the mean absolute
difference between measured and predicted values, and the
modified index of agreement (Legates and McCabe, 1999),
which uses absolute instead of squared differences to reduce
sensitivity to extreme values.
SENSITIVITY ANALYSIS
A sensitivity analysis was conducted to assess the effect
of uncertainties in the crop and soil parameters on the
seasonal crop evapotranspiration and irrigation needs, using
a 27‐year daily climate record from ICARDA's Tel Hadya
climate station. The parameter values for the base scenario
were similar to those used for the Tel Hadya field trials, which
represent fairly average conditions for the 300‐ to 400‐mm
rainfall zone in northern Syria. However, similar to farmer's
practices, if the soil moisture in the top layer (20‐cm) had not
exceeded the soil moisture stress level (RAW) within 1 week
of planting, an initial irrigation of 50 mm was given.
To estimate the soil moisture conditions at planting,
simulations were started on 1 September, which is usually
before the beginning of the rainy season. The initial soil
moisture content on this date, after the hot and dry summer
months, was assumed to be 11% in the initial root zone
(0‐20 cm) and 24.5% in the soil profile below. Initial moisture
conditions depend not only on the soil but also on the
preceding crop, management and irrigation practices, but the
values represent typical conditions for the clay soils in the
region, which generally exhibit dry, cracking conditions in
the top 30 to 45 cm of the profile, higher moisture contents
(26% to 32%) in the layers below 75 to 90 cm, and wilting
point in between. The use of an average moisture content for
such a variable profile is more or less compensated for by the
fact that the dryer layers in the upper part of the profile are
recharged first.
Selected parameters were changed, one at the time, to
their expected minimum and maximum values. The relative
sensitivities (RS) of the 27‐year averages were computed as:
RS = [(O ‐ O b )/O b ] / [(I ‐ I b )/I b ] (10)
where O and Ob are model output parameter values of the
current test run and the base run, respectively, and similarly
I and Ib are the input parameter values of the test run and the
base run. A fortran‐version of the spreadsheet program was
made to facilitate long‐term analysis and data processing.
IRRIGATION SCHEDULING
To evaluate the use of the model for providing real‐time
irrigation scheduling advice for a larger area surrounding a
climate station, the effect of applying a generic schedule to
soils with different physical properties was tested. When
applying the same irrigation advice to a variety of fields, it
is more appropriate to develop a schedule for a soil with low
water storage capacity and apply it to soils with higher
storage capacities so that the irrigations will not cause
drainage from the profile. Therefore, considering the soils in
the area, an irrigation schedule was simulated for each of the
27 years of climate data for a relatively shallow soil (90 cm)
with a relatively low water holding capacity (12%, FC = 36%
and WP = 24%), thus providing a total available water
capacity of 108 mm. The initial SWC below the initial root
zone and evaporation layer was set at 27%, halfway between
the wilting point and the crop stress level (30%). All other
conditions were kept similar to those of the previous
simulations. This is referred to as the test case.
The irrigation schedule developed for the test case was
applied to the soils at Tel Hadya, which have a greater water
holding capacity. Assuming a profile depth of 120 cm and a
water holding capacity of 16% (FC of 38% and WP of 22%),
the total available water capacity is 192 mm (78% larger than
the test case). To also assess the effect of dryer or wetter
initial SWC in the profile below the initial root zone, SWC
from 20 to 120 cm was set either at 22% (wilting point) or at

208 APPLIED ENGINEERING IN AGRICULTURE
30% (crop stress level). These are referred to as the dry and
wet cases, respectively. Irrigation schedules were simulated
for the dry and wet cases, and also for the dry and wet cases
but imposing the same schedule developed for the test case.
RESULTS AND DISCUSSION
MODEL EVA L U AT I O N USING FIELD TRIALS
A summary of ETa as estimated by the model, along with
precipitation and irrigation for the four cropping seasons, is
shown in table 2. The largest total irrigation amount was in
the 1992/1993 season, during which rainfall was low.
However, the individual irrigations were likely excessive
because the model predicted that three of the four irrigations
resulted in significant drainage from the profile (178 mm
over the season). No drainage was predicted in the other three
seasons. In all four seasons the sum of irrigation and rainfall
exceeded estimated ETa . In the 1993/1994, 1994/1995, and
1995/1996 seasons the excess was in the range 70 to 140 mm,
and this was stored within the profile at the end of the season.
In the 1992/1993 season the excess was about 300 mm, which
could not all be stored within the profile and so resulted in
drainage.
Measured and predicted values of SWC in the top 1.2 m
of soil for full supplemental irrigation are presented in
figure 3 for each year. The mean absolute error of 23 mm
appears acceptable considering potential sources of error
from other factors, such as in the soil water measurements
made by the neutron probe, and in the model inputs, such as
amount of rain or irrigation water applied. For example, a
systematic error in the volumetric moisture content reading
of 0.02 m3/m3 would result in an error of 24 mm over the
entire 1.2‐m profile at the site. Similarly, a stationary
sprinkler irrigation system with good performance and
operating under low wind conditions may have a coefficient
of uniformity (CU) of 85%. This implies that, for 100 mm of
irrigation, there will be an average absolute deviation of
15 mm in the amount of water applied anywhere in the field.
Measured and predicted values of SWC in the entire
profile during both the rain‐fed and irrigated periods in the
1993/1994 season are presented in figure 4. This relatively
wet season (360‐mm rain between 1 November and 31 May)
caused SWC to increase to close to FC by the end of February.
The stored soil water was then extracted by
evapotranspiration until the first of three irrigations
replenished the profile to close to FC again. By this time the
rate of soil water extraction was increasing as a result of
higher evapotranspiration rates. Soil water content before the
final irrigation had declined to below the desired p level of
50%, likely causing some stress. At the end of the season the
model predicted some residual water in the profile that the
crop may not have been able to use but which might be
available to a subsequent crop.
Table 2. Estimated ETa, precipitation (PPT), and irrigation (IRR)
from 1 November to 31 May during the four cropping seasons.
Cropping Season
1992/93 1993/94 1994/95 1995/96
ETa (mm) 520 610 513 503
PPT (mm) 276 360 294 395
IRR (mm) 548 323 354 180
Figure 3. Measured and predicted volumetric soil water content under full
supplemental irrigation during the 1992/93, 1993/94, 1994/95 and
1995/96 cropping seasons.
Soil water content in the current root zone of the crop is
important for irrigation scheduling. However, measuring the
root zone over the course of the season is difficult and so is
rarely done. Measurements of SWC with depth at intervals
over time can sometimes help in determining root depth. For
example, a decrease in SWC in the upper part of the profile
between two dates, but not at deeper depths between the same
dates, can indicate water extraction by roots from the upper
zone. However, this becomes more difficult to quantify when
rain or irrigation also adds water to the same upper part of the
profile.
An example from the 1994/1995 season is shown in
figure 5, in which the change in measured SWC is plotted (at
the midpoint of the depth increment) for a period during
which there was neither irrigation nor significant rainfall but
which was also long enough for significant changes in SWC
from crop uptake. The change in measured SWC between the
beginning of the period on 22 April 1995, and the two
subsequent measurement dates before the next irrigation,
29 April and 4 May is shown. A positive change at the same
depth between the dates means that SWC at the beginning of
the period was higher than on subsequent measurement dates.
A reduction in SWC (positive change) should primarily be
due to extraction by roots. Where the change is close to zero
there may be no root extraction. However, there may also be
water movement within the profile, but unless data are logged
continuously the two processes cannot be separated. The root
depths predicted by the model at the beginning (0.91 m) and
end of the period (1.04 m) are shown. These predictions are
consistent with the depth at which SWC does not change
significantly.
Because the actual root zone is not known, a direct
comparison between measured and predicted values is not
possible and so the predicted root zone must be used instead.
Examples of measured and predicted SWC within the
predicted root zone for 1992/1993 (a relatively dry season
with 276‐mm rainfall between 1 November and 30 May), and
1993/1994 (a wetter season with 360‐mm rainfall) are
presented in figure 6. Also shown are rainfall and irrigation,
as well as the water content at WP and FC as the predicted
root zone increases over the season. There is reasonable

209Vol. 24(2): 203‐214
Figure 4. Measured and predicted volumetric soil water content in the 1993/94 cropping season during the rain‐fed period and subsequently under full
supplemental irrigation.
agreement between the measured and predicted values, such
that the model would be useful in scheduling supplemental
irrigation. The lower rainfall amounts in the 1992/1993
season did restrict predicted root zone development prior to
the first irrigation.
Figure 5. Change in measured soil water content with depth between
22 April and 29 April, and between 22 April and 4 May 1995, during
which time there was no irrigation and negligible rainfall. The root depths
predicted by the model on these days are also shown.
SENSITIVITY ANALYSIS
The results of the sensitivity analysis are presented in
table 3. The relative sensitivity, calculated using equation 10,
is shown for two essential output values in this model, namely
ET and the number of irrigations scheduled by the model.
Because the input parameter values being tested were set at
their potential extremes, it is also useful to look at the
resulting differences in the average values of the water
balance components for the 27‐year period. As expected,
both ET and the number of irrigations are highly sensitive to
changes in the length of the crop development stage, which
in this case was also assumed to change the total length of the
crop season by the same number of days. The crop
evapotranspiration was also relatively sensitive to the value
of the crop coefficient during mid stage, but the average
difference between values for the 27‐year record remained
relatively small.
Except for TEW, all the other parameters had almost no
effect on ET. However, even for a 50% increase in TEW
(from 54 to 81 mm), with a corresponding increase in the
initial root zone to 30 cm, the maximum soil water deficit in
the evaporation layer never exceeded 71 mm (corresponding
to a SWC of 14%) during the crop season. In half of the years
(13 out of 27) deficits never exceeded 60 mm, and only in
2 years did the deficit exceed 60 mm before May. When TEW

210 APPLIED ENGINEERING IN AGRICULTURE
Figure 6. Predicted and measured soil water content in the predicted root
zone in (a) 1992 (a relatively dry year and (b) 1993 (a relatively wet year),
along with rainfall and irrigation, and the soil water contents at field
capacity and permanent wilting point.
was set equal to 41 mm, deficits in the 15‐cm evaporation
layer did not exceed 36 mm before May in 17 of the 27 years
(corresponding to a SWC of 14%). Allen et al. (2005)
proposed a two‐step linear process to simulate the
evaporation of cracking clay soils, but the low sensitivity of
this parameter indicates that this may not be necessary for
supplemental irrigation of winter crops in this environment.
A change in the net average total irrigation did not always
result in the same change in ETa. Some of the differences in
irrigation amounts were partially compensated for by other
water balance components, as shown in table 3. For the base
run, the SWC in the profile averaged 70 mm higher at harvest
than at the start of the simulation period on 1 September. This
occurred because SWC in the profile at the beginning of the
rainy season was close to WP while at the end of the season
it was higher because crop water use or evaporation had not
yet depleted it back down to its initial value. The simulations
indicated that the change in soil water storage between the
beginning and end of the season was most affected by the
timing and size of the final irrigation. However, there was
also a change in the average seasonal evapotranspiration ratio
(ETa/Etc) for the 27‐year period. The lower this ratio, the
more the crop experienced periods of water stress. The
minimum ratio, 0.88, occurred when the root growth was low
(5 mm/d), while the maximum ratio of 0.97 occurred for
simulations with high initial water content (30%) in the
profile below the initial root zone.
Table 3. Relative sensitivities and differences for the 27‐year averages of the model's
water balance components caused by changes in selected model parameters.
Relative
Sensitivity
Difference [Test‐Base]
Parameter
Base
Value
Test
Value
ETc
(mm)
Irrigation
(mm)
Irrigations
(No.)
Evaporation
(mm)
Leaching
(mm)
Moisture
Change
(mm)
ETc Irrigations
Duration of development stage[a] (d) 90 76 1.04 0.92 ‐101 ‐65 ‐0.6 ‐16 0 13
Duration of development stage[a] (d) 90 104 1.24 1.22 120 72 0.8 17 ‐1 ‐14
Crop coefficient at mid stage 1.1 1 0.23 0.26 ‐13 ‐11 ‐0.1 23 1 ‐2
Crop coefficient at mid stage 1.1 1.2 0.34 0.26 19 11 0.1 ‐17 ‐1 ‐3
Available water content (vol %)[b] 16 12 ‐0.02 ‐1.05 4 ‐9 1.1 4 6 ‐3
Available water content (vol %)[b] 16 20 ‐0.02 ‐0.67 ‐3 1 ‐0.7 ‐3 ‐4 ‐1
Depletion level at which stress occurs (% TAW)[c] 50 40 0.01 0.24 ‐1 ‐9 ‐0.2 ‐1 1 3
Depletion level at which stress occurs (% TAW)[c] 50 60 0.01 0.24 1 9 0.2 1 0 ‐1
Readily evaporable water (mm) 8 6 0.01 0.00 ‐2 2 0.0 ‐2 1 2
Readily evaporable water (mm) 8 12 0.01 0.00 4 5 0.0 4 ‐1 1
Total evaporable water[d] (mm) 54 40.5 0.06 0.10 ‐10 ‐10 ‐0.1 ‐10 ‐1 ‐2
Total evaporable water[d] (mm) 54 81 0.03 ‐0.05 10 12 ‐0.1 10 1 ‐1
Maximum potential root zone (m) 1.2 0.9 ‐0.02 ‐0.29 3 ‐3 0.3 3 10 ‐19
Maximum potential root zone (m) 1.2 1.5 0.00 ‐0.10 ‐1 ‐1 ‐0.1 ‐1 ‐7 10
Root growth (mm/d) 10 5 ‐0.02 ‐0.95 6 ‐18 2.0 6 ‐2 4
Root growth (mm/d) 10 18 0.00 ‐0.15 ‐2 8 ‐0.5 ‐2 4 1
Root growth moisture threshold (% TAW) 50 30 0.01 0.36 ‐2 12 ‐0.6 ‐2 5 2
Root growth moisture threshold (% TAW) 50 70 0.01 0.77 3 ‐16 1.3 3 ‐1 ‐3
Initial soil moisture below root zone (vol %) 24.5 22 ‐0.01 ‐0.47 1 ‐2 0.2 1 ‐5 14
Initial soil moisture below root zone (vol %) 24.5 30 ‐0.01 ‐0.85 ‐1 ‐5 ‐0.8 ‐1 15 ‐47
Base run (results in mm) 620 320 4.2 167 8 70
[a] The change in the length of the development stage also changed the total length of the crop season by the same number of days.
[b] Both the wilting point and the field capacity were adjusted by 2%. The initial volumetric soil moisture content was kept similar to that of the base
run.
[c] A lower depletion level (p) indicates that the crop transpiration becomes stressed at a lower level of soil moisture depletion, which is a higher level
of available soil moisture (the fraction is taken backwards from field capacity). Irrigation, in the third crop stage, was applied as soon as the soil
moisture fell below the depletion level.
[d] With the TEW also the depth of the evaporation layer and initial root zone was changed to 15 and 30 cm, respectively. The initial soil moisture in
this layer was set equal to the TEW, but the total soil moisture in the profile remained the same as for the base run.

211Vol. 24(2): 203‐214
The average total number of irrigations required per year
during the 27‐year period was highly sensitive to almost all
changes in the soil's water holding properties as well as root
growth and crop parameters. The change in the number of
irrigations required was highest when the root growth rate
was low (5 mm/d); the soil water threshold required for root
growth was high (70% of TAW); and when the TAW was low
(12%). These parameters all affect the depth of the root zone,
with a smaller root zone requiring more irrigations (but lower
water amounts per irrigation). Increases in the root growth
rate beyond 10 mm/d had a less dramatic effect on the depth
of the root zone and the resulting irrigation schedule. The
simulated average depth of the root zone by the time of the
first irrigation in early April was 91 cm for the base
simulation, but was only 53 cm for the low root growth rate,
and 67 cm for the high root growth moisture threshold. These
last two depths may be low considering the measured soil
moisture extraction patterns, but this cannot be confirmed
without appropriate measurements of SWC.
The average amount of leaching from the profile
simulated by the model for the 27‐year period varied between
2 and 15 mm per year for all runs. In 4 out of the 27 years,
leaching occurred during the January to March period.
One‐third of the total leaching occurred during the irrigation
season, but this could probably have been prevented if the
model (or the user) could have taken the weather forecast into
account and thus prevented irrigations when the probability
of rainfall is sufficiently high. A sound management practice
to reduce the chance of leaching is to not fully refill the root
zone to FC but to leave some storage capacity as warranted
by the probability of rainfall.
The results of this long‐term sensitivity analysis have
implications for the potential use of this model as a real‐time
irrigation scheduling tool:
SThe estimation of irrigation needs is clearly very sensitive
to the duration of the crop stages. Considering the
variability of the climatic characteristics that affect crop
growth, it is important to select the crop stage durations
based on current field observations of crop development.
SKnowledge of the behavior of the root system under soil
moisture limiting conditions is important for improving
the efficiency of supplemental irrigation.
SThe threshold at which the crop starts experiencing stress
(p) was not the most sensitive parameter for irrigation
scheduling, but a better understanding of this parameter is
needed to quantify the effect of delaying irrigation beyond
the optimal date.
IRRIGATION SCHEDULING
The results of the irrigation scheduling simulations are
presented in table 4 as averages of the individual simulations
in each of the 27 years of climate data. In 25 of the 27 years,
roots in the test case simulation (90‐cm soil profile with TAW
of 12%) had reached the bottom of the profile by the time of
the first irrigation. The application of the same irrigation
schedule to a soil with a higher water storage capacity and
dryer or wetter initial SWC in the lower profile (dry and wet
cases, respectively) resulted in a maximum change in the
evapotranspiration ratio of 0.02 (from 0.96 to 0.94) for the
dry case and 0.03 (from 0.93 to 0.90) for the wet case. Apart
from the amount of irrigation water applied, an important
consideration for farmers is the number of irrigations because
this directly affects their labor input.
For the dry case the number of irrigations was reduced or
stayed the same in 9 of the 27 years compared to the test case.
In 12 of the 27 years one additional irrigation was needed, and
in 6 of the years two extra irrigations were needed. For the wet
case, one additional irrigation was required in 12 of the
27 years, two additional irrigations were required in 13 of the
years, and three additional irrigations in 2 of the years.
Soil moisture content in the root zone for the test case, the
dry and wet cases, and the dry and wet cases with the test case
irrigation schedule are presented in figure 7 for the 1990/91
season. This season ranked among the top 5 seasons in terms
of the decrease in the evapotranspiration ratio as a result of
imposing the test case irrigation schedule on both the dry and
the wet cases. For the dry case (fig. 7a) the schedule applied
by the simulations had an evapotranspiration ratio of 0.87,
while applying the schedule from the test case resulted in a
ratio of 0.84. For the wet case (fig. 7b) the evapotranspiration
decreased from 0.97 to 0.95. Imposing the test case irrigation
schedule on the dry case (fig. 7a) clearly caused SWC to fall
below 30%, (the crop stress level, p) during the second half
of April.
The 2001/02 season ranked among the top five seasons for
both the dry and the wet case in terms of the increase in the
evapotranspiration ratio as a result of imposing the test case
irrigation schedule. Figure 8 shows SWC in the root zone for
the test case, the dry and wet cases, and the dry and wet cases
with the test case irrigation schedule for this season. For the
dry case (fig. 8a) the evapotranspiration ratio increased from
0.94 for the simulated schedule to 0.95 for the imposed test
case schedule. For the wet case (fig. 8b), the
evapotranspiration ratio correspondingly increased from
0.97 to 0.99. However, the reduction in the ratio in both the
dry and wet cases when irrigations are scheduled by the
model rather than imposing the test case irrigation schedule
does not take into account the timing of the increased stress.
The main difference occurred at the end of the season due to
the timing of the irrigation just before maturity when the crop
is not very sensitive to stress.
Table 4. Comparison of simulated average number of irrigations, irrigation amount, and evapotranspiration ratios for a test case and for
simulations with a deeper potential root zone (RZp), higher total available water fraction (TAW), and different initial water content
below the initial root zone (SWCL) for irrigation schedules as simulated and using the same schedule as the test case. [a]
Case
Irrigation
Schedule
RZp
(cm)
FC
(%) WP
TAW
(%)
SWCL
(%)
Irrigations
(no.)
Irrigation
(mm) ETa/ETc
Test Simulated 90 36 24 12 27 5.1 303 0.92
Dry Simulated 120 38 22 16 22 4.4 318 0.90
Dry As test 120 38 22 16 22 5.1 303 0.89
Wet Simulated 120 38 22 16 30 3.4 315 0.97
Wet As test 120 38 22 16 30 5.1 303 0.97
[a] The simulations are based on daily climate data at Tel Hadya for the 27 winter cropping seasons from 1979/80 to 2005/06.

212 APPLIED ENGINEERING IN AGRICULTURE
Both figures 7 and 8 illustrate the lower soil water contents
in the test case after the initial irrigation or rain at the
beginning of the season, due to the lower water storage
capacity. Once the root zone starts expanding, root growth is
generally more limited for the dry case than for the test case,
because the smaller TAW of the test case also results in higher
volumetric soil moisture percentages (figs. 7a and 8a).
Conversely, root growth for the wet case is more rapid than
for the test case. By growing into a relatively wet sub soil,
moisture contents generally remained higher for the wet case
than for the test case (figs. 7b and 8b).
SUMMARY AND CONCLUSIONS
A model in spreadsheet form, based on the spreadsheet
program of Allen et al. (1998) but accounting for the dynamic
development of the root zone in dry soils, was developed for
computing net irrigation water needs for supplemental
irrigation of winter crops in a homogeneous soil in
Mediterranean environments. The dual crop coefficient
approach that splits evapotranspiration into crop
transpiration and evaporation from the soil was used.
Evaluation of the model with a data set on wheat showed that
(a)
(b)
Figure 7. Average soil water contents in the root zone for the test case (90cm soil depth, TAW 12%, initial SWC 27% in lower profile); and (a) dry case
(120‐cm soil depth, TAW 16%, initial SWC 24% (WP) in the lower profile along with the dry case with the test case irrigation schedule; and (b) the test
case, the wet case (120‐cm soil depth, TAW 16%, initial SWC 30% in the lower profile) and the wet case with the test case irrigation schedule, for the
1990/91 season. Rainfall and irrigations are also shown. Note: the test, dry, and wet cases all received the same first irrigation at the beginning of the
season in November.

213Vol. 24(2): 203‐214
(a)
(b)
Figure 8. Average soil water contents in the root zone for the test case (90‐cm soil depth, TAW 12%, initial SWC 27% in lower profile); and (a) dry case
(120‐cm soil depth, TAW 16%, initial SWC 24% (WP) in the lower profile along with the dry case with the test case irrigation schedule; and (b) the test
case, the wet case (120‐cm soil depth, TAW 16%, initial SWC 30% in the lower profile) and the wet case with the test case irrigation schedule, for the
2001/02 season. Rainfall and irrigations are also shown. Note: the test, dry and wet cases all received the same first irrigation at the beginning of the
season in November.
it could predict SWC within acceptable error limits.
Sensitivity analysis showed that, within the expected
parameter value range, computed evapotranspiration for a
typical supplemental irrigation scenario for a clay soil in
northern Syria was not sensitive to changes in the soil
evaporation parameter values over the 27 year climate data
record. The resulting computed irrigation schedule was
sensitive to changes in all other soil, crop, and root
parameters. However, the use of the model to develop
irrigation schedules for application to areas where the
weather is the same but soil properties and initial moisture
conditions may be different produced schedules that resulted
in limited crop stress for the different cases.
In Syria, as in many other dry areas, water is the resource
that most limits crop production, with irrigation being the
major consumer of this resource. In lieu of measuring SWC
before making an irrigation decision, this model could be a
useful tool to help with the decision because it is simple yet

214 APPLIED ENGINEERING IN AGRICULTURE
appears to have sufficient accuracy to generate realistic
guidelines. Using the model with real‐time data from climate
stations and rain gauges could help agricultural research and
extension services to provide information that would enable
farmers to make improved irrigation management decisions.
Application of the model for various conditions and scenarios
could aid in the development of policies to use irrigation
water more effectively in Mediterranean climates. The
model is available at www.rec.udel.edu/TopLevel/Research
_staff_and_programs.htm or www.icarda.cgiar.org.
ACKNOWLEDGMENTS
The authors would like to acknowledge the discussions
with our ICARDA colleagues Hamid Farahani, Ahmed
Hachum and Bogachan Benli and their help in shaping the
current version of this paper and providing internal reviews.
We thank the assistants of ICARDA's Integrated Water and
Land Management Program for the management of the field
trials, and Dr. Eddy De Pauw and his staff for the collection
and provision of online access to the ICARDA climate data.
The contribution of the first author was supported by
USAID‐CGIAR linkage funds.
REFERENCES
Allen, R. G., L. S. Pereira, D. Raes, and M. Smith. 1998. Crop
evapotranspiration. Guidelines for computing crop water
requirements. FAO Irrigation and drainage paper 56. Rome,
Italy: FAO.
Allen, R. G., L. S. Pereira, M. Smith, D. Raes, and J. L. Wright.
2005. FAO‐56 dual crop coefficient method for estimating
evaporation from soil and application extensions. J. Irr. Drain.
Eng. 131(1): 2‐13.
Barraclough, P. B., and R. A. Leigh. 1984. The growth and activity
of winter wheat roots in the field: the effect of sowing date and
soil type on root growth of high yielding crops. J. Agric. Sci.,
Camb. 103(1): 59‐74.
Bastiaanssen, W. G. M., R. G. Allen, P. Droogers, G. D'Urso, and P.
Stewduto. 2007. Twenty‐five years modeling irrigated and
drained soils: State of the art. Agric. Water Management 92(3):
111‐125
Brown, S. C., D. H. Keatinge, P. J. Gregory, and P. J. M. Cooper.
1987. Effects of fertilizer, variety and location on barley
production under rain‐fed conditions in northern Syria. 1. Root
and shoot growth. Field Crops Research 16(1): 53‐66.
Bruggeman, A., I. McCann, M. Pala, and T. Oweis. 2005. Improved
decision making for deficit irrigation of wheat in northern Syria.
ASAE Paper No. 052217. St. Joseph, Mich.: ASAE.
Cooper, P. J. M., P. J. Gregory, D. H. Keatinge, and S. C. Brown.
1987. Effects of fertilizer, variety and location on barley
production under rain‐fed conditions in northern Syria. 2. Soil
water dynamics and crop water use. Field Crops Research 16(1):
67‐84.
Ellis, F. B., and B. T. Barnes. 1980. Growth and development of
root systems of winter cereals grown after different tillage
methods including direct drilling. Plant and Soil 55(1):
283‐295.
Gregory, P. J., McGowan, P. V. Biscoe, and B. Hunter. 1978. Water
relations of winter wheat. I. Growth of the root system. J. Agric.
Sci., Camb. 91(1): 91‐102.
Gregory, P. J., D. Tenant, and R. K. Belford. 1992. Root and shoot
growth, and water and light use efficiency of barley and wheat
crops grown on a shallow duplex soil in a Mediterranean‐type
environment. Australian J. Agric. Res. 43(3): 555‐573.
Izzi, G, H. J. Farahani, A. Bruggeman, and T. Y. Oweis. 2007.
In‐season wheat root growth and soil water extraction in the
Mediterranean environment of northern Syria. Agric. Water
Management 95(3): 259‐270.
Legates, D., and G. McCabe. 1999. Evaluating the use of
“goodness‐of‐fit” measures in hydrologic and hydroclimatic
model validation. Water Resources Res. 35(1): 233‐241.
MAAR (Ministry of Agriculture and Agrarian Reform). 2006. The
annual agricultural statistical abstract 2005. Department of
Planning and Statistics, Division of .Agricultural Statistics
Computer Center. Damascus, Syria: MAAR.
Maheshwari, B., M. Plunkett, and P. Singh. 2003. Farmer's
perceptions about irrigation scheduling in the
Hawkesbury‐Nepean catchment. Proc. of the Australasia Pacific
Extension Network. Available at www.apen.org.au.
NAPC (National Agricultural Policy Center). 2003. The state of
food and agriculture in the Syrian Arab Republic 2002.
Damascus, Syria: MAAR.
Oweis, T. Y., and A. Hachum. 2006. Water harvesting and
supplemental irrigation for improved water productivity of dry
farming systems in West Asia and North Africa. Agric. Water
Management 80: 57‐73.
Oweis, T., M. Pala, and J. Ryan. 1999. Management alternatives for
improved durum wheat production under supplemental
irrigation in Syria. European Journal of Agronomy 11(3‐4):
255‐266.
Oweis, T., P. N. Rodrigues, and L. Pereira. 2003. Simulation of
supplemental irrigation strategies for wheat in near east to cope
with water scarcity. In Tools for Drought Mitigation in
Mediterranean Regions, eds. G. Rossi et al., 259‐272. The
Netherlands: Springer.
Pala, M, C. O. Stockle, and H. C. Harris. 1996. Simulation of
Durum wheat (Triticum turgidum ssp. Durum) growth under
different water and nitrogen regimes in a Mediterranean
Environment using CropSyst. Agricultural Systems 51(2):
147‐163.
Ryan, J., S. Garabet, S. Masri, J. Diekman and H. Habib. 1997.
Soils of ICARDA's agricultural experiment stations and sites.
ICARDA, Aleppo, Syria.
Sato, S., O. S. Abdalla, T. Y. Oweis, and T. Sakuratani. 2006. The
validity of predawn leaf water potential as an irrigation‐timing
indicator for field‐grown wheat in northern Syria. Ag. Water
Management 82(1‐2): 223‐236.
Schweers, W., A. Bruggeman, A. Rieser, and T. Oweis. 2004.
Irrigation practice and water‐use efficiency in Khanasser valley,
northwest Syria. J. Applied Irrig. Sci. 39(2): 241‐252.
Zhang, H., and T. Oweis. 1999. Water‐yield relations and optimal
irrigation scheduling of wheat in the Mediterranean region.
Agric. Water Management 38(3): 195‐211.