Agricultural Water Management 77 (2005) 159–179

www.elsevier.com/locate/agwat

GISAREG—A GIS based irrigation scheduling

simulation model to support improved
water use
P.S. Fortes a, A.E. Platonov b, L.S. Pereira a,*
a
Agricultural Engineering Research Center, Institute of Agronomy, Technical University of Lisbon,
Tapada da Ajuda, 1349-017 Lisbon, Portugal
b
Scientific Information Center of the Interstate Commission for Water Coordination (SIC-ICWC),
Karasu-4, 700187 Tashkent, Uzbekistan
Accepted 1 September 2004
Available online 23 March 2005

Abstract

To support improved irrigation scheduling in the Syr Darya basin, Uzbekistan, the ISAREG model
was selected due to previous applications in similar climates. It is a conceptual non-distributed water
balance model for simulating crop irrigation schedules at field level and to compute irrigation
requirements under optimal and/or water stressed conditions. To provide for the use of the model at
the project scale, a new version of the model, integrated with a Geographical Information System
(GIS), was developed. This GIS based application is aimed at supporting the implementation of
improved farm irrigation management and, in a later phase, to also help project management. The
integration concerns the creation of spatial and weather GIS databases usable by ISAREG, the
models operation for different water management scenarios, and the production of crop irrigation
maps and time dependent irrigation depths at selected aggregation modes, including the farm scale.
The resulting information on alternative irrigation schedules is, therefore, spatially distributed and
shall be used both to support irrigation scheduling advising and to help in the identification of
practices that may lead to water saving and provide for salinity control. The paper includes brief
descriptions of the models, the databases and the integration of the models, as well as main features of
the GISAREG application. Results relative to the comparative analysis of irrigation scheduling
scenarios show the capabilities of the model to support the selection of water saving alternatives. The
appropriateness of implementing 15–20-day time intervals between irrigations is shown. In addition,

* Corresponding author. Tel.: +351 21 3653480; fax: +351 21 3621575.

E-mail addresses: p.fortes@clix.pt (P.S. Fortes), a.lexp@sicicwc.aral-sea.net (A.E. Platonov),
lspereira@isa.utl.pt (L.S. Pereira).

0378-3774/$ – see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.agwat.2004.09.042
160 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

results relative to compare net irrigation requirements for wet and dry years show the usefulness of
the model in handling time series of weather data.
# 2005 Elsevier B.V. All rights reserved.

Keywords: Water balance; Irrigation requirements; Irrigation water savings; Aral Sea basin; Modeling and
simulation

1. Introduction

Irrigation scheduling is the farmers decision process relative to ‘‘when’’ to irrigate and
‘‘how much’’ water to apply at each irrigation event. It requires knowledge of crop water
requirements and yield responses to water, the constraints specific to the irrigation method
and respective on-farm delivery systems, the limitations of the water supply system relative
to the delivery schedules applied, and the financial and economic implications of the
irrigation practice (Smith et al., 1996). When adequately applied, irrigation scheduling
favors improved water use, i.e. reduced irrigation demand and higher water and land
productivity (Pereira et al., 2002).
A large number of tools are available to support field irrigation scheduling, from sensors
to simulation models (Martı́n Santa Olalla et al., 1999; Tarjuelo and de Juan, 1999).
Irrigation scheduling models are particularly useful to support individual farmers and
irrigation advisory services (Ortega et al., 2005). A main interest in using models results
from their capabilities to simulate alternative irrigation schedules relative to different
levels of allowed crop water stress and to various constraints in water availability (Pereira
et al., 1995).
The irrigation scheduling simulation model ISAREG is after long in use in several parts
of the World for evaluating current irrigation schedules, selecting the most appropriate
irrigation scheduling for several crops, and for computing crop irrigation requirements
using weather data time series (Teixeira and Pereira, 1992; Liu et al., 1998). The model is
able to generate irrigation scheduling alternatives that are evaluated from the relative yield
loss produced when crop evapotranspiration is below its potential level. Examples of those
applications to winter and summer irrigated crops are presented by Oweis et al. (2003) and
Zairi et al. (2003) for surface irrigation in the Mediterranean region, and Liu et al. (2000)
and Campos et al. (2003) for surface irrigated crops in North China.
The main limitation of this type of simulation models is that model computations are
performed at the crop field scale for specific soil, crop, and climate conditions, which
characterize that crop field and the respective cropping and irrigation practices. When the
computation procedure is applied at the region scale it becomes heavy and slow due to the
need to consider a large number of combinations of field and crop characteristics to be
aggregated at sector or project scales (e.g. Teixeira et al., 1996), or adopting geostatistical
tools such as krigging (Sousa and Pereira, 1999). However, the spatially distributed
characteristics of the input data required by ISAREG makes their integration with a
Geographical Information System (GIS) particularly attractive and useful as proved with
other water balance models (Burrough and McDonell, 1998). The interest of this
approach is increased when remote sensed crop data is available as for this application.
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 161

The GIS treat the input water balance data as a set of layers (Tim, 1996) and is able to
identify for each cropped field the respective set of predominant characteristics. Former
GIS applications include regional evapotranspiration (Hashmi et al., 1995), crop
evapotranspiration (Ray and Dadhwal, 2001), crop yields estimation (Qing and Linvill,
2002), delivery scheduling analysis (Rowshon et al., 2003), or the assessment of recharge
from irrigated fields (Chowdary et al., 2003). Many other applications to irrigation are
reported but most of them do not refer to an integration of the models with the GIS, as it is
the case of previous ISAREG applications with GIS (San-Payo and Teixeira, 1996; Pereira
et al., 1999).
Assuming that each field is homogeneous, it is possible to automatically call the model
for each cropped field represented in the GIS crop fields theme and then up-scaling the
results produced using a variety of attributes. This is the basic procedure adopted in
GISAREG, which is an effective integration of the ISAREG model with a GIS. This
application is developed in the framework of a research project aimed at water saving and
salinity control in selected irrigation areas of the Syr Darya River Basin, in Uzbekistan,
Kyrgyzstan and Tajikistan, Aral Sea Basin.
A general view on the problems, possible solutions and prospects to the Aral Sea Basin
is presented by Dukhovny (2003). Issues for water saving at farm level in the area are
discussed in other papers in this issue: Horst et al. (2005) discuss improvements relative to
surface irrigation systems and the need to combine them with improved irrigation
scheduling aimed at water saving; Stulina et al. (2005) use the RZWQM model to simulate
alternative crop management practices, including a combination with ISAREG results,
aimed at water saving and improved water productivity.
The main interest in using GISAREG results from the capability of the model to
simulate alternative irrigation schedules relative to different levels of allowed crop water
stress as well as to various constraints in water availability. The irrigation scheduling
alternatives are evaluated from the relative yield loss produced when crop evapotranspira-
tion is below its potential level. The specific objectives of this paper include the
presentation of the model development and examples of its application to the computation
of spatially distributed crop irrigation requirements and to simulate the irrigation demand
aggregated at main nodes of the distribution system. The application herein refers to the
farming area Gafura Gulyama, Syr Darya Province, Uzbekistan.

2. The ISAREG model

The ISAREG model is an irrigation scheduling simulation model that performs the soil
water balance at the field scale. The model is described with detail by Teixeira and Pereira
(1992), Liu et al. (1998) and Pereira et al. (2003). The water balance is performed for a
multilayered soil and follows the classical approach referred by Doorenbos and Pruitt
(1977). Various time step computations are adopted, from daily up to monthly, depending
on weather data availability. Inputs are precipitation, reference evapotranspiration (ETo),
total and readily available soil water, soil water content at planting, and crop factors
relative to crop growth stages, crop coefficients, root depths and water-yield response
factors (Fig. 1).
162 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

Fig. 1. Diagram of ISAREG model and links with programs KCISA and EVAP56.

ISAREG has two auxiliary programs, EVAP56 to compute ETo, and KCISA to compute
the crop factors (Fig. 1). EVAP56 performs the calculation of ETo with any of the
alternative methods proposed in the guidelines FAO 56 (Allen et al., 1998), depending on
the availability of weather data. KCISA uses the FAO methodology (Allen et al., 1998) to
compute the time averaged crop coefficients for the initial, mid and end season
(Kc,ini, Kc,mid and Kc,end), the soil moisture depletion fraction for no stress ( p), and the
effective rooting depths (Zr) for each crop development stage (Rodrigues et al., 2000). Four
crop development periods are considered: initial, crop development, mid season and end
season.
The ISAREG model computes the actual evapotranspiration from the potential crop
evapotranspiration (ETc = KcETo) depending upon the soil water availability in the root
zone. The model starts the soil water simulations with an initial soil water content provided
by the user or simulated from an antecedent period of fallow, where simulation starts at end
of summer, when most of soil water is consumed, or by the winter, when replenishment of
soil water may be assumed. Irrigation depths and dates are set in various ways depending
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 163

on the available data and the user options relative to the application depths and the soil
water thresholds. The impacts of water stress on yields are assessed using the approach
proposed by Stewart et al. (1977), where the relative yield losses are a function of the
relative evapotranspiration deficit through the water-yield response factor, Ky. The model
includes an algorithm for improved computation of the groundwater contribution (GW)
and the percolation (Fernando et al., 2001; Liu et al., 2001), where GW is a function of the
groundwater table depth, soil water storage, soil characteristics influencing capillarity and
ETc. An algorithm to take into consideration the salinity impacts on ETc and yields is also
added (Campos et al., 2003).
The ISAREG model performs the irrigation scheduling simulations according to user-
defined options such as:

 to define an irrigation scheduling to maximize crop yields, i.e. without crop water stress;
 to define an irrigation scheduling using selected irrigation thresholds, including for
allowed water stress and responding to water restrictions imposed at given time periods;
 to evaluate yield and water use impacts of a given irrigation schedule;
 to test the model performance against observed soil water data and using actual
irrigation dates and depths;
 to execute the water balance without irrigation;
 to compute the net crop irrigation requirements. Including to perform the frequencial
analysis of irrigation requirements when a weather data series is considered.

The model input data can be provided at run-time by keyboard or trough pre-defined
ASCII Files. The input data required for ISAREG and the auxiliary program KCISA are
summarized in Table 1. Each item concerns a specific input file.
The ISAREG model has been validated in numerous applications (e.g. Liu et al., 1998;
Oweis et al., 2003; Zairi et al., 2003). For the Syr Darya basin it was validated using
appropriate meteorological and soil water data sets relative to observations performed with
cotton in Gafura Gulyama, and with cotton and wheat in Fergana Valley, Uzbekistan
(Cholpankulov et al., 2004).

3. GISAREG, the model integration with a GIS

3.1. GISAREG feature

The integration of ISAREG and KCISA with GIS follows a close coupling strategy
developed in a commercial GIS (ArcView 3.2) using Avenue script language. ISAREG and
KCISA where converted into dynamic link libraries (DLL), i.e. compiled collections of
procedures or functions, that can be called from the new application and linked to it at run-
time. This allows a smooth integration between GIS and the simulation models. Differently
to the ISAREG Windows version, where the programs KCISA and EVAP56 are integrated
with the model (Fig. 1), in the GISAREG distinct links with the GIS are adopted, so the
EVAP56 computations are performed out of the GIS and the ETo data are input to
GISAREG.
164 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

Table 1
Model input data
Input data for ISAREG
Meteorological data
Effective precipitation, Pe (mm)
Reference evapotranspiration, ETo (mm)

Crop data (when KCISA not used)

Dates of the crop development stages
Crop coefficients, Kc
Root depths, Zr (m)
Soil water depletion fraction for no stress, p
Seasonal water-yield response factor, Ky

Soil data (multi-layered soil)

Layer depths
Soil water at field capacity, uFC (mm mm 1)
Soil water at wilting point, uWP (mm mm 1)
Initial soil water content, ui (mm mm 1)

Ground water contribution

Potential groundwater contribution, G (mm day 1), or groundwater table and soil data

Irrigation data
Irrigation thresholds
Options to define application depths
Restrictions on water availability (time periods and water volumes)

Input data for KCISA (Kc’s computation)

Meteorological data
Precipitation, Pe (mm)
Reference evapotranspiration, ETo (mm)
Number of rainfall events, n ( )
Wind speed, U2 (m s 1)
Minimum relative humidity, RH (%), or maximum and minimum temperature, Tmax and Tmin (8C)

Crop data
Crop planting or season initiation date
Dates or lengths of the crop development periods
Root depths, Zr (m)
Soil water depletion fraction for no stress, p; seasonal water-yield response factor, Ky

Soil data (evaporation layer)

Layer depth, Ze (m)
Initial soil water content, uie (mm mm 1)
Soil textural percentages
Soil water content at field capacity, uFC (mm mm 1)
Soil water at wilting point, uWP (mm mm 1)

Soil and water salinity data, ECe (dS m 1)

Irrigation data (initial period, for computing, Kc,ini),

Fraction of soil surface wetted by the irrigation, fw
Foreseen number of irrigation events
Foreseen depth of irrigation events, I (mm)
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 165

GISAREG do not implement all the ISAREG simulation options but only required to
satisfy the objectives of the study:

1. to schedule irrigation aiming at maximum yields;

2. to simulate an irrigation schedule with allowed water stress depending upon the users
selected irrigation thresholds and water restrictions;
3. to execute the water balance without irrigation;
4. to compute the net crop irrigation requirements, including to perform the frequencial
analysis of irrigation requirements when a weather data series is considered.

The GIS component of the application allows the simulation of different user defined
simulation scenarios and the disposal of specific tools for their creation and management.
A simulation scenario concerns a given spatial distribution of crops, irrigation methods,
irrigation scheduling options, and water restrictions. The simulation units correspond to the
cropped fields.
The GISAREG application is performed adopting the following sequential procedures:

1. Loading the spatial and non-spatial database;

2. GIS overlay procedures to identify the main soil and climate characteristics of each
cropped field, which produces the ‘‘Default’’ simulation table;
3. Creation of ISAREG and KCISA input files relative to items listed in Table 1;
4. Calling of KCISA and ISAREG for each cropped field to compute its crop irrigation
requirements and irrigation scheduling relative to a user selected crop scenario;
5. GIS reading of ISAREG outputs for mapping and integrating the results.

The essential feature of GISAREG referring to data and models is presented in Fig. 2:
each field is characterized by a combination of crop, soil and weather characteristics;
KCISA computes then the respective crop and soil input files for ISAREG; these files
together with the ET and rainfall data files are used for ISAREG simulations following a

Fig. 2. GISAREG structure: GIS overlay operation to identify the simulation units, creation by KCISA of the crop
and soil files, and ISAREG simulation relative to a crop scenario producing respective irrigation depths and
timings per a field or aggregated to a selected area.
166 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

selected crop scenario; finally, simulation results relative to a given date or summed up for
a selected period of time are mapped for the selected areas.

3.2. Database

The GIS database is constituted by point and polygon themes data, the first relative to
weather data and the second to soil, crops and fields data. Weather data refers to one or
more years. When weather data series are used, multiple simulations are performed to
determine the frequency of crop water and irrigation requirements, or to perform an
irrigation planning analysis relative to selected years such as dry, average and wet years.
The meteorological stations are identified and the respective weather data are stored in
ASCII files formatted according to ISAREG and KCISA requirements (Table 1).
A MS Access database is used to store the non-spatial data that will be coupled with the
respective polygon GIS themes (Fig. 3). This database is constituted by the following
tables:

 The crop successions table, which defines the annual crops succession (two crops that
successively occupy the same field) and concerns data on these crops and on the
respective irrigation method used. It includes the codes for identification of the crop
succession and respective designation, the winter crop and the respective irrigation
method, and the summer crop and irrigation method.
 The crop table, including: the crop identification code and its designation; the crop type
code (1 = ‘‘bare soil’’, 2 = ‘‘annual crop’’ and 3 = ‘‘annual crop with a soil frozen
period’’), which identifies the time period to be used by the model to estimate the initial
soil moisture; and the crop input data referred in Table 1 required for computing the crop
evapotranspiration using the recent FAO methodology (Allen et al., 1998).
 The irrigation methods table, containing the code and designation of the irrigation
method, and the respective fraction of soil surface wetted by irrigation.
 The soils table, including the soil identification code and designation and the soil data
described in Table 1 required for performing the water balance of the evaporative layer
(Allen et al., 1998; Rodrigues et al., 2000) and the soil water balance of the cropped soil
(Pereira et al., 2003).

The dominant characteristics of each simulation unit (cropped field) are identified by the
GIS through spatial data relative to crops, soils and meteorological stations concerning the
following input themes (Fig. 3):

 A polygon theme with the delimitation of the crop fields, which is associated with a table
relative to field attributes, including the identification code of each crop field, the
identification of the annual crop succession, and other relevant information as desired by
the user;
 A polygon theme with the delimitation of soil types having associated a table of soil
attributes, and where soils are identified by an appropriate code;
 A point theme with the location of the meteorological stations, which is also associated
with a table of attributes where the meteorological stations are identified by a code.
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179
Fig. 3. Schematic representation of the spatial and non-spatial databases and respective inter-relations.

167
168 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

The spatial databases relate with the non-spatial MS Access database and meteor-
ological ASCII files through the above-mentioned attributes tables as described in
Fig. 3.

3.3. Operation with spatial data

GISAREG operation initiates by loading the spatial and non-spatial database followed
by GIS overlay procedures which allow the identification of the soil, climate and cropping
characteristics of each cropped field (Fig. 2). The GIS generates a Thiessen polygons theme
that defines the geographical influence of each meteorological station (Fig. 4). Thus,
overlaying the meteorological Thiessen polygons theme with the field polygons theme, it
results assigning the climate data to each field (Fig. 4). In the following, a selection of the
fields to be considered for the simulation is performed. Non-cropped fields are excluded.
The last overlay procedures consist of the intersection between the soils themes and the
cropped fields, which is followed by the identification of the dominant soil type in each
field polygon, so assigning only one soil data set to each field (Fig. 4).
The weather data utilized in this application refer to the meteorological stations of Syr
Darya and Djizak and to the period 1970–1999, consisting of 10-day data on precipitation,
maximum and minimum temperature, average relative humidity, sunshine duration, and
wind speed. Data were provided by the Central Asian Scientific Research Institute of

Fig. 4. Above: intersecting cropped fields with the climatic Thiessen polygons and identification of the climatic
data dominant in each field; below: intersecting soils with cropped fields and identification of the soil dominant in
each field. For both operations, also included the identification of fields to be simulated.
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 169

Irrigation (SANIIRI), Tashkent. The crop fields’ polygons were obtained from
digitalization and interpretation of satellite images at SIC-ICWC, Tashkent. The crops
assigned to each field were obtained from comparing the NDVI relative to two images
taken in different dates (April and August) using a automated computer classification. The
soil maps are originated from the SIC-ICWC database and soil attributes were provided by
SANIIRI (G. Stulina, personal communication). Because fields are large, generally of
several hectares each, corresponding to those of former state and cooperative farms, the
accuracy of remote sensed data was adequate. The soil data were obtained from past soil
surveys in the area, which were complemented with laboratory analysis, also having
adequate quality for the aimed application.
Once crop, climatic and soil characteristics are assigned to each cropped field, a
simulation table (ST) is built. This ST is a *.dbf table relative to each ‘‘Simulation
Scenario’’, which stores the data required to perform the corresponding simulation for all
crop fields in the selected area (Table 2). In the ST, every row corresponds to a cropped field
but fields having a crops succession (a winter and a summer crop) are represented by two
rows.
The ST columns are filed by default with the values stored in the non-spatial database
referred above, but columns relative to the dominant soil and meteorological station are
filled through the GIS overlay procedures described above. New simulation scenarios are
built by editing the cells of the ST and saving that them with a user defined descriptive
name. Each created simulation scenario is stored in the hard disk in a form *.dbf file, which
may be easily imported to the ST for consulting and modification, or to execute a
simulation with the selected scenario.

Table 2
Description of the simulation table
Codes
Field identification
Crop
Irrigation method,
Dominant soil in the field,
Meteorological station
Irrigation scheduling option
Water restrictions

Data for building a simulation scenario

Crop seeding or planting date

Lengths of the crop stages

Initial period, Lini (days)
Development, Ldev (days)
Mid stage, Lmid (days)
End stage, Lend (days)

Harvest date, Harv-date = Seed-date + Lini + Ldev + Lmid + Lend

Initial soil moisture in the top and deeper layers
Crop field area
170 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

Fig. 5. General view of the GISAREG application interface including the crops map and respective legend, the
simulation table (ST) and the scenarios management window (below on right). On top left, ST contextual menus
and, on top right, buttons and tools for scenario editing and chart outputs.

3.4. Application interface

The main window for the GISAREG application consists of three areas (Fig. 5): one for
mapping, including the legend; another for the simulation table described above, or for
presentation of results in table format; and the third for scenarios management. In addition
to the ArcView’s default commands, there are several contextual menus, buttons and tools
that where purposefully built to ease data input, building scenarios, and commanding the
simulation operations and chart outputs. The example of Fig. 5 includes on the top left the
contextual menus ‘‘Scenario’’, ‘‘Water Restrictions’’ and ‘‘Irrigation Scheme’’, which
command the ISAREG options referred before, and on the top right the buttons and tools
for scenario editing and chart presentation.
The definition of the irrigation options and water availability restrictions to be
considered when a simulation is performed are made through auxiliary windows. Irrigation
scheduling options refer to the selection of application depths and the definition of soil
water thresholds, equal or below the optimal one, which is defined for the depletion fraction
for no stress p (Allen et al., 1998). The application depths may be defined in several ways
such as fixed irrigation depths (D) depending on the irrigation method used, or the amount
of water required to refill the root zone storage up to the total available soil water (TAW) or
to a percentage of TAW. The water restrictions apply to selected time periods and refer to a
minimum interval between two successive irrigations, or to the total water depth that may
be used for irrigation during a given time interval. Different water restrictions and
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 171

Fig. 6. Mapping a simulation scenario where crop systems are randomly assigned to each field but subject to a
selected crop distribution coverage and to restrictions on soil type.

irrigation scheduling options may be assigned to selected fields, crop systems or cropped
areas through the simulation table.
GISAREG allows to simulate a scenario using a crop systems distribution pattern
different from that observed in the field by editing the ‘‘crop’’ column of the ST, or through
a specific window aimed at randomly assigning the crop systems within the project area
according to some user-defined coverage percentages. That spatial distribution of the crop
systems may be submitted to restrictions on the use of some soil type by a given a crop, e.g.
not allowing to assign a rice crop system to sandy soils. A map may be created showing the
spatial distribution of the crop systems that satisfy the user-defined criteria (Fig. 6), which
may be later imported to the ST and be used for a simulation scenario.
The simulation of multiple scenarios created as referred above allows visualizing the
impacts of alternative crop systems and irrigation management options on water use and
productivity, thus selecting the best alternatives for further implementation.

3.5. Outputs

GISAREG outputs can be presented in tabular, graphical or mapping formats and may
concern a single field, all fields inside a selected area, or the total area under analysis. In
addition, results may refer to a single date, e.g. crop water deficits at a selected day, or the
total simulation period, e.g. the crop irrigation requirements relative to a selected scenario.
Annual results are stored in the results table (RT), described in Table 3, which is a *.dbf
table named after the respective simulation scenario and simulation year. For any
172 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

Table 3
Data produced in the results table (RT)
Codes
Field
Crop

Simulation results
Annual crop irrigation requirements (mm)
Readily available water (RAW) at the beginning and end of the irrigation period (mm)
Total available water (TAW) in deep soil layers at the beginning of the irrigation (mm)
Percolation due to excess of irrigation (mm)
Precipitation during the irrigation period (mm) and non-used precipitation (mm)
Cumulated actual and maximum evapotranspiration (mm), ETa and ETm, respectively
ETa/ETm ratio
Relative yield loss, Qy ( )
Critical unit flow rate Qc (l s 1 ha 1) corresponding to the largest application depth
Irrigation date when Qc was computed
Critical period length (days) when Qc is required
Monthly irrigation requirements (mm) given in 12 columns

simulation scenario, there will be as many results tables (RT) as many years the user
decided to simulate.
Using ArcView mapping capabilities, any of the outputs included in RT may be mapped
for the entire region or for a selected area. In addition, results of the water balance for any
crop field may be visualized in a chart format (Fig. 7) clicking over that crop field record in

Fig. 7. Simulation results presented as a map of crop irrigation requirements, and water balance charts for two
fields where different irrigation schedules were applied, for no stress on left, and for an allowed water deficit on
right.
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 173

Fig. 8. Seasonal irrigation demand hydrograph relative to a selected node of the distribution system.

the RT and then pressing the auxiliary water balance button. This allows appropriate
analysis of water balance results for any simulation unit, similarly with that provided with
ISAREG. The example in Fig. 7 shows comparative results for two fields where different
irrigation management options were applied, one aiming at maximizing yields without
allowing water stress, the other relative to deficit irrigation, thus adopting a lower threshold
for timing irrigations.
Selecting the appropriate area on the map, or the respective crop fields’ records in the
simulation table, selected results aggregated to that area may be computed and displayed.
The example in Fig. 8 refers to the computation of the demand hydrograph relative to a
main node of the distribution system. The fact that the base remote sensing image is used in
the GIS makes easier to identify the irrigation distribution network that concerns the area
under analysis. This procedure is relevant to the management of irrigation projects when
both the demand and the delivery are simulated.
Further developments concern the on-going application to the Fergana Valley, also in
the Syr Darya basin, Uzbekistan. For this area, the GIS database includes a groundwater
table depth raster GIS layer having a time scale of 10 days, which allow for better
considering groundwater contribution to the crop water requirements. A soil salinity raster
GIS layer is also included for considering impacts of salinity on evapotranspiration and
crop yields. For the Fergana Valley, also a decision support tool for demand and delivery is
being developed by adapting the SEDAM model (Gonçalves et al., 2003) to operate with
GIS. In this application, GISAREG is used interactively with a surface irrigation model and
a variety of scenarios may be analyzed; the respective results are used to create attributes
relative to irrigation performances, water and land productivity, and economic indicators.

4. Application

GISAREG was applied to assess the crop irrigation requirements at Gafura-Gulyama

comparing different irrigation scenarios aiming at water savings. The area has an arid
climate, with cold winter and hot summer, where only storm rains occur from May to
174 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

Table 4
Observed crop pattern and respective crop irrigation requirements for 1999
Summer crop Area Crop irrigation requirements
(ha) (%) (mm) (m3)
Bare soil 772 25 0 0
Cotton 2,195 71 511 11,205,690
Maize (grain) 32 1 558 180,950
Maize (silage) 87 3 309 269,170
Vegetables 11 0 213 22,989
Total 3,097 100 11,678,799

September. The soils are generally heavy and deep, with high water holding capacity,
similar to those described by Stulina et al. (2005). The main summer crop is cotton and
winter crop is wheat, generally furrow irrigated. Despite land privatization is underway,
farms and fields are very large as commonly adopted in state and cooperative farms.
Simulations refer to the crop data observed in the year 1999 and scenarios are the
following:

 OY, aiming at attaining the optimal yield, i.e. adopting the p depletion fraction for no
stress as irrigation threshold;
 RES15, using the same irrigation threshold but imposing a 15 days time interval between
successive irrigations;
 RES20, as above but with a time interval of 20 days;
 STRESS, adopting a threshold 50% below that for OY.

Data in Table 4 refers to the crop distribution pattern observed in 1999 and the
corresponding net irrigation requirements. The highest demand was from the grain maize
(558 mm) followed by cotton (511 mm).
Simulation results are presented in Table 5 relative to the four scenarios. Results
indicate that a water saving representing 9% of the net irrigation demand corresponding to

Table 5
Comparing net irrigation requirements NIR (m3) for summer crops and relative yield losses Qy (%) referring to
several irrigation scheduling scenarios aiming at water saving
Summer crops Simulation scenarios
OY RES15 RES20 STRESS
NIR (m3) Qy NIR (m3) Qy NIR (m3) Qy NIR (m3) Qy
(%) (%) (%) (%)
Cotton 10,596,166 2.0 9,696,851 7.3 9,192,325 11.5 7,362,381 21.8
Maize (corn) 180,205 0.0 154,813 17.9 120,861 34.4 106,625 44.2
Maize (silage) 279,549 0.0 248,062 8.3 203,927 20.5 163,281 34.9
Vegetables 22,406 0.1 21,456 2.7 18,380 8.1 9,929 35.0
Total 11,078,326 – 10,121,182 9,535,493 7,642,216
Water saving 957,144 (8.6%) 1,542,833 (13.9%) 3,436,110 (31%)
relative to OY
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 175

the OY scenario is attainable when adopting RES15, which leads to relative yield losses of
7.3% for the cotton crop but >17% for grain maize. The RES20 scenario produces 14%
water savings in relation to the OY net demand but the relative yield losses increase to
11.5% for cotton crop and >30% for grain maize. The STRESS scenario leads to a very
high demand reduction (3.4 million m3, i.e. 31% relative to OY) but to quite large yield
losses (Table 5). Consequently, the scenario STRESS could only be considered if severe
restrictions on water supply are enforced because not only yield losses are high but also it
does not allow appropriate control of salinity, which is required in this area.
The scenarios RES15 and RES20 are generally feasible and allow for appropriate
application of controlled leaching fractions with every irrigation event. However, RES20 is
questionable for maize. Adopting RES15 produces an appreciable reduction of the net
demand volumes during the peak months of July and August. For July, the demand could
reduce from 4 to 3.45 million m3 (13.8%), and for August it could decrease from 4 to 3.3
million m3 (17.8%). These results are compatible with those presented by Horst et al.
(2005) relative to improving irrigation performances, for another location but the same
crop and similar soils. They clearly state that higher efficiencies are only attainable when
the irrigation intervals be increased.
GISAREG was also applied to assess the inter-annual variation of crop irrigation
requirements. With this purpose, the model was run for all the years of the climatic data set,
and the wet, dry and average demand years were identified. Simulations with the scenarios
OY, RES20 and STRESS were performed for the crop distribution pattern given in Table 4.
The results for the average demand year are presented in Fig. 9. When visualizing such
results, comparing Fig. 9a and b relative to OY and RES20, small differences are apparent

Fig. 9. Mapping crop irrigation requirements for the average demand year adopting three irrigation scheduling
scenarios: (a) OY, without water restrictions, (b) RES20, for 20 days minimum interval between irrigations, and (c)
STRESS, adopting a low irrigation threshold.
176 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

Fig. 10. Mapping irrigation water requirements for the dry (a) and the wet (b) years.

but for maize, that has a higher demand. On the contrary, the STRESS scenario (Fig. 9c)
shows evident differences with the other scenarios indicating that respective impacts are
large for most of the field crops.
The season net irrigation requirements (NIR) for the wet and the dry year are compared
for the scenario OY through the GIS maps in Fig. 10. Results show that NIR differences
generally do not exceed 100 mm and are mainly apparent for the cotton crop because, for
such arid climate, with negligible rains during late spring and summer, differences in
irrigation demand from a wet to a dry year mainly reflect differences in early spring
rainfall. This behavior indicates the possibility for establishing generic irrigation calendars
for farmers’ advice that apply in most of the years.

5. Conclusions

The model GISAREG, resulting from the integration of ISAREG with the GIS ArcView
3.2, allows the easy handling of an irrigation scheduling simulation model over a large area
or irrigation project, as well as the visualization of the spatial distribution of the water
demand inside that region or project. Its capabilities for performing a large number of
irrigation scheduling simulations, including for comparing different scenarios aiming at
irrigation water saving, were tested for an irrigated area in the Syr Darya basin, Uzbekistan.
Among results analyzed, the advantage for adopting 15–20-day time intervals between
irrigation was identified when focusing water saving.
The possibility for easily creating and changing scenarios, including the adoption of
crop, field or aerial specific irrigation management options, allows tailoring irrigation
management according to identified requirements. Using the visualization capabilities of
GISAREG, it becomes easy to localize where discrepancies in water management may
occur, how large they are, and which technical interventions may be required. It also allows
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 177

an easy analysis of crop water and irrigation requirements, their spatial distribution, and to
compare target with actual values. In addition, it allows computing the demand hydrograph
at main nodes of an irrigation network, or for the entire project area.
Further on-going developments concern the application to the Fergana Valley, where the
GIS database includes groundwater table depth and soil salinity raster GIS layers which
allow for better considering the groundwater contribution and impacts of salinity on
evapotranspiration and crop yields. Other on-going improvements concern the use of
GISAREG with other models constituting a decision support tool applicable at both the
farm and the irrigation system scales.

Acknowledgments

This research is funded by the research contract ICA2-CT-2000-10039 with the

European Union. Thanks are due to Dr. G. Stulina and Dr. M. Horst for providing data, to
Prof. Dr. J. Matos for his advice on the development of the GIS application, and to Prof. Dr.
J.L. Teixeira, Ms. M.J. Calejo and Ms. P. Rodrigues for the support provided relative to the
DLL of the ISAREG and KCISA models.

References

Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration. Guidelines for Computing Crop
Water Requirements, FAO Irrig. Drain. Pap. 56. FAO, Rome, 300 pp.
Burrough, P.A., McDonell, A., 1998. Principles of Geographic Information Systems for Land Resources
Assessment. Clarendon Press, Oxford, 333 pp.
Campos, A.A., Fabião, M.S., Gonçalves, J.M., Pereira, L.S., Paredes, P., 2003. Improved surface irrigation
and scheduling for upland crops in the Huinong Irrigation District: water saving issues. In: Pereira, L.S., Cai,
L.G., Musy, A., Minhas, P.S. (Eds.), Water Savings in the Yellow River Basin. Issues and Decision Support
Tools in Irrigation. China Agriculture Press, Beijing, pp. 107–130.
Cholpankulov, E.D., Inchenkova, O.P., Pereira, L.S., 2004. Adaptation of the irrigation scheduling model ISAREG
to conditions of Central Asia. In: Food Production and Water (Proceedings of the Inter-Regional Conference,
Moscow, September 2004). ICID and RuCID, Moscow, CD-ROM.
Chowdary, V.M., Rao, N.H., Sharma, P.B.S., 2003. GIS based decision support system for groundwater assessment
in large irrigation project areas. Agric. Water Manage. 62, 229–252.
Dukhovny, V.A., 2003. The Aral Sea basin-rumors, realties, prospects. Irrig. Drain. 52, 109–120.
Doorenbos, J., Pruitt, W.G., 1977. Crop Water Requirements. Irrig. Drain. Paper 24. FAO, Rome, 193 pp.
Fernando, R.M., Pereira, L.S., Liu, Y., 2001. Simulation of capillary rise and deep percolation with ISAREG. In:
Wang, M.H., Han, L.J., Lei, T.W., Wang, B.J. (Eds.), International Conference on Agricultural Science and
Technology. Information Technology for Agriculture, ICAST, vol. 6. Ministry of Science and Technology,
Beijing, China, pp. 447–455.
Gonçalves, J.M., Pereira, L.S., Campos, A.A., Fabião, M.S., 2003. Demand and delivery simulation and multi-
criteria analysis for water saving. In: Pereira, L.S., Cai, L.G., Musy, A., Minhas, P.S. (Eds.), Water Savings in
the Yellow River Basin. Issues and Decision Support Tools in Irrigation. China Agriculture Press, Beijing, pp.
247–272.
Hashmi, M.A., Garcia, L.A., Fontane, D.G., 1995. Spatial estimation of regional crop evapotranspiration. Trans.
ASAE 38 (5), 1345–1351.
Horst, M.G., Shamutalov, Sh.S., Pereira, L.S., Gonçalves, J.M., 2005. Field assessment of the water saving
potential with furrow irrigation in Fergana Aral Sea Basin. Agric. Water Manage. 77, 210–231.
178 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179

Liu, Y., Teixeira, J.L., Zhang, H.J., Pereira, L.S., 1998. Model validation and crop coefficients for irrigation
scheduling in the North China Plain. Agric. Water Manag. 36, 233–246.
Liu, Y., Li, Y.N., Pereira, L.S., Fernando, R.M., Teixeira, J.L., 2000. Irrigation management strategies for water
saving in North China plain. In: The XIV Memorial CIGR World Congress, Tsukuba, Japan. CIGR, CD-Rom
paper R 1105.
Liu, Y., Fernando, R.M., Pereira, L.S., 2001. Water balance simulation with ISAREG considering water table
interactions. In: Zazueta, F.S., Xin, J.N. (Eds.), World Congress on Computers in Agriculture and Natural
Resources (Foz do Iguaçu, Brasil). ASAE, St. Joseph, MI, pp. 857–863.
Martı́n Santa Olalla, F.M., Fabeiro, C., Brasa, A., Legorburo, A., 1999. Irrigation scheduling techniques. In: van
Lier, H.N., Pereira, L.S., Steiner, F.R. (Eds.), CIGR Handbook of Agricultural Engineering, Vol. I: Land and
Water Engineering. ASAE, St. Joseph, MI, pp. 294–297.
Ortega, J.F., de Juan, J.A., Tarjuelo, J.M., 2005. Improving water management: the irrigation advisory service of
Castilla la Mancha (Spain). Agric. Water Manag. 77, 37–58.
Oweis, T., Rodrigues, P.N., Pereira, L.S., 2003. Simulation of supplemental irrigation strategies for wheat in Near
East to cope with water scarcity. In: Rossi, G., Cancelliere, A., Pereira, L.S., Oweis, T., Shatanawi, M., Zairi,
A. (Eds.), Tools for Drought Mitigation in Mediterranean Regions. Kluwer, Dordrecht, pp. 259–272.
Pereira, L.S., van den Broek, B.,Kabat, P., Allen, R.G. (Eds.), 1995. Crop–Water Simulation Models in Practice.
Wageningen Pers, Wageningen, p. 339.
Pereira, L.S., Carreira, D., Teixeira, J.L., 1999. Besoins en eau d’irrigation: modélisation et application SIG
(Invited Lecture). In: F., Lebdi, N., Lamaddalena (Eds.), Modélisation, Techniques Informatiques et SIG pour
la Gestion de l’Eau et des Réseaux Hydrauliques (Advanced Short Course, Tunis, May 1999). Inst. Agron.
Mediter., Bari, pp. 153–193.
Pereira, L.S., Oweis, T., Zairi, A., 2002. Irrigation management under water scarcity. Agric. Water Manag. 57,
175–206.
Pereira, L.S., Teodoro, P.R., Rodrigues, P.N., Teixeira, J.L., 2003. Irrigation scheduling simulation: the model
ISAREG. In: Rossi, G., Cancelliere, A., Pereira, L.S., Oweis, T., Shatanawi, M., Zairi, A. (Eds.), Tools for
Drought Mitigation in Mediterranean Regions. Kluwer, Dordrecht, pp. 161–180.
Qing, Z., Linvill, D.E., 2002. Integrating crop models into GIS regional crop yield. In: 2002 ASAE Annual
International Meeting/CIGR XVth World Congress. Chicago, IL, July 28–31, ASAE and CIGR, CD-Rom
Paper 021196.
Ray, S.S., Dadhwal, V.K., 2001. Estimation of crop evapotranspiration of irrigation command area using remote
sensing and GIS. Agric. Water Manag. 49, 239–249.
Rodrigues, P.N., Pereira, L.S., Machado, T.G., 2000. KCISA, a program to compute averaged crop coefficients.
Application to field grown horticultural crops. In: Ferreira, M.I., Jones, H.G. (Eds.), Proceedings of the Third
International Symposium on Irrigation of Horticultural Crops (Estoril, Jun–July 1999). Acta Horticult. No.
537, ISHS, Leuven, 535–542.
Rowshon, M.K., Kwok, C.Y., Lee, T.S., 2003. GIS based scheduling and monitoring of irrigation delivery for rice
irrigation system. Part I. Scheduling. Agric. Water Manage. 62, 105–116.
San-Payo, M.M., Teixeira, J.L., 1996. Use of GIS and a\water balance model to estimate irrigation requirements.
In: Ragab, R., El-Quosy, D.E., van den Broek, B., Pereira, L.S. (Eds.), Crop–Water–Environment Models
(Proceedings of the International Workshop). Egypt Nat. Com. ICID, Cairo, pp. 199–208.
Smith, M., Pereira, L.S., Berengena, J., Itier, B., Goussard, J., Ragab, R., Tollefson, L., van Hoffwegen, P.
(Eds.), 1996. Irrigation Scheduling: From Theory to Practice. Water Report 8. FAO, Rome, 384 pp.
Sousa, V., Pereira, L.S., 1999. Regional analysis of irrigation water requirements using krigging. Application to
potato crop (Solanum tuberosum l.) at Trás-os-Montes. Agric. Water Manag. 40, 221–233.
Stewart, J.L., Hanks, R.J., Danielson, R.E., Jackson, E.B., Pruitt, W.O., Franklin, W.T., Riley, J.P., Hagan, R.M.,
1977. Optimizing crop production through control of water and salinity levels in the soil. Utah Water Research
Laboratory Re PRWG151-1, Utah State University, Logan, 206 pp.
Stulina, G., Cameira, M.R., Pereira, L.S., 2005. Using RZWQM to search improved practices for irrigated maize
in Fergana Uzbekistan. Agric. Water Manage. 77, 263–281.
Tarjuelo, J.M., de Juan, J.A., 1999. Crop water management. In: van Lier, H.N.,Pereira, L.S., Steiner, F.R. (Eds.),
CIGR Handbook of Agricultural Engineering, Vol. I: Land and Water Engineering. ASAE, St. Joseph, MI, pp.
380–429.
 P.S. Fortes et al. / Agricultural Water Management 77 (2005) 159–179 179

Teixeira, J.L., Pereira, L.S., 1992. ISAREG, an irrigation scheduling model. ICID Bull. 41 (2), 29–48.
Teixeira, J.L., Paulo, A.M., Pereira, L.S., 1996. Simulation of irrigation demand at sector level. Irrig. Drain. Syst.
10, 159–178.
Tim, U.S., 1996. Emerging technologies for hydrologic and water quality modeling research. Trans. ASAE 39 (2),
465–467.
Zairi, A., El Amami, H., Slatni, A., Pereira, L.S., Rodrigues, P.N., Machado, T., 2003. Coping with drought:
deficit irrigation strategies for cereals and field horticultural crops in Central Tunisia. In: Rossi, G., Can-
celliere, A., Pereira, L.S., Oweis, T., Shatanawi, M., Zairi, A. (Eds.), Tools for Drought Mitigation in
Mediterranean Regions. Kluwer, Dordrecht, pp. 181–201.